Properties and Antioxidant Capacity of Anchovy ( Engraulis encrasicholus ) By-Product Protein Films Containing Thyme Essential Oil

Due to the environmental impact of the excessive quantity of non-degradable waste materials, packaging studies have recently focused on developing biodegradable and/or edible packaging fi lms obtained from natural sources, such as by-products of agriculture, fi shing or livestock raising (1). Among these natural sources, proteins from fi sh processing by-products have been successfully utilised in the development of edible fi lms, as reported by various researchers (2–4). A small sea fi sh, anchovy (Engraulis encrasicholus) is nearly 12 cm long with the highest catch rate among the marine fi sh found in Turkey (5,6). Fish head, frame and viscera, as by-products, are generated by anchovy processing, and have been generally turned into by-products with low value (for example fi sh fl our powder to be added to animal feed) or directly disposed into the environment, which in turn causes pollution problems (6). Therefore, it is important that fi sh processing by-products are eff ectively evaluated for the prevention of environmental pollution and obtaining high value-added products (4).


Introduction
Due to the environmental impact of the excessive quantity of non-degradable waste materials, packaging studies have recently focused on developing biodegradable and/or edible packaging fi lms obtained from natural sources, such as by-products of agriculture, fi shing or livestock raising (1).Among these natural sources, proteins from fi sh processing by-products have been successfully utilised in the development of edible fi lms, as reported by various researchers (2)(3)(4).A small sea fi sh, anchovy (Engraulis encrasicholus) is nearly 12 cm long with the highest catch rate among the marine fi sh found in Turkey (5,6).Fish head, frame and viscera, as by-products, are generated by anchovy processing, and have been generally turned into by-products with low value (for example fi sh fl our powder to be added to animal feed) or directly dis-posed into the environment, which in turn causes pollution problems (6).Therefore, it is important that fi sh processing by-products are eff ectively evaluated for the prevention of environmental pollution and obtaining high value-added products (4).
Food products can be protected by edible fi lms and coatings against moisture migration or diff usion of the gases that have substantial role in food deterioration, such as O 2 or CO 2 .Moreover, fi lms and coatings can act as carriers for antioxidants, antimicrobials, nutrients, spices and herbs (1).However, due to their hydro philic character, fi sh protein fi lms are generally insuffi cient barriers to prevent water vapour migration, as reported by various researchers (2,3,7,8).Water vapour permeability (WVP) and antioxidant properties of the fi lms can be improved by adding active compounds such as essential oils (2,3,9,10).Among these essential oils, thyme essential oil is a key essential oil containing active compounds such as thymol and carvacrol with known antioxidant capacity (11).Various properties of chitosan fi lms containing thyme essential oil were investigated by Altiok et al. (12), and they indicated that oxygen transmission and WVP rate of the fi lm barely increased, while its mechanical properties decreased with the addition of thyme essential oil.Moreover, they observed an increase in the antioxidant capacity of fi lms with the addition of thyme essential oil.Pires et al. (2) studied the characteristics of biodegradable fi lms containing thyme essential oil and hake proteins.Their study showed that the addition of thyme essential oil to the fi lms reduced the WVP and increased the antioxidant activity.Kavoosi et al. (13) found that the addition of thyme essential oil to gelatin fi lms led to an important decrease in tensile strength (TS), an increase in water solubility, elongation at break (EAB) and WVP.In addition, they indicated that the fi lms containing thyme essential oil showed perfect antioxidant properties.Jouki et al. (14) found that thyme essential oil addition to quince seed mucilage fi lms increased the WVP and oxygen permeability (OP), whereas decreased the glass transition temperature.
As seen from the above results, the incorporation of thyme essential oil into chitosan, hake protein, gelatin or quince seed mucilage fi lms aff ected WVP, OP, water solubility, mechanical properties, glass transition temperature and antioxidant capacity.As far as we know, there is no data on the addition of thyme essential oil to anchovy by--product protein fi lms.Therefore, this work aims to evaluate the eff ect of thyme essential oil, at various volume fractions, on the properties and antioxidant capacity of the anchovy by-product protein fi lms.Thickness, solubility, transparency, WVP, OP, mechanical and other properties were determined by diff erential scanning calorimetry, scanning electron microscopy and Fourier-transform infrared spectroscopy.Ferric reducing antioxidant power (FRAP) and 2,2-diphenyl-1-picrylhydrazyl (DPPH) scavenging activity values of the anchovy by-product protein fi lms containing thyme essential oil were also assessed.

Anchovy by-product protein powder and essential oil preparation
Anchovy (Engraulis encrasicholus) processing by-products (head, frame and viscera) used as material in the fi lm production were obtained from SASTAS A.S. (Samsun, Turkey) and minced with an electronic meat mincer (model MEW 613; Mado Primus, Eilenburg, Germany).Anchovy by-product proteins were prepared, with some modifi cations, as reported by Zavareze et al. (4).Firstly, minced by-products were mixed with water at 1:2 ratio (by mass per volume) in a blender (Model SHB 3062; Sinbo, Istanbul, Turkey) for 1 min.For protein solubilisation, 2 M NaOH (pH=11.2) were added and the mixture was homogenised by continuous agitation with a magnetic stirrer (model HS15-03P; MTOPS, Yangju City, Kyunggido, Korea) for 1 h.The sample was centrifuged with a laboratory centrifuge (model NF 800R; Nuve, Ankara, Turkey) at 9000×g and 4 °C for 30 min in order to obtain three phases (soluble proteins, insoluble proteins and lipids).The supernatant (soluble proteins) was collected and acidifi ed with 2 M HCl (pH=5.35)for protein precipitation.Then, it was stirred for 5 min and centrifuged with a laboratory centrifuge again at 9000×g and 4 °C for 15 min.Finally, the precipitated proteins were oven-dried with forced air (model JSOF-50; JSR, Gongju City, Korea) at 35 °C for 24 h, milled and stored in glass jars at 4 °C until fi lm preparation.The protein mass fraction of anchovy by-product protein powder was 68.58 %, determined according to standard AOAC method 960.52 (15).Dried thyme (Thymus vulgaris L.) was obtained from a local market (Samsun, Turkey), powdered in a coff ee grinder (model 2909; Sinbo) and hydrodistilled for 3 h with 50 g of dry plant material and 500 mL of distilled water in a Clevenger-type device (Sesim Kimya Laboratuvar, Ankara, Turkey).Amber coloured vials were used to collect the extracted essential oil, which was dehydrated using anhydrous sodium sulphate, then stored in the dark at 4 °C until use.

Preparation of fi lms
The fi lms were prepared according to Limpan et al. (16), with some modifi cations.A mass of 4.0 g of anchovy by-product protein powder was mixed with 100 mL of distilled water, the solution pH was set to 11.5 using 5 M NaOH, and glycerol (Merck, Darmstadt, Germany) was then added at 40 % (by mass of anchovy by-product protein) as plasticiser.The solution was homogenised for 5 min at 11 600×g with an Ultraturrax homogeniser (model T25; IKA, Staufen, Germany).Aft erwards, it was gently stirred with a magnetic stirrer at 85 °C for 60 min to completely dissolve anchovy by-product protein powder, centrifuged at 2900×g for 10 min and fi ltered to eliminate undissolved residuals and cooled to (40±2) °C.Tween 80 (Merck) at 0.05 % (by volume) of essential oil was combined with the fi lm-forming solution to help the dissolution of the essential oil, then thyme essential oil was added at 0 (as control), 0.5, 1.0 and 1.5 % (by volume).These solutions were emulsifi ed for 3 min at 11 600×g with an Ultraturrax homogeniser.The fi lm-forming emulsion (50 g) was placed into acrylic plates 15 cm in diameter, and then oven-dried (model JSOF-50; JSR) for 24 h at 40 °C under air circulation.Samples of dried fi lms were taken out of the plates and conditioned to 54 % relative humidity (by using a solution of saturated magnesium nitrate) within a desiccator at room temperature for 3 days.All samples were prepared in triplicate.

Determination of physicochemical properties
A digital micrometer (model 3101-25A; Insize, Jiangsu, PR China) reading to the nearest 0.001 mm was used to determine the fi lm thickness.Values of each triplicate fi lm sample were measured at 10 diff erent locations, and then water vapour permeability, mechanical properties and transparency value of the fi lms were calculated based on the mean thickness value.
A method used by Gennadios et al. (17) was slightly modifi ed and applied to determine the fi lm solubility in water.In order to determine initial dry mass, precondi-tioned fi lm pieces (20 mm×20 mm) were left to dry for 24 h at 105 °C.Following the agitation for 24 h in 50 mL of distilled water with sodium azide (0.1 %, by mass per volume) at room temperature, the insoluble matt er was separated carefully and left to dry for 24 h at 105 °C for the determination of fi nal dry mass.The following equation was used to calculate the solubility in water: /1/ where m initial and m fi nal are the initial and fi nal dry mass (g), respectively.All tests were carried out in triplicate.

Determination of optical properties
Transparency value of the fi lm was calculated by measuring the absorbance at 600 nm using UV-Vis spectrophotometer (model Cary 60; Agilent Technologies, Mulgrave, Australia) according to Kurt and Kahyaoglu (18).The samples were cut into rectangular pieces and placed in the spectrophotometer test cell.The reference value was determined from an empty test cell.The following equation was used to determine the transparency value of the fi lm: where A 600 nm is the absorbance at 600 nm, and δ is the mean fi lm thickness (mm).All measurements were performed in triplicate at 5 diff erent locations.

Determination of mechanical properties
A texture analyser (TA-XT2 texture analyser; Stable Micro Systems Co., Ltd., Godalming, UK) was utilised to determine the tensile strength (TS) and elongation at break (EAB) of the fi lms, based on ASTM D882-12 standard method (19).Sample fi lms were sliced into strips (1 cm×4 cm) and conditioned in 54 % relative humidity (RH) for 3 days.During the extension of the strips, force and distance were determined at 1.5 mm/s up to break.The load at break was divided by the fi lm cross-sectional area to calculate TS.Cross-sectional area was calculated based on the mean thickness of fi lm strips.Elongation at the moment of rupture was divided by the initial gauge length, and then multiplied by 100 in order to calculate the EAB (%).The results show the average of fi ve samples for each type of fi lms.

Determination of barrier properties
The water vapour permeability (WVP) values were determined according to ASTM E96 / E96M-10 standard method (20).The fi lms (d=14 mm) were sealed with paraffi n on glass permeation cups containing silica gels (0 % RH) and left in desiccators containing distilled water (100 % RH) at 25 °C.The mass of the glass permeation cups was measured for 8 h at 1-hour intervals and the following equation was used to determine the WVP ((g•mm)/ (m 2 •h•kPa)) of the fi lms: /3/ where m is the mass gain (g), t is the time of gain (h), δ is the fi lm thickness (mm), Δp is the partial pressure diff er-ence through the fi lm at 25 °C (kPa) and A is the fi lm area exposed to moisture transfer (1.539•10 −4 m 2 ).The m/t was determined via linear regression (R 2 >0.99) based on the mass of water absorbed by the system at the steady state.All values were determined in triplicate for each fi lm type.
Oxygen permeability (OP) was determined based on the peroxide value (PV) using AOCS Cd 8-53 method (21).Briefl y, a 25-mL tube was fi lled with 10 mL of sunfl ower oil, covered with fi lms containing thyme essential oil, sealed using paraffi n and sticky tape, and stored at a controlled temperature (60 °C) for 10 days.The PV of the sunfl ower oil samples was determined by titration with sodium thiosulfate.All values were determined in triplicate for each fi lm type.

Fourier-transform infrared spectroscopy measurements
Fourier-transform infrared (FTIR) spectra of the samples were measured using a FTIR spectrometer (model Spectrum Two; PerkinElmer, Akron, OH, USA) at 4 cm -1 resolution within a range of 650-4000 cm -1 .Before the analysis, the samples were left at room temperature in a desiccator containing silica gel for two weeks to obtain the highest dehydration.For each spectrum 32 scans were run.

Scanning electron microscopy determination
Scanning electron microscope (model JSM-7001F; JEOL, Tokyo, Japan) was utilised at 10 kV to determine the surface and cross-section characteristics of the samples.Before the visualisation process, sputt er coater with gold/palladium (model SC7620; Quorum Technologies, Laughton, UK) was used to coat the fi lms, and surface and cross-section photographs were taken at 1500 and 4000× magnifi cation, respectively.

Diff erential scanning calorimetry measurements
Thermal properties of the samples were measured via a diff erential scanning calorimeter (model DSC 4000; PerkinElmer).Before the analysis, fi lm samples were stored at room temperature for two weeks in a desiccator with silica gel to dehydrate.A sample weighing 5 mg was sealed in aluminium pans, and then scanned under a nitrogen atmosphere at 10 °C per min over the range of -50 to 120 °C.
For the FRAP determination, 1.5 mL of methanol was used to dissolve 0.15 g of anchovy by-product protein fi lm, and then 50 μL of fi lm extract solution were added to 0.95 mL of FRAP reagent prepared by mixing 300 mM acetate buff er (pH=3.6),20 mM FeCl 3 and a solution of 10 mM ferric-2,4,6-tripyridil-s-triazine (TPTZ) in 40 mM HCl at the ratio of 10:1:1.The absorbance was determined at 593 nm by UV-Vis spectrophotometer (Model Cary 60; Trolox standard curves were used to calculate FRAP values, which were expressed in mg of Trolox per mL of fi lm. For the DPPH scavenging activity determination, 10 μL of fi lm extract solution, obtained by dissolving 0.15 g of fi lm in 1.5 mL of methanol, were added to 1 mL of DPPH in methanol solution (60 μM) and 40 μL of methanol.Aft er fi rmly shaking the mixture, it was left at room temperature in the dark for 30 min.The absorbance was determined at 517 nm.All reagents, except for the sample, were used to prepare a negative control.Following equation was employed to calculate the DPPH scavenging activity: /4/ where A control is the absorbance of the control, and A sample is the absorbance of the sample.

Statistical analysis
All experiments were performed in triplicate and the values were presented as mean±standard deviation.Data of physical, optical, mechanical and barrier properties, and antioxidant capacity were determined using the analysis of variance (ANOVA) and the comparison of the mean values was done using Duncan's multiple range test.SPSS statistical package program v. 17.0 for Windows (SPSS Inc., Chicago, IL, USA) was utilised with a signifi cance level at p=0.05.

Physicochemical properties of fi lms
Table 1 shows the physicochemical properties of anchovy by-product protein fi lms containing thyme essential oil at various volume fractions.The sample thickness ranged from 0.197 to 0.202 mm, and the addition of thyme essential oil did not have a signifi cant (p>0.05)eff ect on the resulting fi lm thickness.Altiok et al. (12) reported similar results in the chitosan fi lms with thyme essential oil, and Bahram et al. (10) also obtained similar results in the whey protein fi lms prepared with 0.8 and 1.5 % cinnamon essential oil.However, Jouki et al. (14) showed that the addition of thyme essential oil increased the thickness of quince seed mucilage fi lms, but signifi cant increase was only obtained at the highest thyme essential oil level used (2 %).These diff erent results could be att ributed to the solid content of fi lm-forming solutions.
Solubility is one of the important properties of edible fi lms, due to its eff ect on the fi lm resistance to water, in particular in humid environments (24).As seen in Table 1, the fi lms containing thyme essential oil were less soluble than the control fi lm (p<0.05),but diff erences between solubility values of fi lms containing diff erent volume fractions of thyme essential oil were not signifi cant (p>0.05).A decrease in the hydrophilic nature of the sample may have led to these results.According to Kavoosi et al. (13), the type and concentration of compounds and their inherent hydrophilicity and hydrophobicity indices alter their additive eff ect on the fi lm solubility.Decreased solubility with the addition of diff erent essential oils was also determined by various researchers (3,10,(24)(25)(26).

Optical properties of fi lms
Consumer preference is directly aff ected by optical properties of edible fi lms used as food packaging materials (27).As seen in Table 1, the addition of thyme essential oil aff ected the transparency values of anchovy by-product protein fi lms, and only the addition of 1.5 % thyme essential oil reduced transparency (p<0.05), which was verifi ed by higher transparency value.The decrease in transparency could be att ributed to the increase in light scatt ering caused by oil droplets in the fi lm network.In addition, the colour components in thyme essential oil probably contributed to the decrease of fi lm transparency and fi lms containing thyme essential oil became less transparent.Similar results were obtained by Pires et al.
(2) for hake protein fi lms containing thyme essential oil, by Tongnuanchan et al. (28) for fi sh skin gelatin fi lms containing plai, turmeric and ginger essential oils and by Hosseini et al. (29) for fi sh gelatin chitosan biocomposite fi lms containing oregano essential oil.

Mechanical properties of fi lm samples
Mechanical properties of the samples were determined based on the TS and EAB values, the key indicators of fi lm fl exibility and strength (25).Table 1  The results represent mean values±standard deviation of three replicates.Values in the same row with diff erent lett ers in superscript are signifi cantly diff erent (p<0.05) containing thyme essential oil at diff erent volume fractions (0.5, 1.0 and 1.5 %).According to the results, control fi lm had the highest TS and the lowest EAB (p<0.05) compared to others.On the other hand, no changes were observed in the TS and EAB values in fi lms containing 0.5, 1.0 and 1.5 % thyme essential oil (p>0.05).Results showed that a 64 % reduction in the TS values (1.46-0.52MPa) was obtained by the addition of 1.5 % thyme essential oil to the anchovy by-product protein fi lms.The addition of essential oils into fi lms may reinforce the development of heterogeneous fi lm matrix, which may in turn lead to discontinuity in the fi lm network (9,28).Addition of essential oils to protein-based fi lms could prevent protein-protein interaction and produce fl exible domains within the fi lm (9).Moreover, these changes could be att ributed to plasticising eff ect of the thyme essential oil (14).As a result, fi lms containing thyme essential oil became weaker, with lower TS, but more stretchable.Jouki et al. (14) observed an increase in the EAB and a decrease in TS of quince seed mucilage fi lms containing thyme essential oil, which is in line with the results of this study.Similarly, Shojaee--Aliabadi et al. (26) found that the addition of Satureja hortensis essential oil into ĸ-carrageenan fi lm caused a signifi cant (p<0.05)decrease of TS and increase of EAB.Increases of EAB and decreases of TS are typical results of essential oil addition, which have been largely handled in the research on various edible fi lms (9,25,(27)(28)(29).

Barrier properties of fi lms
Water and oxygen are important compounds whose migration needs to be controlled, because physical and chemical changes in foods are usually created by the transfer of moisture and oxygen between the food and its surrounding atmosphere (29).Table 1 shows the WVP values of anchovy by-product protein fi lms containing various volume fractions of thyme essential oil.The addition of thyme essential oil to anchovy by-product protein fi lms increased the WVP values from 1.5 to 2.12 (g•mm)/(m 2 •h•kPa) (p<0.05),but signifi cant diff erences were not detected between the fi lms containing various volume fractions of thyme essential oil (p>0.05).Many factors such as the ratio of hydrophilic and hydrophobic fi lm components, fi lm thickness, water sensitivity and crystallinity can change the WVP values of the fi lms (14).The increase in the WVP values of the fi lms containing essential oil can be att ributed to the discontinuities, caused by lipid droplets in the polymer network leading to a loss in fi lm cohesion and therefore enhancing the transport of water vapour through the fi lm (29,30).Additionally, this may also be due to the ratio of hydrophilic and hydrophobic fi lm constituents, directly infl uencing the transfer of water vapour through the fi lm network (29).Altiok et al. (12) showed that WVP of chitosan fi lms containing thyme essential oil barely increased compared to the control fi lm, Jouki et al. (14) found a signifi cant increase in the WVP of quince seed mucilage fi lms containing thyme essential oil and in another study, Hosseini et al. (29) found that the addition of oregano essential oil into fi sh gelatin/chitosan fi lms also increased WVP values of the fi lms, all of which is consistent with the present results.In contrast, some researchers found that the incorporation of essential oils into edible fi lms improved WVP of the fi lms (3,9,25,26).These diff erences could be att ributed to the above-mentioned factors.
In this study, OP value of the control fi lm, measured as PV, was determined to be 14.0 mmol of O 2 per kg.The incorporation of thyme essential oil into the anchovy by--product protein fi lms signifi cantly increased (p<0.05) this value, which can be explained by higher oxygen solubility in non-polar oil phase that contributes to the increase of the transfer rate of oxygen molecules into plasticised polymer matrix (14).Additionally, this behaviour could be att ributed to the irregularities in the fi lm matrix.Altiok et al. (12) reported similar results in chitosan fi lms containing various amounts of thyme essential oil, Jouki et al. (14) also found similar results in quince seed mucilage fi lms containing 1.5 and 2 % thyme essential oil, and Ghasemlou et al. (25) reported similar results in corn starch fi lms containing 3 % Zataria multifl ora Boiss essential oil.

Results of Fourier-transform infrared spectroscopy
Anchovy by-product protein and thyme essential oil interaction was investigated with FTIR spectroscopy.The FTIR spectra of control fi lm and fi lms containing 0.5, 1.0 and 1.5 % thyme essential oil are shown in Fig. 1.The peak around 1043 cm -1 was observed in all fi lms, corresponding to the OH group, primarily from glycerol incorporated as a plasticiser (28,29).Increasing thyme essential oil volume fraction probably resulted in the dilution of glycerol and decrease of the peak amplitude.Tongnuanchan et al. (28) also found similar spectra for fi sh skin gelatin fi lms containing ginger, turmeric and plai essential oils, where OH group was observed at the wavenumber (ν) of 1031-1035 cm -1 .All fi lms had major peaks at ν=1624-1628 cm -1 (amide I, representing C=O stretching/hydrogen bonding coupled with COO), 1542-1547 cm -1 (amide II, resulting from bending vibration of NH groups and stretching vibrations of CN groups) and 1235-1237 cm -1 (amide III, illustrating the in-plane bending vibrations of CN and NH groups of bound amide or vibrations of CH 2 groups of glycine) (9,28).Similar results were also found by Arfat et al. (9) for fi sh protein isolate/fi sh skin gelatin fi lms with zinc oxide nanoparticles and basil leaf essential oil, where amide I, amide II and amide III peaks were observed at ν=1631, 1537 and 1238 cm -1 , respectively.In general, the Fig. 1.FTIR spectra of anchovy by-product protein fi lms containing thyme essential oil ϕ(TEO)/%: a=control, b=0.5, c=1.0 and d=1.5 amplitudes of the fi lms decreased as the thyme essential oil volume fraction increased and the highest amplitudes for amides I, II and III were determined in control fi lms, probably because of the highest protein content.
For all fi lm samples, an amide A peak (representing the NH stretching coupled with hydrogen bonding) was found at ν=3273-3276 cm -1 and an amide B peak (representing CH stretching and NH 3 + ) was found at ν=2919-2923 cm -1 .The amplitude of amide A peak decreased when thyme essential oil was added to the fi lms, especially at 1.5 % volume fraction.This certifi ed the lower interaction among anchovy by-product proteins, shown by the lower TS of the fi lm with the increasing EAB, when higher volume fractions of thyme essential oil were added.Similar results for fi sh skin gelatin fi lms containing ginger, turmeric and plai essential oils were also reported (28).Peaks at ν=2851 cm -1 (methylene symmetric CH stretching) and 2871 cm -1 (methyl asymmetric CH stretching) were also found in all fi lm samples.The amplitude of the peak at ν=2871 cm -1 decreased, while the amplitude of the peak at ν=2851 cm -1 increased when thyme essential oil was incorporated into fi lms.It was reported that symmetric methylene stretching bands approx.at ν=2851 cm -1 were present in most of the lipids, while asymmetric methyl stretching bands approx.at ν=2871 cm -1 were present in most of the proteins (9,28,31).Similar results in gelatin fi lms containing ginger, turmeric and plai essential oils and in fi sh gelatin fi lms containing basil and citronella essential oils have been reported (28,31).Hence, the FTIR spectrum results indicated that the addition of thyme essential oil into anchovy by-product protein fi lms aff ected the molecular organisation and intermolecular interaction in polymer matrix.

Results of scanning electron microscopy observations
Scanning electron microscopy (SEM) images of the surface and cross-section of anchovy by-product protein fi lms containing thyme essential oil are shown in Fig. 2. The control fi lm showed a homogeneous and continuous surface (Fig. 2a), while the anchovy by-product protein fi lms containing thyme essential oil had a heterogeneous surface (Figs.2b-d).In addition, many granules were observed on the surface of anchovy by-product protein fi lms containing thyme essential oil, and in general, the size of the granules increased as the essential oil volume fraction increased.This showed that the essential oil droplets were still unstable in emulsion systems stabilised by Tween 80. Consequently, the permeability of the surface containing granules was facilitated and resulted in an increase in WVP and OP values of anchovy by-product protein fi lms containing thyme essential oil.
The control fi lm had a heterogeneous, rough and compact cross-section structure (Fig. 2e), while fi lms containing thyme essential oil had a relatively smooth cross-section structure (Figs.2f-h).With the addition of thyme essential oil, cross-section structure probably soft ened, which could be responsible for the higher WVP and OP values of the fi lms containing thyme essential oil.Various researchers reported that the surface and cross-section structure of edible fi lms were aff ected by both addition and volume fraction of essential oils (12,29,30).Jouki et al. (14) also reported similar results when thyme essential oil was added into the quince seed mucilage fi lms.

Diff erential scanning calorimetry analysis
Film transitions were determined by diff erential scanning calorimetry (DSC) when thermal changes occurred in the fi lm (31).DSC thermograms of anchovy by-product protein fi lms containing thyme essential oil are shown in Fig. 3.All thermograms showed endothermic peaks at the phase transition temperature (t m ) ranges of 53.53-58.04°C.These peaks are probably related to the denaturation temperature of myosin and actin proteins (7).Da Rocha et al. (7), in their studies with Argentine anchovy (Engraulis encrasicholus) protein isolate, observed the endothermic peak at 62.2 °C.Altiok et al. (12) reported that chitosan fi lms with or without thyme essential oil had wide endothermic peaks nearly at 35-140 °C.Temperature diff erences of endothermic peaks could be att ributed to the type of fi lm or essential oil, the oil quality and volume The highest glass transition temperature (t g ) of -24.27 °C was obtained in control fi lm and the addition of 0.5 % thyme essential oil into the anchovy by-product protein fi lms caused a decrease in t g to -32.82 °C (data not shown).However, t g of fi lms containing 1.0 and 1.5 % thyme essential oil most likely became too low, thus could not be observed in the tested temperature range.Lower t g is better for fi lm fl exibility, and results showed that fi lms containing thyme essential oil had lower t g , indicating bett er fl exibility than the control sample.Moreover, Jouki et al. (14) found that pure quince seed mucilage fi lm had t g of -35.2 °C, which is higher than of the same fi lm containing thyme essential oil (t g =-43 to -57 °C).

Antioxidant capacity of fi lms
FRAP and DPPH scavenging activity values of anchovy by-product protein fi lms containing various volume fractions of thyme essential oil are shown in Fig. 4. The lowest FRAP value (0.91 mg/mL, expressed as Trolox equivalents, Fig. 4a) was observed in the control fi lm, and it signifi cantly increased (p<0.05) with the increase of the volume fraction of thyme essential oil from 0.5 to 1.5 %.Like FRAP value, the lowest DPPH scavenging activity value (0.06 %) was determined in the control fi lm, and it signifi cantly increased (p<0.05) with the increase of the thyme essential oil volume fraction.However, no significant diff erences were detected between the fi lms containing 1.0 and 1.5 % thyme essential oil (p>0.05)(Fig. 4b).Lower FRAP and DPPH scavenging activity values could be att ributed to free sulfh ydryl groups and amino acids of anchovy by-product protein such as tryptophan, methionine and tyrosine (32).Eff ect of thyme essential oil on the antioxidant capacity of chitosan fi lms was discussed by Altiok et al. (12) and Bonilla et al. (30), by Espitia et al. (33) in edible açai fi lms and Jouki et al. (14) in quince seed mucilage fi lms.Similar results were also shown for hake protein fi lms containing thyme essential oil (2).The antioxidant capacity of thyme essential oil may be explained by its thymol and carvacrol content, two phenolic compounds with known antioxidant capacity (11).The results indicated that anchovy by-product protein fi lms containing thyme essential oil could be used for the packaging of foods as active fi lms with antioxidant activity.

Conclusions
Addition of thyme essential oil aff ected the solubility, transparency, mechanical properties, water vapour permeability (WVP), oxygen permeability (OP), FTIR spectra, texture and thermal properties of the fi lms.The addition of thyme essential oil into anchovy by-product protein fi lms decreased the solubility and tensile strength values, but increased the elongation at break, WVP and OP values.The increase in WVP and OP values of the fi lms might limit their application in foods.However, increasing the volume fraction of thyme essential oil did not lead to signifi cant changes in these properties.Film containing 1.5 % thyme essential oil became more opaque than other fi lms.The antioxidant capacity of anchovy by--product protein fi lms was considerably improved by the addition of thyme essential oil.These results showed that the addition of thyme essential oil into the anchovy by--product protein fi lms is suitable for developing antioxidant fi lms.However, research on the assessment of the antimicrobial properties of these fi lms is also required for their practical applicability.

Fig. 4 .Fig. 3 .
Fig. 4. Antioxidant capacity of anchovy by-product protein fi lms containing thyme essential oil (TEO) measured by: a) FRAP as Trolox equivalents, and b) DPPH scavenging activity.Bars represent mean values±standard deviation of three replicates, diff erent lett ers on the bars indicate signifi cant diff erences (p<0.05)

Table 1 .
shows the TS and EAB values of anchovy by-product protein fi lms Physicochemical, optical, mechanical and barrier properties of anchovy by-product protein fi lms with thyme essential oil