Preparation and Characterization of UV-Curable Acrylic Membranes Embedding Natural Antioxidants

We examine the behaviour of acrylic resin-based membranes containing natural anti-oxidants, such as Galla chinensis tea powder extract (TP) and Taiwanese green propolis (TGP), in different concentrations ranging between 5 and 20 wt %. Membrane morphology was investigated by means of Environmental Scanning Electron Microscopy (ESEM), while the UV-curing reaction was monitored by Fourier-Transform Infra-red (FTIR) spectroscopy. In most cases Thermogravimetric (TG), Differential Scanning Calorimetric (DSC) and Dynamo-mechanical Thermal (DMT) analyses showed that the desirable characteristics of the UV-cured acrylic resin are not substantially altered by the presence of the organic fillers. The release kinetics of polyphenols and flavonoids, determined in water for TP-containing membranes (ETx) and in ethanol/water mixture (7:3 v/v) for TGP-containing ones (EPx), was satisfactory, reaching a plateau after 24 h. Finally, preliminary antibacterial tests against S. epidermidis were performed on the membranes with higher additive amount and gave positive results for ET-type; on the contrary, no inhibitory effect was observed for the tested EP-type membranes.


Introduction
In recent years scientists have been putting growing attention to natural products, such as tea leaves and honeybee product called propolis; actually, these products have shown considerable antibacterial, antifungal, antiviral, anti-inflammatory and antioxidant activity, due to their high content of polyphenols and flavonoids as well as their derivatives [1,2]. Propolis is a natural resin collected by honeybees from plant parts and subsequently elaborated by them in order to seal and protect beehives [3]. Propolis composition comprises around 45% resins, 30% waxes and fatty acids, 10% essential oils, 5% pollens and 10% organic compounds and minerals [4]. It is well known for its anti-microbial properties, thus it can reduce biofilm generation resulting in accelerated healing processes [5]. Moreover, it may be used to develop anti-inflammatory drugs, i.e., for leukocyte recruitment inhibitory [6]. In particular, Taiwanese green propolis (TGP) reveals a broad spectrum of biological activities: propolins from Taiwanese green propolis inhibited the growth of Gram-positive bacterial strains like Staphylococcus aureus, Listeria monocytogenes, Bacillus subtilis and Paenibacillus larvae [7]. Moreover, Tea extract was obtained from commercially available Galla chinensis by mixing for 24 h in distilled water in the ratio of 5:1 (v/w). The solution was filtered and the obtained crude product was dried for 72 h at room temperature. Taiwanese green propolis (TGP) was harvested by a propolis collector in 2014 in Taiwan and then stored at 5 °C in darkness at the Institute of Organic Chemistry with Center of Phytochemistry at the Bulgarian Academy of Science in Sofia, Bulgaria. Before the membrane preparation the propolis sample was crushed using a rotating blade coffee grinder Bosch MKM6003 type KM13 (BSH Electrodomésticos España S.A, Madrid, Spain) to the dimension ≤ 100 µm and used as obtained.

Membrane Preparation
To produce the polymer network, 4 wt %. of the photoinitiator (Irgacure 1173) was added to the Ebecryl 150 resin. Afterwards, various amount of fillers (TP or TGP) were added to produce the expected percentage ratio of the filler in the UV-curable resin. The blends were mixed with a glass baguette to obtain homogenous dispersions. After the procedure, the mixtures were coated on glass slides and spread using a wire wound applicator (thickness set to 200 µm). To start the process of curing, slides were exposed to the UV radiation (300W; FusionUV Systems, Inc., Gaithersburg, MD, USA) in Qurtech machine model QSB-6 (Aspelt, Luxembourg) working in the horizontal configuration. The lamp-sample distance was fixed at 10 cm. Applied time of 15 s was sufficient to complete the photo-polymerization reaction. The obtained UV-cured membranes were carefully peeled off from the glass substrate and used for further analysis. Tea extract was obtained from commercially available Galla chinensis by mixing for 24 h in distilled water in the ratio of 5:1 (v/w). The solution was filtered and the obtained crude product was dried for 72 h at room temperature. Taiwanese green propolis (TGP) was harvested by a propolis collector in 2014 in Taiwan and then stored at 5 • C in darkness at the Institute of Organic Chemistry with Center of Phytochemistry at the Bulgarian Academy of Science in Sofia, Bulgaria. Before the membrane preparation the propolis sample was crushed using a rotating blade coffee grinder Bosch MKM6003 type KM13 (BSH Electrodomésticos España S.A, Madrid, Spain) to the dimension ≤ 100 µm and used as obtained.

Membrane Preparation
To produce the polymer network, 4 wt %. of the photoinitiator (Irgacure 1173) was added to the Ebecryl 150 resin. Afterwards, various amount of fillers (TP or TGP) were added to produce the expected percentage ratio of the filler in the UV-curable resin. The blends were mixed with a glass baguette to obtain homogenous dispersions. After the procedure, the mixtures were coated on glass slides and spread using a wire wound applicator (thickness set to 200 µm). To start the process of curing, slides were exposed to the UV radiation (300W; FusionUV Systems, Inc., Gaithersburg, MD, USA) in Qurtech machine model QSB-6 (Aspelt, Luxembourg) working in the horizontal configuration. The lamp-sample distance was fixed at 10 cm. Applied time of 15 s was sufficient to complete the photo-polymerization reaction. The obtained UV-cured membranes were carefully peeled off from the glass substrate and used for further analysis.

Membrane Characterizations
Optical micrographies were taken with optical microscope Nikon Eclipse LV (European Headquarters Nikon Instruments Europe BV, Amsterdam, Netherlands). The membrane's morphology was characterized by environmental scanning electronic microscopy (ESEM), using a FEI Quanta 600 (Thermo Fisher Scientific, Waltham, MA, USA). The samples were cryofractured into liquid nitrogen and then fixed on the support suitable for the cross-section analysis. Membrane thickness was measured by ESEM on the sample cross-section at least in three points.
The organic powders and the blended-resin systems were examined by infra-red attenuated total reflectance spectroscopy (IR-ATR) spectroscopy using a spectrometer (Nicolet 5700, Thermo Fisher Scientific, Waltham, MA, USA). Spectra were collected using 32 scans within the spectral range of 700-4000 cm -1 wavenumber at resolution of 1 cm -1 . Thermal stability of the films was monitored by using a thermogravimetric analyser Discovery TGA (TA Instruments-Div. di Waters SpA, Sesto San Giovanni, Italy). All experiments were performed with samples of 10-12 mg in open alumina crucibles in the temperature range of 50-700 • C under a heating rate of 10 • C/min in nitrogen or air atmosphere. The values of T 10% (the temperature corresponding to 10% weight loss) and T max (the temperature of maximum weight loss) were obtained by analysing the plots (TG) and computing the first derivative (dTG).
Differential scanning calorimetry (DSC) measurements were performed on TA Instruments Q20 DSC TGA (TA Instruments-Div. di Waters SpA, Sesto San Giovanni, Italy) using standard cell FC under nitrogen atmosphere with a 50 mL/min flow rate. Samples of 6-8 mg were sealed in aluminum crucibles. Three cycles were performed: first, samples were heated from 0 to 120 • C at a constant rate of 10 • C/min; secondly, samples were cooled down to 0 • C at 10 • C/min; finally, samples were again heated from 0 to 120 • C at a constant rate of 10 • C/min.
The effect of the temperature on the storage modulus (log E') and tan δ (loss tangent = E"/E') was evaluated from dynamic-mechanical analyses (DMTA) were carried out on a DMA Q800 (TA Instruments -Div. di Waters SpA, Sesto San Giovanni, Italy) apparatus. The samples were cut to the rectangular dimension of 0.5 cm × 3 cm. The analysis was carried out in tensile configuration from room temperature to 120 • C with a heating rate of 5 • C/min; 1 Hz tension frequency and 0.05% of oscillation amplitude in strain-controlled mode were used. For each sample DMTA tests were repeated at least 3 times on different films.

Determination of Polyphenols Content
A total of 0.5 mL of the sample was mixed with 0.5 mL of Folin-Ciocalteu's reagent and 10 mL of water. After 5 min, 8 mL of 7.5% (w/v) aqueous Na 2 CO 3 solution was added to the mixture and stored for two hours in the dark. The concentration of released phenols was determined by UV-Vis spectroscopy at 765 nm using a UV-1800 Shimadzu apparatus (Thermo Fisher Scientific, Waltham, MA, USA) [28]. The results were calculated as Gallic Acid Equivalents (GAE), using a standard curve obtained from previous experiments: Abs = 0.1214x + 0.1142, where Abs is the absorbance of the sample at 765 nm and x is the phenolic concentration in µgGAE/mL (R 2 = 0.9988). All samples were analysed in triplicate.

Determination of Flavonoids Content
A total of 0.5 mL of the sample was mixed with 0.5 mL of 2% (w/v) ethanolic solution of AlCl 3 . The mixture was stored for one hour in the dark. The concentration of released flavonoids was determined by UV-Vis spectroscopy at 420 nm using a UV-1800 Shimadzu apparatus (Thermo Fisher Scientific, Waltham, MA, USA). The results were calculated as quercetin equivalents (QE), using a standard curve obtained from previous experiments: Abs = 0.0674x + 0.0182, where Abs is the sample Polymers 2020, 12, 358 5 of 16 absorbance at 420 nm and x is the phenolic concentration in µgQE/mL (R 2 = 0.9955). All samples were analysed in triplicate.

Release Tests of Polyphenols and Flavonoids
A set of membranes cut to the dimension of 0.5 cm × 0.5 cm were placed in glass vials containing 4 mL of water. The physical parameters of the release depend on several factors, i.e., not only on properties of the films such as thickness of the material, diffusion through pores, degradation of polymer matrix, but also for mechanical cracks or physical hindrance. To exclude the last ones, each membrane was cut with particular attention and hang in the vial to have the maximum contact with the liquid. The films were supported with a plastic ring that allowed them to get into maximum contact with the liquid. For each filler, a set of four vials with different equilibrium time were analysed: 0.5, 1, 6 and 24 h. After the desired time, each membrane was discarded and the solution analysed spectrophotometrically according to the procedures explained below. As a reference of total release, a vial with the propolis amount corresponding to the one contained in 5%, 10% and 15% samples, respectively, was exposed to the mixture EtOH-H 2 O (7:3 v/v) for 24h. Filtered solutions were used for further analysis.

Antibacterial Activity
The antibacterial activity for the best performing membranes was studied against Staphylococcus epidermidis (S. epidermidis, strain CECT231). The experiments consisted in the cultivation of bacteria at 37 • C during 24 h in 5 mL Nutrient broth (NB, CECT1) medium. After that time, the cultures were 1/5 diluted in NB medium. The NB was prepared by dissolving 5.0 g/L of beef extract, 10.0 g/L of peptone and 5.0 g/L of sodium chloride in distilled water. One square centimetre membranes ET20, EP10 and EP15 were immersed in NB medium. S. epidermidis was inoculated and incubated for 1, 2, 3, 4, 6, 8 and 24 h at 37 • C. Positive and negative controls were also prepared. After each time 25 µL sample of bacteria solution was placed on the surface of NB-agar solid medium plate and incubated at 37 • C for 24 h. NB-agar solid medium was prepared by dissolving 3.75 g of beef extract, 7.5 g of peptone, 3.75 g of sodium chloride and 11.25 g of agar in distilled water. Solid medium was poured onto a plastic Petri dishes and left for gelation. Experiments were performed in duplicate.

Results and Discussion
A set of eight membranes were prepared as stated in Table 1. Two different sources of biologically active compounds were used: Galla chinensis tea powder (TP) and Taiwanese green propolis (TGP). Membranes ET5-ET20 were prepared with 5, 10, 15 or 20 wt % of TP, while membranes EP5, EP10, EP15 were obtained from Ebecryl 150 resin loaded with 5, 10 and 15 wt % of TGP, respectively. Membrane E0 was fabricated without a filler. Unfortunately, due to characteristic of propolis [29], which contain high amount of wax, films containing propolis weight percentage higher than 15 exhibited very poor mechanical properties and were not suitable for further analysis. Optical microscope studies Polymers 2020, 12, 358 6 of 16 put into evidence that in all cases both fillers are homogenously distributed in the membrane structures. As a matter of example, Figure 2 presents photos of ET10 and EP10 film surfaces captured by an optical microscope.  Figure 3 shows membrane surfaces and cross-sections captured by means of ESEM microscope. E0 membrane possesses dense structure, while in case of the composite films, the presence of the fillers can be easily evidenced in their cross-sections. Furthermore, the ESEM cross-section micrographs show that both TP and TGP particles are well incorporated in the membrane structures and they form a part of compact films. Surfaces of the membranes prepared with tea powder seem to be smooth while in case of EP membranes the TGP particles can be observed on the membrane surfaces. It is noteworthy that the particle size of the TP filler is in a range of 5-30 µm and it is twice smaller than TGP size. Moreover, the particle sizes of the fillers influence the membrane thicknesses. Indeed, thickness of the EP membranes is approximately 6% larger than the EP membranes. In the case of EP15, the ESEM cross-section image also suggests the existence of larger TGP aggregates and a less homogenous distribution in EP15.   Figure 3 shows membrane surfaces and cross-sections captured by means of ESEM microscope. E0 membrane possesses dense structure, while in case of the composite films, the presence of the fillers can be easily evidenced in their cross-sections. Furthermore, the ESEM cross-section micrographs show that both TP and TGP particles are well incorporated in the membrane structures and they form a part of compact films. Surfaces of the membranes prepared with tea powder seem to be smooth while in case of EP membranes the TGP particles can be observed on the membrane surfaces. It is noteworthy that the particle size of the TP filler is in a range of 5-30 µm and it is twice smaller than TGP size. Moreover, the particle sizes of the fillers influence the membrane thicknesses. Indeed, thickness of the EP membranes is approximately 6% larger than the EP membranes. In the case of EP15, the ESEM cross-section image also suggests the existence of larger TGP aggregates and a less homogenous distribution in EP15.
IR analysis was performed to check the occurrence of UV curing and the filler presence inside the membranes. As a matter of example, Figure 4 shows the IR spectra of tea powder, ET10 sample uncured and after UV curing, respectively, while Figure 5 shows the IR spectra of propolis powder, EP10 sample uncured and after UV curing, respectively. From the comparison of the spectra of the resins before (red lines) and after UV-curing process (green lines), it is noticeable that the peak at 1638 cm -1 , ascribed to the acrylic C=C stretching and the peaks at 983 cm -1 and at 807 cm -1 , characteristic of =C-H stretching in acrylic resins, completely disappear on curing or it is strongly reduced (see inset in Figures 4 and 5). For sake of comparison, Figure S1 reports the FTIR spectra of neat Ebecryl before and after UV curing in the same conditions, respectively. This observation confirms that the UV-curing process was driven to completion [30] in the case of ET systems; nevertheless, some residual reactivity can be still supposed in the case of the membranes containing TGP. No evident changes in other peaks of the original resin could be observed, which suggests that only physical interactions are established between Ebecryl and the fillers; this is desirable if their efficient release from the membrane is sought.
It is well known that both tea and propolis are rich in polyphenols and flavonoids, which are clearly visible in their FTIR spectra [16,31]. In the black lines of Figure 4 (TP) and Figure 5 (TGP), the wide peak situated between 3000 and 3600 cm -1 is related to the -OH group in molecules forming interand intramolecular hydrogen bonds. Additionally, propolis powder ( Figure 4, black line) contains discrete bands situated around 2880-2930 cm -1 , attributed to -OH groups which stay in accordance with previously reported data [16,24]. The fingerprint region of both compounds is similar and possess IR bands at about~1600 cm -1 , 1450 cm -1 and 1100 cm -1 , referring to C=C and C=O groups due to aromatic rings stretching vibrations of polyols, flavonoids and amino acids [16,24,32]. Both systems after UV-curing (green lines) possess bands belonging to polyphenols (3000 and 3600 cm −1 ) and acrylic resin (700-1800 cm -1 ). From the results, we can conclude that both fillers were successfully incorporated into the cured resin. and they form a part of compact films. Surfaces of the membranes prepared with tea powder seem to be smooth while in case of EP membranes the TGP particles can be observed on the membrane surfaces. It is noteworthy that the particle size of the TP filler is in a range of 5-30 µm and it is twice smaller than TGP size. Moreover, the particle sizes of the fillers influence the membrane thicknesses. Indeed, thickness of the EP membranes is approximately 6% larger than the EP membranes. In the case of EP15, the ESEM cross-section image also suggests the existence of larger TGP aggregates and a less homogenous distribution in EP15. IR analysis was performed to check the occurrence of UV curing and the filler presence inside the membranes. As a matter of example, Figure 4 shows the IR spectra of tea powder, ET10 sample uncured and after UV curing, respectively, while Figure 5 shows the IR spectra of propolis powder, EP10 sample uncured and after UV curing, respectively. From the comparison of the spectra of the resins before (red lines) and after UV-curing process (green lines), it is noticeable that the peak at 1638 cm -1 , ascribed to the acrylic C=C stretching and the peaks at 983 cm -1 and at 807 cm -1 , characteristic of =C-H stretching in acrylic resins, completely disappear on curing or it is strongly reduced (see inset in Figures 4 and 5). For sake of comparison, Figure S1 reports the FTIR spectra of neat Ebecryl before and after UV curing in the same conditions, respectively. This observation confirms that the UV-curing process was driven to completion [30] in the case of ET systems; nevertheless, some residual reactivity can be still supposed in the case of the membranes containing TGP. No evident changes in other peaks of the original resin could be observed, which suggests that only physical interactions are established between Ebecryl and the fillers; this is desirable if their efficient release from the membrane is sought. It is well known that both tea and propolis are rich in polyphenols and flavonoids, which are clearly visible in their FTIR spectra [16,31]. In the black lines of Figure 4 (TP) and Figure 5 (TGP), the wide peak situated between 3000 and 3600 cm -1 is related to the -OH group in molecules forming inter-and intramolecular hydrogen bonds. Additionally, propolis powder ( Figure 4, black line) contains discrete bands situated around 2880-2930 cm -1 , attributed to -OH groups which stay in accordance with previously reported data [16,24]. The fingerprint region of both compounds is similar and possess IR bands at about ~1600 cm -1 , 1450 cm -1 and 1100 cm -1 , referring to C=C and C=O groups due to aromatic rings stretching vibrations of polyols, flavonoids and amino acids [16,24,32]. Both systems after UV-curing (green lines) possess bands belonging to polyphenols (3000 and 3600 cm −1 ) and acrylic resin (700-1800 cm -1 ). From the results, we can conclude that both fillers were successfully incorporated into the cured resin.  Thermogravimetric (TGA) analysis was used to get information about the amount and rate of decomposition of the samples as a function of temperature in a controlled atmosphere (nitrogen or air). T 10% , Tmax and the percentage of the residue at 700 • C of all the samples in both atmospheres are gathered in Table 2, while the corresponding TG and dTG graphs are depicted in Figures S3 and S4. Thermogravimetric (TGA) analysis was used to get information about the amount and rate of decomposition of the samples as a function of temperature in a controlled atmosphere (nitrogen or air). T10%, Tmax and the percentage of the residue at 700 °C of all the samples in both atmospheres are gathered in Table 2, while the corresponding TG and dTG graphs are depicted in Figure S3 and S4. As far as the fillers (tea and propolis) are concerned, the difference in the percentage of residues with or without access to oxygen is clearly visible ( Table 2 and Figures S1 and S2). In both cases, an air atmosphere accelerates the combustion over two times. In the set of synthesized films, the higher percentage of the residue with respect to the neat resin E0, both in aerobic and anaerobic conditions, is connected with the increased filler content.
In nitrogen atmosphere, it seems that the presence of tea does not noticeably influence the degradation process of the resin ( Figure S1). On analysing dTG graphs and comparing their Tmax values (Table 2), it can be seen that the addition of 20% tea (ET20 sample) lowers the maximum degradation temperature of Ebecryl (E0 sample) only of 4 °C, whereas the presence of 15% TGP Figure 5. FTIR spectra of Taiwanese green propolis (TGP) powder (black line), EP10 before UV-curing (red line) and EP10 UV cured (green line). In the inset, the 790-1650 cm -1 region is zoomed. Table 2. Gathered values of T 10% , T max and the percentage of the residue at 700 • C obtained from thermogravimetric analysis for neat Ebecryl resin, ET and EP films. As far as the fillers (tea and propolis) are concerned, the difference in the percentage of residues with or without access to oxygen is clearly visible ( Table 2 and Figures S1 and S2). In both cases, an air atmosphere accelerates the combustion over two times. In the set of synthesized films, the higher percentage of the residue with respect to the neat resin E0, both in aerobic and anaerobic conditions, is connected with the increased filler content.
In nitrogen atmosphere, it seems that the presence of tea does not noticeably influence the degradation process of the resin ( Figure S1). On analysing dTG graphs and comparing their Tmax values (Table 2), it can be seen that the addition of 20% tea (ET20 sample) lowers the maximum degradation temperature of Ebecryl (E0 sample) only of 4 • C, whereas the presence of 15% TGP (EP15) causes a decrease of 12 • C. On the other hand, T 10% is remarkably influenced, as expected on the basis of lower decomposition temperatures of the fillers.
More complex changes are observed in thermo-oxidative process ( Figure S2). First peak observed in dTG curves situated between approximately 200 and 300 • C is most probably connected with the removal of moisture. The main decomposition steps occur between 300 and 600 • C. Actually, degradation of the films in aerobic atmosphere takes place in three steps [33]. The first step usually refers to the decomposition of small molecular weight compounds (i.e., dimers or trimers); the second step, the main one, is connected with the major degradation of the polymer network due to cross-links breaking. The last degradation step can be characterized as the oxidation of the degradation products created in the previously reported steps [33]. The presence of the organic fillers increases the onset temperature of the first degradation step as it can be easily noticed by comparing T 10% and T max1 values. Therefore, addition of both fillers slightly thermally stabilizes the polymer networks. However, Polymers 2020, 12, 358 9 of 16 on analysing the full set of membranes, higher fillers content leads to lowering of the temperature values corresponding to the following degradation steps.
Overall, incorporation of the organic fillers keeps the desirable characteristics of the UV-cured resin as far as thermal stability is concerned. However, the results obtained for EP15 film put into evidence the lack of T max1 , whereas in case of tea this situation is observed for ET20 sample (Table 2), thus showing a change in the decomposition mechanism. This may suggest the maximum amount of loading which does not affect the polymer network in relevant extent. Lei et al. as well noticed a strong addition-dependent pattern [18]. In particular, the incorporation of tea polyphenols to pectin and konjac glucomannan films increased their mechanical and water-resistant properties but only to some extent. Higher concentrations (beyond 2%) had a negative impact on their performance.
Polymer glass transition temperature (T g ) is a diagnostic value for determining compatibility or incompatibility of polymer blends. Table 3 presents the values of glass transition temperature (T g ) determined by DSC for E0, ET and EP samples, while Figures S3 and S4 show their enthalpy relaxation curves. All samples were heated twice, as reported in the experimental section. On analysing the first heating run, it is visible that addition of both fillers does not influence T g of neat Ebecryl (E0 sample). Glass transition temperature of pure resin was determined as 49 • C and stays in agreement with previously reported data (50 • C) [33]. However, in the presence of the fillers, qualitative differences in plots are unquestionable ( Figures S3 and S4). On the other hand, stronger changes are visible in the second heating run. In this case, all T g 's for polymers with various percentages of fillers are characterised by lower temperatures than neat resin sample E0 (Table 3). Actually, in the second heating scan, glass transition of E0 slightly increases (53 • C) whereas T g of all blends decreases; furthermore, no pronounced thermal transition is observed. This suggests that after the first heating scan, the segments between cross-link points in the matrix were re-arranged. However, in this case the reduced T g value of all blends in comparison with pure resin indicates higher matrix mobility: this could be related with higher amount of filler, that can act as a plasticizer for Ebecryl films. In any case, one should underline that no blend compositions evidenced multiple glass transition temperatures from DSC analysis over the entire range of scanning, thus suggesting that the tested samples were homogeneous.
DMTA technique is often used to investigate the compatibilities between filler and polymer matrix [34,35]. In fact, by this technique it is possible to identify particle/polymer interactions and evaluate the reinforcement or labefaction effects on the viscoelastic properties of matrix. The dynamic-mechanical thermal properties of Ebecryl blends are shown in Figure 6 and the values of the determined T g and storage modulus (E') at 32 • C are reported in Table 3. As expected, the addition of both fillers (tea and TGP) affects the UV-cured acrylic resin: in each formulation, the storage modulus decreases and tan δ peak broadens and is slightly shifted. In both cases, the stiffness and the T g of the blends decrease as a consequence of increasing percentage of the filler. dynamic-mechanical thermal properties of Ebecryl blends are shown in Figure 6 and the values of the determined Tg and storage modulus (E') at 32 °C are reported in Table 3. As expected, the addition of both fillers (tea and TGP) affects the UV-cured acrylic resin: in each formulation, the storage modulus decreases and tan δ peak broadens and is slightly shifted. In both cases, the stiffness and the Tg of the blends decrease as a consequence of increasing percentage of the filler. In the case of propolis, the influence on the viscoelastic properties of Ebecryl matrix is characterized by a gradual decrease in storage modulus over the whole temperature range and by a significant broadening and decrease in tan δ, as expected based on the plasticizing effect of TGP. Moreover, FTIR results reported in Figure 5 for EP10 suggest that the presence of increasing amounts of TGP interferes somewhat in the completion of crosslinking reaction. When the blend reaches 15 wt %, DMTA test does not provide clear results. Repeated analysis on different samples gave lower storage modulus and tan δ peak located at lower temperature as depicted in Figure 6C,D (yellow solid and dashed lines). In order to understand this evidence, we should take into account that, unarguably, higher content of TGP softens the polymer network. This situation is understandable when keeping in mind that natural propolis contains not only flavonoids and polyphenols, but also In the case of propolis, the influence on the viscoelastic properties of Ebecryl matrix is characterized by a gradual decrease in storage modulus over the whole temperature range and by a significant broadening and decrease in tan δ, as expected based on the plasticizing effect of TGP. Moreover, FTIR results reported in Figure 5 for EP10 suggest that the presence of increasing amounts of TGP interferes somewhat in the completion of crosslinking reaction. When the blend reaches 15 wt %, DMTA test does not provide clear results. Repeated analysis on different samples gave lower storage modulus and tan δ peak located at lower temperature as depicted in Figure 6C,D (yellow solid and dashed lines). In order to understand this evidence, we should take into account that, unarguably, higher content of TGP softens the polymer network. This situation is understandable when keeping in mind that natural propolis contains not only flavonoids and polyphenols, but also waxes. Indeed, several melting points were reported [29]. The two values reported for the maximum of tan δ peaks from tests on different samples ( Figure 6D and Table 3) interestingly stay in agreement with the two transitions evidenced by DSC for TGP ( Figure S4, blue line). On analysing the morphology of the membrane by ESEM (Figure 3), it can be seen that this formulation is characterized by less homogenous distribution and larger aggregates of filler in comparison to the blend obtained with tea. Therefore, it is reasonable to attribute the mismatch between different DMTA tests on EP15, to the heterogeneities present in this sample.
Summing up, DMTA results are in good agreement with DSC analysis debated earlier. Conducted experiments confirm that the mechanical properties of the prepared films strongly depend on their composition, as expected. In particular, the highest percentage of both fillers remarkably softens the polymer network. Table 4 gathers the total concentrations of polyphenols and flavonoids obtained for tea and TGP powder in the determinant test and treated as 100% of release. The release kinetics of polyphenols and flavonoids for ET samples is shown in Figure 7.
Conducted experiments confirm that the mechanical properties of the prepared films strongly depend on their composition, as expected. In particular, the highest percentage of both fillers remarkably softens the polymer network. Table 4 gathers the total concentrations of polyphenols and flavonoids obtained for tea and TGP powder in the determinant test and treated as 100% of release. The release kinetics of polyphenols and flavonoids for ET samples is shown in Figure 7. Tea extract is soluble in water and the tea powder incorporated in the membrane passes fast into environment, as expected. As it can be observed in Figure 7, all presented graphs adopt the exponential shape and seem to reach a saturation. The fastest "burst" release is observed in the first 5 h and can be attributed to the filler trapped closer to the membrane surface; as shown by Figure 5, 24 h seem to be sufficient to reach a plateau for polyphenols measurement. In the case of ET20, release is the fastest one. In the case of ET5, experimental determination of released polyphenols gave low values affected by too high an error, and were therefore considered not significant. On the other hand, Tea extract is soluble in water and the tea powder incorporated in the membrane passes fast into environment, as expected. As it can be observed in Figure 7, all presented graphs adopt the exponential shape and seem to reach a saturation. The fastest "burst" release is observed in the first 5 h and can be attributed to the filler trapped closer to the membrane surface; as shown by Figure 5, 24 h seem to be sufficient to reach a plateau for polyphenols measurement. In the case of ET20, release is the fastest one. In the case of ET5, experimental determination of released polyphenols gave low values affected by too high an error, and were therefore considered not significant. On the other hand, the membranes ET10, ET15 and ET20 after 96 h of experiment reached 26%, 39% and 48% of release, respectively. These results most likely arise because of the membrane loosening consequent to higher amount of filler, which eases solvent penetration into the membrane and the subsequent filler extraction. Indeed, the release kinetics of flavonoids shows that, after 96 h, as much as 98% of components is released from ET20 sample. Visibly lower released concentration was found from membranes containing lower filler amount: 67%, 57% and 23%, respectively, for ET15, ET10 and ET5. Very recently Gao et al. [36] presented a study of time release tests for polysaccharide films based on pectin and chitosan. They incorporated tea polyphenols and created formulations of 2.5%, 5% and 15% w/w. The highest ratio of 45% release was registered for 15% loading after 24 h. Figure 8 reports the release kinetics of polyphenols and flavonoids in the mixture EtOH-H 2 O (7:3 v/v) for EP samples. membranes containing lower filler amount: 67%, 57% and 23%, respectively, for ET15, ET10 and ET5. Very recently Gao et al. [36] presented a study of time release tests for polysaccharide films based on pectin and chitosan. They incorporated tea polyphenols and created formulations of 2.5%, 5% and 15% w/w. The highest ratio of 45% release was registered for 15% loading after 24 h. Figure 8 reports the release kinetics of polyphenols and flavonoids in the mixture EtOH-H2O (7:3 v/v) for EP samples. All EP films show similar kinetic release pattern of antioxidants through the whole time of the experiment (96 h), which is also analogous to the one observed for ET samples. Actually, also for EP samples the fastest release is observed within the first five hours. A similar time-dependent mode was observed for propolis extract release in natural rubber latex membrane [16]. In the same way as ET5, in the case of EP5 the experimental values of released polyphenols were affected by great uncertainty and were, therefore, discarded. On the other hand, the determination of flavonoids released form EP5 gave coherent values. During the first hour, EP10 membrane releases 28% of polyphenols and 32% of flavonoids; after 24 h, this amount rose to 67% and 64%, respectively. Increased amount of the filler to 15% increases the release of the active substance, i.e., after 24 h, 91% of polyphenols and 83% of flavonoids are being exposed to the environment. Similarly to ET systems, the results suggest that increased percentage of filler facilitates solvent penetration and enhances release. However, while in EP5 and EP10 after 96 h of experiment a plateau can be envisaged, this is not the case for EP15, since the release from this sample significantly decreased at this time, especially for polyphenols. It stays unclear why the experiment ended with lowered release value than expected. It could be explained with either spontaneous sedimentation or unexpected stacking effects of aromatic moieties of antioxidants. Nevertheless, obtained results are satisfactory from time-release point of view. Moreover, this unclear result somehow overlaps with the adverse effect of 15 wt % formulation onto mechanical and thermal properties of Ebecryl matrix.
Presented release results stay in agreement with already published data. Hydroalcoholic extract of propolis was incorporated into the poly(vinyl alcohol) (PVA) to create nanofiber mat (60% w/w) [23]. The initial burst within the first 2-3 h released around 45% of the substance. Later, a release of 69% and 83% for 24 h and 96 h, respectively, was observed.
Antibacterial activity was preliminarly tested against S. epidermidis. This is a staphylococcus bacterium, consisting of Gram-positive coconuts arranged in groups; it is a component of the normal All EP films show similar kinetic release pattern of antioxidants through the whole time of the experiment (96 h), which is also analogous to the one observed for ET samples. Actually, also for EP samples the fastest release is observed within the first five hours. A similar time-dependent mode was observed for propolis extract release in natural rubber latex membrane [16]. In the same way as ET5, in the case of EP5 the experimental values of released polyphenols were affected by great uncertainty and were, therefore, discarded. On the other hand, the determination of flavonoids released form EP5 gave coherent values. During the first hour, EP10 membrane releases 28% of polyphenols and 32% of flavonoids; after 24 h, this amount rose to 67% and 64%, respectively. Increased amount of the filler to 15% increases the release of the active substance, i.e., after 24 h, 91% of polyphenols and 83% of flavonoids are being exposed to the environment. Similarly to ET systems, the results suggest that increased percentage of filler facilitates solvent penetration and enhances release. However, while in EP5 and EP10 after 96 h of experiment a plateau can be envisaged, this is not the case for EP15, since the release from this sample significantly decreased at this time, especially for polyphenols. It stays unclear why the experiment ended with lowered release value than expected. It could be explained with either spontaneous sedimentation or unexpected stacking effects of aromatic moieties of antioxidants. Nevertheless, obtained results are satisfactory from time-release point of view. Moreover, this unclear result somehow overlaps with the adverse effect of 15 wt % formulation onto mechanical and thermal properties of Ebecryl matrix.
Presented release results stay in agreement with already published data. Hydroalcoholic extract of propolis was incorporated into the poly(vinyl alcohol) (PVA) to create nanofiber mat (60% w/w) [23]. The initial burst within the first 2-3 h released around 45% of the substance. Later, a release of 69% and 83% for 24 h and 96 h, respectively, was observed.
Antibacterial activity was preliminarly tested against S. epidermidis. This is a staphylococcus bacterium, consisting of Gram-positive coconuts arranged in groups; it is a component of the normal microbiota of human skin and surface mucous membranes [37,38]. Nevertheless, it has also turned as the most important pathogen in infections related to some biomedical materials, such as joint prostheses and heart valves [39,40]. S. epidermidis provokes the growth of biofilms in plastic devices that are inserted in the body. Bacteria possess an outer layer of polysaccharides that adheres strongly to plastic, which also hinders antibiotic treatment, by preventing their penetration. For instance, together with other Staphylococcus bacteria, is responsible for 47% of joint prosthetic infections and probably represents the most common species found in laboratory tests [39].
First, we tested antibacterial activity of antioxidants (polyphenos and flavonoids), extracted from TP and TGP, respectively. Figure 9 shows the results obtained at time 0 ( Figure 9A) and 48 h ( Figure 9B), respectively. As it can be seen, both TP and TGP extracts show significant inhibitory activity against S. epidermidis, as evidenced by the clear holes surrounding the well after 48 h. It can be also noticed that TP exhibits remarkably higher diffusion into the medium. that are inserted in the body. Bacteria possess an outer layer of polysaccharides that adheres strongly to plastic, which also hinders antibiotic treatment, by preventing their penetration. For instance, together with other Staphylococcus bacteria, is responsible for 47% of joint prosthetic infections and probably represents the most common species found in laboratory tests [39].
First, we tested antibacterial activity of antioxidants (polyphenos and flavonoids), extracted from TP and TGP, respectively. Figure 9 shows the results obtained at time 0 ( Figure 9A) and 48 h ( Figure 9B), respectively. As it can be seen, both TP and TGP extracts show significant inhibitory activity against S. epidermidis, as evidenced by the clear holes surrounding the well after 48 h. It can be also noticed that TP exhibits remarkably higher diffusion into the medium. Afterwards, we tested whether the antibacterial activity is sustained when the antibacterial agents are entrapped into a polymeric matrix, i.e., in the previously prepared membranes. For this scope, we selected the membranes with higher antioxidant content that is ET20, EP10 and EP15. Controls were also prepared: the positive one, containing the liquid medium + S. epidermidis, while the negative one contained the liquid medium without bacteria. In addition, the test was also performed with neat E0 membrane. The results are shown in Figure 10. It can be noticed that, after 24 h incubation, E0 membrane shows no antibacterial effect, as expected, which also suggests that no component of UV formulation possesses any antibacterial activity after UV-curing. Actually, in the literature several examples about UV-cured films containing Afterwards, we tested whether the antibacterial activity is sustained when the antibacterial agents are entrapped into a polymeric matrix, i.e., in the previously prepared membranes. For this scope, we selected the membranes with higher antioxidant content that is ET20, EP10 and EP15. Controls were also prepared: the positive one, containing the liquid medium + S. epidermidis, while the negative one contained the liquid medium without bacteria. In addition, the test was also performed with neat E0 membrane. The results are shown in Figure 10.
together with other Staphylococcus bacteria, is responsible for 47% of joint prosthetic infections and probably represents the most common species found in laboratory tests [39].
First, we tested antibacterial activity of antioxidants (polyphenos and flavonoids), extracted from TP and TGP, respectively. Figure 9 shows the results obtained at time 0 ( Figure 9A) and 48 h ( Figure 9B), respectively. As it can be seen, both TP and TGP extracts show significant inhibitory activity against S. epidermidis, as evidenced by the clear holes surrounding the well after 48 h. It can be also noticed that TP exhibits remarkably higher diffusion into the medium. Afterwards, we tested whether the antibacterial activity is sustained when the antibacterial agents are entrapped into a polymeric matrix, i.e., in the previously prepared membranes. For this scope, we selected the membranes with higher antioxidant content that is ET20, EP10 and EP15. Controls were also prepared: the positive one, containing the liquid medium + S. epidermidis, while the negative one contained the liquid medium without bacteria. In addition, the test was also performed with neat E0 membrane. The results are shown in Figure 10.  It can be noticed that, after 24 h incubation, E0 membrane shows no antibacterial effect, as expected, which also suggests that no component of UV formulation possesses any antibacterial activity after UV-curing. Actually, in the literature several examples about UV-cured films containing additives with antibacterial activity were reported; nevertheless, when the films containing no additive were investigated, no inhibitory effect could be found [41][42][43].
Differently, a clear inhibitory effect can be evidenced in the case of ET20 membrane. On the other hand, no inhibitory effect can be detected for EP10 and EP15 systems. In fact, this apparently surprising result can be reasonably explained: first, it must be pointed out that fractionation of propolis to get compounds is difficult due to its complex composition [44]. Traditionally, the soluble fraction in alcohol, called 'propolis balsam' is extracted and used for therapeutic scopes. As a matter of fact, we could get TGP relatively fast released, when 7/3 v/v ethanol/water mixture was used as medium. Even in this case, from Table 4 one can observe that the obtained content of polyphenols and flavonoids from EP10 and EP15, is far lower than the one released from ET membranes containing the same amount of additive and extracted from water, thus putting into evidence lower solubility of TGP with respect to TP in their respective extraction media. Actually, it was reported that water extract from TGP revealed no antibacterial activity against S. aureus [7]. Therefore, since the antibacterial activity tests are performed in aqueous medium, the lack of inhibition of EP systems could be due to an insufficient amount of antioxidants diffused into it.

Conclusions
We prepared acrylic membranes, based on Ebecryl 150 acrylic resin, filled with tea extract from Galla chinensis (TP) or Taiwanese green propolis (TGP) as organic fillers with potential antibacterial activity. The organic additive was added in different percentages, ranging from 5% to 20%. In all cases the presence of TP and TGP did not remarkably interfere with the UV-curing process and homogeneous, self-sustainable membranes could be obtained, as evidenced by FTIR and ESEM analyses; however, in the case of 20 wt % TGP, the resulting membrane showed poor mechanical properties and was discarded. The desirable characteristics of the UV-cured acrylic resin were sustained in most cases, as shown by TGA, DSC and DMTA. However, the concentration dependent motif is clearly visible. Results obtained for 15% TGP film and 20% TP suggest that this may be the maximum amount of the filler, which does not affect the polymer network and its mechanical properties in higher extent. Detailed release tests of polyphenols and flavonoids revealed that all sets of membranes show similar kinetic release pattern of antioxidants through the experiment. The fastest release is observed within first 5 h and for all systems 24 h seem to be sufficient to achieve a plateau. For all the investigated membranes, the release rate increase with increasing the tea or propolis content in the matrix. Finally, preliminary antibacterial tests were performed against Staphylococcus epidermidis on the membranes containing higher additive content and revealed a satisfactory inhibitory effect in the case of 20% TP membrane. On the other hand, no effect was detected for TGP containing membranes, probably due to the limited solubility of the additive into the aqueous medium.