Preparation and characterization of interface-modified PLA/starch/PCL ternary blends using PLLA/triclosan antibacterial nanoparticles for medical applications

Abstract

In this study, the interface-modified ternary blends based on polylactic acid/starch/polycaprolactone were prepared via melt blending. On one hand, the addition of triclosan-loaded polylactic acid (LATC30) nanoparticles imparted antibacterial properties to the blend, and on the other hand, the interfacial affinity between hydrophilic starch and hydrophobic polyesters was improved by establishing interactions between –Cl groups of triclosan and ester groups of PCL. AFM and rheological results demonstrated that the starch particles are very coarse in PLA/starch 50:50 (PLAS50) but as the content of PCL increased, the extent of chemical interactions between starch and PCL was increased which resulted in a much better dispersion and homogeneity of starch within the ternary blends. Tensile analysis indicated that the interface modification of starch by the hydrophobic PCL and triclosan significantly improved the elongation at break and tensile strength values of PLAS50, and the mechanical character of blends was transformed from brittle to ductile. Biodegradability was also investigated and the triclosan release rate was tuned, while the blends are also antibacterial and their cell-viability was approved. Overall, the ternary blending approach of PLA/starch/PCL using triclosan is a promising technique to extend the property range of PLA suitable for many medical applications.

---

1. Introduction

Biodegradable materials are widely used in packaging, agriculture, medicine and many other areas. In recent years, the interest in biodegradable polymers has increasingly grown. However, biodegradable polymers cannot be widely used because of the limitations induced by their prices and mechanical properties.¹ Polylactic acid (PLA) is a biodegradable polyester which is used in a wide range of applications such as medical usages,² packaging^3,4 etc. PLA has good thermoplasticity, processability, biocompatibility and physical properties, such as high strength and modulus. However, this polymer is very brittle having serious physical aging problems during its long service life.⁵ Poly(ε-caprolactone) (PCL) is a well-known synthetic, biodegradable and semi-crystalline polyester which is famous for its high elongation at break. Its low melting point at around 60 °C is regarded as an impediment for its application as a common thermoplastic. PCL is an appropriate candidate for moderating brittleness of PLA whereas PLA has been recommended for increasing the low strength of PCL.⁶ Both PLA and PCL show hydrophobicity as compared to the other polyesters^5,6 which is not desirable for medical applications; however, the main limitation of these polymers is their high price which is arisen from their complicated production process. On the other hand, starch is a low cost material which can be abundantly found in nature,⁷ but mostly it cannot be used in practical applications because of its brittleness and high water absorption.⁸ Starch is a partially crystalline polymer, with a melting temperature (T_m) between 220 and 240 °C, and degradation temperature around 220 °C. Therefore, it has to be modified in order to be melt-processed as a thermoplastic.⁹ Starch is moisture-sensitive and exhibits weak mechanical properties as compared with the other thermoplastic polymers. Blending of starch with other polymers is a practical route to surmount such limitations.¹⁰ Numerous studies have been conducted on the blending of starch with different biodegradable polyesters. The main goal of those researches was to develop a biodegradable blend with desired properties.^1,5,8,11,12 Previous studies demonstrated that the addition of starch into PLA could decrease modulus, elongation at break, tensile strength and impact resistance of the obtained blends with respect to the neat PLA.^5,13 The obtained poor mechanical properties of these blends were attributed to the poor interfacial affinity of starch and PLA, which is originated from the incompatibility of hydrophilic starch and hydrophobic polymers like PLA¹⁴ and PCL. Jang et al. reported that the maleated thermoplastic starch is not as appropriate as maleic anhydride (MA) to be used as a reactive compatibilizer for PLA/starch blends.¹² They claimed that the crystallinity of PLA/starch blends were enhanced by the addition of MA due to the dual effects of nucleation of starch and plasticization of MA. An interfacial modification was performed by Huneault et al. via free-radical grafting of MA onto the PLA chains followed by a reaction with starch macromolecules. MA-grafted PLA/starch blends showed much finer dispersed phase of starch and exhibited a dramatic improvement in ductility.¹⁵ Mohammadi-Rovshandeh et al. used triethyl citrate as a plasticizer for PLA/starch/rice husk blend.⁸ A PLLA/starch blend with composition of 50/50 wt% compatibilized by 10 wt% PLLA-gr-starch was prepared by Chen et al., and exhibited a tensile strength of 24.7 MPa and an elongation at break of 8.7% while the non-compatibilized 50/50 blend showed a tensile strength of 11.3 MPa and an elongation at break of 1.5%.¹ In another work, significant improvements in mechanical properties were confirmed as PLA was reactively blended with starch by using diethylene triamine, 1,6-diisocyanatohexane.⁵ Wang et al. deduced that water may act as an effective plasticizer for starch; however, it could depolymerize PLA in the course of extrusion processing deteriorating the properties of TPS/PLA blends.¹⁴ Wu demonstrated that hydrophobicity of starch was increased, and the water resistance of acrylic acid (AA)-gr-PLA blend was higher than that of PLA/starch blend due to the improved interaction between AA-gr-PLA and starch.¹⁶ In another study, PCL and granular starch were found to be mechanically compatible, however, they were phase-separated. Moreover, starch and its corresponding derivatives can be chemically incorporated into PCL composites to increase the modulus.⁷ In this regard, chlorinated starch (starch-Cl) was blended with PCL by Kweon et al. in order to enhance the miscibility of PCL/starch blends. The increased level of miscibility for this blend was ascribed to a chemical reaction between the chloride group in starch-Cl and the ester group in PCL. This phenomenon was further proved by the results of FTIR, DSC and SEM.¹⁷ Starch is known to form capillary, surface sorption and/or inclusion complexes with inorganic and organic guest molecules which are reported in previous studies,^18,19 in which the results point to the importance of van der Waals and dispersion forces in holding hydrocarbons and inorganic molecules by starch.¹⁸ Lay Ma et al. also reported that ester inclusion complexes may have potential for the delivery of bioactive compounds. However, the formation of these complexes may be limited by the solubility and molecular structure of the specific compound.¹⁹ Sarazin et al. showed that the ternary blend of PLA/PCL/TPS could exhibit great ductility with a greatly enhanced elongation at break of ∼55% as compared to 5% for the pure PLA.¹⁰ Liao et al. lowered the price of PLA/PCL blends by means of adding starch into these blends. In fact, the problem of poor compatibility between PLA/PCL blend and starch was surmounted by grafting acrylic acid to PLA/PCL blend such that a chemical reaction occurred between –OH group of starch and –COOH group of AA-gr-PLA/PCL.²⁰ Up to now, a wide range of nanomaterials have been developed for efficient antibiotic drug delivery, and their efficacy has been demonstrated. In the last decade, several compounds such as titanium dioxide (TiO₂), chlorinated compounds, selenium, quaternized poly(2-dimethylamino ethyl methacrylate), silver ion and nanoparticles of silver and triclosan (TC) have been considered as the antibacterial and antimicrobial agents for polymeric matrices.²¹ TC is a well-known commercial and a Food and Drug Administration (FDA) approved-synthetic-non-ionic-broad-spectrum antimicrobial agent.²¹ Moreover, TC is chemically stable and can be heated up to 200 °C for two hours while showing no degradation, so it is suitable for incorporation into various reinforced plastic materials.^21–23 In one of our previous studies, PLLA/TC nanoparticles were prepared by the emulsification–diffusion process, and the PLLA/TC with composition of 70/30 wt% (LATC30) showed larger encapsulation efficiency, better molecular dispersion and the best release profiles according to the presented fitted data by the different models.²⁴ In another study by our group, the encapsulated TC in low molecular weight PLLA (LATC30) was dispersed in PLLA with higher molecular weight via melt blending to increase the overall properties, and in particular, the antibacterial activity of the system. The proposed method resulted in an utterly homogenous composite for which 5 wt% of LATC30 filler showed the optimum results.²¹ In the current study, the optimum amount of LATC30 in our previous work (5 wt% LATC30) was added during the blending of PLA, starch and PCL via an internal mixer. An attempt was made to fully characterize the prepared blends. The main goal of this study is to examine overall properties of PLA/starch/PCL ternary blends prepared via melt processing operation and investigate the effect of antibacterial TC on the final properties which, to the best of our knowledge, has not been reported elsewhere. Moreover, it is also anticipated that upon addition of TC into this system, chlorination would occur for –OH groups of starch followed by a chemical reaction between the as-formed new site and the ester group of PCL, which improves the miscibility of the prepared ternary blend. The main novelty of this work is use of triclosan as an antibacterial material and compatibilizer for improving the miscibility of the blend. The modified ternary blend can be ultimately used in different applications such as food packaging, membranes and many medical applications.

2. Materials and methods

2.1. Materials

PLLA was synthesized from the purified L-lactide prepared by L-lactic acid.^25,26 The ring opening polymerization was performed for the dried L-lactide monomers via reactive extrusion method fully described elsewhere.²⁴ The molecular weight (M_w) and poly dispersity index (PDI) of the PLLA were 120 000 g mol⁻¹ and 1.78, respectively. The PLLA/TC nanoparticles (LATC30) were prepared according to the procedure mentioned in our previous work,²⁴ based on which the PLLA/TC nanoparticles were synthesized by the modified emulsification–diffusion process. It was found that the sample containing 30% TC and 70% PLLA showed the best results exhibiting the highest encapsulation efficiency with an approximate size of 100–300 nm. However, since only 30% of the nanoparticles consists of TC, the real dosage of TC is ca. 1.5%.²⁴ Henceforth, PLLA will be referred as PLA. TC was supplied from KAF Co. (Branch of DAROUGAR group, Iran) originally bought from Dekaben TC Premium, Jan Dekker. The technical grade corn starch supplied by Glucosan Co. was used in this study. The moisture content in the as-received starch was ca. 12 wt%, which was measured by drying the starch in an oven heated up to 105 °C for 24 h. Poly(ε-caprolactone) with a molecular weight of 80 000 (M_n) was purchased from Sigma-Aldrich (Germany). All the other chemical and solvents were reagent grades procured from Merck (Darmstadt Germany).

2.2. Sample preparation

Prior to the blending process, the PLA and starch granules were vacuum dried at 80 °C for 4 h, and PCL granules were also vacuum dried at 30 °C for 12 h. Afterwards, they were melt-blended by an internal mixer (Brabender 50 EHT at a rotational speed of 60 rpm for 15 min). All of the components were simultaneously loaded into the mixing chamber in several compositions. The stock temperature of the internal mixer was maintained at 170 °C. The final blends were hot-pressed at the same temperature to prepare specimens for all measurements, and then, cooled to the room temperature. Finally, the prepared samples were sterilized by ethylene oxide. The samples were named as PLASCL followed by a number representing the weight percent of PCL in the blend while the remained content was consisted of 50 wt% PLA and variant starch percentages. For instance, PLASCL10 contains 10 wt% PCL, 40 wt% starch and 50 wt% PLA. Based on previous studies, 5 wt% LATC30 was added to the 100 wt% of blends.²¹ For contact angle measurements, SCL5 (90 wt% starch, 10 wt% PCL and small amount of LATC30, which were in the same ratio of starch and PCL in PLASCL5) and SCL25 (50 wt% starch, 50 wt% PCL and small amount of LATC30 similar to PLASCL25) were separately prepared using internal mixer. Moreover, 0.2 g of TC powder at room temperature was pressed for 1 min under 50 bar loading to prepare a flat tablet for contact angle measurement.

2.3. Characterization

The infrared spectroscopy was performed using the FTIR-ATR, Bruker Equinox 55LS 101 series for determining the functional groups. The measurements of mechanical behavior of the ternary blends were carried out by means of a Gotech Universal AI-7000-LA according to ASTM D638. The measurements were conducted at room temperature (25 °C). The cross-head speed was 5 mm min⁻¹. At least three specimens of every composition were tested. The impact strength of the blends was obtained using a FRANK Baldwin-Model-BLI pendulum impact testing machine to measure the edgewise un-notched rectangular specimens (50 × 10 × 1.5 mm³) according to ISO 179/3 e U. At least three specimens were tested for each sample. The hardness of the samples was measured by a Shore D durometer (FRANK testing machine). For each sample, at least five specimens were tested at different points according to DIN53505. The rheological properties of blends were measured at 170 °C using an Anton Paar Physica MCR102 (Graz, Austria) with parallel-plate geometry according to ASTM D4440-07 under air atmosphere. The dynamic experiments were performed under oscillatory shear mode with dimensions of 25 mm (diameter) and 0.5 mm (gap). The frequency sweep measurements were carried out in the angular frequency range of 0.1 to 500 rad s⁻¹. Prior to frequency sweep measurement, the strain sweep test was carried out to confirm that the applied strain did not exceed the limit of linear viscoelasticity. The time sweep was carried out to ensure the physical properties during the frequency sweep have not been changed. The differential scanning calorimetry (DSC) was carried out using a Mettler Toledo DSC 1 Star System equipped with a low-temperature accessory. The temperature scale was calibrated with the high-purity standards. The DSC measurements were performed at a heating rate of 10 °C min⁻¹, in the nitrogen atmosphere and at temperatures ranging from −60 to 260 °C. The glass transition temperatures (T_g) of samples were taken from the midpoint of the stepwise specific heat increment. The repeated heating scans were performed in order to verify the reproducibility of results. All of the adjustments were carried out according to ASTM D3418. The degree of crystallinity (X_c) for samples was determined according to the following equation:²⁷where ΔH_m and ΔH_c are the enthalpies of melting and cold crystallization, respectively. W and ΔH⁰_m are the weight fraction of poly(L-lactide) and melting enthalpy of 100% crystalline poly(L-lactide) or PCL, respectively. The topography of specimens was detected using a DS 95 AFM from DME Nanotechnologie GmbH in the tapping mode. Silicon tips and a spring with constant of ∼20–40 N m⁻¹ were used. The samples were molded into disc shapes with 2.5 cm in diameter and thickness of 0.5 mm. After the preparation of the surface by a cryogenic Leica microtome with a glass knife, the specimens were reached to the room temperature, and then, AFM was performed. The antibacterial properties of blends were carried out against Gram-positive Staphylococcus aureus (S. aureus) and Gram-negative Escherichia coli (E. coli) bacteria, according to the disc diffusion method. This method was performed in a medium solid agar Petri dish. The samples were molded into disc shapes 2.5 cm in diameter, sterilized with ethylene oxide for 2 h, and placed on E. coli and S. aureus cultured agar plates. Then, they were incubated for 24 h at 37 °C, and the inhibition zones were recorded.²¹ The MTT assay was performed as the proof of cell viability, according to our previous works with the same routes on osteoblast cells for the disk-shaped samples.^21,28 The UV-visible spectrophotometer was used to determine the amount of TC release from the blends according to our previous work and the results were validated with the following models:

Zero order: Q_t = Q₀ + K₀t (3)

First order: ln Q_t = ln Q₀ + K₁t (4)

Higuchi: Q_t = K_Ht^1/2 (5)

Hixon–Crowell: Q₀^1/3 − Q_t^1/3 = K_st (6)

where M_t is the accumulative amount of drug released at time t, M_∞ shows the initial drug loading, K represents the constant characteristic of the drug–polymer system, and n denotes the diffusion exponent which implies the nature of the release mechanism. Q_t is the amount of drug dissolved at time t, Q₀ is the initial amount of drug in the solution (most of the time, Q₀ = 0), K₀ is the zero-order release constant, K₁, the first-order release constant, K_H, the Higuchi dissolution constant, and K_s is a constant incorporating the surface–volume relation.^21,29 The Kruss G10 apparatus was implemented for measurement of water contact angle to obtain hydrophilicity of the samples. The water contact angle of samples was measured by dropping droplets of ∼5 μL at room temperature. The distance of vibrating syringe was ∼5 cm and all contact angles were obtained at the initial stage (i.e., <10 s). The water contact angle (CA) measurements of each sample were conducted at least three times across the sample surface. In order to detect the localization of not-reacted TC in the blends, knowledge of thermodynamic equilibrium facilitates determination of the final location of free TC since the time of melt mixing was sufficiently long for TC immigration. A set of test liquids with determined dispersive and polar surface energy (γ^d and γ^p) were used for determination of the surface free energy of every discrete component of samples. Relation of total free surface energy of solid surface to polar and dispersive surface energy is based on eqn (7) and by knowing the contact angles of a set of liquid (with determined γ^d and γ^p) on a solid surface, γ^d and γ^p for the surface can be obtained according to eqn (8).

γ_s = γ^p_s + γ^d_s (7)

where s and l represent the solid and liquid surfaces, respectively, and θ is the contact angle of a liquid droplet on the surface. The interfacial energy of every two components can be obtained from their dispersive and polar parts of surface energies using different methods like harmonic-mean and geometric-mean equations.^30–34

In these equations, γ_i is the surface energy of component i, γ₁₂ is the interfacial energy of components 1 and 2, and d and p superscripts stand for dispersive and polar parts of surface free energy of the components, respectively. Wetting coefficient, ω_a, has been widely used to predict the thermodynamic equilibrium distribution of nanofillers in the polymer blends^30,31 and is defined as follows:

3. Results and discussion

3.1. FTIR spectra

The FTIR spectra of the PLASCL5 and PLASCL25 as representatives are depicted in Fig. 1. The C=O stretching vibrations and vibrations of C–O bonds in ester groups (PLA and PCL) show peaks at 1749, 1182 cm⁻¹ and 1725, 1177 cm⁻¹ for PLASCL5 and PLASCL25, respectively. Upon addition of PCL, the intensity of these peaks is increased. In addition, the peaks at 1365 and 1453 cm⁻¹ in both blends are indicative of the CH₃ and CH₂ groups of PLLA and PCL respectively, implying that the PCL is physically dispersed within the PLA to form a partially miscible blend. All the characteristic peaks of PLA and PCL, appeared in Fig. 1, reveal that the main structure of PLA and PCL has not been altered during the blending procedure.

Fig. 1 FTIR spectra of PLASCL5 and PLASCL25.

The interesting observation in FTIR results is the occurrence of a chemical reaction between PCL and starch in the presence of LATC30 which consists of three Cl groups in each molecule. According to Kweon et al.¹⁷ and the other reports on starch inclusion complexes,^18,19 the Cl group can be attached to the starch and form the chlorinated starch, and then the Cl group can be easily replaced by PCL. The C–Cl stretching bond shows a peak at 756 cm⁻¹ for PLASCL5. The intensity of this peak weakens as the PCL content is increased in the blend, which indicates that the Cl group was attached to the starch and reacted with PCL generating a grafted phase of starch/PCL (SCL). Since the C–Cl peak intensification shows a decrease, one can infer that a portion of Cl has been released during the blending. The peak at 1129 cm⁻¹ belongs to C–OH of starch, which can be clearly observed in PLASCL5 with a higher content of starch, however, with an increase in the amount of PCL (accompanied by a decrease in starch concentration) in the case of PLASCL25, this peak almost disappeared. On the other hand, a peak at 1420 cm⁻¹ clearly confirms the reaction between starch and PCL by showing CH₂ bending. The peaks which appeared in 1082 and 1084 cm⁻¹ belong to C–O–C stretching vibrations of starch, and upon the increase in PCL content, the intensity of these peaks is decreased representing the ether linkage contribution in the chemical reaction between starch and PCL. A broad O–H stretching absorbance at 3200–3700 cm⁻¹ is slightly increased upon using higher PCL content.²⁰ The CH stretching and C=C ring-related vibrations occur at ∼2900 cm⁻¹, and they typically exhibit multiplicity of bonds which show the existence of TC. Moreover, with an increase in PCL content, these multiple peaks are intensified implying the attachment of TC to the starch and PCL. The results clearly state that the performed blending has been reactive, and as a result, PCL–starch–TC was physically dispersed in PLLA.

3.2. Blend morphology and image analysis (AFM results)

The AFM micrographs of PLASCL5 and PLASCL25 were shown in the 3D images of Fig. 2a and b, respectively. As depicted in Fig. 2a, there can be found globular starch particles distributed within the PLA matrix. These starch particles are more rigid and stiffer than PLA due to the generation of brighter domains in the AFM image,³⁵ which further reveals separately dispersed particles of starch–PCL in the matrix. This sea-island morphology is stable even by changing the concentration of PCL, but the size of starch particles became smaller when PCL concentration increased to 25 wt%. Moreover, the voltage of tapping was reduced when the PCL concentration increased from 5 to 25 wt% meaning a reduction in the starch elastic modulus. The reduced modulus can be due to the plasticization effect of PCL on the starch phase.¹⁷ These results are in accordance with the FTIR results since the chlorinated globular starch (by TC) was chemically reacted and surrounded with PCL. It is reasonable to deduce that the thickness of the surrounding PCL increases, but the diameter of dispersed starch particles decreases. Hence, the sensible modulus of dispersed SCL (starch–PCL) phase decreased.

Fig. 2 3D visual morphologies of (a) PLASCL5 and (b) PLASCL25 and 2D grayscale morphologies of (c) PLASCL5 and (d) PLASCL25 obtained using AFM.

To better visualize the morphologies of these blends, the AFM grayscale 2D up-view of PLASCL5 and PLASCL25 microscopies are also depicted in the Fig. 2c and d. These 2D schemes clearly show the contrast of the PLA and SCL phases. In these ternary blends, the attached PCL to starch could act as a plasticizer for the starch phase, and thus, the dispersed starch granules are partially plasticized, but cannot completely flow in the melt state similar to thermoplastic materials.¹⁰ For the PLASCL25 blend, the effect of the microtome knife is rather evident (unclear borders of SCL phases) because of the presence of slight deformation of SCL phase. This result strongly supports the notion that the PLASCL25 is more ductile than PLASCL5.¹⁰

3.3. Thermal properties

To investigate the thermal behavior of PLASCL blends, DSC measurements were carried out. DSC curves of pure PLA and PCL, PLA/starch 50:50 blend (PLAS50) and PLASCL blends are presented in Fig. 3.

Fig. 3 DSC thermograms of PLASCL blends and their corresponding counterparts.

Thermal properties of these materials, such as T_g, T_m, melting enthalpy (ΔH_m) and (X_c) are reported in Table 1. As is observed, the addition of starch to PLA decreased T_g due to the enhanced mobility of PLA chains.⁸ In the PLASCL5 blend, the T_g of PLA could not be detected since it overlaps with T_m of PCL. It can be deduced that the T_g of PLA phase in the blends is increased by the addition of PCL, which is because of the plasticization effect of this material for the starch phase. In fact, the PCL plasticization effect causes the more flow-ability of starch during melt-blending inducing a higher level of reactions between starch hydroxyl groups and Cl groups in TC followed by reaction of these chlorinated sites with PCL. Altogether, the addition of more PCL causes a higher level of chemical reaction in the produced blend which causes finer and more dispersed SCL phase possessing physical interactions with the PLA phase (discussed in rheometry section). Hence, finer SCL particles lead to the creation of more interface with PLA, and thus, more interactions occur which might restrict the chain movements.⁵ Therefore, the T_g of PLA phase increases by the addition of more PCL. T_m and X_c in PLAS50 blend were remarkably lower than those of the neat PLA. This is probably due to the impending role of starch on the movement of polymer segments causing the polymer chain arrangement more difficult. Another reason might be the poor adhesion between hydrophilic starch and hydrophobic PLA leading to a steric effect.¹⁶ The increase of crystallinity when PCL is the dispersed phase within the PLA matrix was also reported in previous studies,^36,37 since the interface of such system can serve as a nucleating agent during the crystallization stage of PLA.³⁸ However, as the PCL content was increased in ternary blends, SCL disrupted the regularity of the chain structures, and thus, free space between chains was increased, and as a result, T_m and X_c were reduced.²⁰ The exothermal peaks right before T_m of PLA phase can be attributed to the recrystallization of lower perfection crystals of PLA into α crystals of higher perfection.^39,40 Reduction of melting temperature and also crystallinity of PCL in the blends respect to the neat PCL can be attributed to the process in which the thermal history of the blends was removed by heating of samples up to 120 °C; as a result, PCL phase in blends was melted, while the neat PCL was heated up to 40 °C in order to remove thermal history, and in this case, PCL was not melted. Therefore, due to the slow rate of crystallization in PCL,⁴¹ its chains had not enough time to form ordered chains. To make sure the occurrence of a chemical reaction between PCL and starch through the triclosan Cl group which was previously seen by FTIR and AFM results occurred, DSC analysis was also performed for PCL/starch 50:50 with triclosan (PCS50).

Table 1 Thermal and mechanical properties of PLASCL blends

		PCL	PLA	PCS50	PLAS50	PLASCL5	PLASCL10	PLASCL15	PLASCL20	PLASCL25
a By considering the melting enthalpy of 100% crystalline PCL139.3 J g^−1  ^7 and PLLA 93 J g^−1.^24
Thermal properties	Tg (°C) (PCL)	−62.7	—	—	—	—	—	—	—	—
	Tm (°C) (PCL)	65.9	—	51.8	—	53.2	52.7	52.4	53.7	53.7
	Tg (°C) (PLLA)	—	57.9	—	47.0	—	48.3	48.3	50.1	50.9
	Tm (°C) (PLLA)	—	164.5	—	146.4	164.6	163.1	162.1	163.0	161.8
	ΔHm (J g^−1) (PCL)	87.6	—	67.2	—	4.0	7.7	11.2	14.8	18.3
	Xt^a (%) (PCL)	62.9	—	48.3	—	57.4	55.2	53.6	53.1	52.5
	ΔHm (J g^−1) (PLLA)	—	38.2	—	8.4	23.7	21.9	21.8	21.5	21.3
	Xt^a (%) (PLLA)	—	41.1	—	9.0	50.9	47.1	46.8	46.2	46.5
Mechanical properties	Elastic modulus (MPa)	469	3770	—	1920	1145	657	397	326	229
	Tensile strength (MPa)	14.4	66.2	—	11.1	18.9	20.9	23.2	24.9	26
	Elongation at break (%)	229	2.7	—	2	4.3	11.8	19.3	25.1	20.2
	Impact (kJ m^−2)	—	—	—	—	21.07 ± 0.45	22.08 ± 0.38	23.04 ± 0.51	26.82 ± 0.59	28.79 ± 0.46
	Hardness (Shore D)	—	—	—	—	71.4 ± 0.89	68.6 ± 0.55	63.6 ± 0.55	61 ± 1	59.4 ± 0.86

---

Upon blending with PCL, starch and LATC30, PCL melting temperature peak shifted to lower temperatures. Moreover, a weak peak appeared around 154 °C which was also reported by Kweon et al. They described this peak as a weak glass transition peak of starch attached in PCL, due to the chemical reaction between triclosan-starch Cl group and the ester group in PCL.¹⁷ According to ternary blend DSC thermograms, this weak transition overlaps with PLA melting temperature and consequently was not detected distinctly. Therefore, the lower T_m, ΔH_m and X_c of PCS50 comparing to PCL, and also, lower amounts of the mentioned parameters in the blends can be attributed to the chemical reactions between the chloride group in starch-Cl and ester group in PCL, and the reciprocal action between the crystalline region of PCL and amorphous region of starch indicating the high compatibility between these two polymers.¹⁷

3.4. Mechanical properties

Stress-strain curves of neat PLA and PCL, PLAS50 and PLASCL ternary blends are presented in Fig. 4 and the related data are summarized in Table 1. It is shown that by the addition of starch into PLA, elongation at break was decreased accompanied by a remarkable reduction in the elastic modulus and tensile strength. The fairly weak mechanical properties of this blend may be explained in terms of lack of affinity between the two phases.^13,14,42,43

Fig. 4 Stress vs. strain of PLASCL blends.

By adding PCL to PLAS50, as discussed above, PCL is grafted to the starch's chlorinated sites which were generated by the reaction of –Cl groups of TC and –OH groups of starch. By reaction of PCL with starch, compatibility of this blend was enhanced, and as a result, the tensile strength was improved about 134% for PLASCL25, and elongation at break showed an 11.5-fold increase comparing to PLAS50. As the PCL content increased in the ternary blends, the chemical reaction between PCL and starch increased and the elongation at break was further improved. Generally, in starch containing blends, the level of dispersion and diameter of the dispersed starch are the two fundamental factors affecting the mechanical properties of the blend.¹⁶ Therefore, upon interface modification of hydrophilic starch particles with the hydrophobic PCL, the diameter of SPCL domains became smaller, and the level of dispersion was improved, and thus, a compatibilized blend with better interfacial adhesion between the blend components was obtained. Upon grafting of PCL to the blend, the brittle behavior of PLA/starch blend changed into a ductile behavior and toughness came into play. SPCL acts as a compatibilizer and not only improves the level of dispersion of starch within the matrix, but also promotes the adhesion between the fillers and the matrix.¹ The Charpy impact strength of the blends is represented in Table 1. As can be seen, with increasing the PCL concentration from 5 to 25 wt% in the matrix, the Charpy impact strength is enhanced. Actually, the increased impact strength is the result of PCL aggregation around the starch domains, and the consequent grafting generates the SCL dispersed phase within the PLA matrix. Enhancement of Charpy impact strength can also be considered as a sign of toughness improvement.⁴⁰ As is shown in Table 1, it is obvious that the hardness values of samples were decreased by increasing the PCL concentration in the ternary blend which can be attributed to the finer dispersion of SCL phase with lower hardness. In fact, these fine SCL particles are more dispersed having a higher PCL content residing on the shell of SCL particles in the matrix which leads to a softer blend. According to the attained mechanical properties, one can infer that PLASCL20 shows excellent mechanical properties alongside a ductile behavior making it an appropriate candidate for a wide variety of applications.

3.5. Rheological investigation

The rheological properties of PLASCL blends are shown in Fig. 5. The strain sweep (Fig. 5a) was carried out to confirm that the applied strain did not exceed the limit of linear viscoelasticity. This test was performed at 170 °C, which is similar to the blending temperature, within the strain range of 0.01 to 100.

Fig. 5 Rheological properties of PLASCL blends (a) G′ and G′′ vs. strain (b) G′ and G′′ vs. frequency (c) complex viscosity vs. frequency (d) time sweep.

Eventually, the strain of 0.1% was opted for the consequent frequency sweep tests. As can be seen, all the blends show a more profound elastic behavior since the storage modulus values are higher than the corresponding loss modulus values. Moreover, with the increase in PCL content, due to its lower modulus as compared to PLLA and starch, both storage and loss modulus are decreased. The dynamic frequency sweep measurements were also carried out to investigate the dynamics of samples as shown in Fig. 5b and c. Unlike the strain sweep test, these results showed a different trend. Upon increasing the PCL content (decrease in starch content), loss modulus was decreased while the storage modulus and complex viscosity showed an unexpected increasing trend. According to the FTIR and AFM results, the chemical reaction between PCL, starch and TC made the starch particles finer and more physically dispersed in the PLLA matrix, and the reaction between PCL and starch increased, making the strongly dispersed SCL phase. It can be postulated that this phenomenon could be the reason for the unexpected increased elasticity upon the addition of more PCL in the blends. The intersection point shows a U-trend with the increase of PCL, or in other words, with increasing the PCL inclusion up to 15%, the intersection frequency decreases from 18 to 11 Hz, and finally, it increases to 194 Hz for PLASCL25, showing that the abovementioned structure in the blends has made them partially compatible.³⁵ The time sweep test was carried out for PLASCL5 and PLASCL25 as representatives (Fig. 5d) to ensure that there is no change in overall properties with time. While PLASCL 25 shows an utterly stable behavior during the understudied period of time, an increase in rheological parameters of PLASCL5 was observed. This behavior suggests that PCL addition in the presence of TC and starch can make the blends rheologically stable during the period of 30 minutes. The relaxation spectra of ternary blends were obtained using the plots of G′ and G′′ data versus the angular frequency at the reference temperature. From the relaxation spectra, the zero shear viscosity (η₀ = ∑λ_iG_i), plateau modulus (G⁰_N = ∑G_i), mean relaxation time and entanglement density (ν_e = ρ_a/M_e, M_e = ρRT/G⁰_N) of the blends were calculated.⁴⁴ Table 2 shows the obtained parameters from the relaxation spectra. ν_e, the entanglement density, is defined as the number of entanglement junctions per unit volume. R denotes the ideal gas constant, T represents the temperature, ρ is the melt density and ρ_a is the amorphous density. ρ and ρ_a for PLLA are 1.248 and 1.290 g cm⁻³, for starch are 1.5 and 1.498 g cm^−3 44 and for PCL are 1.2 and 1.02 g cm⁻³,⁴⁵ respectively. For the PLASCL blends, ρ and ρ_a are calculated based on the mixture law and reported in Table 2. The zero shear viscosity and plateau modulus show larger values upon the addition of PCL. The calculated entanglement density also exhibits the same trend, and PLASCL25 has the highest level of entanglement among the samples indicating that PCL/starch chemical reaction and fine particles may have physical interactions with PLLA matrix.

Table 2 Viscoelastic parameters of blends

Sample	η0^a (Pa s) × 10^−4	G^0N^a (Pa) × 10^−4	[small lambda, Greek, macron]^a (s)	νe^a (g mol cm^−3)	ρ^b (g cm^−3)	ρa^b (g cm^−3)
a Obtained from rheology.b Calculated using mixture law.
PLASCL5	3.1	3.91	3.51	10.7	1.36	1.37
PLASCL10	5.3	5.04	3.78	13.7	1.34	1.35
PLASCL15	8.5	5.61	3.88	14.36	1.33	1.32
PLASCL20	11	6.06	3.91	16.2	1.31	1.3
PLASCL25	22.5	7.56	4.17	20.1	1.3	1.27

---

The rheometry results are clearly in conformity with the FTIR and AFM observations. According to Polaczek et al.,¹⁸ starch granules exhibited nanometric needles on their surface resembling hedgehogs. These needles were formed by the content of granule the interior expelled by the water removal due to the high temperature in melting process and deposited on the granule surface and this process was responsible for the formation of empty domains inside the granules which provide space for Cl groups and PCL to interact with the starch. The mean relaxation time, which is the ratio of the elastic to viscous response, also shows an increasing trend due to the physical and chemical interactions, although a small increase upon addition of PCL is observed. While other similar studies show that the PLA, starch and PCL are physically mixed, use of TC antibacterial agent has made the blending to a reactive process in which all the properties are improved while the antibacterial effect is seen as well.

3.6. Antibacterial activity, cell viability and in vitro degradation of PLASCL blends

In order to investigate the antibacterial activity of TC in PLASCL blends, PLASCL20 sample was placed on the seeded agar plates to measure the effectiveness of an antibacterial agent by determining the zone of inhibition around samples.²¹ Fig. 6a and b show the inhibitory zone formation around the PLASCL20 samples containing TC in agar plates of E. coli (2 mm) and S. aureus (4.5 mm). As can be seen, the inhibitory zone against S. aureus was larger than the zone against E. coli, which confirms while TC has in vitro activity against a broad range of both Gram-negative and Gram-positive bacteria, it also has greater activity against Gram-positive species.^21,46 TC, reduces the bacterial growth and polymer adherences directly from the polymer surface with minimum antimicrobial release. Also, secondary bonding between bacteria and the polymer is considered to be interrupted by TC vibrational fluctuating molecular bond rotations as a possible mechanism to prevent microbial surface attachments.⁴⁷ The in vitro biocompatibility, the effect of PCL/starch content change on the release of TC and the final effect on osteoblast growth were determined within three and seven days after seeding, using the MTT assay. As shown in Fig. 6c, with the increase of PCL, the cell numbers are decreased, due to the weaker attachment of the cells on the surface of the blends. As previously mentioned, with the increase in PCL content, TC can facilitate the attachment of starch and PCL, which means that even though blends show antibacterial effect due to their phenol group, the negative effect of TC²¹ on the cells is decreased. However, as the surface shows more hydrophobic behavior, the cell attachment weakens, and therefore, with increasing the PCL content the MTT assay results show a slight reduction comparing to PLASCL5 which contains more hydrophilic starch and TC. For in vitro degradation investigation, the ternary blends were immersed in phosphate buffer at pH of 7.4 and temperature of 37 °C. Normally, the estimated degradation time for PLA is between 18 and 22 months.^2,21 As PLA has a semi-crystalline nature, diffusion of buffer or any other liquids into its bulk is quite difficult, however, PLAS50 blend shows degradation in 10 months, showing that upon addition of highly hydrophilic starch, PLA degrades much faster. Fig. 6d shows the in vitro degradation of PLASCL blends. PLASCL blends exhibit lower degradation times comparing to PLA, but with increasing the PCL content, due to its hydrophobic nature and also its modifying effect on starch, the degradation time shows an increasing trend, in which for PLASCL5 and PLASCL25, 12 and 16 month degradation time can be seen, respectively.

Fig. 6 PLASCL blends (a) antibacterial activity against E. coli (b) antibacterial activity against S. aureus (c) MTT assay after 3 and 7 days (d) hydrolytic degradation at 37 °C (e) accumulative release profile in 6 months (f) contact angle.

3.7. Drug release rate studies and kinetics

According to the triclosan calibration curve in our previous work,²⁴ the drug release of the PLASCL blends at 37 °C in buffer solution are shown in Fig. 6e. The blends show a burst release at the first 24 h which overall is less than 1 μg ml⁻¹, and is in accordance with the previous studies that used LATC30.^21,24 The amount of this release is decreased by increasing the PCL content. The amount of release for the samples with different PCL contents is 0.16, 0.14, 0.12, 0.1 and 0.09 μg ml⁻¹ every day. The minimum inhibitory concentration (MIC) of triclosan is between 0.025 and 1 μg ml⁻¹ for medical and packaging applications and the obtained results are acceptable.⁴⁸ The drug release was monitored up to six months in order to observe the mechanism of the drug release for long term treatments. It needs to be mentioned that a continuous release during the application is desirable, therefore, the degradation time and release time need to show similar profiles. Based on the results, PLASCL5 releases the total drug in nearly 10 months, while shows a degradation time of 12 months which means it releases the drug prior to its complete degradation.

On the other hand, PLASCL25 releases the drug in 17 months and its degradation time is 16 months, therefore, it releases a large amount of the drug at the time of degradation due to the higher load of the drug, and in both cases, the results are not desirable. Based on the obtained results, only PLASCL20 releases the drug during its degradation period. For medical applications, a slower release profile is preferable. To investigate the release mechanism of triclosan, the accumulative release profiles for PLASCL blends were analyzed by the Peppas–Korsmeyer equation, zero-order and first-order, Higuchi, and Hixon–Crowell models. Table 3 illustrates the regression coefficients (R²) as well as their equation constants calculated and fitted with the obtained data. As observed in Table 3, according to the Peppas–Korsmeyer model, the drug transport mechanism with the increase in PCL content changes from non-fickian diffusion or anomalous transport to fickian diffusion. PLASCL25 and then PLASCL20 after approximation by rounding both show a fickian diffusion. According to their model, only the release curves, where is less than 0.6, should be used and ‘n’ values ca. 0.5 may be used for the fickian diffusion release of non-well-known pharmaceutical polymeric dosage.²⁴

Table 3 Regression coefficients and constants of models fitted to the release of triclosan in blends

Samples	Peppas–Korsmeyer		Zero order		First order		Higuchi		Hixon–Crowell
	R^2	n	K0	R^2	K1	R^2	KH	R^2	Ks	R^2
PLASCL5	0.9561	0.63	0.0064	0.9987	0.001	0.6835	0.3922	0.9443	−0.0005	0.8893
PLASCL10	0.9611	0.61	0.0057	0.9985	0.001	0.7093	0.3501	0.9390	−0.0005	0.8987
PLASCL15	0.9603	0.60	0.0047	0.9973	0.001	0.7269	0.2905	0.9301	−0.0005	0.9075
PLASCL20	0.9589	0.56	0.0045	0.9986	0.001	0.7331	0.2905	0.9424	−0.0005	0.9063
PLASCL25	0.9610	0.55	0.0039	0.9974	0.0009	0.7722	0.2366	0.9282	−0.0005	0.9016

---

The regression coefficients for different blends showed a great fitness with the zero-order model, which describes the release of a low-soluble drug into the release environment. Due to the absorption of triclosan in phosphate buffer solution,⁴⁹ all the samples were fitted with the Higuchi model showing a relatively high regression coefficient, which is mainly due to the drug release by diffusion since drug diffusion is faster than the matrix degradation. Because of nearly long-time degradation of PLASCL blends, the release mechanism could not be well fitted by the first order and Hixon–Crowell models. According to the release studies, the zero order, Peppas–Korsmeyer and Higuchi models show higher regression coefficients, indicating that these models are the best to describe the blends release behavior. All in all, based on the release profiles and matrix hydrolytic degradation, PLASCL20 is recommended as the best candidate with the optimum results.

3.8. Contact angle

As depicted in Fig. 6f, the water contact angle of the blends has an increasing trend by enhancing PCL concentration. PLASCL25 with high PCL concentration results in much more hydrophobicity as compared to PLASCL5. This behavior is originated from the structure of polymers since PCL has less concentration of hydrophilic groups (ester group) due to the more repeatedly hydrophobic CH₂ part in the backbone as compared to PLA. On the other hand, the starch structure contains many hydrophilic hydroxyl groups making it more hydrophilic in comparison with PLA and PCL. PCL gathers and reacts with the functional groups residing on the outside of starch dispersed domains, so as a result, PLASCL5 shows a lower water CA compared with PLASCL samples with higher PCL as represented in Fig. 6f. The AFM schemes obviously showed that all of the samples exhibit typical ‘sea-island’ morphologies, where discrete spherical domains consist of starch–PCL uniformly dispersed in the PLA matrix, which was also confirmed by the FTIR, rheometry and other test results. As discussed in the previous sections, some of the TC reacted with starch and created a suitable site for the reaction of PCL. Therefore, PLA is single phase, and for simplification, the starch-reacted PCL is considered the second phase named as SCL. Deionized water and diiodomethane were selected as the test liquids for measuring contact angles. γ^d for deionized water and diiodomethane was reported to be 22.1 and 44.1 mN m⁻¹, respectively, and γ^p for the mentioned solvents was determined 50.7 and 6.7 mN m⁻¹ in the same order.³⁰ Water and diiodomethane contact angles of PLA, triclosan (TC), SCL5 and SCL25 were measured three times and the mean of contact angles was listed in Table 4. By knowing contact angles for each surface and using eqn (7) and (8), dispersive part, polar part and the total surface free energies were determined for every component^32,33,50 and listed in Table 4.

Table 4 The measured contact angles (at 25 °C) and calculated surface energies of sample components

Sample	Contact angle (°)		Surface energy (mN m^−1)
	Water	Diiodomethane	γ	γ^d	γ^p
PLA	72.5	55.5	34.11	20.95	13.16
Triclosan	78	24	47.02	43.75	3.27
SCL5	59	12.5	51.14	37.38	13.76
SCL25	61	12	50.7	38.49	12.21

---

The calculated interfacial energies of component couples based on harmonic and geometric mean equations are summarized in Table 5.

Table 5 The calculated values of component couple surface energies and sample wetting coefficients

Component couple	γ12^a (mN m^−1)	γ12^b (mN m^−1)
a Based on the harmonic mean equation.b Based on the geometric mean equation.
PLA/triclosan	13.96	7.43
PLA/SCL5	4.63	2.36
PLA/SCL25	5.20	2.65
SCL5-triclosan	6.94	3.84
SCL25-triclosan	5.48	2.99
ω^PLASCL5a	1.52	1.52
ω^PLASCL25a	1.63	1.68

---

The wetting coefficient is mostly used to predict the thermodynamic stable phase of nanofillers in polymer blends,^30,31 and in this work, TC consists of nanoparticles as the same size of usual nanofillers. Based on the definition of ω_a, for ω_a > 1, TC is only found in SCL phase; for ω_a < −1, TC is present only in the PLA and for −1 < ω_a < 1, TC is located at the interface of phases.³⁰

ω^PLASCL5_a and ω^PLASCL25_a are the wetting coefficients of PLASCL5 and PLASCL25, respectively, which are further clarified based on the mother eqn (11). All of the ω_a for both samples using the two calculation methods are larger than unity (the values from 1.52 to 1.63), therefore, TC was thermodynamically driven into the starch grafted PCL or SCL phase (like SCL5 or SCL25) during the melt mixing process. Altogether, contact angle results indicated that the thermodynamically stable place for TC localization is the SCL phase, which is beneficial for TC releasing i.e. slowing down drug releasing by entrapping it in SCL phase which is further distributed in PLA matrix. This conclusion was previously confirmed by UV-visible release, which showed that the samples with more PCL show lower release rate due to the more reactions between starch, PCL and the entrapped TC.

4. Conclusion

This study revealed that high price of PLA could be reduced by the addition of inexpensive corn starch, and thereby, the resulted brittle behavior could be transformed into ductile by incorporating PCL via melt blending method. In this work, triclosan antibacterial agent was introduced to the system for imparting antibacterial properties which also caused surface modification of hydrophilic starch by hydrophobic PCL through chemical reactions and further compatibilizing the PLA–starch–PCL ternary blends. This compatibilization method was confirmed by FTIR measurements, DSC analysis and AFM studies. PLA–starch–PCL blend exhibited a sea-island morphology in which the size of dispersed starch–PCL became smaller by the increase of the ratio of PCL to starch. Smaller starch–PCL particles have created more interface with PLA leading to more interactions, which might restrict the chain movements, and as a result, the T_g of PLA phase increased by the addition of higher content of PCL. It was also found that as PCL content increases in the ternary blends, starch–PCL could disrupt the regularity of the chain structures, and thus, free space between the chains may become larger, and as a result, T_m and X_c were reduced. Moreover, the antibacterial studies proved the antibacterial properties of the prepared blends against Gram-positive and Gram-negative bacteria, and the MTT assay was performed as a proof of cell viability of the blends. Contact angle results indicated that the thermodynamically stable place for triclosan localization is the starch–PCL phase, which slows down the drug releasing by triclosan entrapping in small dispersed starch–PCL phase placed in the PLA matrix. This conclusion was also confirmed by UV-visible release data, which showed that the release rate in the samples with more PCL was reduced due to the entrapped triclosan in starch–PCL dispersed phase (because of more reactions). Rheological properties as well as calculated entanglement density show higher values upon addition of PCL and increase the entanglement indicating that PCL/starch chemical reaction and fine particles may have physical interaction with the PLLA matrix. The rheometry results clearly show the reason behind the FTIR and AFM observations. While other similar studies showed that the PLA, starch and PCL are physically mixed, use of the TC antibacterial agent has made the blending a reactive process, in which all the properties are improved, while the antibacterial effect is seen. Significant improvements in the impact strength, hardness and elongation at break of the samples were also achieved. A substantial increase in the notched Charpy impact energy is also observed with blends demonstrating high energy dissipation potential of the blends. Elongation at break for PLA/starch/PCL 50/30/20 wt% reached 25.1%, as compared to 2.7% for the pure PLA. The reported results reveal that the blends, especially PLA/starch/PCL 50/30/20 wt% became ductile, compatibilized, cell-viability approved, antibacterial while their drug release rate and degradation rate are fully desirable which can be used in a wide range of applications.

References

1. L. Chen, X. Qiu, Z. Xie, Z. Hong, J. Sun, X. Chen and X. Jing, Carbohydr. Polym., 2006, 65, 75–80.
2. S. M. Davachi and B. Kaffashi, Polym.-Plast. Technol. Eng., 2015, 54, 944–967.
3. R. G. Sinclair, J. Macromol. Sci., Part A: Pure Appl.Chem., 1996, 33, 585–597.
4. E. Oliaei, B. Kaffashi and S. Davoodi, J. Appl. Polym. Sci., 2016, 133, 43104.
5. C. L. Jun, J. Polym. Environ., 2000, 8, 33–37.
6. B. Torabinejad, J. Mohammadi-Rovshandeh, S. M. Davachi and A. Zamanian, Mater. Sci. Eng., C, 2014, 42, 199–210.
7. M. F. Koenig and S. J. Huang, Polymer, 1995, 36, 1877–1882.
8. J. Mohammadi-Rovshandeh, P. Pouresmaeel-Selakjani, S. M. Davachi, B. Kaffashi, A. Hassani and A. Bahmeyi, J. Appl. Polym. Sci., 2014, 131, 41095.
9. L. Avérous, J. Macromol. Sci., Part C: Polym. Rev., 2004, 44, 231–274.
10. P. Sarazin, G. Li, W. J. Orts and B. D. Favis, Polymer, 2008, 49, 599–609.
11. G. Biresaw and C. J. Carriere, J. Polym. Sci., Part B: Polym. Phys., 2001, 39, 920–930.
12. W. Y. Jang, B. Y. Shin, T. J. Lee and R. Narayan, J. Ind. Eng. Chem., 2007, 13, 457.
13. O. Martin and L. Averous, Polymer, 2001, 42, 6209–6219.
14. N. Wang, J. Yu, P. R. Chang and X. Ma, Carbohydr. Polym., 2008, 71, 109–118.
15. M. A. Huneault and H. Li, Polymer, 2007, 48, 270–280.
16. C.-S. Wu, Macromol. Biosci., 2005, 5, 352–361.
17. D.-K. Kweon, N. Kawasaki, A. Nakayama and S. Aiba, J. Appl. Polym. Sci., 2004, 92, 1716–1723.
18. E. Polaczek, F. Starzyk, K. Maleńki and P. Tomasik, Carbohydr. Polym., 2000, 43, 291–297.
19. U. V. Lay Ma, J. D. Floros and G. R. Ziegler, Carbohydr. Polym., 2011, 83, 1869–1878.
20. H.-T. Liao and C.-S. Wu, Mater. Sci. Eng., A, 2009, 515, 207–214.
21. S. M. Davachi, B. Kaffashi, A. Zamanian, B. Torabinejad and Z. Ziaeirad, Mater. Sci. Eng., C, 2016, 58, 294–309.
22. I. Makarovsky, Y. Boguslavsky, M. Alesker, J. Lellouche, E. Banin and J.-P. Lellouche, Adv. Funct. Mater., 2011, 21, 4295–4304.
23. S. M. Davachi, B. Kaffashi, B. Torabinejad, A. Zamanian, J. Seyfi and I. Hejazi, Polymer, 2016, 90, 232–241.
24. S. M. Davachi and B. Kaffashi, Int. J. Polym. Mater. Polym. Biomater., 2015, 64, 497–508.
25. S. M. Davachi, B. Kaffashi and J. M. Roushandeh, Polym. Adv. Technol., 2012, 23, 565–573.
26. S. M. Davachi, B. Kaffashi, J. M. Roushandeh and B. Torabinejad, Mater. Sci. Eng., C, 2012, 32, 98–104.
27. N. López-Rodríguez, A. López-Arraiza, E. Meaurio and J. R. Sarasua, Polym. Eng. Sci., 2006, 46, 1299–1308.
28. S. M. Davachi, B. Kaffashi, B. Torabinejad and A. Zamanian, Polymer, 2016, 83, 101–110.
29. P. Costa and J. M. Sousa Lobo, Eur. J. Pharm. Sci., 2001, 13, 123–133.
30. H. Xiu, H. W. Bai, C. M. Huang, C. L. Xu, X. Y. Li and Q. Fu, eXPRESS Polym. Lett., 2013, 7, 261–271.
31. D. Wu, Y. Sun, D. Lin, W. Zhou, M. Zhang and L. Yuan, Macromol. Chem. Phys., 2011, 212, 1700–1709.
32. V. Karde and C. Ghoroi, Int. J. Pharm., 2014, 475, 351–363.
33. E. Schwach and L. Avérous, Polym. Int., 2004, 53, 2115–2124.
34. A. Popelka, I. Novák, M. Lehocký, I. Chodák, J. Sedliačik, M. Gajtanska, M. Sedliačiková, A. Vesel, I. Junkar and A. Kleinová, Molecules, 2012, 17, 762–785.
35. B. Imre, D. Bedő, A. Domján, P. Schön, G. J. Vancso and B. Pukánszky, Eur. Polym. J., 2013, 49, 3104–3113.
36. N. Noroozi, L. L. Schafer and S. G. Hatzikiriakos, Polym. Eng. Sci., 2012, 52, 2348–2359.
37. K. Chavalitpanya and S. Phattanarudee, Energy Procedia, 2013, 34, 542–548.
38. H. Kusanagi, Y. Chatani and H. Tadokoro, Polym. J., 1997, 29, 65–68.
39. J. Zhang, Y. Duan, H. Sato, H. Tsuji, I. Noda, S. Yan and Y. Ozaki, Macromolecules, 2005, 38, 8012–8021.
40. F. Feng and L. Ye, J. Appl. Polym. Sci., 2011, 119, 2778–2783.
41. M. J. Jenkins and K. L. Harrison, Polym. Adv. Technol., 2006, 17, 474–478.
42. T. Ke, S. X. Sun and P. Seib, J. Appl. Polym. Sci., 2003, 89, 3639–3646.
43. J. Ren, H. Fu, T. Ren and W. Yuan, Carbohydr. Polym., 2009, 77, 576–582.
44. S. M. Davachi, S. Bakhtiari, P. Pouresmaeel-Selakjani, J. Mohammadi-Rovshandeh, B. Kaffashi, S. Davoodi and A. Yousefi, Adv. Polym. Technol., 2015 DOI:10.1002/adv.21634.
45. D. Homminga, B. Goderis, I. Dolbnya and G. Groeninckx, Polymer, 2006, 47, 1620–1629.
46. R. D. Jones, H. B. Jampani, J. L. Newman and A. S. Lee, Am. J. Infect. Control, 2000, 28, 184–196.
47. R. C. Petersen, AIMS Mol. Sci., 2016, 3, 88–103.
48. E. Díaz, I. Sandonis, I. Puerto and I. Ibáñez, Polym. Eng. Sci., 2014, 54, 2571–2578.
49. M. T. E. Suller and A. D. Russell, J. Antimicrob. Chemother., 2000, 46, 11–18.
50. T. Białopiotrowicz, Food Hydrocolloids, 2003, 17, 141–147.

---