Journal of Food Engineering 98 (2010) 294301

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Diffusivity of propolis compounds in Polylactic acid polymer for the development

of anti-microbial packaging lms
Erika Mascheroni a, Valrie Guillard b,*, Federico Nalin c, Luigi Mora a, Luciano Piergiovanni a
a

DISTAM, Department of Food Science and Microbiology, University of Milan, Via Celoria 2, 20133 Milan, Italy
Joint Research Unit Agropolymers Engeenering and Emerging Technologies, UMR 1208, Montpellier SupAgro, INRA, UM2, CIRAD, CC023 Place Eugne Bataillon,
34095 Montpellier cedex 5, France
c
Specchiasol srl, Via Bruno Rizzi 1/3, 37012 Bussolengo (VR), Italy
b

a r t i c l e

i n f o

Article history:
Received 22 June 2009
Received in revised form 23 December 2009
Accepted 27 December 2009
Available online 4 January 2010
Keywords:
Propolis
Polylactic acid
Diffusion kinetics
Mathematical models

a b s t r a c t
A major research gap is the lack of packaging materials that can provide the release of active compounds
at rates suitable for a wide range of food packaging applications. For this reason an anti-microbial/antioxidant release system for food packaging applications was realized by incorporation of propolis into
Polylactic acid (PLA) lm. The composition of the lms was modied by adding polyethylene glycol
(PEG) and calcium bentonite (CB) to the initial PLA casting solution; dispersed structures in fact open
the molecular network and increase migration rates. The presence of the anti-microbial compound is
required essentially at the food surface where the microorganisms are numerous and where they are
intended to grow. The diffusivity of four polyphenols was measured in water and ethanol as food simulating liquids (FSL) and the concentration of additives at the interface PLA/Food Simulant was calculated
using Fickian models. The external mass transfer coefcient at the interface polymer/FSL could be
neglected (with Bi number higher than 200). This is due to the low diffusivity values of propolis polyphenols in the PLA matrix (0.030.83  1013 m2/s) which lead to a predominant internal mass transfer phenomenon compared to the external one in the system PLA/water. The concentration at interface at
equilibrium was different for each substance and depended of the thermodynamical parameter K. Such
a delivery system for direct contact with liquid aqueous medium would be a very efcient delivery system because some active agents (polyphenols acids) would be released in relevant quantity in the food
whereas others (avonoids) would remain in the polymer to act at the polymer/food interface.
2010 Elsevier Ltd. All rights reserved.

1. Introduction
Promising active packaging systems are based on the incorporation of substances, in particular natural anti-microbial and antioxidant agents, in food biopackaging materials (Padgett et al., 1998;
Vermeiren et al., 1999; Appendini and Hotchkiss, 2002; Quintavalla and Vicini, 2002; Cooksey, 2005). Anti-microbial packaging systems incorporate additives and release them slowly into the food
in order to prevent microbial growth. Their incorporation into
the packaging matrix preserves them from oxidation or other degradation reactions and keeps them active all long the shelf life of
the food product. This design allows to reduce the quantity of additives added to the product as previously assessed by Guillard et al.
(2009) on a system made of an active wheat gluten-based edible
coating including sorbic acid as anti-microbial compound put on
a model food. Anti-microbials in active packaging system could
be distinguished in two types according to their volatility, i.e. (i)

* Corresponding author. Tel.: +33 380 63 32 77; fax: +33 380 63 32 27.
E-mail address: guillard@univ-montp2.fr (V. Guillard).
0260-8774/$ - see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2009.12.028

volatile compounds prevent anti-microbial growth by direct or

indirect contact between the packaging and the food whereas,
(ii) non volatile compounds require direct contact between the
food and the packaging to be efcient.
Polylactic acid (PLA) lms are of attracting, increasing interest
for developing biomaterial for food packaging applications, not
only because of the need to replace many petroleum-based polymers (Gontard and Guilbert, 1993), but also because of its useful
physical and mechanical characteristics (Drumright et al., 2000;
Auras et al., 2004; Holm et al., 2004). The performance of PLA as
food packaging material is undergoing progressive improvement
concerning, for example, the gas barriers and mechanical properties, as a result of recent extensive developments (Auras et al.,
2003; Sinha Ray et al., 2003; Lagaron et al., 2007). Moreover, PLA
was evaluated for its retention properties of various molecules
such as bacteriocins or other proteins and its application as a material for anti-microbial food packaging. Salmasoa et al. (2004)
loaded nisin into PLA particles and evaluated the sustained antimicrobial activity system. Bacteriocins were coated directly onto
PLA lm surface (Liu et al., 2007) or lysozyme was incorporated

E. Mascheroni et al. / Journal of Food Engineering 98 (2010) 294301

into the PLA lms (Del Nobile et al., 2008) and their anti-microbial
effectiveness was evaluated against food borne pathogens (Jin and
Zhang, 2008). PLA was suitable as a polymer material for antimicrobial food packaging with proteins (nisin, bacteriocins and
lysozyme) as anti-microbial agents, but further research is need
to understand its suitability to incorporate and release vegetable
compounds as organic acids and avonoids.
A natural active agent with high polyphenol content is propolis.
Propolis is a resinous substance gathered by Apis mellifera and
known for its biological properties, having antibacterial, antifungal,
antioxidant, antiviral activities (Kooa et al., 2000). Its anti-microbial properties against a widely variety of microorganisms like
Staphylococcus aureus, Enterococcus faecalis, Candida utilis, have
been reported and these activities are generally associated with
its polyphenolic fraction (Marcucci et al., 2000; Mendes da Silva
et al., 2006; Bankova et al., 2000). Research has been performed
on using edible polymer as retention matrices such as mucoadhesive polymeric lms based on blends of poly(acrylic acid) and
(hydroxypropyl) cellulose and vehicles of propolis extracts (dry,
ethanolic and glyceric) in the treatment of oral cavity and other
diseases and for pharmaceutical applications (Juliano et al., 2007;
Shin et al., 2005). Although its potentialities, propolis has not been
explored yet for anti-microbial food packaging applications, so research and development of its incorporation in food packaging
materials is needed. Mathematical modeling of mass transfer is
necessary to in-depth understanding and optimization of such
propolis delivery system, as it has been already done for drug release systems (Charlier et al., 2000; Siepmann and Peppas, 2001)
or for other active packaging (Buonocore et al., 2003; El Kouali
et al., 2003). Most of the mathematical models used for modeling
the release kinetics are based on analytical solutions of the Ficks
second law given for restrictive conditions of geometry and specic
boundary conditions. For instance, the more frequent solution employed is of value for a plane sheet of material immersed in a so
volume of liquid to be considered as innite, and assuming that
the external mass transfer coefcient at the solidliquid interface
is negligible (Rosca and Vergnaud, 2007; Buonocore et al., 2003;
Hernandez-Munoz et al., 2001). However, in most cases, the volume of the solution could not be considered as innite and the liquid becomes saturated as a function of time; the partition
coefcient between the polymer and the solution is thus a required
parameter to model and predict the release at equilibrium.
The present work aimed at studying the PLA/propolis system
and the migration of propolis into food simulating liquid (FSL).
Kinetics of selected propolis polyphenols from PLA into water
and ethanol were experimentally measured and mathematically
modeled. Three mathematical models were developed: (i) one considering an innite solution and negligible external mass transfer
coefcient, (ii) one considering a nite solution and negligible
external mass transfer coefcient, and (iii) one considering an innite solution and a non negligible external mass transfer coefcient. The inuence of the three mathematical models on the
value of diffusivity identied was discussed for both water and
ethanol as food simulating liquid. Based on the release kinetics obtained, the availability of PLA as retention matrix for propolis in the
objective of developing anti-microbial packaging was discussed.

295

Italy) while pinocembrin (P) was purchased from Extrasynthese

(Genay, France). Methanol, acetonitrile, formic acid, ethanol (EtOH)
and ethyl acetate were from Carlo Erba (Milan, Italy). Water was
obtained from a MilliQ apparatatus (Millipore, Milford, MA). Calcium bentonite (CB) was from Laviosa Chimica Mineraria (Italy)
and polyethylene glycol (PEG) 4000 was from Univar (Milan, Italy).
2.2. Film preparation
PLA-polymer lms were prepared by a solvent casting method
for prototyping in a laboratory scale. PLA granules were put in
nitrogen liquid and ground (IKA-Universalmuhle M20); after that,
two mixtures were prepared: for both the solutions, 3 g of ground
PLA were put in 30 mL of ethyl acetate under magnetic stirring at
40 C until the polymer was totally dissolved. While the rst contain only PLA, in the second one, 3 g of PLA were mixed in 30 mL of
ethyl acetate with CB (10%) as opening polymer structure agent
and PEG (15%) as plasticizer, and the solution was mixed again
for 30 min at the temperature of 40 C. Then propolis (13% w/wPLA),
previously dissolved in 5 mL of ethyl acetate, was added to all mixtures, that were stirred by a magnetic bar until the additives and
propolis were homogeneously dispersed; the mixtures were subsequently casted onto plexiglass plates using a thin layer applicator
(Mod. 1117, Sheen Instruments Ltd., Kingston, England) and the
solvent was allowed to evaporate at room temperature under a
chemical hood for an hour. After that, the two types of lms
(PLA/propolis and modied PLA/propolis) were peeled from plexiglas plate and placed in a sealed container containing silica gel until time of use. The lm thickness was 43.4 0.5 lm
2.3. Kinetics of propolis release in solvents
Films of 100 cm2 surface area and with 43.5 0.5 lm average
thickness were cut into 2.5  10 cm pieces, the weight and the
thickness of each piece was measured and all the pieces were
put vertically into 50 mL of distilled water and EtOH and stored
at 25 C under stirring. Aliquots of 100 lL were removed at prescribed time intervals, centrifuged at 13000 rpm for 5 min and
50 lL were injected in the LC-system as reported in Section 2.5.
2.4. Extraction and determination of propolis polyphenols in PLApolymer lm
The PLA lms containing propolis before and after kinetic release in solvents were cut in small pieces (about 1 cm2) and
100 mg sonicated for 10 min in 20 mL of ethanol. The mixture
was centrifuged at 13000 rpm for 5 min and the solid residue
was extracted one more using the same procedure. The clear
supernatants were mixed and the nal volume adjusted to 50 mL
by ethanol.
One aliquot of the sample was ltered through a Millipore
0.2 lm lter and 20 lL injected in the LC-system as reported in
Section 2.5. The extraction method was validated by recovery tests
for phenolic acids and avonoids from PLA lm samples and the
precision (intra- and inter-day) of the assay was veried.
2.5. LC-DAD-MS analysis for propolis identication and quantication

2. Materials and methods

2.1. Materials
PLA polymer (semi-crystalline, 2100D-Cargill, USA) was used as
biopolymer. Standardized and puried propolis extract (EPID)
was kindly provided by Specchiasol S.r.l (Bussolengo, Italy). Chrysin (C), p-coumaric acid (pC), were from SigmaAldrich (Milan,

Identication and quantication of propolis polyphenols in the

solvents were carried out by the LC-DAD-MS analysis as reported
by Gardana et al. (2007). The chromatographic system consisted
of an Alliance 2695 (Waters, Milford, MA) equipped with a model
2996 (Waters) photodiode array detector and a triple quadrupole
mass spectrometer mod. Quattromicro (Micromass, Beverly, MA).
Calibration curves were obtained from mother solutions prepared by dissolving 10 mg of standard powder in 10 mL methanol

296

E. Mascheroni et al. / Journal of Food Engineering 98 (2010) 294301

and the working solutions prepared in the range of 0.550 lg/mL.

Pinobanksin (Pb), pinobanksin-5-methyl-ether (Pb5ME) were assayed using pinocembrin calibration curve and their amounts were
normalized by the molecular mass ratios.
2.6. Models development
Depending of the assumptions made for the boundary condition
at the interface lm/solution and on the volume of the solution itself, three alternative models were developed in order to predict
migration inside the lm and to determine the propolis compounds diffusivity:
Model 1: innite volume of solution and negligible external
mass transfer coefcient.
Model 2: innite volume of solution and non negligible external mass transfer coefcient.
Model 3: nite volume of solution and negligible external
mass transfer coefcient.

this additive is brought to the surface by internal diffusion through

the polymer:

t > 0;

x L;

D

oC
kC L;t  C 1 ;
ox

where CL; t is the concentration in migrant at the polymer surface at any time and C 1 is the concentration in migrant at equilibrium in the polymer (in kg/m3) and k is the external mass transfer
coefcient (m/s).
The kinetics of transport of the substance out of the sheet of
thickness 2L is expressed by the following relationship (Vergnaud
and Rosca, 2006):



1
X
Mt
2R2
Dt

 exp b2n 2 ;
1
2
2
2
M1
L
n1 bn bn R R

where bn are the positive roots of

b tan b R;

and the dimensionless number R (Bi number) is given by

For all the models, a Fickian formalism with a constant diffusivity is used. Whatever the boundary hypothesis used for describing
the system, the mass conservation equation in the lm can be written as

t > 0;

L < x < L;

oC
o2 C
D 2;
ot
ox

where x is the distance (m), t is time (s), C the lm concentration in

migrant (kg additive/m3 of polymer), D the migrant diffusivity within the packaging (m2/s) and L the half-thickness of the packaging. It
is considered that the initial concentration of the migrant in the lm
is homogeneously distributed and equal to a constant:

t 0;

L < x < L;

C C in ;

2.6.1. Model 1: innite volume of solution and negligible external mass

transfer coefcient
If the volume of the solution is larger than that of the lm and/
or if, at anytime, the concentration in the solution remains lower
than the concentration in the solution at equilibrium and if the
solution is strongly stirred, the boundary condition is

x L;

C C1;

Mt
1
M1
p

n0

1
2n 12

exp 

The value of the external mass transfer coefcient k largely depends on the type of convection: it is rather low in motionless liquid and very large when the liquid is strongly stirred.
2.6.3. Model 3: nite volume of solution and negligible external mass
transfer coefcient
If the volume is nite and the liquid of nite volume is strongly
stirred, the diffusion depends only on time, and the total amount of
diffusing substance in the sheet and the liquid remains constant
when diffusion proceeds. The balance equation is

t > 0;

2n 1
4L2

p2 Dt ;

where Mt is the quantity of migrant entering the solution at time t

and M1 is the corresponding quantity in the liquid at equilibrium.
2.6.2. Model 2: innite volume of solution and non negligible external
mass transfer coefcient
At the lmliquid interface, the rate at which the additive is
transferred into the liquid is constantly equal to the rate at which

x L;

D

oC V L oC

;
ox AP ot

where AP is the lm area, V L is the volume of solution.

With a constant diffusivity and the condition of the uniform
concentration of the diffusing substance in the lm, the kinetics
of transport of the substance out of the sheet of thickness 2L is expressed by the following relationship (Vergnaud and Rosca, 2006):

 2 
1
X
Mt
2a1 a
q Dt
1
exp
 n2 ;
2 q2
1

M1
a

a
L
n
n1

10

P
V
with a kV
represents the volume of polymer and K the
P where V
partition coefcient of the migrant in the system between the polymer and the solution, which can be assumed as constant for low
concentrations:

where C1 is the concentration in migrant at equilibrium in the polymer (in kg/m3).

The kinetics of transport of the substance out of the sheet of
thickness 2L is expressed by the following relationship:
1
8X

kL
:
D

where Cin is the initial homogeneous migrant concentration in the

polymer.
The hypothesis chosen to describe the external mass transfer
phenomena inuence the nature of the boundary condition at
x = L. For the three models developed, the expressions for the
boundary condition at the interface lm/solution as the mathematical solving can be found in the following:

t > 0;

C polymer;1
;
C liquid;1

11

qn are the non-zero positive roots of

tan qn aqn : Crank; 1975:

12

2.7. Estimation of the external mass transfer coefcient (k)

The external mass transfer coefcient (k) is calculated from the
equation of the migrant ux at the beginning (rst 4 min) of the
migration experiment, where the concentration of the migrant in
the polymer is close to the initial value and where the concentration of the migrant in the solution is low and far from the value obtained at equilibrium (Kondoyan and Daudin, 1993; Roca et al.,
2007):

E. Mascheroni et al. / Journal of Food Engineering 98 (2010) 294301


 V L dC
L
/ k C L; xL  C L P
;
A dt

13

where U is the migrant ux in the liquid (kg/kg of liquid), AP is the

surface of polymer (m2), CL the concentration in migrant in the liquid at the beginning of the experiment at innite distance from
the lm (this concentration could be neglected), C L;xL is the concentration in the liquid at the polymer/liquid interface in equilibL
rium with the initial concentration in the polymer and dC
dt
represents the time derivative of the concentration present in the
solution.

2.8. Calculation procedure

2.9. Statistical analysis

The root mean square error (RMSE) was used to estimate the
quality of model tting and was calculated as follows:

RMSE

N  p

3. Results and discussion

The outline of the present study is to design active PLA-based
lms containing propolis as anti-microbial compounds. In order
to in-depth understand the behavior of such active PLA-based lms
and for further mathematical modeling of propolis diffusion, it is
necessary to identify the diffusivity of propolis components in
PLA. The diffusion kinetics of four different polyphenols contained
in propolis was then experimentally measured in two different
food simulating liquids (FSL): water and ethanol. Water and ethanol were chosen as two liquids with different polarity and solubility characteristics.
3.1. Suitability of PLA for anti-microbial purpose

Simulations of polyphenols release were performed using the

models developed and elaborated with MATLAB software (The
Mathworks Inc., Natick, MA, USA). Theoretical kinetic release and
concentration proles were calculated with a half-thickness L of
21.7 0.6 lm. The polyphenols initial quantity was deduced from
the lm preparation procedure and the initial quantity of each
polyphenols in propolis extract and was expressed in kg of polyphenol/kg of polymer as reported in Table 1. The partition coefcient of polyphenol in polymer/solvent corresponded to the ratio
of concentrations of each polyphenol in the lm and in the FSL at
equilibrium and was obtained at least after 20 days of storage. Positive roots qn and bn in respectively Eqs. (12) and (7) are calculated
by using a routine programmed in MATLAB software.

v
u _ 2
u y y
t

297

Release kinetics of propolis components were rst measured for

PLA/propolis system immersed in water as FSL. Among the propolis
polyphenols some of the major representative components were
evaluated; in particular: pinobanksin, pinobanksin-5-methyl-ether
as avanonols, chrysin as avon and p-coumaric acid as phenolic
acid as shown in Fig. 1.
Microbiological tests carried out for modied PLA/propolis lms
in a previous study (Mascheroni et al., 2007), indicated that a
signicant level of propolis molecules were released from the modied PLA-polymer lm. In this study, to support the microbiological test, the concentration of the Pb5ME, one of the most abundant
avonoids in propolis, was monitored in water, at 25 C. The resulting release kinetics was expressed using the ratio kg additive/ kg
water as shown in Fig. 2. For a same percentage of Pb5ME in
PLA, signicantly higher quantity of Pb5ME was released from
the modied PLA/propolis lm with PEG (15%) and CB (10%) respect to the native PLA lm (1  105 kg/kg versus 1  109 kg/
kg for the native PLA/propolis lm). The modied PLA/propolis lm
was chosen in the following for studying the release kinetics of the
four selected propolis polyphenols.

14

3.2. Partition coefcient between PLA and water

where y and y are respectively the experimental and predicted

residual value, N is the number of experimental measurements
and p the number of identied parameters. If RMSE tends toward
0 or is very close to the experimental error, it means that model
is able to represent the experimental data.

2.10. Scanning electron microscopy (SEM)

Observation of the fractured surface of the resulting lms was
performed using a Scanning Electron Microscope, to assess the lm
micro-structure (JEOL JSM - 6380 LV) at magnication ranging between 0.65 and 4 K. The samples were previously sputter-coated
with a thin layer of gold (Automatic Sputter Coater, Cressington
108 Auto).

Table 1
Initial concentration of propolis components in PLA and partition coefcient of
polyphenols between PLA and water.
Propolis
component

Percentage in
standardized and puried
propolis extract (%)

Initial
concentration
(kg/kg of
polymer)

Partition
coefcient
PLA/water (K)

Pb
Pb5ME
pC
C

3.62
1.68
0.44
3.54

4.7  103
2.0  103
0.5  103
4.69  103

205 20
47 6
25 5
1884 130

Experimental release kinetics in water of pinobanksin 5 dimethylether (Pb5ME), pinobanksin (Pb), chrysin (C) and p-coumaric acid
(pC) were determined at 25 C and compared in Fig. 3. At equilibrium (i.e. after 20 days of contact), remaining content in PLA lm
were measured for all the tested polyphenols. From these concentrations at equilibrium, partition coefcients (K) were calculated
according to Eq. (11). The resulting K values are reported in Table 1.
This coefcient compares the relative afnity of the molecule studied between PLA and water.
To assess if the propolis principal constituents were completely
dissolved into PLA, SEM pictures of the fractured surface of the
modied PLA were realized. The modied PLA lm does not show
separation phase phenomena between the components of the matrix (Fig. 4b) also at higher magnication (Fig. 4c). The PLA lm was
also tested as reference (Fig. 4a).
In our cases, according to the obtained K, the following classication of the four polyphenols as regard their afnity for PLA could
be achieved: C > Pb > Pb5ME > pC. With a K value of 25 pC was the
most hydrophilic and it had the most afnity towards water. This
can be explained by the metilation of hydroxyl in C5 that confers
more hydrophilicity to the molecule, as conrmed by the chromatogram of the propolis extracts, where the retention time of
Pb5ME is always lower than that of Pb. Moreover, the higher afnity of C (K = 1884) for the PLA in respect to Pb, its related avanonol, could be basically explained by the absence of the OH in
position 3 (Fig. 1). These characteristics confer to C less hydrophilicity, which could be able to slow down the release of C in distilled

298

E. Mascheroni et al. / Journal of Food Engineering 98 (2010) 294301

Fig. 1. Chemical structure of propolis components studied: chrysin (a), pinobanksin-5-methyl ether (b), pinobanksin (c) and p-coumaric acid (d).

kg of migrant/kg of simulant

1.5E-05

Pb5ME
from modified PLA
1.0E-05

5.0E-06

Pb5ME
from PLA
0.0E+00
0

Time (Days)

0.2

Fig. 2. Experimental Pb5ME content obtained in food simulating liquid (water) after migration from PLA/propolis lm (h) and modied PLA/propolis lm (D) at 25 C.

water at 25 C compared with the release of the other active

compounds.
3.3. Diffusivity values of propolis components in PLA
In order to go ahead in the quantication of the release of the
selected propolis polyphenols, the diffusivity values of each polyphenols in PLA were identied from the experimental data presented in Fig. 3.
To be done, a mathematical model was required. The most frequent mathematical model used in this case is an analytical solution of Ficks second law given for a plane sheet of material
immersed in a great liquid volume considered as innite, i.e. Model
1 (Eq. (4)).
Model 1 (case of a negligible external mass transfer coefcient
and innite volume of solution) was used in a rst approach to obtain an approximate value of the diffusivity. Resulting diffusivity
values (called D*) are gathered in Table 2 with the RMSE obtained
for each t of the experimental data. According to Table 2, the dif-

fusivity values D* obtained are not signicantly different except for

Pb5ME (higher value with 1.03  1013 m2/s) and pC (lower value
with 0.74  1013 m2/s).
Diffusivities identied with Model 1 (D*) are of value providing
that: (i) the external mass transfer coefcient at the polymer/solution interface could be neglected, and (ii) the volume of solution
could be considered as innite. The hypothesis of a negligible
external mass transfer coefcient, which is the case for a wellmixed food or a Bi number greater than 100 (Vergnaud, 1998;
Vergnaud and Rosca, 2006) was rst tested. The Bi number was
then tentatively calculated in the polymer/liquid system studied
here. To do this, the external mass transfer coefcient (k) was rst
evaluated using Eq. (13). The resulting value of k and Bi are in Table 3. It is obvious from Table 3 that the external mass transfer
coefcient could be neglected (Bi > 200) conrming the effectiveness of the stirring. Conrming this, by using Model 2 with Bi numbers of Table 3 and diffusivity values of Table 2, the simulation
curves obtained for each polyphenol were not signicantly different than those obtained by using Model 1 (curves not shown).

299

E. Mascheroni et al. / Journal of Food Engineering 98 (2010) 294301

Polyphenol content in water (kg/kg)

1.6

x 10

-5

1.4
1.2
1
0.8
0.6
0.4
0.2
0

0.2

0.4

0.6

0.8

1.2

Time (days)
Fig. 3. Comparison between experimental and predicted p-coumaric acid (}),
pinobanksin (D), pinobanksin 5 dimethylether (s) and chrysin (x) content obtained
at 25 C in food simulating liquid (water) after migration from destabilized PLA
containing semi puried propolis (ts obtained with Model 3).

In most cases, the volume of the solution could not be considered as innite and the liquid becomes saturated as a function of
time. In this case, Eq. (4) does not apply and the diffusivity identied by this simplied model is limited.
In order to rene the approximate diffusivity values obtained
(D*), the hypothesis of a nite vs. innite volume of solution was
tested (Model 3 vs. Model 1). Diffusivity values were then identied by using Model 3 which required as input parameter the partition coefcient (K) given in Table 1. The resulting diffusivity
values (D) are given in Table 4. The new diffusivity values (D) obtained were not signicantly different from the D* ones except
for Pb and C propolis compounds which displayed signicant lower
diffusivity values with Model 3 than those indentied with Model
1. These two polyphenols presented the higher partition coefcient
(Table 1). This result tended to demonstrate that in the case studied here of a high afnity of the migrant for the polymer (K > 200)
and in the conditions of lm surface/FSL volume ratio applied in
this study (20 dm2 of polymer for 1 L of food simulating liquid),
it would be necessary to consider a solution of a nite volume. In
Fig. 3, examples of model tting obtained by using Model 3 were
shown. Model 3 succeeded in simulating experimental release
kinetics of propolis components in water with RMSE values close
or lower than 10% (Table 4). These RMSE values were all close to
that obtained with Model 1 except for chrysin where the RSME value obtained with Model 3 was higher than that obtained with
Model 1 (13.6% instead of 6.2%).
Literature show that diffusivity value of aroma compounds from
PLA lms are very low (263  1013 m2/s for d-Limonene at 2 C;
Auras et al., 2006), while the D values obtained in this case of study
for polyphenol compounds with a Mw from 160 to 280 g/mol are
comparable for example with the D values of potassium sorbate
from whey proteins lm (Ozdemir and Floros, 2001) or the D value
the aroma compound tymol from zein lms (both substances with
Mw of 150.22 g/mol) (Mastromatteo et al., 2008).
Differences between D values (Table 4) could be related to the
afnity of the polyphenol of concern for the PLA matrix. For instance, the lowest D was obtained for chrysin and pinobanksin
which displayed the highest K coefcients (1884 and 205 respectively). In this case, that means that the higher the K was, the lower
the D value would be. Specic bonding between the active molecules and the PLA should be created which yielded to a high K value

Fig. 4. Fractured surface of PLA lms visualized by SEM after treatment with gold:
(a) PLA lm magnication X65, (b) modied PLA lm magnication X65, and (c)
modied PLA lm magnication X4000.

Table 2
Approximate diffusivity values (D*) obtained by using the simplied approach (Model
1).
Propolis component
Pb
Pb5ME
pC
C

D* (1013 m2/s)

RMSE** (%)
a,b

0.89 (0.800.98)
1.03 (0.910.15)a
0.74 (0.580.90)b
0.755 (0.6350.875)a,b

4.2
4.5
8.8
6.2

Values between brackets are the 95% condence interval on the identied values
(calculated by a MATLAB built-in function).
a,b
Values with the same letters are not signicantly different.
**
To be compared to the experimental error 3.75% (average value of condence
intervals obtained on all experimental points).

but also prevented the molecules to easily migrate in the matrix. It

is worth noticing that Model 1 was not precise enough to be discriminative on the identied diffusivity values (Table 2). The different diffusion behaviors of propolis components could be put in
evidence only by using the appropriate modelling assumption.

300

E. Mascheroni et al. / Journal of Food Engineering 98 (2010) 294301

Table 3
External mass transfer coefcient and corresponding Bi number for each propolis
polyphenols tested.
Propolis component

k (106 m/s)

Bi number

Pb
Pb5ME
pC
C

2.13
1.14
0.72
3.29

515
237
210
936

Table 4
Rene diffusivity values (D) obtained by using Model 3 (hypothesis of a nite volume
of solution).
Propolis component
Pb
Pb5ME
pC
C

D (1013 m2/s)
0.42
0.83
0.66
0.03

RMSE** (%)
a,c

(0.360.48)
(0.730.93)b
(0.420.90)b,c
(0.010.05)d

5.5
4.6
8.6
13.8

Values between brackets are the 95% condence interval on the identied values
(calculated by a MATLAB built-in function).
ad
Values with the same letters are not signicantly different.
**
To be compared to the experimental error 3.75% (average value of condence
intervals obtained on all experimental points).

According to the obtained D values (Table 4), the following classication of the four polyphenols as regard their ability to diffuse
into the PLA matrix could be achieved: C < Pb < pC Pb5ME. This
classication was similar than that obtained based on the K values.
Interest of a mathematical model is that some simulations
could be carried out to predict and compare concentration in active
agent in the polymer/model system. For example, using Model 3
and diffusivities of Table 4, the concentration of additive at the
interface PLA/liquid could be calculated. The concentration at
interface was calculated for the four polyphenols: Pb, Pb5ME, C,
pC. The concentration at interface at equilibrium was different
for each substance and depended of the thermodynamical parameter K. Due to the high afnity of pinobanksin and chrysin for PLA
(high K values), high contents of avonoids remain in the polymer.
Polyphenols, like pC acid and Pb5ME, have medium value of K and
are released in higher quantity (around 8090% of the initial quantity). Such a delivery system for direct contact with liquid aqueous
medium would be a very efcient delivery system because some
active agents (polyphenols acids) would be released in relevant
quantity in the food whereas others (avonoids) would remain in
the polymer to act at the polymer/food interface. These complementary behaviors of propolis components would make this compound very interesting for anti-microbial applications such as
active packaging with controlled delivery system.
3.4. Kinetic release of propolis components in ethanol
Experimental release kinetics of pinobanksin 5 dimethylether
(Pb5ME), pinobanksin (Pb), chrysin (C) and p-coumaric acid (pC)
were also determined in pure ethanol at 25 C. The behavior of
PLA in ethanol is very interesting to be studied because of the
growing interest of PLA for packaging liquid. It is particularly interesting to see if PLA could be used for alcoholic beverages. Moreover, alcoholic solution (10% ethanol) is considered as food
simulating liquid for the European Legislation FS (EC No. 1935/
2004).
The kinetic release was very fast whatever the polyphenol and
equilibrium was reached in a few minutes (Fig. 5). From concentrations at equilibrium, the partition coefcient PLA/ethanol for each
polyphenols was calculated (Table 5). Pb and C display a higher
afnity for ethanol than for water with a K coefcient harshly de-

Fig. 5. Comparison between experimental and predicted p-coumaric acid (}),

pinobanksin (D), pinobanksin 5 dimethylether (s) and chrysin (x) content obtained
in food simulating liquid (ethanol) after migration from destabilized PLA containing
semi puried propolis (ts obtained with Model 3).

Table 5
Partition coefcient PLA/ethanol of propolis components.
Propolis component

Partition coefcient PLA/ethanol (K)

Pb
Pb5ME
pC
C

19 8
23 10
22 9
0.04 0.03

creased (for instance, from 1884 to 0.04 for chrysin). On the contrary, Pb5ME and pC display only a slight decrease in their
partition coefcient between PLA/water and PLA/ethanol system
proving that these two molecules had less afnity for ethanol than
C and Pb. By using the K values of Table 5, and the diffusivity values
of polyphenols in PLA (Table 4), the release kinetic of the four polyphenols studied from PLA into ethanol was tentatively modeled.
Modeling curves were shown in Fig. 5 and compared with experimental data. It was obvious from Fig. 5 that the diffusivity of Table 4 and the mathematical model used (Model 3) did not
succeeded in predicting the release kinetic of propolis polyphenols
in pure ethanol, especially for Pb and C. The equilibrium being
reached very quickly, the diffusivity value should be higher than
those of Table 4 for the experimental curves tting well the experimental data. The greatest discrepancy between simulated and
experimental data was observed for chrysin which displayed a totally different behavior when ethanol was used as FSL compare to
water (its afnity for PLA harshly decreased). The inability of the
model to predict the release kinetics in ethanol could be due to
structural modications of the PLA matrix under ethanol sorption.
Ethanol is considered as a poorer solvent for PLA as compared to
acetone or methanol but it has been observed an interaction between ethanol and poly(l)lactice and ethanol is also used as a
cosolvent for preparation of PLA nano- and microparticles drug
delivery systems (Peltonen et al., 2002; Lassalle and Ferreira,
2007). Ethanol would play as a plasticizer for the PLA matrix and
would open the structure creating void spaces favouring the polyphenols migration. For instance, according to Model 3 and the optimization procedure used to identify parameter, the diffusivity of
Pb in PLA in contact with ethanol would be of 1.42  1010 m2/s,
i.e. more than 3000 folds higher than that found in PLA in contact
with water. This particular behavior, which is of high interest in

E. Mascheroni et al. / Journal of Food Engineering 98 (2010) 294301

the case of commercial use of PLA matrix for packing alcoholic beverages, would be worth being studied in a more extensive work.
4. Conclusion
The results of this study suggested that the incorporation of
propolis in the biopolymer PLA correctly modied could provide
a possible delivery system for food packaging applications. The
method used for the determination of polyphenolic substances release allowed evaluating the diffusion of single natural active compounds to water. Analytical models allowed understanding the
mass transfer phenomena of different propolis polyphenols; in particular, such tools succeeded in predicting polyphenols release kinetic from modied PLA/propolis lm into water and the
polyphenols distribution prole in PLA lm as a function of time.
However, only mathematical model taking into account the selling
out of the solution permitted an overall and comprehensive study
of the kinetic release and a classication of propolis polyphenols
according to their behaviour. The diffusivity values obtained for
different propolis compounds are usually not signicantly different
whereas the concentration at interface at equilibrium was different
for each substance and depended of the thermodynamical parameter K. Such a delivery system for direct contact with liquid aqueous medium would be a very efcient delivery system because
some active agents (polyphenols acids) would be released in relevant quantity in the food whereas others (avonoids) would remain in the polymer to act at the polymer/food interface.
Moreover, preliminary results showed that PLA polymer could
have interactions with liquid foods containing ethanol probably
because ethanol migrate into PLA and modify its structure, so it
would be important, before using PLA/propolis to pack foods containing alcohol, to in-depth study the evolution of PLA structure.
Acknowledgement
We wish to thank Silvio Farag from Stazione Sperimentale
Seta, Italy that allowed us to include the SEM analyses in this work.
References
Appendini, P., Hotchkiss, J.H., 2002. Review of antimicrobial food packaging.
Innovative Food Science and Emerging Technologies 3, 113126.
Auras, R., Harte, B., Selke, D., Hernandez, R., 2003. Mechanical, physical, and barrier
properties of poly(lactide) lms. Journal of Plastic Film & Sheeting 19, 123135.
Auras, R., Harte, B., Selke, S., 2004. An overview of polylactides as packaging
materials. Macromolecular Bioscience 4, 835864.
Auras, R., Harte, B., Selke, S., 2006. Sorption of ethyl acetate and d-limonene in
poly(lactide) polymers. Journal of the Science of Food and Agriculture 86, 648
656.
Bankova, V.S., Castro, S.L., Marcucci, M.C., 2000. Propolis: recent advances in
chemistry and plant origin. Apidologie 31 (1), 315.
Buonocore, G.G., Del Nobile, M.A., Panizza, A., Corbo, M.R., Nicolais, L., 2003. A
general approach to describe the antimicrobial agent release from highly
swellable lms intended for food packaging applications. Journal of Controlled
Release 90, 97107.
Charlier, A., Leclerc, B., Couarraze, G., 2000. Release of mifepristone from
biodegradable matrices: experimental
and theoretical evaluations.
International Journal of Pharmaceutics 200, 115120.
Cooksey, K., 2005. Effectiveness of antimicrobial food packaging materials. Food
Additives and Contaminants 22 (10), 980987.
Crank, J., 1975. The Mathematics of Diffusion. Clarendon Press, Oxford. p. 414.
Del Nobile, M.A., Conte, A., Buonocore, G.G., Incoronato, A.L., Massaro, A., Panza, O.,
2008. Active packaging by extrusion processing of recyclable and biodegradable
polymers. Journal of Food Engineering. doi:10.1016/j.jfoodeng.2008.12.022.
Drumright, R.E., Gruber, P.R., Henton, D.E., 2000. Polylactic acid technology.
Advanced Materials 12, 18411846.
EC No. 1935/2004. Regulation (EC) No. 1935/2004 of the European Parliament and
of the Council of 27 October 2004 on materials and articles intended to come
into contact with food. <http://ec.europa.eu/food/food/chemicalsafety/
foodcontact/legisl_list_en.htm>.
El Kouali, M., Salouhi, M., Labidi, F., El Brouzi, M., Vergnaud, J.M., 2003. Theoretical
study of the release and consumption of an agent in food from active polymer
packaging. Polymers & Polymer Composites 11 (4), 301310.

301

Gardana, C., Scaglianti, M., Pietta, P., Simonetti, P., 2007. Analysis of the
polyphenolic fraction of propolis from different sources by liquid
chromatographytandem mass spectrometry. Journal of Pharmaceutical and
Biomedical Analysis 45 (3), 390399.
Gontard, N., Guilbert, S., 1993. Bio-packaging: technology and properties of edible
and/or biodegradable material of agricultural origin. In: Mathlouthi, M. (Ed.),
Food Packaging and Preservation. Blackie Academic and Professional. Chapter 9.
Guillard, V., Issoupov, V., Redl, A., Gontard, N., 2009. Food preservative content
reduction by controlling sorbic acid release from a supercial coating.
Innovative Food Science and Emerging Technologies 10, 108115.
Hernandez-Munoz, P., Catala, R., Gavara, R., 2001. Food aroma partition between
packaging materials and fatty food simulants. Food Additives and Contaminants
18 (7), 673682.
Holm, V.K., Ndoni, S., Risbo, J., 2004. The stability of poly(lactic acid) packaging lms
as inuenced by humidity and temperature. Journal of Agricultural and Food
Chemistry 50, 43644370.
Jin, T., Zhang, H., 2008. Biodegradable polylactic acid polymer with nisin for use in
antimicrobial food packaging. Journal of Food Science 73 (3), M127M134.
Juliano, C., Pala, C.L., Cossu, M., 2007. Preparation and characterisation of polymeric
lms containing propolis. Journal of Drug Delivery Science and Technology 17
(3), 177181.
Kondoyan, A., Daudin, J.D., 1993. Determination of transfer coefcients by
psychometry. International Journal of Heat and Mass Transfer 36, 1807
1818.
Kooa, H., Gomes, B.P.F.A., Rosalena, P.L., Ambrosanoa, G.M.B., Parkb, Y.K., Curya, J.A.,
2000. In vitro antimicrobial activity of propolis and Arnica montana against oral
pathogens. Archives of Oral Biology 45 (2), 141148.
Lagaron, J.M., Gimenez, E., Sanchez-Garcia, M.D., Ocio, M.J., Fendler, A., 2007. Novel
nanocomposites to enhance quality and safety of packaged foods. In: Food
Contact Polymer. Rapra Technologies, Shawbury, Shrewsbury, UK. Paper 19.
Lassalle, V., Ferreira, M.J., 2007. PLA nano- and microparticles for drug delivery: an
overview of the methods of preparation. Macromolecular Bioscience 7, 767
783.
Liu, L.S., Finkenstadt, V.L., Liu, C.K., Jin, T., Fishman, M.L., Hicks, K.B., 2007.
Preparation of poly(lactic acid) and pectin composite lms intended from
applications in antimicrobial packaging. Journal of Applied Polymer Science
106, 801810.
Marcucci, M.C., Ferreres, F., Custo dio, A.R., Ferreira, M.M.C., Bankova, V.S., GarcaViguera, C., 2000. Zeitschrift fur Naturforschung 55c (12), 7681.
Mascheroni, E., Mora, L., Trinetta, V., Piergiovanni, L., 2007. Kinetic diffusion of
antimicrobials from PLA lms. In: Poster I Biopol Conference, Alicante, Spain.
Mastromatteo, M., Barbuzzi, G., Conte, A., Del Nobile, M., 2008. Controlled release of
thymol from zein based lm. Innovative Food Science and Emerging
Technologies 10 (2), 222227.
Mendes da Silva, J.F., de Souza, M.C., Matta, S.R., de Andrade, M.R., Vidal, F.V.N.,
2006. Correlation analysis between phenolic levels of Brazilian propolis
extract and their antimicrobial and antioxidant activities. Food Chemistry 99,
431435.
Ozdemir, M., Floros, J.D., 2001. Analysis and modeling of potassium sorbate
diffusion through edible whey protein lms. Journal of Food Engineering 47,
149155.
Padgett, T., Han, I.Y., Dawson, P.L., 1998. Incorporation of food-grade antimicrobial
compounds into biodegradable packaging lms. Journal of Food Protection 61,
303307.
Peltonen, L., Koistinen, P., Karjalainen, M., Hkkinen, A., Hirvonen, J., 2002. The effect
of cosolvents on the formulation of nanoparticles from low-molecular-weight
poly(l)lactide. Journal of Pharmaceutical Science and Technology 3 (4), 32.
Quintavalla, S., Vicini, L., 2002. Antimicrobial food packaging in meat industry. Meat
Science 62, 373380.
Roca, E., Broyart, B., Guillard, V., Guilbert, S., Gontard, N., 2007. Controlling moisture
transport in a cereal porous product by changing structural or formulation
parameters. Food Research International 40, 461469.
Rosca, I.D., Vergnaud, J.M., 2007. Problems of food protection by polymer packages.
Journal of Chemical Health and Safety (March/April).
Salmasoa, S., Elvassoreb, N., Bertuccob, A., Lantec, A., Calicetia, P., 2004. Nisin-loaded
polyl-lactide nano-particles produced by CO2 anti-solvent precipitation for
sustained antimicrobial activity. International Journal of Pharmaceutics 287,
163173.
Shin, J.E., Kim, J.M., Bae, E.A., Hyun, Y.J., Kim, D.H., 2005. In vitro inhibitory effect of
avonoids on growth, infection and vascuolation of Helicobacter pylori. Journal
of Medical Microbiology 71 (3), 197201.
Siepmann, J., Peppas, N.A., 2001. Modeling of drug release from delivery systems
based on hydroxypropyl methylcellulose (HPMC). Advanced Drug Delivery
Reviews 48 (23), 139157.
Sinha Ray, S., Yamada, K., Okamoto, M., Ueda, K., 2003. New polylactide/layered
silicate nanocomposite, 2: current improvements of material properties,
biodegradability and melt rheology. Polymer 44, 857866.
Vergnaud, J.M., 1998. Problems encountered for food safety with polymer packages:
chemical exchange, recycling. Advances in Colloid and Interface Science 78,
267297.
Vergnaud, J.M., Rosca, I.D., 2006. Assessing Food Safety of Polymer Packaging. Rapra
Technology. pp. 2136.
Vermeiren, L., Devlieghere, F., Van Beest, M., De Kruijf, N., Debevere, J., 1999.
Development in the active packaging of foods. Trends in Food Science and
Technology 10, 7786.