Food Hydrocolloids 87 (2019) 270–286

Contents lists available at ScienceDirect

Food Hydrocolloids
journal homepage: www.elsevier.com/locate/foodhyd

Solvent strength and biopolymer blending effects on physicochemical T

properties of zein-chitosan-polyvinyl alcohol composite films
Stephen Gitonga Giterua,b, M. Azam Alia, Indrawati Oeya,b,∗
a
Department of Food Science, University of Otago, PO Box 56, Dunedin, 9054, New Zealand
b
Riddet Institute, Palmerston North, New Zealand

A R T I C LE I N FO A B S T R A C T

Keywords: The aim of this research was to optimise and develop edible films made from zein (ZN), chitosan (CS), poly(vinyl
Zein alcohol) (PVOH) and poly(ethylene glycol) (PEG400) using Box Behnken Design (BBD). Four factors viz. ZN:CS
Chitosan ratio (w/w) (1:3, 1:1, 3:1), PVOH wt.% (in dry film) (0, 0.25, and 0.5), PEG400 wt% (0.13, 0.21, 0.31) and
Poly(vinyl alcohol) ethanol (40–80% v/v in water), were investigated. The effects on tensile strength (TS), Young's modulus of
Blending
elasticity (EM), elongation at break (%EAB), water vapour permeability (WVP), solubility in water (Ws), and
Edible films
Solvent properties
optical properties were evaluated. The films were further characterised using differential scanning calorimetry
(DSC), thermogravimetric analysis (TGA), scanning electron microscopy (SEM), x-ray diffraction (XRD) and
swelling behaviour. SEM micrographs showed that high levels of ZN and PEG400 resulted in rough surfaces with
major film defects, but these were minimised at high levels of CS or PVOH. XRD analysis indicated potential
interactions through hydrogen bonding between the polar groups of ZN and hydroxyl moieties of PVOH and CS.
Stress-strain curves showed that ZN and CS composite films possessed high TS but were brittle. Incorporation of
PVOH increased the ductility of the composite films. Optimum films were obtained using ZN/CS/PVOH/PEG400
ratio of 0.35/0.29/0.13/0.23 (wt.%) and 60% ethanol, giving dry state properties of TS, EM, and EAB as
24.21 MPa, 356.62 MPa and 84.23%. respectively. This study demonstrated the influence of critical interactions
between solvent features (polarity) and interfacial properties of ZN, CS, PVOH and PEG400. The observed
crosslinking behaviour suggested the effectiveness of blending technique in improving the compatibility of
biopolymers and overall functionality of edible films.

1. Introduction
Protein-polysaccharide blends or combinations with biosynthetic
polymers have increasingly attracted attention for biomaterial appli-
cations (Escamilla-García et al., 2013; Sun, Dai, & Gao, 2017a; Yuan
et al., 2016). The functionality and cross-linking mechanisms for pro-
teins -polysaccharides composite films have been reviewed (Azeredo &
Waldron, 2016). Among the proteins used in the formation of edible
films, zein (ZN), a prolamin protein from the corn endosperm is an
interesting material due to its amphiphilicity and self-assembly cap-
abilities through hydrophobic/hydrophilic interactions (Shukla &
Cheryan, 2001). Owing to the high proportion (50%) of non-polar
amino acid residues, zein does not dissolve in water but can be solu-
bilised in acetone, ethanol, anionic detergents and high concentration
of urea or alkali (pH ≥ 11) (Shukla & Cheryan, 2001). It is highly re-
garded for its potential in film formation (Wang & Padua, 2010), de-
velopment of delivery systems (Liang et al., 2015) and fabrication of
fibres and fabrics (Selling, Woods, & Biswas, 2011). Compared to other
proteins, zein films present better water vapour barrier properties
(Ghanbarzadeh & Oromiehi, 2008). However, they are constrained by
fragility and poor extensibility (Argüello-García et al., 2014), limiting
their use in tension-based applications.
Chitosan (CS), a copolymer of α-(1 → 4) glucosamine, (C6H11O4N)
n, is a cationic polysaccharide obtained by alkaline deacetylation of
chitin. Besides biodegradability and renewability, its biocompatibility
and non-toxic features permit applications in the food industry, parti-
cularly in the formulation of food materials, films and coatings
(Escamilla-García et al., 2013). The presence of amino groups on the
structure of chitosan enables it to undergo ionic or hydrophobic mod-
ifications under varying pH (Sun, Zhang, & Chu, 2007). As a widely
investigated material for edible film production, notable properties of
its products include clear, flexible and robust film materials with an
adequate oxygen barrier and may possess antimicrobial properties. Like
zein films, chitosan films lack flexibility due to high brittleness and

∗
Corresponding author. Department of Food Science, University of Otago, PO Box 56, Dunedin, 9054, New Zealand.
E-mail address: indrawati.oey@otago.ac.nz (I. Oey).

https://doi.org/10.1016/j.foodhyd.2018.08.006
Received 28 March 2018; Received in revised form 31 July 2018; Accepted 4 August 2018
Available online 09 August 2018
0268-005X/ © 2018 Elsevier Ltd. All rights reserved.
S.G. Giteru et al.
exhibit weak water resistance characteristics (Suyatma, Tighzert, &
Copinet, 2005).
The weak structural integrity and poor functionality of ZN and CS
have challenged the development of edible and biodegradable films.
Modification approaches, including chemical and physical techniques,
continue to remedy the shortcomings of the food contact materials
through various crosslinking mechanisms (Azeredo & Waldron, 2016).
Attempts by previous studies to involve ZN/CS blends in the formula-
tion of edible films (Cheng, Wang, & Weng, 2015; Escamilla-García
et al., 2013) or delivery systems involving nanoparticles (Liang et al.,
2015) showed that additional cross-linking techniques such as the use
of dicarboxylic acids or heat were required for improvement of me-
chanical and functional properties. However, the continued use of
chemicals and energy inefficient modification techniques render these
materials less attractive for food applications (Ashogbon & Akintayo,
2014). A major drawback to the use of chemicals is the occurrence of
undesired changes and potential toxic effects (Sun et al., 2007). In
addition, the majority of chemically and physically modified films and
coating still exhibit poor mechanical and barrier properties (Azeredo &
Waldron, 2016).
Poly(vinyl alcohol) (PVOH) is a biodegradable and non-toxic bio-
synthetic polymer (Chiellini, Corti, D'Antone, & Solaro, 2003) used as
an ingredient in the food and pharmaceutical applications, for example,
in the production of coatings and films (Nakano et al., 2007; Wright,
Andrews, McCoy, & Jones, 2013). The linear carbon backbone of PVOH
is rich in hydroxyl (–OH) groups making it highly hydrophilic and
dispersible in water. The chemical structure of PVOH permits various
reactions through cross-linking between the −OH groups and other
polar molecules to form entanglements and aggregates (Bercea,
Morariu, & Teodorescu, 2016). Its excellent film forming ability has
been used in combination with numerous biopolymers to enhance their
mechanical properties and chemical resistance (Senna, Salmieri, El-
naggar, Safrany, & Lacroix, 2010; Srinivasa, Ramesh, Kumar, &
Tharanathan, 2003).
Although there have been wide applications of ZN/CS blends in the
formulation of delivery systems and nanoparticles (Luo, Zhang, Cheng,
& Wang, 2010; Müller et al., 2011), applications in edible film for-
mulations have only recently emerged (Cheng et al., 2015; Escamilla-
García et al., 2013, 2017). Available literature shows that these mate-
rials are characterised by low TS and elastic modulus (5.4–5.6 MPa and
65 to 7 MPa respectively) (Escamilla-García et al., 2013), which may
pose a challenge in food applications. Similarly, investigations invol-
ving ZN/PVOH suggested that unmodified blends could not be handled
for food applications (Lacroix et al., 2014; Senna et al., 2010). Blending
is a promising technique for improvement of functional properties of
protein-polysaccharide edible films. The main advantages of the tech-
nique are the utilisation of distinct functional properties of each com-
ponent to improve the overall functionality of the final product (Senna
et al., 2010). Additionally, plasticization of biomacromolecules is an
established and acceptable approach to alter their physicochemical
properties and functionality during application. PEG400 is a widely
investigated plasticizer for various food and pharmaceutical applica-
tions. It is a water-soluble, non-toxic and biodegradable synthetic
polymer known for its colloidal stability and plasticising effects
(Tillekeratne & Easteal, 2000).
To date, little is known concerning the functionality of ZN/CS/
PVOH blends in composite film formulations or whether the use of
plasticizers such as polyethylene glycol (PEG) 400 can mitigate in-
compatibility and general weaknesses of these materials for food ap-
plications (Escamilla-García et al., 2017; Senna et al., 2010). Initially,
understanding the solvent properties of these combinations is essential
since the solvent determines the unfolding and dispersion of macro-
molecular chains, whereas its removal accelerates molecular aggrega-
tion into ordered microphases (Wang & Padua, 2010). The phase be-
haviour properties of macromolecules during self-assembly are
determined by the solution properties including ionic strength, pH and
271
Food Hydrocolloids 87 (2019) 270–286
the proportions of various compounds (Gaona-Sánchez et al., 2015).
Therefore, varying the type and strength of the solvent can influence
the microstructure and functionality of the formed aggregates (Wang &
Padua, 2010).
The aim of this study was to evaluate the effect of blending using
ZN:CS ratio, PVOH, PEG400 and the concentration of ethanol on me-
chanical and physicochemical properties of composite films. The opti-
misation approach used in this study enabled the understanding of the
effects of individual factors on the physical characteristics of the films
using scanning electron microscopy, which were then related to the
tensile, water interaction and permeability properties. Changes due to
cross-linking and inherent interactions were evaluated using differ-
ential scanning calorimetry, thermogravimetric analysis and X-ray dif-
fraction.
2. Materials and methods
2.1. Materials
Corn zein (ZN) (Z 3625, molecular weight (Mw) 22–24 kDa) and
98% (w/w) protein, poly(vinyl alcohol) (PVOH) (Mw
89,000–98,000 Da, 99+% hydrolysed) and polyethylene glycol
(PEG400) were obtained from Sigma-Aldrich (St. Louis, MO, USA).
Chitosan (CS) (Mw 890 kDa, 91% degree of deacetylation (DD)) was
obtained from Waseta Int'l trading company (Shanghai, China). Glacial
acetic acid (Univar, USA) and ethanol (99.5%) (Anchor ethanol limited,
Auckland, NZ) were of analytical grade. The biopolymer/biosynthetic
polymers used in the study are Generally Recognized as Safe (GRAS)
(U.S. code of federal regulations) (USFDA, 2018) under the sections; ZN
(part 184.1984), CS (GRN 443) (FDA, 2013), PVOH (part 177). All
reagents and chemicals were used as received.
2.2. Experimental design and optimisation
Film formulations were optimised to verify the synergistic influence
of the variables using Box-Behnken experimental design (BBD). Four
independent variables were fixed at three equidistant levels; ZN:CS
ratio (X1), (1:3, 1:1, 3:1), ethanol (X2) (40, 60, 80% v/v), PVOH (X3) (0,
25, 50% w/v), and PEG400 (X4) (15, 30, 45% w/w of total solids). The
number of experiments (N = 27) (Table 1) was determined by N=2k (k
– 1) + C0 (where k is the number of factors and C0 is the number of
central points) (Yetilmezsoy, Demirel, & Vanderbei, 2009). The re-
sponse variables were taken as the average values of tensile strength
(TS) (Y1), (MPa), Young's modulus of elasticity (EM) (Y2) (MPa),
elongation at break (%EAB) (Y3), water vapour permeability (WVP)
(g.mm/h.m2. kPa) (Y4), water contact angle (WCA) (Y6) and solubility
in water (Ws) (Y5). Optical properties including; overall colour change
(ΔE) (Y7) and opacity (Op) (AU nm/mm) (Y8) were also determined.
Film preparation and response characterisation for each formulation
type was independently repeated three times.
2.3. Preparation of film-forming solutions
Stock solution of CS (1.5%) (w/v) in 2% (v/v) acetic acid was
prepared by stirring at 22 °C and 400 rpm for 12 h. PVOH, 10% (w/v) in
deionised (DI) water was stirred (400 rpm) for 2 h at 75 ± 3 °C. ZN
solution 1.2% (w/v) in 80% (v/v) ethanol-water solution was prepared
by stirring (400 rpm) for 1 h at 22 °C. The solutions were equilibrated at
4oC for 6 h then centrifuged (Beckman J-1, Beckman, USA) at
10,000 × g and 4oC for 15 min.
2.4. Film casting
Film forming solutions (FFS) were prepared by mixing the stock
solutions as shown in Table 1. This was followed by stirring for 20 min
and homogenisation at 8000 rpm for 2 min (IKA® T 50 ULTRA-
 S.G. Giteru et al.

Table 1
Box–Behnken design matrix with four independent variables expressed in coded, natural units and final composition of biopolymers in the dry composite film.
Formulation of film-forming solutions Mass fraction (wt % in dry film)*

Formulation no. ZN:CS ratio Ethanol (%) PVOH (%) PEG400 (%) pH of FFS ZN CS PVOH PEG400 Blend characteristics†

x1 (coded) X1 (uncoded) x2 (coded) X2 (uncoded) x3 (coded) X3 (uncoded) x4 (coded) X4 (uncoded)

1 −1 1:3 −1 40 0 25 0 30 4.47 ± 0.05 0.13 0.39 0.25 0.23 - translucent, yellowish, separation
2 0 1:1 −1 40 0 25 +1 45 4.52 ± 0.11 0.22 0.22 0.25 0.31
3 0 1:1 −1 40 0 25 −1 15 4.58 ± 0.07 0.31 0.31 0.25 0.13
4 0 1:1 −1 40 +1 50 0 30 4.57 ± 0.07 0.13 0.13 0.50 0.23
5 0 1:1 −1 40 −1 0 0 30 4.51 ± 0.09 0.38 0.38 0.00 0.23
6 +1 3:1 −1 40 0 25 0 30 4.48 ± 0.19 0.39 0.13 0.25 0.23
7 −1 1:3 0 60 0 25 −1 15 5.04 ± 0.11 0.15 0.46 0.25 0.13 - precipitation, no separation
8 −1 1:3 0 60 −1 0 0 30 4.91 ± 0.1 0.19 0.58 0.00 0.23 - translucent, yellowish, no separation
9 −1 1:3 0 60 +1 50 0 30 4.97 ± 0.07 0.07 0.20 0.50 0.23 - uniform weak gel formation, clear white
10 −1 1:3 0 60 0 25 +1 45 5.05 ± 0.04 0.11 0.33 0.25 0.31 - firm gel separation, clear top layer
11 0 1:1 0 60 +1 50 −1 15 4.97 ± 0.12 0.18 0.18 0.50 0.13
12 0 1:1 0 60 0 25 0 30 5.00 ± 0.15 0.26 0.26 0.25 0.23 - translucent, yellowish, separation

272
13 0 1:1 0 60 +1 50 +1 45 5.11 ± 0.03 0.09 0.09 0.50 0.31 - firm uniform gel
14 0 1:1 0 60 −1 0 +1 45 4.99 ± 0.09 0.34 0.34 0.00 0.31 - translucent, yellowish, no separation
15 0 1:1 0 60 0 25 0 30 5.09 ± 0.1 0.26 0.26 0.25 0.23 - precipitation, yellowish, separation
16 0 1:1 0 60 −1 0 −1 15 5.06 ± 0.06 0.43 0.43 0.00 0.13 - precipitation, lumps, separation
17 0 1:1 0 60 0 25 0 30 5.07 ± 0.14 0.26 0.26 0.25 0.23 - precipitation, yellowish, separation
18 +1 3:1 0 60 +1 50 0 30 5.19 ± 0.03 0.20 0.07 0.50 0.23 - precipitation, lumps, separation
19 +1 3:1 0 60 0 25 +1 45 5.11 ± 0.09 0.33 0.11 0.25 0.31 - precipitation, yellowish, separation
20 +1 3:1 0 60 0 25 −1 15 5.08 ± 0.03 0.46 0.15 0.25 0.13
21 +1 3:1 0 60 −1 0 0 30 5.11 ± 0.1 0.58 0.19 0.00 0.23 - translucent, yellowish, separation
22 −1 1:3 +1 80 0 25 0 30 5.39 ± 0.06 0.13 0.39 0.25 0.23 - cloudy, white, separation
23 0 1:1 +1 80 −1 0 0 30 5.35 ± 0.07 0.38 0.38 0.00 0.23 - firm gel
24 0 1:1 +1 80 0 25 −1 15 5.43 ± 0.11 0.31 0.31 0.25 0.13
25 0 1:1 +1 80 +1 50 0 30 5.49 ± 0.03 0.13 0.13 0.50 0.23 - white precipitate, separation
26 0 1:1 +1 80 0 25 +1 45 5.33 ± 0.07 0.22 0.22 0.25 0.31 - white precipitate, no separation
27 +1 3:1 +1 80 0 25 0 30 5.54 ± 0.05 0.39 0.13 0.25 0.23 - precipitation, yellowish, separation

*Assumption that ethanol content was 0% v/v in dry film; n = 3.

*ZN, zein; CS, chitosan; PVOH, poly(vinyl alcohol); PEG400, poly(ethylene glycol) 400.
†
Separation means visible settling of insoluble complex, precipitate formation means uniformly suspended aggregates.
Food Hydrocolloids 87 (2019) 270–286
S.G. Giteru et al.
TURRAX®, Krackeler Scientific, NY, USA). Thereafter degassing was
carried out for 2 min at 25 ± 3 °C in degassing mode (37 kHz) using
ultrasonic machine (Elmasonic S 30 H, Elma Schmidbauer GmbH,
Singen, Germany). The pH of the FFS increased from ∼4.50 with in-
creasing volume fraction ethanol to ∼5.40 in FFS containing 80%
ethanol. Films were cast on 10 cm × 10 cm polystyrene Petri dishes
using a constant rate (∼0.04 g/cm2 total solids) in order to minimise
variations in film thickness. The drying conditions were kept constant
at 20 ± 2 °C using forced draft (airflow rate v = 0.5 m/s). Dried films
were conditioned at 22 ± 2 °C in a vacuum desiccator equilibrated
using magnesium nitrate hexahydrate (Mg (NO3)2·6H2O) (53 ± 3%
RH) for a minimum of seven days. The films were weighed, and the
average mass of each component was calculated and reported as a
weight fraction (wi) (wt. % dry basis).
2.5. Physicochemical properties of composite films
2.5.1. Determination of film thickness and moisture content
The thickness (T) of the films was measured at six random locations
after seven days of conditioning at 53 ± 3% RH. A hand-held digital
micrometre (Thorlabs, NJ, USA) was used to the nearest 0.01 mm. The
moisture content was determined gravimetrically by drying ∼0.5 g
sample in an oven at 105 ± 2 °C for 24 h (AOAC, 2005).
2.5.2. Scanning electron microscopy
The surface morphology of the composite films was determined
using field emission scanning electron microscopy (JEOL Ltd, Tokyo,
Japan). The samples were mounted on stabs in triplicate, gold-coated
(2 min, 2 mbar) and the micrographs were observed at 10 kV in a va-
cuum of 9.75 × 10−5 Torr.
2.5.3. Determination of tensile properties
Tensile properties were determined on the conditioned films using a
universal testing machine TA-HD plus (Stable Micro Systems Ltd,
Godalming, UK) following a standard method (ASTM D638-02a (V))
(ASTM, 2003). Briefly, films were cut into a dumbbell shape (overall
length (L0) 63.5 mm, gage length l = 7.62 mm, the width of the narrow
section W = 3.18 mm). The test conditions included; a load cell of 50 N,
a cross-head speed of 5 mm/min, pre-test speed of 4.8 mm/min, the
post-test speed of 8 mm/min, initial separation of the grips 25.4 mm
and a target distance of 30 mm. The cross-sectional areas of the test
samples were determined immediately prior to the test using the
thickness (T) of individual samples obtained at five arbitrarily chosen
positions. The tests were conducted at 22 ± 2 °C and 45 ± 3% RH.
The maximum stress (tensile strength, TS) (MPa), Young's modulus of
elasticity (EM) (MPa) from the stress-strain curves and film extensibility
(elongation at break, %EAB) from the force-distance curves were de-
termined using Eqs. (1)–(3) respectively. At least six film samples were
tested for each treatment.
TS (MPa) = Fmax / Ao (1)
EAB (%) = [Δl/ l 0] × 100 (2)
EM = LO (F2 − F1) AO (δ2 − δ1) (3)
where, Fmax is the maximum load for breaking the film (N), Ao is the
initial specimen cross-sectional area (T × W mm2), Δl is the change in
length from the original length (lo) of the specimen between the grips.
The force (F1, F2) and displacement (δ1, δ2 ) values were obtained from
the linear portion of a plot of displacement against the force.
2.5.4. Determination of water vapour permeability
The water vapour permeability (WVP) was determined following a
standard method (ASTM E96/E96-10) (ASTM, 2013b). Circular discs,
40 mm in diameter, were cut from independently prepared films. The
discs were fitted in the closures of Schott bottle (100 mL) containing a
273
Food Hydrocolloids 87 (2019) 270–286
33 mm diameter hole and held in place using an O-ring (25 mm) dia-
meter. The thickness of each film was measured at four random loca-
tions during the analysis. The Schott bottle was filled with 90 mL of
water, and the assembly was weighed at a precision of 0.001 g to obtain
w0. The assembly was then placed in a chamber containing silica gel
(10–15% RH) at 25 °C. The chamber was fitted with a fan to provide
strong driving force across the film for water vapour diffusion. The
shiny side of the films was always exposed to the atmosphere at the
lowest RH. The RH was maintained constant by replacing the silica gel,
after every reading using freshly dried gel. Six readings were taken at
an 8-h interval for 48 h and the weight difference was plotted in a
scatter plot (R2 = 0.99). The water vapour transmission rate (WVTR)
was calculated from the slope of the weight loss vs time plot (Eq. (4))
and WVP was calculated according to Eq. (5).
WVTR (g / h·m2) = (Δw /Δt )/ A (4)
WVP (g ·mm / m2 ·h·kPa) = Δw·T / A·ΔP (5)
where Δw is the weight change from the straight line (g), Δt is the time
during which Δw occurred (h), A is the test area (cup mouth area, m2). T
is the film thickness (mm), ΔP is the differential water vapour partial
pressure across the film [ΔP = (P0 (kPa) × (RH1-RH2)/100: P0 is the
saturation vapour pressure at the test temperature (P0 of water at
25 °C = 3.159 kPa). RH1 is the relative humidity inside the bottle and
RH2 is the relative humidity outside the bottle. An assumption was
made that the relative humidity inside the bottle (RH1) was 100%
(ASTM, 2013b). The relative humidity outside the bottle (RH2) was
measured using a temperature and RH logger, iButton (Maxim In-
tegrated, CA, USA).
2.5.5. Determination of water solubility
The films solubility in water was determined as described by Yuan
et al. (2016) using films that were equilibrated at 53% RH for seven
days. The films were dried at 60 °C and 40% RH for 24 h, then samples
of about 3 × 3 cm were weighed (W0) into 50 mL centrifuge tubes then
the tubes were filled with 35 mL of deionised water containing 0.02%
sodium azide. The tubes were sealed and incubated with shaking at
25 ± 2 °C for 24 h followed by centrifugation at 3300 × g for 15 min,
and the supernatant was discarded. The final weight of the total in-
soluble solids (W1) was obtained after drying the sediments at 60 °C
until they attained a constant weight. The films solubility in water (Ws)
was calculated according to Eq. (6).
WS = (W0 − W1/ W0) × 100 (6)
2.5.6. Contact angle measurement
The surface hydrophobicity of the composite films was estimated
using goniometer (OCA 15 AMP, Dataphysics instruments, Stuttgart,
Germany) with the help of the sessile drop method, based on optical
contact angle technique described by Ferreira, Nunes, Delgadillo, and
Lopes-da-Silva (2009). In this method, a droplet of deionised water
(∼2 μL) was deposited on the airside (upper side during casting) of the
film using a precision syringe installed on the instrument. The image of
the drop was recorded at the point of contact with the surface to obtain
static contact angle data, which was processed and numerically fitted to
La Place-Young equation. Triplicate measurements were conducted for
each film.
2.5.7. Determination of film colour
The colour of the composite films was determined with a Minolta
colourimeter, MiniScan EZ 4500 L (HunterLab, Virginia, USA) using
instrumental colour values L* (lightness), a*(redness), b* (yellowness).
A standard white tile (L* = 92.86, a* = 1.91, b* = −14.98) was used
to calibrate the instrument and as a background during the measure-
ments. An average of nine measurements recorded for each film and
used to calculate the total colour difference (ΔE) according to Eq. (7)
S.G. Giteru et al.
(Gutiérrez & Álvarez, 2016).
ΔE = (ΔL∗)2 + (Δa∗)2 + (Δb∗)2 (7)
where ΔL*, Δa* and Δb* are the differential colour values between the
sample colour parameters and those of the standard white tile used as
the film background.
2.5.8. Determination of light transmission properties
Opacity (Op) of the composite films was determined on rectangular
strips (20 mm × 4 mm) directly placed in a UV–Visible spectro-
photometer test cell (Ultraspec 3300 Pro; Amersham Biosciences,
Uppsala, Sweden) as described by Pena-Serna and Lopes-Filho (2013).
The absorbance of the films was obtained at 600 nm in a UV–Visible
spectrophotometer using air as a reference material. The Op of the films
was expressed as A600/x, where A600 is the absorbance at 600 nm, and
x is the film thickness (expressed as Absorbance Units/millimetres (AU/
mm).
2.5.9. Determination of phase transition properties of the composite films
Differential scanning calorimetry (DSC) was performed according to
a standard method D3418 (ASTM, 2013a) and as described by Shen and
Kamdem (2015). Samples conditioning was carried out at 0% RH, 22 °C
for seven days, then (∼5 mg) were weighed and sealed into a ten μL
DSC Tzero standard pan, and the test was conducted using DSC Q100
(TA Instrument, New Castle, DE, USA). An empty pan was used as an
inert reference, and the calibration was performed using the indium
standard. The instrument was maintained at an isothermal temperature
for 2 min before heating the samples at 5 °C/min using a temperature
regime of 25–225 °C and nitrogen flow rate of 20 mL/min. The glass
transition temperature (Tg), melting transition (TM) and enthalpy of
melting (ΔH) were analysed using the Thermal Analysis Universal 2000
version 4.5A software (TA Instrument, New Castle, DE, USA). The
melting (TM) temperatures (Tc) were taken as the peak of the en-
dothermic and exothermic curves, during heating. DSC analysis was
conducted in triplicate for the selected formulations.
2.5.10. Thermogravimetric analysis of the composite films
Thermogravimetric analysis (TGA) was conducted using (TGA,
Q500, TA Instruments, New Castle, DE, USA) to record the weight-loss
data of the film with changes in temperature. Samples of about 10 mg
were heated by ramping the temperature between 25 and 750 °C at a
heating rate of 10 °C/min and nitrogen flow rate of 20 mL/min. Each
test was analysed in duplicate from two independently prepared films
of the selected formulation type.
2.5.11. Determination of crystal structure using X-ray crystallography
The distribution of the crystal structure in the matrix of the com-
posite film was determined using X-ray diffraction on a PANalytical
X'Pert PRO MPD PW3040/60 diffractometer (Malvern PANalytical,
Almelo, Netherlands) in the range of 5° < 2θ < 35° diffraction angles.
Cu Kα (λ = 1.5406 Å) irradiation was used at 40 kV and 30 mA. Data
were collected on a continuous scan mode with a step size of 0.008° and
10.075 s per step. All data analysis and processing were conducted
using the X'pert accompanying software program X'Pert HighScore
v4.0.
2.6. Validation of regression models and film optimisation
2.6.1. Validation of regression models
The validity of the models was determined by conducting additional
experiments using conditions selected within the range of the BBD
parameters. Triplicate experiments were performed using five for-
mulations to compare experimental results with the predicted values.
274
Food Hydrocolloids 87 (2019) 270–286
2.6.2. Film optimisation
Simultaneous optimisation of the multiple responses was conducted
using the desirability function of Minitab. Obtaining desirability values
enables optimisation by resolving multiple responses into a single re-
sponse (Yuan et al., 2016). The maximum, minimum, and average va-
lues of the response variables experimentally achieved in the BBD ex-
perimental design (Table 2) were applied for the calculation of
composite desirability (Cd) function. Cd is the weighted geometric
average of the individual desirability (di) and ranges between d = 0, for
an undesirable response, and d = 1, for an entirely desired response,
above which further improvements would have no importance (Pena-
Serna & Lopes-Filho, 2013). Five experimental conditions using four
responses (TS, EM, EAB and Ws) having a Cd of 0.82–0.94, were ob-
tained and considered for the validation and optimisation. Optimisation
of the formulations was conducted on the basis of attaining maximum
TS, medium EM, adequate EAB, and minimal Ws (Ortega-Toro,
Jimenez, Talens, & Chiralt, 2014; Yuan et al., 2016). Since the selected
regions for the responses did not lie on the same optimisation slope,
constraints were set by setting targets, minimum and maximum values,
for all independent variables and responses.
2.6.3. Study on swollen state properties of optimised films
To study the behaviour of the optimised composite films in a wet
environment, the films were further characterised using the swelling
behaviour following the method described by Gutiérrez & Álvarez.
(2016). Films were cut into circular discs (10 mm diameter), then im-
mersed in tubes containing 10 mL of either hydrochloric acid (HCl) (pH
2, 0.1 M), potassium phosphate/sodium hydroxide (K2HPO4/NaOH)
buffer (pH 7.5, 0.1 M), or deionised water at 23 °C. Sodium azide
(0.02% w/v) was added to the solutions to prevent microbial growth.
The tubes were covered and left undisturbed at 23 °C for 24 h. Since the
films rapidly disintegrated when dipped in HCl, the immersion period
was reduced to a maximum of 2 min to allow the recording of differ-
ences between various formulations. The swelling behaviour (Sw) was
determined according to Eq. (8).
S w (%) = (D2 − D1 /D1)×100 (8)
where D1 and D2 are the diameters of the films before (ø = 20 mm) and
after immersion in the standard solutions respectively.
Swollen state mechanical properties of the optimum films in the dry
and wet conditions were also determined. Prepared films were further
conditioned at 25 °C and 30 ± 3% RH for 24 h in a forced convention
laboratory incubator (IL-11A, Biotechnical Services, CA, USA). These
conditions allowed some degree of crosslinking to occur, which reduced
the level of film disintegration in the wet environment and allowed
physical handling. Dumbbell-shaped filmstrips were then obtained and
dipped into phosphate/sodium hydroxide (K2HPO4/NaOH) buffer (pH
7.5, 0.1 M), for 2 ± 0.25 h before the tensile tests.
2.7. Statistical analysis
Regression analysis was carried out to test the significance of the
regression relationship between the responses (Yi) and the independent
variables. Experimental data from each response was fitted to a second-
degree polynomial model using backwards stepwise elimination (Chen
et al., 2016) in Minitab (Version 17) (PA, USA). A non-linear quadratic
model obtained from the design was in the form shown in Eq. (9).
Yi = β0 + βi ∑ Xi + βii ∑ Xi2 (i ≠ j ) + βij ∑ Xi Xj + ε (9)
where Yi is the predicted response value; β0, βi, βij and βii are the re-
gression coefficients of the model for the intercept, linear, quadratic
and interaction coefficients respectively; Xi and Xj are the independent
variables. Analysis of variance (ANOVA) was carried out to determine
the significance of the models and the fit for the polynomial equations
using F test and probability (p-values) (Chmiel, Kupska, Wardencki, &
 S.G. Giteru et al.

Table 2
Mean values of experimental responses (uncoded) for the composite films.
Formulation Mass fraction (wt % in dry film) Independent variables∗∗
no.
Ethanol ZN CS PVOH PEG400 Thickness (mm) TS (MPa) EM (MPa) EAB (%) WVP (g·mm/ Ws (%) WCA (O) ΔE (Y7) Opacity (AU/
(%) (Y1) (Y2) (Y3) m2·h·kPa) (Y4) (Y5) (Y6) mm) (Y8)

1 40 0.13 0.40 0.25 0.23 0.04 ± 0.003 30.07 ± 0.24 391.69 ± 1.98 108.9 ± 7.63 0.30 ± 0.02 43.10 ± 1.62 84.87 ± 0.95 7.67 ± 0.16 9.33 ± 0.09
2 40 0.22 0.20 0.25 0.31 0.03 ± 0.002 14.22 ± 0.72 139.77 ± 10.28 112.32 ± 4.02 0.29 ± 0.02 46.09 ± 1.76 58.95 ± 0.47 10.42 ± 0.55 15.42 ± 0.04
3 40 0.31 0.30 0.25 0.13 0.04 ± 0.009 31.6 ± 0.95 530.72 ± 23.84 59.3 ± 2.55 0.17 ± 0.01 49.41 ± 0.20 71.82 ± 0.39 12.8 ± 0.17 11.58 ± 0.01
4 40 0.13 0.10 0.50 0.23 0.03 ± 0.002 23.3 ± 0.60 196.15 ± 11.16 221.22 ± 2.84 0.16 ± 0.01 50.72 ± 0.76 61.23 ± 0.64 6.81 ± 0.51 11.75 ± 0.02
5 40 0.38 0.40 0.00 0.23 0.04 ± 0.004 25.72 ± 0.42 383.92 ± 6.83 33.59 ± 3.40 0.36 ± 0.00 34.65 ± 1.04 70.63 ± 1.52 16.9 ± 0.4 17.49 ± 0.67
6 40 0.39 0.10 0.25 0.23 0.05 ± 0.003 15.94 ± 0.19 164.9 ± 18.15 138.79 ± 0.64 0.32 ± 0.01 45.74 ± 2.67 56.38 ± 1.2 15.83 ± 0.25 11.06 ± 0.08
7 60 0.15 0.50 0.25 0.13 0.04 ± 0.007 37.47 ± 0.39 534.19 ± 3.76 103.15 ± 1.17 0.25 ± 0.01 39.60 ± 1.03 83.48 ± 2.09 8.86 ± 0.34 5.01 ± 0.04
8 60 0.19 0.60 0.00 0.23 0.05 ± 0.005 30.67 ± 0.04 447.07 ± 18.54 97.52 ± 5.06 0.45 ± 0.01 35.04 ± 0.68 87.4 ± 2.27 11.59 ± 0.28 13.42 ± 0.08
9 60 0.07 0.20 0.50 0.23 0.03 ± 0.005 25.72 ± 0.64 286.34 ± 7.06 154.29 ± 2.11 0.33 ± 0.02 44.90 ± 0.69 74.02 ± 2.91 3.97 ± 0.42 2.28 ± 0.11
10 60 0.11 0.30 0.25 0.31 0.05 ± 0.013 14.49 ± 0.17 180.35 ± 4.17 97.45 ± 4.79 0.40 ± 0.02 49.41 ± 2.31 69.03 ± 0.42 6.23 ± 0.31 4.93 ± 0.03
11 60 0.18 0.20 0.50 0.13 0.03 ± 0.001 39.2 ± 0.05 411.66 ± 33.38 156.1 ± 1.48 0.16 ± 0.01 44.84 ± 1.46 83.94 ± 0.54 8.95 ± 0.68 9.93 ± 0.13
12 60 0.26 0.30 0.25 0.23 0.04 ± 0.001 20.38 ± 0.37 259.87 ± 16.31 94.37 ± 1.58 0.30 ± 0.03 35.77 ± 0.42 69.44 ± 0.71 12.66 ± 0.31 8.56 ± 0.12
13 60 0.09 0.10 0.50 0.31 0.03 ± 0.001 19.71 ± 0.12 208.83 ± 18.39 200.5 ± 1.31 0.28 ± 0.01 61.28 ± 3.65 40.81 ± 0 4.83 ± 0.22 6.86 ± 0.12
14 60 0.34 0.30 0.00 0.31 0.04 ± 0.005 13.94 ± 0.22 149 ± 8.43 74.34 ± 3.25 0.37 ± 0.01 36.49 ± 1.62 76.35 ± 2.11 15.24 ± 0.49 12.58 ± 0.17
15 60 0.26 0.30 0.25 0.23 0.04 ± 0.002 20.32 ± 0.10 259.87 ± 16.30 94.37 ± 1.58 0.30 ± 0.03 41.24 ± 0.02 70.29 ± 0.49 11.7 ± 0.41 8.49 ± 0.12
16 60 0.43 0.40 0.00 0.13 0.04 ± 0.003 37.06 ± 0.31 613.08 ± 19.83 31.22 ± 0.98 0.24 ± 0.02 32.16 ± 0.87 76.4 ± 0.91 19.67 ± 0.18 19.32 ± 0.18
17 60 0.26 0.30 0.25 0.23 0.04 ± 0.001 20.26 ± 0.10 259.87 ± 16.30 94.37 ± 1.58 0.30 ± 0.03 38.16 ± 1.78 68.08 ± 2.52 12.16 ± 0.68 8.55 ± 0.07

275
18 60 0.20 0.10 0.50 0.23 0.03 ± 0.01 19.84 ± 0.30 150.12 ± 7.55 233.64 ± 3.35 0.23 ± 0.02 62.65 ± 1.39 53.7 ± 1.32 7.22 ± 0.29 9.85 ± 0.12
19 60 0.33 0.10 0.25 0.31 0.03 ± 0.003 7.85 ± 0.16 62.36 ± 3.89 131.9 ± 1.61 0.31 ± 0.01 49.77 ± 1.64 53.57 ± 0.64 11.84 ± 0.58 6.93 ± 0.18
20 60 0.46 0.20 0.25 0.13 0.03 ± 0.003 25.08 ± 0.28 396.69 ± 9.09 58.89 ± 3.74 0.14 ± 0.01 44.31 ± 0.87 73.27 ± 0.16 17.97 ± 0.29 8.26 ± 0.10
21 60 0.58 0.20 0.00 0.23 0.06 ± 0.002 7.01 ± 0.19 112.89 ± 4.15 25.94 ± 0.88 0.38 ± 0.01 24.48 ± 0.83 66.81 ± 2.31 22.83 ± 1.03 17.37 ± 0.11
22 80 0.13 0.40 0.25 0.23 0.04 ± 0.001 26.49 ± 0.26 328.78 ± 19.65 72.94 ± 1.34 0.44 ± 0.01 31.50 ± 0.55 68.94 ± 1.41 7.94 ± 0.31 8.37 ± 0.22
23 80 0.38 0.40 0.00 0.23 0.04 ± 0.001 20.44 ± 0.63 240.48 ± 10.59 79.28 ± 3.23 0.32 ± 0.04 30.97 ± 0.26 77.9 ± 0.4 17.7 ± 0.59 15.91 ± 0.39
24 80 0.31 0.30 0.25 0.13 0.04 ± 0.001 28.63 ± 0.80 466.45 ± 24.02 55.11 ± 1.50 0.21 ± 0.01 36.51 ± 0.99 74.75 ± 1.24 15.26 ± 0.23 13.26 ± 0.34
25 80 0.13 0.10 0.50 0.23 0.03 ± 0.009 16.56 ± 0.08 222.05 ± 11.88 111.24 ± 0.57 0.33 ± 0.02 48.46 ± 2.06 62.42 ± 1.62 6.8 ± 0.49 9.96 ± 0.29
26 80 0.22 0.20 0.25 0.31 0.03 ± 0.012 11.59 ± 0.05 108.21 ± 13.44 78.58 ± 2.05 0.35 ± 0.01 46.64 ± 1.07 61.42 ± 0.73 9.13 ± 0.37 8.72 ± 0.22
27 80 0.39 0.10 0.25 0.23 0.04 ± 0.009 8.07 ± 0.10 81.77 ± 7.01 79.88 ± 2.84 0.29 ± 0.03 45.13 ± 1.10 64.9 ± 0.83 15.28 ± 0.12 11.54 ± 0.34
Analysis of variance for Model (df, F-value, p-value) 8, 138.51, 10, 257.25, < 0.0001 13, 291.48, 8, 147.09, 12, 79.55, 7, 114.17, 8, 881.67, 10, 69.50,
determination of model < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001
fitting† Residual (df, ssq, msq) 72, 383.70, 70, 47,754, 682 67, 3868, 58 72, 0.03, 0 65, 390.69, 48, 320.66, 72, 18.48, 0.26 70, 119.03,
5.33 6.01 6.68 1.70
Lack of fit (df, F-value, p-value) 16, 152.13, 14, 12.09, < 0.0001 11, 37.01, < 0.0001 16, 2.8, < 0.002 12, 7.90, 17, 7.58, 16, 1.94, 14, 187.7,
< 0.0001 < 0.0001 < 0.0001 < 0.036 < 0.0001
Pure error (df, ssq, msq) 56, 8.63, 0.15 56, 118,873, 212 56, 468, 8 56, 0.02, 0 53, 140.14, 31, 62.19, 2.01 56, 11.9, 0.21 56, 2.48, 0.04
2.64
2 2
R , Ra (%) 93.90, 93.22 97.35, 96.97 98.26, 97.93 94.23, 93.59 93.63, 94.33, 93.51 98.99, 98.88 90.85, 89.54
92.45

∗
ZN, zein; CS, chitosan; PVOH, poly(vinyl alcohol); PEG400, poly(ethylene glycol) 400.
∗∗
All the values are means ± standard deviation, n = 3; TS, tensile strength; EM, Young's Modulus of elasticity; EAB, elongation at break; Ws solubility in water, ΔE, overall colour change.
†
ssq, sums of square; msq, mean square; df, degrees of freedom; p-values < 0.05 were considered to be significant; R2 coefficient of determination, Ra2 adjusted coefficient of determination.
Food Hydrocolloids 87 (2019) 270–286
S.G. Giteru et al.
Fig. 1. (a) SEM micrographs of composite films from selected formulations taken from the smooth (polystyrene contact side), (b) typical stress-strain curves depicting
the effect of blending on tensile properties. Numbers represent respective blend formulation.
Namieśnik, 2017). The coefficient of multiple determination R2
(squares of error/total sum of squares (SSR/SST) was used to describe
the proportion of the variation in Yi explained by the independent
variables X1 to X4 in the regression models. The presence of significant
differences between the predicted and the experimental values was
statistically determined using non–parametric Mann–Whitney U test
(Wilcoxon rank) (Yetilmezsoy et al., 2009) in SPSS (version 21, USA).
3. Results and discussions
3.1. Preparation of film-forming dispersions
The physical characteristics of the blends after 24 h storage at room
temperature (22 °C) are summarised in Table 1. The findings show the
presence of both segregated (repulsive interactions) and associated
(attractive interactions) phenomena across the blends (Fischer, 2013).
ZN may have settled in the form of yellow complex upon the reduction
of ethanol content (40%) (Wang & Padua, 2010), whereas PVOH, CS
and PEG400 phase separated as white lumps or gelling complexes at
high ethanol (80%) content. Meanwhile, despite the signs of separation
attributable to chitosan, PVOH and PEG400, the use of 60% ethanol
resulted in predominantly yellowish solution phase. This was a possible
indication of proper dispersion and solubilisation of ZN in the com-
plexes (Shukla & Cheryan, 2001). Indeed, stable and uniform solution
matrix were observed in complexes containing 60% ethanol and low
content of the molecules of the more hydrophilic compound such as
PVOH.
276
Food Hydrocolloids 87 (2019) 270–286
The thickness of the composite films ranged from 0.04 ± 0.008 mm
(Table 2) and significantly increased with an increase in the chitosan
levels in the film forming solutions. However, increasing the levels of
PVOH resulted in thinner films, which can be attributed to the devel-
opment of a compact and crystalline film microstructure at the high
PVOH content (Senna et al., 2010).
3.2. Characterisation of the composite films
3.2.1. The effect of blending on films morphology and mechanical properties
All blend formulations enabled preparation of continuous free-
standing composite films that did not present any apparent defects. The
surface morphology of the films was observed using FE-SEM as shown
in Fig. 1a. Incorporating PVOH or the combined effect of PVOH-
PEG400 resulted in a smooth morphology, but the surface acquired
groves and pinholes whose characteristics varied with the blend com-
position (Fig. 1a). For instance, the density of the pinholes and depth of
the groves increased at low CS and PVOH or high ZN and PEG400. The
compact structure observed at the high levels of PVOH (50% w/w of
total solids) (Film 11) and low ZN or CS levels were related to the films
exhibiting higher flexibility and low WVP. However, increasing weight
fraction (wi) PEG400 at the low or total absence of PVOH resulted in
more roughness, groves, and amorphous structure, bearing major
morphological defects and poor mechanical properties (Film 19). No-
tably, the defects were evident at the extremes of low CS or high ZN
contents. Moderate levels of ZN and CS seemed to interact in the ab-
sence of PVOH (Film 16) to acquire a compact and less defective
S.G. Giteru et al.
structure, which showed high TS but low %EAB. The findings on the
morphology of the films revealed the inner relationship between the
microstructure, physicochemical and mechanical properties of the
films.
Mechanical properties depict the effect of various components on
inter-/intra-molecular interactions and molecular chain mobility of the
film matrix. The findings on mechanical properties were described
using TS, EM and %EAB (Table 2). The data showed the existence of
two major categories of stress-strain curves (Fig. 1b). One category was
depicted by high brittleness as shown by linear stress-strain curves until
the point of fracture, and the second category showed various degrees
of toughness as evidenced by a yield point followed by an extensive
elongation at approximately constant stress levels (Senna et al., 2010).
Second order polynomial surface equations for predicting TS, EM and %
EAB (Y1, Y2, and Y3 respectively) were constituted as shown in Eqs.
(10)–(12). The linear-terms of all the independent variables were sig-
nificant (p < 0.05) and the overall fit (R2) accounted for ca. 93.90,
97.35 and 98.26% of the variation in TS, EM and %EAB respectively
(Table 2). The prediction models pointed to a decrease in the magnitude
of TS and EM with increasing levels of individual compounds. The
findings further pointed to the reliance on the interaction between
various factors (ZN:CS/PVOH and ZN:CS/PEG400) for improvements in
mechanical properties.
Y1 = 89.56 − 56.34X1 − 0.121X2 − 0.571X3 − 1.725X 4 + 0.005X32
+ 0.015X42 + 0.711X1 X3 + 0.384X1 X 4 (10)
Y2 = 1485.2 − 291X1 − 3.614X2 − 17.14X3 − 33.06X 4 − 307.2X21
− 0.0261X32 + 0.284X42 + 7.92X1 X3 + 0.085X2 X3 + 0.174X3 X 4 (11)
Y3 = −36.2 − 407.2X1 + 4.875X2 + 2.075X3 + 2.049X 4 + 174.3X21
− 0.020X22 + 0.041X23 − 0.032X24 − 1.148X1 X2 + 6.037X1 X3
+ 5.248X1 X 4 − 0.077X2 X3 − 0.025X2 X 4 (12)
Fig. 2 shows that increasing weight fraction zein (wiZN) (X1) at low
levels of other compounds resulted in a brittle film matrix, thus lower
TS. However, this effect was minimised by reducing ZN or increasing
the levels of the other blend components (Fig. 2a). At low wiZN and high
weight fraction chitosan (wiCS), the films acquired high crystallinity due
to the contribution from the rigid CS molecules and the interaction with
ZN, thus high TS and EM (Srinivasa et al., 2003). In the same manner,
PVOH (X3) molecules and the interaction between the polar regions of
zein, PVOH and CS could have resulted in the formation of crystalline
phases (Senna et al., 2010), thus the increasing TS at high PVOH and
CS. Such interactions increased polymer-polymer chain bonding
forming a more compact film structure (Escamilla-García et al., 2013).
On the contrary, TS and EM decreased due to PEG400 (X4) as shown in
Fig. 2(a and b). This may be related to the high levels of hydrophilic
moieties in the backbone of PEG400 molecular chains (Han &
Aristippos, 2005, pp. 239–262), which attracted water molecules to
form hydrodynamic macromolecule-water complexes leading to the
plasticization of the film. Plasticization of films increases their flex-
ibility due to internal lubrication that weakens polymer-polymer in-
teractions (e.g. hydrogen bonding, van der Waals, electrostatic or ionic
interactions) (Han & Aristippos, 2005, pp. 239–262).
The mean values for the %EAB are also summarised in Table 2. The
findings revealed the effect of blends formulations on %EAB, where
increasing levels of both PVOH and PEG400 resulted in higher flex-
ibility (Fig. 2c). As expected, decreasing levels of plasticizers resulted in
lower %EAB, in particular when high wiZN was included in the for-
mulations. The lack of films flexibility at high wiZN (low wiCS) may be
related to the brittleness due to the enhancement of intramolecular
hydrophobic interactions among asymmetric rods that make up the ZN
meshwork structure (Arcan & Yemenicioğlu, 2013). High levels of ZN
established stronger intramolecular hydrophobic interactions while
277
Food Hydrocolloids 87 (2019) 270–286
reducing intermolecular covalent cross-linking and molecular chain
mobility, thus forming a compact but brittle film structure (Masamba
et al., 2016; Suyatma et al., 2005).
Regarding the contribution of ethanol (X2) to the overall mechanical
properties, higher levels of the solvent led to lower TS and EM (Eqs.
(10) and (11)). As a major solvent for zein (Shukla & Cheryan, 2001),
ethanol was incorporated in the blends to allow the dispersion and
unfolding of the protein and increase its interaction with other mac-
romolecules. As expected, various levels of ethanol influenced the ionic
environment, solubility and aggregation behaviour of ZN molecules,
hence affecting the morphologies of ZN microphases (Wang & Padua,
2010). A high concentration of ethanol (∼80%) increased the hydro-
phobicity of the environment in the film forming dispersions (Wang &
Padua, 2010), which was depicted as cloudiness at wiZN > 0.40. The
precipitate was attributable to the separation of macromolecules re-
lying on hydrophilic interactions such as PVOH and CS (Bercea et al.,
2016). At low ethanol concentration (≈0.40 wt%), a more hydrophilic
environment and modification of the hydrophilic-hydrophobic char-
acter of the solution ensued (Zhang et al., 2007), thus influencing ag-
gregation of ZN molecules (Wang & Padua, 2010). Thereafter, gelation
of these compounds occurred as depicted by the formation of lumps in
the film forming dispersions and later occurrence of groves in the films
depicting low TS and EM (Fig. 1b). An increase in %EAB at approxi-
mately 60% ethanol (Fig. 2c) could be related to the improvement in
the solubility characteristics of ZN (Wang & Padua, 2010). Adequate
solubilisation of ZN created more inter-/intra-molecular interactions
between similar ZN-chains or with other compounds (Escamilla-García
et al., 2013; Masamba et al., 2016) resulting in homogenous and less
brittle film matrices Fig. 1a (Film 16). These findings indicate that the
solvent concentration, biopolymer ratio, and plasticization interacted to
influence the mechanical properties of the films.
3.2.2. Effect of blending on water vapour permeability
Water vapour permeability (WVP) is the diffusion of water mole-
cules against the exposed film area, thickness and vapour pressure
driving force. The WVP of biodegradable films is regulated by the re-
lative humidity (RH), plasticizer type and amount, the chemical
structure, molecular mass and degree of cross-linking of macro-
molecules. The thickness of the film influences the partial pressure on
the high humidity side of the film, thus impacting on the diffusion of
water molecules across the films (Masamba et al., 2016). This influence
was minimised by controlling the thickness at approximately
0.04 ± 0.008 mm. The second-order polynomial regression model
predicting WVP (Y4) was developed as shown in Eq. (13). The model
significantly (p < 0.05) explained 94.23% (R2) of the variation in
WVP.
Y4 = 0.059–0.223X1 + 0.003X2 – 0.008X3 + 0.018X4 + 0.556X12 –
2.200 × 10−4X42 – 0.008X1X2 + 1.04 × 10−4X2X3 (13)
High wiZN at any combination of the independent factors resulted in
lower WVP (Fig. 2d). Due to the predominantly hydrophobic nature of
ZN microstructures (Wang & Padua, 2010), the wetting of the films’
surface on the high humidity side was minimised at the high wiZN, thus
decreasing the rate of water vapour transmission (Table 2). Also at ca
60–80% ethanol (at high wiZN), a favourable hydrophilic-hydrophobic
environment was established allowing adequate dispersion of ZN and
the interaction between various compounds (Wang & Padua, 2010).
This resulted in the formation of a compact film structure with low
WVP. An increase in PVOH (Eq. (13)) can reduce the WVP, which is
related to the increase in the crystalline regions in the film. Crystallinity
originates from the PVOH molecules or its interaction with molecular
chains of other macromolecules (Argüello-García et al., 2014). Higher
WVP permeability rates were observed at high levels of PEG400
(Fig. 2d), perhaps due to its contribution to the hydrophilic nature and
reduced crystallinity of the film (Han & Aristippos, 2005, pp. 239–262).
S.G. Giteru et al.
Fig. 2. Response surface plots for (a) TS (Y1), (b) EM (Y2), (c) EAB (Y3) and (d) WVP (Y4) as a function of interactions between independent factors. Factors not
represented in each plot were kept constant at the centre levels.
The WVP values of the composite films obtained in the current study
ranged from 0.14 to 0.45 g mm/m2·h·kPa. These values agreed with the
earlier observed range 0.44–1.16 g mm/m2·h·kPa for ZN-oleic acid films
and transglutaminase treated ZN-oleic acid composite films respectively
(Masamba et al., 2016). Another study by Lacroix et al. (2014) had
earlier reported a range of 0.18–0.24 g mm/m2·h·kPa for unmodified
zein poly(vinyl alcohol) (PVA) films. The findings from the current
study showed that low WVP could be attained at low weight fraction
PEG400 (wiPEG400) irrespective of the combination with other blend
components.
3.2.3. Moisture content, solubility in water and surface hydrophobicity
Determination of the moisture content (MC) provides critical in-
formation on the stability of the composite films under varying relative
humidity environments. It also relates to the ability of the films to
maintain the moisture of the foods. Fig. 4a shows the relationship
278
Food Hydrocolloids 87 (2019) 270–286
between the moisture content, mechanical properties and blend for-
mulations. A high correlation (R2 89.13, p < 0.05) existed between the
blend components and the final MC of the composite films The obtained
values of MC (5–26%) were within the range observed for plasticised
biodegradable films (Suyatma et al., 2005). The findings by Suyatma
et al. (2005) reported an MC of 14.4–24.3% from chitosan-based films.
It was apparent that increasing ZN, PEG400 and PVOH components
resulted in lower moisture content in the films. The decline in MC at
high wiZN may be related to the higher amount of hydrophobic portions
from the protein, which reduced the film matrix-water interactions
(Hopkins, Chang, Lam, & Nickerson, 2015). However, such protein-
plasticizer combination resulted in higher plasticization effect (Liu,
Adhikari, Guo, & Adhikari, 2013) giving lower TS. On the contrary,
higher levels of MC resulted in higher TS (Fig. 4a), which was contrary
to the expected decrease in the magnitude of TS on plasticization with
water (Braga & Cunha, 2004). Additionally, increasing the levels of
S.G. Giteru et al.
PVOH and ZN (Fig. 4b) resulted in lower MC and higher %EAB. These
findings reveal possible crosslinking provided by zein and other com-
pounds in the blends. In addition, there was an evident contribution of
the small amounts of plasticizer in the internal lubrication of the film
matrix, to minimise the brittleness (Arcan & Yemenicioğlu, 2013).
The solubility of composite films in water is an important factor
determining the rate of breakdown during consumption and biode-
gradation (Laohakunjit & Noomhorm, 2004). Water solubility (Ws)
values (Table 2) represent the water-resistance capacity, where high
values indicate low water-resistance. The polynomial model for pre-
dicting Ws (Y5) (Eq. (14)) was highly significant (p < 0.0001) and
explained ∼ 93.63% (R2) of the variation in the observations.
Y5 = 125.600–93.9X1 – 1.205X2 – 0.532X3 – 1.940X4 + 42.1X12 +
0.004X22 + 0.002X32 + 0.023X42 + 0.550X1X2 + 1.132X1X3 +
0.010X2X4 + 0.008X3X4 (14)
Ws of the composite films decreased due to the linear terms of in-
dividual compounds but showed enhancement due to the interactions
and quadratic effects of various compounds. Considering the interac-
tions between the blend components, the surface diagram (Fig. 3a)
indicated that increasing wiZN at low (40%) or high (80%) ethanol led
to high Ws. At low ethanol content, there was a decrease in polymer-
polymer network formation from ZN due to the interruptions by the
predominantly reactive hydrophilic moieties (hydroxyl (OH−) or amine
(NH3+) groups) associated with ethanol, CS, PVOH and PEG400 (Yuan
et al., 2016). ZN aggregation and precipitation led to the development
Fig. 3. Response surface plots for (a) Ws (Y5), (b) ΔE (Y9) and (c) Op (Y10) as a function of interactions between factors of the polynomial model. Factors not
represented in each plot were kept constant at the centre levels.
279
Food Hydrocolloids 87 (2019) 270–286
of poor film matrix allowing rapid hydration and disintegration. Low
solubility films (24%) were obtained at ca 60% ethanol, perhaps due to
the resulting high interaction density between adequately dispersed
molecules of ZN with other compounds in the blend. The low solubility
was further enhanced at high wiZN (0.58%), which was attributable to
the high hydrophobic properties of ZN (Pena-Serna & Lopes Filho,
2015). Even though a highly soluble film is beneficial for cooking and
consumption of coated foods, moist foods or those requiring storage in a
moist environment may require a low solubility (Laohakunjit &
Noomhorm, 2004). Therefore, specific solubility attributes may be po-
sitive or negative, depending on the intended use of the edible packa-
ging.
The findings on the moisture content, water solubility and water
vapour permeability of the composite films were further elucidated by
assessing the wetting properties of the surface of the films. Data on
water contact angle (WCA) (Table 2) was used to evaluate the re-
sistance of the film against water and determining the hydrophobic
properties of the film surface. Generally, pure zein films are wetting
materials with WCA-values of 60° to 75° and plasticizers can impart
more hydrophilicity as observed in zein-glycerol films (WCA = 52°)
(Muthuselvi & Dhathathreyan, 2006). On the other hand, chitosan films
are more hydrophobic, WCA of 100° ± 5° (Silva et al., 2007). The
polynomial model predicting the WCA (Y6) (Eq. (15)) showed a de-
crease in the surface hydrophobicity with the increase in ZN component
of the ZN:CS ratio. Previous studies on composite films involving
chitosan in combination with proteins, for example, chitosan-whey
S.G. Giteru et al.
Fig. 4. Overlay of the contour plots to visualize (a) TS as a function of PEG400 and MC (labelled contours), (b) %EAB as a function of PVOH and MC (labelled
contours) and the effects of WCA (o) on WVP (labelled contours) as a function of (c) PEG400 and (d) PVOH at different ZN:CS ratios.
(Ferreira et al., 2009), chitosan-gelatin (Chen, Wang, Mao, Liao, &
Hsieh, 2008) and chitosan-soy (Silva et al., 2007) have reported a de-
crease in WCA with increasing protein content.
Y6 = 105.91–106.7X1 + 0.034X2 + 0.56X3 + 0.12X4 – 0.005X22 +
1.22X1X2 – 0.027X3X4 (15)
The model parameters showed that the interaction terms of ethanol
(X22) and PVOH/PEG400 (X3X4) resulted in lower WCA. The data also
revealed a decrease in WCA on increasing the content of ZN in the
composite films. Even though ZN molecules possesses higher propor-
tions of non-polar amino acid residues, the results showed its tendency
to increase the hydrophilicity of the films. This behaviour has pre-
viously been attributed to the reorientation of the polar chemical
components of the protein towards the hydrophilic (water) interface in
solution and during the self-assembly process (Muthuselvi &
Dhathathreyan, 2006). Fig. 4c shows that the interaction of the com-
pounds at low plasticization and lower levels of ZN gave higher values
of WCA and generally resulted in lower WVP. These findings show that
increasing hydrophobicity generally reduced the wetting ability and
therefore the diffusion rates of water vapour across the films (Han &
Aristippos, 2005, pp. 239–262). As opposed to PEG400, lower levels of
PVOH also resulted in high WVP, which corroborates the positive
contribution of PVOH to WCA according to Eq. (15). Cast PVOH films
were earlier reported to exhibit high WCA, which may have reduced the
wetting ability of the films (Zuo et al., 2013). According to Zuo et al.
(2013), the surface of PVOH films exhibits high hydrophobicity, with a
WCA of up to 84% after treatment in ethanol (deuterated ethanol,
99%).
3.2.4. Colour and light transmittance of the composite films
The colour and transparency of composite films can be used to de-
termine their acceptability to the consumer (Gaona-Sánchez et al.,
280
Food Hydrocolloids 87 (2019) 270–286
2015). Visual examination of the films revealed that they were trans-
lucent and high proportions of ZN resulted in more yellowness. The
overall change in colour attributes (ΔE) (Y7) was summarised in Table 2
and the prediction model in Eq. (16) (p < 0.05, R2 98.99%)
Y7 = –1.700 + 40.000X1 + 0.101X2 – 0.058X3 + 0.1608X4 – 10.11X12
– 0.320X1X3 – 0.233X1X4 – 0.003 X2X4 (16)
ΔE-values ranged from 3.97 to 22.83 across the blends and was
significantly influenced by the linear (X1, X3, X4), quadratic (X12) and
the interaction–terms (X1X3, X1X4 and X2X4) of the independent vari-
ables. ΔE values increased with the increase in wiZN and wiPEG400 and
decreased with the increase in weight fraction PVOH (wiPVOH) ethanol
content did not significantly (p > 0.05) modifying the overall colour
change of the films. The response surface diagrams (Fig. 3b) were su-
perimposable with those of b*-values (yellowness) (results not shown),
implying that most of the observed changes in colour could be ex-
plained by the change in the factors influencing the yellowness of the
composite films. The increase in the total colour difference due to the
increase in wiZN can be attributed to the natural colour and impurities
(carotenoids, xanthophylls, and other colour pigments) of commercial
ZN (Shukla & Cheryan, 2001; Zhang & Zhao, 2017).
Opacity (Op) is an indicator of the light barrier property of a film,
which describes the capacity of a film to shield a product against de-
terioration due to light. The effect of blending on opacity was sum-
marised in Table 2, where low Op values depict higher transparency
(Srinivasa et al., 2003). The polynomial model predicting the Op (Y8)
(Eq. (17)) showed that opacity increased with increasing wiZN, whereas
transparency increased with the increase in the concentration of
ethanol.
Y8 = 26.13 + 30.29X1 – 0.677X2 – 0.548X3 + 0.292X4 – 26.700X
1 + 0.007X2 + 0.005X3 + 0.145X1X3 – 0.007X2X4 + 0.002X3X4(17)
2 2 2
S.G. Giteru et al. Food Hydrocolloids 87 (2019) 270–286

Fig. 5. Typical curves depicting the effect of blending on thermal properties (a)
DSC thermograms (W/g) (temperature range 25–225 °C, heating rate 5 °C/min)
of pure compounds and selected formulations; (b) TGA thermograms weight
(%); (c) TGA thermograms derived weight (%/°C), (temperature range from 25
to 725 °C and heating rate 10 °C/min). Numbers represent respective formula-
tions, (n = 3).

The response surface plots indicated that low Op values were ob-
tainable at ca 60% ethanol (Fig. 3c). Ethanol concentration (60%)
prevented phase separation of ZN (Wang & Padua, 2010) and increased
intramolecular interactions between ZN molecules and intermolecular
associations with other constituents. These interactions resulted in a
continuous sheet of film network that increased the transmission of
light (Pena-Serna & Lopes Filho, 2015). On the contrary, opacity was
high at 80% ethanol suggesting that increasing non-polar environment
in the solution (Zhang et al., 2007) resulted in possible separation and
precipitation of hydrophilic compounds in the blend. This was also
possible at low ethanol content (40%), where ZN molecules separated
from other components in the blend. Both situations resulted in a dense
film matrix doped with aggregated compounds, resembling a hetero-
geneous structure that provided a high barrier to light transmission
(Pena-Serna & Lopes-Filho, 2013). Generally, ZN has been reported to
bestow opacity and yellowness on the films (Zhang & Zhao, 2017),
while its interaction with the solvent and other macromolecules causes
the variation between the formulations (Pena-Serna & Lopes Filho,
2015). Apart from ethanol, PVOH was the next greatest contributor to
the high transparency of the composite films (Fig. 3c). The interactions
between PVOH and PEG400 or ZN may increase the opacity due to the
compaction of the polymeric chain and increased crystallinity, leading
to the modification of the refractive index and interference with the
transmission of light (Ortega-Toro et al., 2014). Besides the inter-/intra-
molecular interactions involving PVOH, the reduction in the opacity of
the films containing low ZN content may be due to the simple dilution
effect of PVOH (Vanin, Sobral, Menegalli, Carvalho, & Habitante,
2005).
The findings of the current study showed that there were no com-
pletely opaque films. The lowest opacity (high transparency) values
(2.28 AU/mm) were observed at the lowest wiZN (0.07) and the highest
wiPVOH. The range of opacity values obtained (2.28–19.32 AU/mm)
were in good agreement with the Op of ZN–oleic acid–xanthan gum
(8.49 AU/mm) and ZN–oleic acid (5.19 AU/mm) films reported by
Pena-Serna & Lopes Filho. (2015).

3.3. The effect of independent variables on thermal properties

Differential scanning calorimetry (DSC) was conducted to char-

acterise the thermal characteristics of the composite films in relation to
their microstructure. The method was also useful in examining the
miscibility and compatibility of macromolecules (Sun et al., 2007).
Typical DSC thermograms of films prepared from pure ingredients and
a selected combination of independent factors are shown in Fig. 5a.
Regarding the films prepared from pure ingredients, the endothermic
peak of CS at 154.86 °C was in the range of documented values (Sun
et al., 2007). Pure ZN (uncentrifuged), gave an endothermic peak at
163.69 °C, whereas centrifuged ZN showed a peak at (178.22 °C), which
were similar to the documented values 165 °C (Shukla & Cheryan,
2001) and 170–180 °C (Senna et al., 2010; Yu et al., 2017) respectively.
The high endothermic peak depicted by centrifuged ZN, imply that
centrifugation (10,000 × g at 5 °C) improved the purity of ZN (Sun
et al., 2017b). Generally, ZN films had higher melting temperatures
(TM) compared to other ingredients, which suggested the existence of a
compact molecular structure arising from the assembly of small sphe-
rical particles that required a higher temperature to destroy (Wang &
Padua, 2010).
Compared to the melting of pure ingredients, endothermic peaks of

281
S.G. Giteru et al. Food Hydrocolloids 87 (2019) 270–286

the blends were shifted towards the higher temperatures (Table 3). The
higher temperature requirements by the composite films demonstrated
the need for more activation energy to destroy the developed compact
structures of the films (Senna et al., 2010). Low ethanol content (40%
v/v), high wiPVOH/wiPEG400 or low wiZN generated shallow curves dis-
playing indistinct endothermic peaks (Fig. 5a). The broader en-
dothermic peaks in these formulations suggested the existence of a
more hydrophilic environment in the FFS due to the low ethanol, high
wiPEG400 and wiPVOH in the blend. The interaction between the film
components in a hydrophilic environment possibly resulted in the for-
mation of interstices during complexing and subsequent development
of amorphous structures (Sun et al., 2007). Consequently, strong inter
-/intra-molecular hydrogen (H-H) linkages and rigid molecular chain
backbone were formed, but these degraded before reaching the melting
point, thus undefined melting peaks (Sun et al., 2007). With regard to
PEG400, reduced enthalpy (ΔH) values suggested the weakening of
intermolecular interactions (e.g. CS-ZN or CS-PVOH) and perhaps in-
creasing the mobility of the macromolecules in the amorphous regions,
thus weakening the film microstructure (Sun et al., 2007).
Nevertheless, at low PEG400 and in the absence of PVOH the
melting peaks shifted towards higher temperatures. In addition, low
levels of PVOH or its absence led to a decrease in the amorphous state
of the films resulting in the formation of denser matrices and distinct
degradation patterns (Senna et al., 2010). DSC traces of the composite
films with low or no PVOH depicted the melting of a homogenous/pure
substance, suggesting proper complexation or miscibility of the in-
gredients (Sun et al., 2007). Increasing wiZN and weight fraction CS
(wiCS) ≥ 0.31 resulted in higher melting temperatures without ne-
cessarily increasing ΔH values. This increase in the TM is attributable to
the strong intermolecular interaction through electrostatic complexa- Fig. 6. XRD diffraction patterns of films from unmodified zein (powdered), CS,
tion of ZN and CS (Park, Park, & Kim, 2015; Yu et al., 2017). PVOH and selected blend formulations.
High ethanol content induced possible precipitation of polar hy-
drophilic macromolecules resulting in poor miscibility as reflected in (Fig. 5b). The initial weight loss at 100 °C may be attributed to the
the multiple melting peaks on the DSC curve. The shifting of melting removal of water and decomposition of PEG400 (Nakano et al., 2007).
temperatures and the variations between formulation types collabo- As far as the blend composition is concerned, the absence of PVOH in
rated the changes in TS, EM, %EAB (Table 2) and changes in the XRD the blends increased the thermal instability (Fig. 5c). The total mass
patterns (Fig. 6). loss (%) shown in Table 4 decreased with decreasing wiPVOH, which was
also related to the increase in the wiZN and wiCS (Table 1). This ob-
servation suggested the development of thermostable film matrices due
3.4. Thermogravimetric analysis of the composite films
to the formation of dense film network between ZN, CS and PEG400
and limited polymer chain mobility due to lack of interferences from
Thermal stability of the composite films was characterised using
PVOH (Nakano et al., 2007). Notably, formulation eight had the highest
thermogravimetric analysis (TGA) as shown in Fig. 5b and c. Thermal
wiCS in the design, which led to an early onset of degradation. However,
decomposition of the films occurred in two distinct phases; the first
the solid network of the film due to ZN-PEG400 interactions reduced
phase occurred at ca.100 °C and the second between 260 and 440 °C

Table 3
Phase transition temperature data from DSC analysis for selected film formulations.a
Formulation no. First peak Second peak

TO (°C) TP (°C) TP (°C) ΔH (J/g)

a
3 180.14 ± 1.49 nd. 207.3 ± 1.65 34.29 ± 3.21a
5 188.47 ± 0.16 188.78 ± 1.89 209.95 ± 0.11ab 93.60 ± 2.03cp
8 174.43 ± 0.62 nd. 174.58 ± 0.63cd 102.9 ± 1.85p
9 191.41 ± 0.29 nd. 212.17 ± 1.08b 86.18 ± 2.61cfg
16 186.31 ± 2.15 186.53 ± 1.08 192.94 ± 1.05e 99.46 ± 1.51p
19 143.22 ± 0.21 159.55 ± 1.13 159.55 ± 2.73g 52.33 ± 2.33d
24 175.12 ± 0.09 nd. 190.46 ± 1.05e 76.89 ± 4.25 fg
26 171.29 ± 0.71 149.16 ± 0.87 171.43 ± 0.64c 55.07 ± 3.45de
Chitosan 151.95 ± 0.84 nd. 154.86 ± 1.37h 273.4 ± 7.25n
PVOH 154.13 ± 2.22 nd. 165.85 ± 2.38f 54.73 ± 3.88d
Zein (centrifuged) 171.50 ± 0.07 172.08 ± 1.74 178.22 ± 1.02d 62.03 ± 4.22de
Zein (uncentrifuged) 162.49 ± 0.06 nd. 163.69 ± 1.58fg 82.78 ± 3.10g
PEG400 99.3 ± 0.13b nd. 143.26 ± 3.21b 21.17 ± 2.39b

a
All values ± standard deviation, n = 3, values in the same column bearing different letter are significantly different (P < 0.05); TO, onset temperature; TP, peak
temperature; ΔH, enthalpy; nd., peak not detected.
b
Exothermic peak.

282
S.G. Giteru et al. Food Hydrocolloids 87 (2019) 270–286

Table 4 changes observed in the physicochemical, mechanical and barrier

Thermal stability data for selected film formulations. properties of the composite films.
Formulation no. Thermal degradation (stage II)†
3.5. The validity of regression models and optimal film formation
To (°C) Te (°C) Mass loss (%)

3 266.84 ± 3.95a 386.58 ± 3.06a
5 265.08 ± 1.53a 377.29 ± 2.84bc
8 261.54 ± 0.91b 353.4 ± 3.91e
12 262.15 ± 0.27b 374.6 ± 2.87b
13 264.59 ± 2.74a 396.44 ± 0.11d
16 272.22 ± 0.42c 373.18 ± 3.77b
26 265.42 ± 1.02a 380.06 ± 0.51c
†
All values ± standard deviation, n = 3, values in the same column bearing
different superscript letters are significantly different (p < 0.05).
To: onset temperature, Te: end temperature.
the overall degradation. These findings concur with the high TS, EM,
and low %EAB and Ws observed for the formulations (Table 2). The
findings from this study suggest that ZN films modified using CS and
PEG400 have more thermo-stability compared to PVOH modified films.
3.4.1. X-ray diffraction (XRD)
Changes to the crystalline structure of the composite films were
determined using X-ray diffraction (XRD). The XRD patterns of films
from pure compounds and selected formulations are shown in Fig. 6. It
was seen that zein did not display apparent diffraction peaks but two
broad regions were observed at 2θ = 8.8° and 20.5° depicting the
amorphous nature of the protein (Sun et al., 2017a). Pure chitosan, a
partially crystalline polysaccharide, showed two diffractions peaks at
2θ = 11.6° and 18.9°, which are attributable to its regular chain. PVOH
showed a diffraction peak at 2θ = 19.6°. A comparison between XRD
patterns of different blends indicates possible crosslinking and altera-
tion of the peak position or replacement with an amorphous region. For
instance, Film 16 (Fig. 6) shows that the interaction of various com-
pounds (without PVOH) resulted in more crystallinity of the blend,
which was exhibited as enhancement of the peaks 11.6° and 18.3°.
These findings suggest the positive contribution of CS to the overall
crystallinity, mechanical and thermal properties of the films. Likewise,
the incorporation of PVOH enhanced the films crystallinity (Film 11, 12
and 21), but the chitosan peak in these films (2θ = 11.6°) changed to a
broad amorphous region (Film 12, 19 and 21), whereas the peak at
18.9° was shifted to the right. In addition, these interactions minimised
the influence of the zein amorphous structure as shown by the non-
existent peak regions of unblended zein. The changes are attributable to
the intermolecular interactions of the polar zein moieties and hydroxyl
residues of CS and PVOH forming intermolecular hydrogen bonding
(Senna et al., 2010). The crystallinity data from XRD validates the
80.68 ± 1.32a Experimental responses of the predicted values indicated the suit-
73.89 ± 1.42b ability of the methodology developed for the optimisation of the com-
66.67 ± 1.24c posite films, and validity of the surface responses obtained by the
81.5 ± 1.8a
Box–Behnken experimental design as shown in Table 5a. The agreement
89.35 ± 0.42d
65.23 ± 0.6c between the experimental data and the predicted values confirmed the
83.45 ± 0.12a predictive ability of the RSM model according to the optimised condi-
tions.
The physicochemical properties of the optimised formulations in-
vestigated in both dry and swollen states were summarised in Table 5(a,
b). Increasing wiCS or reducing wiPVOH caused a rapid increase in the
films swelling index in water (Sw) irrespective of wiZN (Fig. 7). The high
swelling index at high wiCS or in the absence PVOH is attributable to the
increase in the molecular relaxation of the composite film due to a
possible reduction in crystalline regions (Ortega-Toro et al., 2014). It
was interesting to note that the swelling index in deionised water (pH
7) and hydrochloric acid (0.1 M, pH 2) was 6–45 times higher than in
phosphate buffer (0.1 M, pH 7.5) (swelling index in buffer (Sb),
2.9–5.83%). This was attributable to the responses of the compounds to
the negatively charged ions (PO43− and −OH) at high pH (Sağlam,
Venema, de Vries, & van der Linden, 2013). The buffer (pH 7.5) was
above the isoelectric point (pI) of ZN and CS (≈pH 6.2) (Bharmoria,
Singh, & Kumar, 2013; Sun et al., 2017a), which reduced the degree of
ionisation and led to acquisition of negatively charged ions on mole-
cular surfaces (Sağlam et al., 2013). The reduced degree of ionisation or
high net negative charge increased the tendency of self-aggregation
through inter-/intra-molecular hydrophobic or electrostatic interac-
tions, resulting in the development of a more compact film structure
(Wang, Jiang, Duan, & Shao, 2013), therefore less swelling (Fig. 7A and
B). In earlier studies, Wang et al. (2013) observed a decrease in the
swelling index of gelatin films at pH above 8.0. The data showed that
the swelling behaviour of some formulations in acid (Sa) could not be
determined due to their rapid disintegration. Soaking the films in HCl
(high H+) favoured molecular relaxation and disentanglements leading
to the breakdown of film networks, thus rapid swelling and dissolution
of the film (Masamba et al., 2016).
The TS, EM and EAB at the dried state increased with increasing CS
(Table 5a) but TS and EM decreased by 30–70% and 95% respectively
in the swollen state. However, the swollen state elongation at break in
the buffer (%EABb) remained constant or increased by ca 60%
(Table 5b). Films containing high wiCS (wiZN/wiCS 0.25/0.52) showed
the highest swelling (Sb) 6.6% but still resulted in the highest swollen
state tensile strength (TSb) and elongation (%EABb), 24.56 MPa and

Table 5a
Experimental validation of the optimal conditions for the development of edible films.a
Optimised Mass fraction (weight % Predicted (95% confidence limit)c Validated experimental results (n = 3)
films in dry film)b

ZN CS PVOH TS (MPa) EM (MPa) EAB (%) Ws (%) TS (MPa) EM (MPa) EAB (%) Ws (%)
(%) (%)

1opt 0.25 0.39 0.13 28.97 ± 1.36 439.76 ± 16.78 72.69 ± 6.02 34.13 ± 2.13 30 ± 0.88 412.38 ± 28.78 75.15 ± 3.25 35.52 ± 2.90
2opt 0.35 0.29 0.13 23.18 ± 1.18 368.15 ± 16.79 55.77 ± 6.09 32.99 ± 2.22 24.21 ± 0.80 356.62 ± 15.26 84.23 ± 1.63 33.71 ± 3.11
3opt 0.25 0.27 0.25 24.94 ± 1.2 357.51 ± 17.12 84.38 ± 6.63 36.81 ± 2.41 24.84 ± 0.99 349.51 ± 26.65 78.8 ± 4.71 36.56 ± 3.70
4opt 0.25 0.52 0.00 34.72 ± 2.27 528.51 ± 26.8 82.86 ± 7.71 34.4 ± 2.59 34.58 ± 1.11 530.81 ± 25.21 83.38 ± 2.57 34.14 ± 1.85
5opt 0.35 0.42 0.00 28.55 ± 1.67 459.17 ± 22.05 55.21 ± 6.29 30.8 ± 2.28 29.61 ± 0.99 456.76 ± 27.08 55.17 ± 3.67 31.88 ± 2.56

a
All values ± standard deviation, n = 3, ZN, zein; CS, chitosan; PVOH, poly(vinyl alcohol).
b
Ethanol concentration held at 60% (v/v of film-forming solution); poly(ethylene glycol) 400 held at mass fraction 0.23% (w/w in dry film).
c
TS, tensile strength; EM, Young's Modulus of elasticity; EAB, elongation at break; Ws solubility in water.

283
S.G. Giteru et al. Food Hydrocolloids 87 (2019) 270–286

Fig. 7. (A) Swelling characteristics of optimised films in deionised water after 2 h incubation at 20 °C; dry sample (a); 1opt (b); 2opt (c); 3opt (d); 4opt (e). (B) SEM
micrograph of optimised film (2opt).

Table 5b
Swollen state properties of optimised films.
Optimised films Swelling index (%) Swollen state tensile propertiesa,b

Water Phosphate buffer Hydrochloric acid TS (MPa) EM (MPa) EAB %

1opt 136.27 ± 2.285a 2.90 ± 0.794a n.d. 11.48 ± 0.66a 13.80 ± 0.8a 71.17 ± 10.266a
2opt 18.05 ± 0.071b 2.43 ± 0.404a 184.17 ± 8.036a 18.61 ± 0.945b 14.84 ± 1.341a 127.74 ± 3.272b
3opt 31.93 ± 1.361c 3.60 ± 0.656a n.d. 5.03 ± 0.238c 13.63 ± 1.202ab 46.25 ± 2.728c
4opt 230.5 ± 7.089d 6.60 ± 0.648b 198.07 ± 3.002b 24.56 ± 0.704d 11.01 ± 1.327b 144.51 ± 4.115bd
5opt 149.33 ± 5.132e 5.83 ± 0.208b n.d. 10.96 ± 0.436a 10.85 ± 0.648b 86.72 ± 10.709a

abcde
values in the same column bearing different superscripts are significantly different (p < 0.05); n.d. not determined.
a
TS, tensile strength; EM, Young's Modulus of elasticity; EAB, elongation at break.
b
Mean thickness (swollen film) = 0.02 ± 0.005 mm.

144.51% respectively. These findings are likely due to the absence of
PVOH in the blend, since in the absence of PVOH molecules, the films
acquired less hydrophilic surface in the buffer, perhaps easing the polar
tension on the surface of ZN, CS and their cross-linked structures (Park
et al., 2015). As a consequence, the interactions between ZN and CS
through electrostatic linkages became stronger, enabling the formation
of a denser film network (Fig. 7b) (Park et al., 2015).
With regards to the soaking of the composite films in the buffer at
pH near the pI of ZN and CS, the possible burying of the hydrophobic
regions occurred, which minimised water-macromolecule/ions inter-
actions (Escamilla-García et al., 2013). Accordingly, the relaxation and
weakening of the film was minimised, which preserved its integrity thus
a high TS and %EAB, which contrasted dipping in water and HCl (re-
sults not shown). In addition, the high CS content imparted greater
chain mobility and high elongation in the swollen state and subse-
quently reducing EM (11.01 MPa). Optimum films (Fig. 7b) were ob-
tained under the weight fractions ZN/CS/PVOH, 0.35/0.29/0.13 (0.23
wiPEG400 and 60% ethanol). The dry state mechanical properties (TS,
EM, EAB) were 24.21 MPa, 356.62 MPa, 84.23% respectively. The
swollen state mechanical properties of the optimal formulation in buffer
TSb, EMb, EABb, Sb and Ws were 18.61 MPa, 14.84 MPa, 127.74%,
2.43% and 33.71% respectively.
4. Conclusions
Blending and solvent properties had a statistically significant effect
on physicochemical and mechanical properties of the composite films.
Based on the SEM micrographs of the freestanding films, high levels of
ZN and PEG400 resulted in roughness with major film defects, but these
were minimised at high levels of CS or PVOH. In addition, XRD analysis
indicated potential cross-linking interactions through hydrogen
bonding between the polar groups of ZN and hydroxyl moieties of
PVOH and CS. These findings were in agreement with the data from
stress-strain curves, WVP, Ws and WCA. Increasing polar residues in the
film forming solution at low ethanol content or increasing non-polar
environment at high ethanol content showed adverse effects on the
responses. This study demonstrated the influence of critical interactions
between solvent features (polarity) and interfacial properties of ZN, CS,
PVOH and PEG400. The observed crosslinking behaviour suggested the
effectiveness of blending technique in improving the compatibility of
biopolymers and overall functionality of edible films.
Acknowledgements
This research was supported by the University of Otago Doctoral
scholarship awarded to Stephen Giteru and Riddet Institute, a New

284
S.G. Giteru et al.
Zealand Centre of Research Excellence funded by the New Zealand
Tertiary Education Commission.
Appendix A. Supplementary data
Supplementary data related to this article can be found at https://
doi.org/10.1016/j.foodhyd.2018.08.006.
References
AOAC (2005). Official method 925.10: Solids (total) and loss on drying (moisture) in flour.
Official Methods of Analysis of AOAC International.
Arcan, I., & Yemenicioğlu, A. (2013). Development of flexible zein–wax composite and
zein–fatty acid blend films for controlled release of lysozyme. Food Research
International, 51(1), 208–216.
Argüello-García, E., Solorza-Feria, J., Rendón-Villalobos, J. R., Rodríguez-González, F.,
Jiménez-Pérez, A., & Flores-Huicochea, E. (2014). Properties of edible films based on
oxidized starch and zein. International Journal of Polymer Science, 2014, 1–9.
Ashogbon, A. O., & Akintayo, E. T. (2014). Recent trend in the physical and chemical
modification of starches from different botanical sources: A review. Starch - Stärke,
66(1–2), 41–57.
ASTM (2003). Standard D638 − 02a: Standard test method for tensile properties of thin
plastics. PA, USA: ASTM International.
ASTM (2013a). Standard D3418 - 15: Standard test method for transition temperatures and
enthalpies of fusion and crystallization of polymers by differential scanning calorimetry.
USA: ASTM International.
ASTM (2013b). Standard E96/E96M−12: Standard test methods for water vapor transmis-
sion of materials. USA: ASTM International.
Azeredo, H. M. C., & Waldron, K. W. (2016). Crosslinking in polysaccharide and protein
films and coatings for food contact – a review. Trends in Food Science & Technology,
52, 109–122.
Bercea, M., Morariu, S., & Teodorescu, M. (2016). Rheological investigation of poly(vinyl
alcohol)/poly(N-vinyl pyrrolidone) mixtures in aqueous solution and hydrogel state.
Journal of Polymer Research, 23(7).
Bharmoria, P., Singh, T., & Kumar, A. (2013). Complexation of chitosan with surfactant
like ionic liquids: Molecular interactions and preparation of chitosan nanoparticles.
Journal of Colloid and Interface Science, 407, 361–369.
Braga, A. L. M., & Cunha, R. L. (2004). Plasticization and antiplasticization by small
molecules in brittle cellular food: TMDSC and mechanical properties. International
Journal of Food Properties, 7(1), 105–120.
Cheng, S.-Y., Wang, B.-J., & Weng, Y.-M. (2015). Antioxidant and antimicrobial edible
zein/chitosan composite films fabricated by incorporation of phenolic compounds
and dicarboxylic acids. Lebensmittel-Wissenschaft und -Technologie- Food Science and
Technology, 63(1), 115–121.
Chen, C.-H., Wang, F.-Y., Mao, C.-F., Liao, W.-T., & Hsieh, C.-D. (2008). Studies of chit-
osan: II. Preparation and characterization of chitosan/poly(vinyl alcohol)/gelatin
ternary blend films. International Journal of Biological Macromolecules, 43(1), 37–42.
Chen, C., Zhong, M., Li, G., Yang, F., Huang, R., Xiao, W., et al. (2016). Optimization on
preparation conditions of calcium-crosslinked modified chitosan as potential matrix
material for theophylline sustained-release beads and its evaluation of release ki-
netics. Journal of Alloys and Compounds, 658, 348–355.
Chiellini, E., Corti, A., D'Antone, S., & Solaro, R. (2003). Biodegradation of poly (vinyl
alcohol) based materials. Progress in Polymer Science, 28(6), 963–1014.
Chmiel, T., Kupska, M., Wardencki, W., & Namieśnik, J. (2017). Application of response
surface methodology to optimize solid-phase microextraction procedure for chro-
matographic determination of aroma-active monoterpenes in berries. Food Chemistry,
221, 1041–1056.
Escamilla-García, M., Calderón-Domínguez, G., Chanona-Pérez, J. J., Farrera-Rebollo, R.
R., Andraca-Adame, J. A., Arzate-Vázquez, I., et al. (2013). Physical and structural
characterisation of zein and chitosan edible films using nanotechnology tools.
International Journal of Biological Macromolecules, 61, 196–203.
Escamilla-García, M., Calderón-Domínguez, G., Chanona-Pérez, J. J., Mendoza-Madrigal,
A. G., Di Pierro, P., García-Almendárez, B. E., et al. (2017). Physical, structural,
barrier, and antifungal characterization of chitosan-zein edible films with added es-
sential oils. International Journal of Molecular Sciences, 18(11).
FDA (2013). US food and drug administration, generally recognized as safe (GRAS) substance
under the US FDA regulation. GRAS Notice No. GRN 443, FDA response letter, 13
March 2013. Retrieved from https://www.accessdata.fda.gov/scripts/fdcc/index.
cfm?set=GRASNotices&id=443&sort=GRN_No&order=DESC&startrow=1&
type=basic&search=chitosan.
Ferreira, C. O., Nunes, C. A., Delgadillo, I., & Lopes-da-Silva, J. A. (2009).
Characterization of chitosan-whey protein films at acid pH. Food Research
International, 42(7), 807–813.
Fischer, P. (2013). Rheology of interfacial protein-polysaccharide composites. The
European Physical Journal - Special Topics, 222(1), 73–81.
Gaona-Sánchez, V. A., Calderón-Domínguez, G., Morales-Sánchez, E., Chanona-Pérez, J.
J., Velázquez-de la Cruz, G., Méndez-Méndez, J. V., et al. (2015). Preparation and
characterisation of zein films obtained by electrospraying. Food Hydrocolloids, 49,
1–10.
Ghanbarzadeh, B., & Oromiehi, A. R. (2008). Biodegradable biocomposite films based on
whey protein and zein: Barrier, mechanical properties and AFM analysis. International
Journal of Biological Macromolecules, 43(2), 209–215.
Gutiérrez, T. J., & Álvarez, K. (2016). Physico-chemical properties and in vitro
285
Food Hydrocolloids 87 (2019) 270–286
digestibility of edible films made from plantain flour with added Aloe vera gel.
Journal of Functional Foods, 26, 750–762.
Han, J. H., & Aristippos, G. (2005). 15 - edible films and coatings: A review Innovations in
food packaging. London: Academic Press.
Hopkins, E. J., Chang, C., Lam, R. S. H., & Nickerson, M. T. (2015). Effects of flaxseed oil
concentration on the performance of a soy protein isolate-based emulsion-type film.
Food Research International, 67, 418–425.
Lacroix, M., Khan, R., Senna, M., Sharmin, N., Salmieri, S., & Safrany, A. (2014).
Radiation grafting on natural films. Radiation Physics and Chemistry, 94, 88–92.
Laohakunjit, N., & Noomhorm, A. (2004). Effect of plasticizers on mechanical and barrier
properties of rice starch film. Starch - Stärke, 56(8), 348–356.
Liang, H., Zhou, B., He, L., An, Y., Lin, L., Li, Y., et al. (2015). Fabrication of zein/qua-
ternized chitosan nanoparticles for the encapsulation and protection of curcumin.
RSC Advances, 5(18), 13891–13900.
Liu, H., Adhikari, R., Guo, Q., & Adhikari, B. (2013). Preparation and characterization of
glycerol plasticized (high-amylose) starch–chitosan films. Journal of Food Engineering,
116(2), 588–597.
Luo, Y., Zhang, B., Cheng, W.-H., & Wang, Q. (2010). Preparation, characterization and
evaluation of selenite-loaded chitosan/TPP nanoparticles with or without zein
coating. Carbohydrate Polymers, 82(3), 942–951.
Masamba, K., Li, Y., Hategekimana, J., Liu, F., Ma, J., & Zhong, F. (2016). Effect of Gallic
acid on mechanical and water barrier properties of zein-oleic acid composite films.
Journal of Food Science & Technology, 53(5), 2227–2235.
Müller, V., Piai, J. F., Fajardo, A. R., Fávaro, S. L., Rubira, A. F., & Muniz, E. C. (2011).
Preparation and characterization of zein and zein-chitosan microspheres with great
prospective of application in controlled drug release. Journal of Nanomaterials, 2011,
928728.
Muthuselvi, L., & Dhathathreyan, A. (2006). Contact angle hysteresis of liquid drops as
means to measure adhesive energy of zein on solid substrates. Pramana, 66(3),
563–574.
Nakano, Y., Bin, Y., Bando, M., Nakashima, T., Okuno, T., Kurosu, H., et al. (2007).
Structure and mechanical properties of chitosan/poly(vinyl alcohol) blend films.
Macromolecular Symposia, 258(1), 63–81.
Ortega-Toro, R., Jimenez, A., Talens, P., & Chiralt, A. (2014). Properties of starch-hy-
droxypropyl methylcellulose based films obtained by compression molding.
Carbohydrate Polymers, 109, 155–165.
Park, C.-E., Park, D.-J., & Kim, B.-K. (2015). Effects of a chitosan coating on properties of
retinol-encapsulated zein nanoparticles. Food Science and Biotechnology, 24(5),
1725–1733.
Pena-Serna, C., & Lopes Filho, J. F. (2015). Biodegradable zein-based blend films:
Structural, mechanical and barrier properties. Food Technology and Biotechnology,
53(3), 348–353.
Pena-Serna, C., & Lopes-Filho, J. F. (2013). Influence of ethanol and glycerol con-
centration over functional and structural properties of zein-oleic acid films. Materials
Chemistry and Physics, 142(2–3), 580–585.
Sağlam, D., Venema, P., de Vries, R., & van der Linden, E. (2013). The influence of pH and
ionic strength on the swelling of dense protein particles. Soft Matter, 9(18),
4598–4606.
Selling, G. W., Woods, K. K., & Biswas, A. (2011). Electrospinning formaldehyde-cross-
linked zein solutions. Polymer International, 60(4), 537–542.
Senna, M. M., Salmieri, S., El-naggar, A.-W., Safrany, A., & Lacroix, M. (2010). Improving
the compatibility of zein/poly(vinyl alcohol) blends by gamma irradiation and graft
copolymerization of acrylic acid. Journal of Agricultural and Food Chemistry, 58(7),
4470–4476.
Shen, Z., & Kamdem, D. P. (2015). Development and characterization of biodegradable
chitosan films containing two essential oils. International Journal of Biological
Macromolecules, 74, 289–296.
Shukla, R., & Cheryan, M. (2001). Zein: The industrial protein from corn. Industrial Crops
and Products, 13(3), 171–192.
Silva, S. S., Goodfellow, B. J., Benesch, J., Rocha, J., Mano, J. F., & Reis, R. L. (2007).
Morphology and miscibility of chitosan/soy protein blended membranes.
Carbohydrate Polymers, 70(1), 25–31.
Srinivasa, P. C., Ramesh, M. N., Kumar, K. R., & Tharanathan, R. N. (2003). Properties and
sorption studies of chitosan–polyvinyl alcohol blend films. Carbohydrate Polymers,
53(4), 431–438.
Sun, C., Dai, L., & Gao, Y. (2017a). Formation and characterization of the binary complex
between zein and propylene glycol alginate at neutral pH. Food Hydrocolloids, 64,
36–47.
Sun, C., Xu, C., Mao, L., Wang, D., Yang, J., & Gao, Y. (2017b). Preparation, character-
ization and stability of curcumin-loaded zein-shellac composite colloidal particles.
Food Chemistry, 228, 656–667.
Sun, G., Zhang, X. Z., & Chu, C. C. (2007). Formulation and characterization of chitosan-
based hydrogel films having both temperature and pH sensitivity. Journal of Materials
Science: Materials in Medicine, 18(8), 1563–1577.
Suyatma, N. E., Tighzert, L., & Copinet, A. (2005). Effects of hydrophilic plasticizers on
mechanical, themal and surface properties of chitosan films. Journal of Agricultural
and Food Chemistry, 53, 3950–3957.
Tillekeratne, M., & Easteal, A. J. (2000). Modification of zein films by incorporation of
poly(ethylene glycol)s. Polymer International, 49(1), 127–134.
US FDA (2018). Title 21-food and drugs: Parts 1 to 1499 (eCFR) electronic code of federal
regulations. Washington, DC: U.S. Food and Drug Adminstration.
Vanin, F. M., Sobral, P. J. A., Menegalli, F. C., Carvalho, R. A., & Habitante, A. M. Q. B.
(2005). Effects of plasticizers and their concentrations on thermal and functional
properties of gelatin-based films. Food Hydrocolloids, 19(5), 899–907.
Wang, Y., Jiang, L., Duan, J., & Shao, S. (2013). Effect of the carbonyl content on the
properties of composite films based on oxidized starch and gelatin. Journal of Applied
S.G. Giteru et al.
Polymer Science, 130(5), 3809–3815.
Wang, Y., & Padua, G. W. (2010). Formation of zein microphases in ethanol-water.
Langmuir, 26(15), 12897–12901.
Wright, E. J., Andrews, G. P., McCoy, C. P., & Jones, D. S. (2013). The effect of dilute
solution properties on poly(vinyl alcohol) films. Journal of the Mechanical Behavior of
Biomedical Materials, 28, 222–231.
Yetilmezsoy, K., Demirel, S., & Vanderbei, R. J. (2009). Response surface modeling of Pb
(II) removal from aqueous solution by Pistacia vera L.: Box-Behnken experimental
design. Journal of Hazardous Materials, 171(1–3), 551–562.
Yuan, X., Wang, Y., Cai, D., Zheng, M., Xiu, L., & Liu, J. (2016). Optimization by response
surface methodology of film-forming properties of whey protein of glycosylation
modifications at dry-heating condition. Journal of Chinese Institute of Food Science and
286
Food Hydrocolloids 87 (2019) 270–286
Technology, 16(7), 105–112.
Yu, H., Li, W., Liu, X., Li, C., Ni, H., Wang, X., et al. (2017). Improvement of functionality
after chitosan-modified zein biocomposites. Journal of Biomaterials Science, Polymer
Edition, 28(3), 227–239.
Zhang, Z.-S., Li, D., Wang, L.-J., Ozkan, N., Chen, X. D., Mao, Z.-H., et al. (2007).
Optimization of ethanol–water extraction of lignans from flaxseed. Separation and
Purification Technology, 57(1), 17–24.
Zhang, S., & Zhao, H. (2017). Preparation and properties of zein-rutin composite nano-
particle/corn starch films. Carbohydrate Polymers, 169, 385–392.
Zuo, B., Hu, Y., Lu, X., Zhang, S., Fan, H., & Wang, X. (2013). Surface properties of poly
(vinyl alcohol) films dominated by spontaneous adsorption of ethanol and governed
by hydrogen bonding. Journal of Physical Chemistry C, 117(7), 3396–3406.