Accepted Manuscript

Metformin hydrochloride microencapsulation by complex coacervation: Study of size

distribution and encapsulation yield using response surface methodology

Daya Mancer, Eric Allemann, Kamel Daoud

PII: S1773-2247(17)30919-X
DOI: 10.1016/j.jddst.2018.03.015
Reference: JDDST 608

To appear in: Journal of Drug Delivery Science and Technology

Received Date: 23 October 2017

Revised Date: 9 March 2018
Accepted Date: 10 March 2018

Please cite this article as: D. Mancer, E. Allemann, K. Daoud, Metformin hydrochloride
microencapsulation by complex coacervation: Study of size distribution and encapsulation yield using
response surface methodology, Journal of Drug Delivery Science and Technology (2018), doi: 10.1016/
j.jddst.2018.03.015.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Metformin hydrochloride microencapsulation by complex coacervation:
Study of Size distribution and encapsulation yield using response surface
methodology

Daya MANCER1, Eric ALLEMANN2, Kamel DAOUD1

1
Laboratory of Transfer Phenomena, Faculty of Mechanical and Process Engineering,

PT
University of Science and Technology Houari Boumediene, El Alia, BP32, Bab Ezzouar,
16111, Algiers, ALGERIA.
2

RI
School of pharmaceutical sciences, University of Geneva, University of Lausanne, 30 quai
E.Ansermet, 1211 Geneva 4, SWITZERLAND

SC
Abstract
Metformin hydrochloride is a biguanide antihyperglycemic agent, widely used in the
management of non-insulin dependent diabetes mellitus (type 2), it has a short biological half-

U
life of 1.5–1.6 h, and its daily requirement is 1.5–3 g/day with an absolute bioavailability of
50–60%, when administered orally. However some gastro-intestinal symptoms, like
AN
abdominal discomfort, nausea and diarrhea may occur during the treatment.
This study was carried out to optimize conditions for metformin microencapsulation
by complex coacervation using response surface methodology (RSM). Sunflower oil was used
M

to obtain the primary emulsion, whereas the coating materials of microencapsulation were
soybean protein isolate (SPI) and pectin. For the complex coacervation process, soy lecithin
and Tween 80 were used as surfactants to enhance W/O/W double emulsion stability. Thus,
D

double emulsion followed by complex coacervation is studied.

TE

The microencapsulation of metformin under the optimized conditions ensures the

mean size of 16 µm of microcapsules combined with the narrow size distribution of 1.3 (span)
and the highest yield reaching up to 84 %.
EP

These results show that complex coacervation using SPI/pectin as wall material was an
efficient method which have never been reported in literature as microencapsulation process
of metformin hydrochloride.
C

Keywords: Metformin hydrochloride; complex coacervation; double emulsion;

AC

response surface methodology (RSM); soybean protein isolate (SPI); pectin,

 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

Metformin hydrochloride microencapsulation by complex coacervation:

Study of Size distribution and encapsulation yield using response surface
methodology

1. Introduction
Metformin hydrochloride (1, 1-Dimethylbiguanide hydrochloride) is an

PT
antihyperglycemic drug used in the management of type 2 diabetes [1]. Some high incidence
of concomitant gastro-intestinal symptoms, such as abdominal discomfort, nausea and
diarrhea may occur during the treatment. Therefore, administration of an extended release

RI
dosage form of microencapsulated metformin hydrochloride, once daily, could reduce the side
effects, therefore the dosing frequency improve patient compliance [2, 3]. Metformin
microencapsulation methods for extended release system that have been mostly reported

SC
include emulsion solvent evaporation using pectin, low permeability eudragit, high
permeability eudragit and ethylcellulose [4-7], ionic gelation using cordia gum/gellan gum,
alginate/(ethylcellulose, hydroxypropylmethylcellulose (HPMC), carbopol, chitosan),

U
alginate/ ispaghula and alginate/ gum karaya [8-11].
In order to distinguish the simple coacervation of a single polymer, Bungenberg de
AN
Jong and Kruyt, study the system of Arabic gum-gelatin, coined the name “complex
coacervation” to the separation of a macromolecular solution composed of two oppositely
charged polymers into two immiscible liquid phases [12, 13]. Complex coacervation is a
M

phase separation process based on the simultaneous desolvation of oppositely charged

polyelectrolytes induced by media modifications [14]. Microcapsules produced by
D

coacervation possess generally controlled release characteristics and heat resistant properties
[15]. In most cases, the two polymers used in the complex coacervation process, include
TE

protein and polysaccharide molecules. The dense liquid phase, which is relatively
concentrated in macromolecules, is called the coacervate, it refers to the metastable
suspension of macroion-rich droplets. [16]
EP

The protein–polysaccharide combinations that have been reported for the

microencapsulation of several drugs by complex coacervation include gelatin/ Arabic gum for
ibuprofen, vitamin A and lutein [17-19], whey protein/ acacia gum to encapsulate dodecyl
C

acetate [20] and many other drugs [21-23]. Pectin and soy protein isolate (SPI) have been
successfully used in other studies to promote the encapsulation of hydrolyzed casein and
AC

propolis [24, 25].

Particle size measuring and size distribution determination at an early stage in drug
forms development are therefore of significant benefit to researchers who need to engineer
specific product attributes, they are often important parameters in defining product
performance. According to Malvern Instruments Experts (2014) and A.F.T. Silva, et al
(2013), particle size measurement can be used as a tool to predict and control properties such
as product stability, uniformity, flowability and appearance, and can also help with
understanding the processability of a new material [26, 27].
Particles size distribution can be expressed by several parameters like 10th percentile
(D10), the median particle size (D50), and 90th percentile (D90), some authors prefer a single

CONFLICTS OF INTEREST: NONE

 ACCEPTED MANUSCRIPT
number answer such as D50. More experienced particle scientists cringe when they hear this
question, knowing that a single number cannot describe the distribution of the sample. A
better approach is to report both a central point of the distribution along with one or more
values to describe the width of distribution, the values of span are often used to answer this
approach [28,29]. In the present study, span gives an indication of how far the 10 percent and
the 90 percent points are apart, normalized with the midpoint.
Pectin is a polymer formed from units of D-galacturonic acid connected by α-(1- 4)
glycosidic bonds, its molecular weight can vary from 50 to 150 Kilo-Daltons [29]. It is a

PT
water soluble branched anionic polysaccharide generally obtained from vegetable cellular
walls like the skin and pulp of citric fruits, apples and also from algae [29]. In the same way,
it has the advantage of being more stable than the majority of other polysaccharides at acidic

RI
pH [31, 32]. The presence of several freely available –COOH and –OH functional groups act
as sites to interact with charged proteins [30]. Then, because of this cross–linking property,

SC
pectin is expected to increase the efficiency of drug encapsulation in a polysaccharide–protein
complex. It is known that pectin is resistant to stomach and intestine enzymes, but completely
degraded by the colonic bacterial enzymes, in the other hand the hydrophilic and swelling

U
nature of pectin may also cause the extended drug release from the delivery device [33].
Soy protein isolate (SPI) is produced from defatted soy meal by alkali extraction
AN
followed by acid precipitation (pH 4.5) [34]. The soy isolate is enriched with more than 90 %
of protein, this protein is amphoteric and has an isoelectric point of 4.5. The soy protein
solubility depends on the pH, therefore, the addition of sodium hydroxide increases the
M

solubility of protein reaching 90% with a pH 7.5 and 95% with a pH 10 [16, 35].
As other hydrophilic compounds, metformin might be incorporated efficiently in a
W/O/W double emulsion and further encapsulated by a complex coacervation process.
D

The relationship between a response and a set of variables of interest affecting a

process can be designed and evaluated using response surface methodology (RSM) which is a
TE

collection of statistical helpful techniques allowing deducing inferences about the

optimization procedure.
Hence, RSM requires minimum experimentation and time, thus showing to be far more
EP

fruitful and economical than the conventional formulation development methods. This
methodology encircles the generation of polynomial equations and of response over the
experimental domain to determine the optimum formulation [36].
C

The aim of this work is to study the metformin microencapsulation by complex

AC

coacervation, which to our knowledge was never reported. In addition, this study was carried
out to optimize this process using RSM and to determine the most efficient conditions for the
microencapsulation of metformin by W/O/W double emulsion followed by complex
coacervation. This study is based on the analysis of encapsulation yield and size distribution
which is reported as span.

2. Material and methods

2.1. Material
Metformin hydrochloride and polysorbate 80 (Tween 80) were obtained as gift from
SAIDAL Pharmaceutical Industry (Algiers, Algeria). Soy lecithin was purchased from
Argolanda B.V (NL). Soy protein isolate (SPI) was purchased from Puritan’s Pride, Inc
 ACCEPTED MANUSCRIPT
(Oakdale, NY, USA). Low methoxyled pectin was purchased from Sigma Aldrich (Saint
Louis, MO, USA). Sunflower oil was purchased from Cévital Spa (Bejaia, Algeria).

2.2. Step of microencapsulation

The method of Mendanha et al. [24] was used to prepare the microcapsules with slight
modifications, adding a double emulsion step at the beginning of the process of complex
coacervation. Thus, two-step method was used for preparing W/O/W emulsions (fig. 1).
In the first step, a 1% (w/v) of metformin aqueous solution was emulsified in

PT
sunflower oil at 1 :1.7(v/v) at 40ºC ±1°C by an Ultra Turrax homogenizer (IKA T25) for 3
min at 15000 rpm, using soybean lecithin as emulsifier at three different concentrations.
In a second step, the primary emulsion was dispersed and emulsified; the outer

RI
aqueous phase is formed by 2.5% (w/v) solution of soy protein isolate adjusted at pH 8 and
containing Tween 80 at three different concentrations; the mixture was homogenized at 15000
rpm for 3 min.

SC
The pectin aqueous solution (2.5 % w/v) was slowly added to the W/O/W emulsion
under magnetic stirring. The SPI/Pectin (w/w) ratio was fixed at 1:1.
To promote coacervation, the pH was adjusted to 4.4 by adding diluted hydrochloric

U
acid under constant magnetic stirring. All these steps were carried out at 40°C ±1°C. The
obtained suspensions were stored at 5°C ±1°C overnight to induce decantation.
AN
2.3. Optimizing the microencapsulation process by RSM
RSM was used to investigate the two responses (variation of microencapsulation yield
M

and size distribution width) with respect to operating parameters including the concentration
of the two surfactants, stirring speed and the stirring duration.
The compositions of four variables were designed by central composite design (CCD)
D

which is 2k factorial experiment with star points and central points [37].
Twenty seven experimental designs consisting of 16 complete factorial experiments, 8
TE

star points and 3 central points were generated with 2 levels for the 4 factors by the principle
of RSM using JMP 9.0.0 (SAS Institute Inc. Cary, NC, USA). A quadratic polynomial
regression model was used to fit the data points.
EP

2.4. The four factors limits determination

Various studies showed that the surfactant concentration needed for the formulation of
a stable emulsion must be lower than the critical micellar concentration (CMC) [38-40].
C

According to Mayya et al. and Bhattacharyya et al. [41, 42], the surfactant
AC

concentration allowing an optimal encapsulation rate is between the critical aggregation

concentration (CAC) and the critical micellar concentration (CMC).
2.4.1 Surface tension limits
In order to determine the CMC of the lipophilic surfactant (soy lecithin), surface
tension measurements were done at 40°C ±1°C by using a platinum ring tensiometer (KRÜSS
K6) according to the method used by several authors [38-40].
The critical aggregation concentration (CAC) represents the beginning of association
between a protein and a surfactant and it represents the concentration to which the fixing of
surfactant on the protein chains becomes co-operative. To consider this concentration, the
protein was dissolved in distilled water (25 g/ L) and adjusted to pH 8. A quantity of
 ACCEPTED MANUSCRIPT
surfactant (100 mmole/ L) was added and several dilutions were carried out using the protein
solution. The surface tension was determined as described above for the CMC of soy lecithin
using the platinum ring tensiometer.

2.4.2 Stirring Speed and duration limits

The lower and higher limits of stirring speed and the stirring duration were tested
based on previous studies [16, 19, 20, 11, 41], so, stirring durations of 30 and 120 min and a
stirring speed corresponding to 400 and 1000 rpm were selected.

PT
2.5. Microencapsulation yield
The obtained suspensions were harvested, diluted then filtered; non encapsulated
metformin in these filtered solutions was analyzed by UV-Visible spectrometry OPTIZEN

RI
2120 at 233 nm (characteristic wavelength of the metformin).
Microencapsulation yield of the various samples was defined as the ratio of
microencapsulated metformin to the total used metformin in the preparation; MEY was

SC
calculated according to the following equation [16]:
–
MEY (%) = ×

U
AN
2.6. Granulometric analysis
A sample of each batch was analyzed by laser diffraction using a Mastersizer 2000
(MALVERN Instruments) [28, 42].
M

Laser diffraction is one of the most used instruments to measure the particle size
distribution, implying an interest in the width or breadth of the distribution. Experienced
scientists typically shun using a single number answer to the question “What size are those
D

particles?”. To describe distribution width, once “model independent” algorithms were

introduced many particle scientists began using different calculations [43, 44].
TE

We can deduct from the granulometric analysis many important parameters like the
volume-surface diameter D[3,2] (Sauter diameter) which is the ratio of the
EP

cube of the volume equivalent diameter to the square of the surface-

equivalent diameter and defined as:
∑!
, =
∑!
C
AC

One of the common values used for laser diffraction results is the span, with the strict
definition shown in the equation below:

"# – "
Span =
"$

For particle size distributions, the median is called the D50 (or Dv0.5: median for a
volume distribution). Dn50 is used for number distributions, and Ds50 is used for surface
distributions. The D50 is the size in microns that splits the distribution with half above and
half below this diameter.
Since the primary result from laser diffraction is a volume distribution, the default D50
cited is the volume median and D50 typically refers to the Dv50 without including the v. This
 ACCEPTED MANUSCRIPT
value is one of the easier statistics to understand and one of the most meaningful for particle
size distributions. [28].
Similarly, 90 percent of the distribution lies below the D90, and 10 percent of the
population lies below the D10 [28].
Span allow us to understand the distribution width, the smaller span value indicates the
narrower size distribution [45-52]. Certain authors use the span value to analyze the
polydispersity in the particle size distribution and characterize the microspheres as
monodisperse, homogenous and heterogeneous systems, then, higher span indicates the high

PT
level of non-uniformity [53].
3. Results and Discussion
3.1. Factors limits for RSM

RI
a- Identification of critical micellar concentration (CMC) of Soy lecithin
The surface tension measurement of sunflower oil containing different concentrations

SC
of soy lecithin was determined, then we observe that without surfactant, the sunflower oil
surface tension was 34.25 mN/m and is in agreement with literature [54]. However, the
addition of soy lecithin to the oil led to a significant drop of surface tension at 1 mmole/ L

U
corresponding to the CMC.
AN
b- Determination of critical aggregation concentration (CAC) and critical
micellar concentration (CMC) of Tween 80 concentration
To enable selection of Tween 80 concentrations for the experimental design,
determination of CAC and CMC were first needed.
M

As presented in fig. 2 the results show that proteins solubilization in water decreases
its surface tension. When the surface tension of water containing a concentration of SPI (25
D

g/L) without Tween 80, the surface tension is 46 mN/m, while in water without protein it is
TE

of 69 mN/m at 40°C, (Blue rhomb and red triangle respectively corresponding to 0 mmole/ L
of Tween 80 in fig. 2). The addition of small quantity of surfactant decreases slightly the
value of SPI solution surface tension, then, we notice an inflection point reaching the value of
5×10-5 mmole/ L synonymous of the critical point of aggregation (CAC), after a tiny increase,
EP

the interfacial tension decreases again until reaching a value of 43 mN/m. The critical
aggregation concentration represents the beginning of association between soy protein isolate
C

and Tween 80; it also represents the concentration to which the fixing of surfactant on the
protein chains becomes co-operative.
AC

Fig. 2 reports the change of the surface tension as function of the evolution of
hydrophilic surfactant (Tween 80) concentration in water. The literature indicates that the
CMC of Tween 80 in water is estimated to 0.012 mmole/ L at 20-25°C. At 40 °C, from fig. 2,
we can deduce the CMC of Tween 80 evaluated to 0.005 mmole/ L. This result is in
agreement with the literature, which stipulates that the value of the critical micellar
concentration of non-ionic surfactant decreases with the increase in the temperature (under
50°C) [55].
Fig. 2 shows the interval which extends from the critical aggregation concentration to
the critical micellar concentration, which was selected as field for the experimental design in
 ACCEPTED MANUSCRIPT
the present study. The limits of the experimental design are customized in table 1 considering
the above CMC and CAC as well as information collected in the literature.

Table 1: Factors and limits for the experimental design

PT
Factors Low limits High limits Center points
(-1) (+1) (0)
X1 : Soy lecithin concentration
0,04 0.8 0.42

RI
(mmole/L)
X2 : Tween 80 concentration
10-4 25×10-4 13×10-4
(mmole/L)

SC
X3 : Stirring speed (rpm) 400 1000 700
X4 : Stirring duration (min) 30 120 75

U
3.2. Microencapsulation yield
AN
Table 2 shows the values of microencapsulation yield which vary from 70.5% to
84.7%, these yields are higher than those obtained by Rocha-Selmi et al. [56] who
encapsulated aspartam by double emulsion followed by complex coacervation using Arabic
M

gum and gelatin (42.2-71.7%), and slightly higher than the encapsulation yields of propolis by
SPI and pectin (66.12-72%) [25]. Our results are in the same range as those reported by Jun-
xia et al. [16] encapsulating the orange oil using SPI with Arabic gum (between 65 and 85%),
D

and those observed by Qv et al. [19] encapsulating lutein using the gelatin and Arabic gum.
TE

Mendanha et al. [24] succeeded to reach yields up to 91.6 % for the encapsulation of casein
hydrolizate by SPI and pectin by double emulsion followed by complex coacervation method.
The study carried out by Maestrelli et al. [57] consisting on the preparation of
EP

metformin particles by dropping alginate in a calcium chloride solution gives an

encapsulation efficiency varying from 8.33% to 16.87%. Swamy et al. [58] obtained the
encapsulation efficiency of 32% to 73% when they encapsulate the metformin with
C

carboxymethylcellulose in a hydrogel beads. The ionotropic gelation method riches an

encapsulation efficiency varying from 69% to 92% using ispaghula husk mucilage and gellan
AC

gum as polymers [59] and from 32% to 94% using tamarind seed polysaccharide and alginate
[60].

Table 2: Central composite design consisting of 27 experiments for the optimization of

microencapsulation of metformin hydrochloride
Ru Coded Microencaps Particles
Configur Process variables
n variables ulation yield Span Size :
ation
no. X 1 X 2 X 3 X 4 X 1 X 2 X 3 X 4 (%) D[3,2]
-4
1 ++−− +1 +1 -1 -1 0,8 25×10 400 30 71.4 1.92 28.0
-4
2 0 +α00 0 +1 0 0 0,42 25×10 700 75 79.7 2.38 12.6
 ACCEPTED MANUSCRIPT
3 −−−+ -1 -1 -1 +1 0,04 10-4 400 120 81.4 2.22 16.5
-4
4 +α000 +1 0 0 0 0,8 10 700 75 77.1 1.94 7.72
-4
5 −−−− -1 -1 -1 -1 0,04 10 400 30 79.0 2.68 22.0
-4
6 −+−− -1 +1 -1 -1 0,04 25×10 400 30 80.3 1.42 13.2
-4
7 0000 0 0 0 0 0,42 13×10 700 75 78.6 2.08 15.8
-4
8 00 +α0 0 0 +1 0 0,42 13.10 1000 75 76.8 1.69 13.8
-4
9 −α 000 -1 0 0 0 0,04 13×10 700 75 75.2 2.22 13.3
-4
10 −++− -1 +1 +1 -1 0,04 25×10 1000 30 76.4 1.83 12.4

PT
11 0 −α00 0 -1 0 0 0,42 10-4 700 75 79.7 2.38 10.8
-4
12 −−++ -1 -1 +1 +1 0,04 10 1000 120 81.5 1.50 14.8

RI
-4
13 0000 0 0 0 0 0,42 13×10 700 75 79.1 2.34 13.6
-4
14 +−−+ +1 -1 -1 +1 0,8 10 400 120 80.5 1.77 16.0
-4
15 0000 0 0 0 0 0,42 13×10 700 75 75.1 2.29 10.5

SC
-4
16 −+−+ -1 +1 -1 +1 0,04 25×10 400 120 84.7 1.41 15.8
-4
17 +−+− +1 -1 +1 -1 0,8 10 1000 30 77.6 1.33 8.80
-4
18 −+++ -1 +1 +1 +1 0,04 25×10 1000 120 84.5 1.88 12.3

U
-4
19 ++++ +1 +1 +1 +1 0,8 25×10 1000 120 80.3 1.43 15.8
AN
-4
20 +−−− +1 -1 -1 -1 0,8 10 400 30 73.7 2.56 0.37
-4
21 +++− +1 +1 +1 -1 0,8 25×10 1000 30 72.6 0.74 6.75
-4
22 000 +α 0 0 0 0 0,42 13×10 700 120 81.4 2.38 9.76
M

-4
23 −−+− -1 -1 +1 -1 0,04 10 1000 30 70.5 1.71 8.23
-4
24 00 −α0 0 0 -1 0 0,42 10 400 75 75.1 2.33 10.3
-4
25 +−++ +1 -1 +1 +1 0,8 10 1000 120 82.4 1.59 16.1
D

-4
26 ++−+ +1 +1 -1 +1 0,8 25×10 400 120 77.1 1.56 11.6
-4
27 000 −α 72.7 2.45 9.51
TE

0 0 0 -1 0,42 13×10 700 30

X1: soy lecithin concentration; X2: Tween 80 concentration; X3: stirring speed; X4: stirring
duration; +: High limit or high level of the factor, −: low limit or low level of the factor, 0: the
center point, α: the distance from the center of the design space to a factorial point (in the case
EP

of CCD, α = 1).

3.3. Microscopic observation

C

The morphology of all the microparticles formulated during the successive steps were
analyzed with optical microscope (OPTIKA MICROSCOPES, ITALY) as shown in fig.3.
AC

The microparticles in the W/O emulsion appeared spherical and distinct from each other with
different size (fig.3.a). However, the microparticles of the W/O/W double emulsion obtained
after adding the first emulsion to the SPI solution are shown in the fig.3.b, where we observe
two types of W/O/W microparticles. The first is consisted of only one globule reservoir which
is a single hollow chamber within the capsule according to the monocored system, when the
second consist of different sized chambers within the particles according to the polycored
system. Fig 3.c and fig 3.d show the particles of the metformin obtained after the complex
coacervation between pectin and SPI. Thus, we notice the size heterogeneity of the obtained
particles.
 ACCEPTED MANUSCRIPT
3.4. Granulometric analysis

As reported in table 2, the granulometric analysis shows the mean sizes D [3,2] from
0.371 to 28.024 µm lower than the particles diameter reported by Rocha-Selmi et al. varying
between 84.22 and 102.38 µm by using gum Arabic and gelatin as polymers[56]. However,
our results close more those obtained by Mendanha et al. between 16.24 and 24.12 µm [24].
Some studies report that the preparation of metformin calcium alginate offers a
particles size varying between 701nm and 1688 nm [57], the particle size of metformin

PT
carboxymethylcellulose hydrogel beads obtained by Swamy et al. gives diameters from 1564
µm to 1954 µm [58]. Ionotropic gelation method allows a particle size varying between 1.28
mm and 1.61 mm using ispaghula husk mucilage and gellan gum [59] and an average of 1.24

RI
mm using tamarind seed polysaccharide and alginate [60].
Regarding the above comparison of different results obtained by several authors
studying metformin encapsulation, we notice that in one hand, high yield encapsulation was

SC
reached but with a corresponding high particles size, in other hand, other methods than
complex coacervation allow to obtain metformin nanoparticles (very small particles size) with
low encapsulation yield. According to our results, complex coacervation let the obtaining of

U
lower sized microparticles with a high encapsulation yield.
As visualized on the microscopic observations (fig.3), the granulometric analysis
AN
reveals heterogeneity of particles size distribution observed for the twenty seventh
experiences; then, four distribution profiles are distinguished in our study (table 3, fig. 4).
M

Table 3: Distribution profiles information.

Distribution profiles Run no. Span interval
For 13 tests: 1, 4, 6, 11, 12,
D

Unimodal distribution 0.74 and 2.45

15, 16, 18, 21, 22, 24 and 26.
TE

Imperfect unimodal
For 6 tests: 2, 8, 9, 10, 17 and
distribution with a spread out 1.33 and 2.38
25.
base
EP

Bimodal distribution with two For 5 tests: 7, 14, 19, 20 and

1.43 and 2.56
distinct peaks 23.
2.22; 2.68 (maximum value
Imperfect bimodal distribution For 3 tests: 3, 5 and 13.
C

for all the tests) and 2.34.

AC

The unimodal distributions with weak span is the needed profiles which concord with
microparticles having or approaching the mean diameter.

3.5. Optimizing microencapsulation yield and span using RSM

For estimating both simple and combined effects of formulation variables (X1, X2, X3,
X4) on metformin encapsulation yield and size distribution, systematic optimization,
separately at first, for each response and then simultaneously for the two responses was
carried out using the central composite design, consisting of 27 experiments. The method of
analysis was carried out in a similar manner as in several previous reports [61-69].

3.5.1. Optimizing the microencapsulation yield

 ACCEPTED MANUSCRIPT
The metformin encapsulation yield obtained by complex coacervation could be
represented in the form of a polynomial second degree equation relating to the defined
experimental field and implying the four studied factors. The estimated effects of each
coefficient are represented in table 4.
MEY = 77.156-1.149 X1+0.024 X2-0.03 X3+3.313 X4-1.659 X1X2+1.427 X1X3-
0.058 X1X4+0.189 X2X3+0.057 X2X4+0.768 X3X4-0.773 X12+2.813 X22-0.965 X32+0.1 X42

The value of the determination coefficient R² is 0.89 and the calculated value of F

PT
(Fisher-Snedecor) is 0.679. Therefore, the high values of the statistical tests (R² and F),
enable to consider the obtained model of the second degree as a valid model to represent the
experimental design results for metformin encapsulation yield.

RI
Table 4 : Coefficients significance of the encapsulation yield quadratic equation.
Regression Standard Significance

SC
Variables t-value
coefficient error (p-value)
Constant 77.156 0.682 113.04 <0.0001*
X1 -1.149 0.437 -2.63 0.0218*

U
X2 0.024 0.437 0.06 0.9569
AN
X3 -0.030 0.437 -0.07 0.9465
X4 3.313 0.437 7.59 <0.0001*
X1X2 -1.659 0.463 -3.58 0.0038*
M

X1X3 1.427 0.463 3.08 0.0095*

X1X4 -0.058 0.463 -0.13 0.9019
X2X3 0.189 0.463 0.41 0.6895
D

X2X4 0.057 0.463 0.12 0.9040

X3X4 0.768 0.463 1.66 0.1230
TE

2
X1 -0.773 1.155 -0.67 0.5158
2
X2 2.813 1.155 2.44 0.0314*
2
X3 -0.965 1.155 -0.84 0.4196
EP

2
X4 0.1 1.155 0.09 0.9320
* Values in boldface represent significant factors (p < 0.05).
C

From the polynomial equation, one must know the significance of simple effects, the
interaction effects (combined) and quadratic effects. Thus, the Pareto diagram as shown in fig.
AC

5, reveals five significant effects respectively ordered according to their t-value and p-value
as stirring duration, interaction between the two surfactants concentrations (soy lecithin
concentration* Tween 80 concentration), soy lecithin concentration* stirring speed, soy
lecithin concentration and finally the quadratic effect of Tween 80 concentration.
The encapsulation yield rises with the augmentation of the stirring duration (positive t-
value), but it decreases with the interaction between the two surfactants concentrations
(negative t-value). Both the interaction between lecithin concentration * stirring speed and the
quadratic effect of Tween 80 concentration affect positively the encapsulation yield, however,
the soy lecithin concentration has a negative effect on the response. Thus, we deduce that the
encapsulation yield is more sensitive to the chemical proprieties (surfactants concentrations)
 ACCEPTED MANUSCRIPT
of the formulation system than the process factors (stirring speed and its duration).

We note that some linear, quadratic and some combined effects are not significant. So,
they are generally neglected in the mathematical model, which will be written as:
MEY= 77.156-1.1496 X1+3.313 X4-1.659 X1X2+1.4279 X1X3 +2.813 X22

The 3D response surface graphs for the microencapsulation yield as a function of the
four selected parameters (X1, X2, X3, and X4) were shown in Fig.6.
Fig.6.a illustrates the response surface of the combined effects of the soy lecithin

PT
concentration as well as Tween 80 concentration on the metformin microencapsulation yield,
whereas the stirring speed and duration are fixed at their central levels (respectively 700 rpm
and 75 min). The value of the maximum microencapsulation yield that we can reach is 81, 4%

RI
for the low level of soy lecithin concentration and the high level of Tween 80 concentration.
Fig.6.b illustrates the combined effects between the concentration of first surfactant

SC
(soy lecithin) and the stirring speed. Microencapsulation yield of 78% was obtained when the
low levels of the two variables were combined, whereas, the variation of the soy lecithin
concentration from his low to his high level and by maintaining the stirring speed at 400 rmp,

U
lead to a reduction of the encapsulation yield down to 73%.
As shown in fig.6.c, fixing the value of stirring duration and varying the value of the
AN
soy lecithin concentration, a slight reduction in the encapsulation yield was observed. When
the stirring duration vary from 30 to 120 min, encapsulation yield increases (81,5% with a soy
lecithin concentration of 0,04 mmole/ L, stirring duration of 120 min, Tween 80
M

concentration of 13×10-4 mmole/ L and stirring speed of 700 rpm). These observations
confirmed the results obtained from the parameters interactions study as those extracted from
Pareto chart which emphasize the importance of the stirring duration.
D

The response surface of the combined effects of Tween 80 concentration with stirring
speed (fig.6.d) has a horse saddle form from which show high values of encapsulation yield
TE

(79% to 80,5%) with both low and high levels of Tween 80 (10-4 mmole/ L and 25×10-4
mmole/ L, respectively) regardless of the stirring speed, but the maximum encapsulation
yield can be reached with the central value of the of stirring speed (700 rmp).
EP

Fig.6.e illustrates response surfaces of the combined effects of the Tween 80

concentration and the stirring duration on the encapsulation yield. The encapsulation yield
increases in a linear and fast way when stirring time passes towards its high level. The
C

quadratic effect of the Tween 80 concentration presents an influence on the encapsulation

AC

yield giving a minimum when it is in the central value. This combination suggests a
maximum of yield when the stirring duration trends to 120 min and the Tween 80
concentration is at the extremities of the experimental field. In these fields, yields higher than
84% are reached.
Increase in soy lecithin concentration allow to a reduction in the encapsulation yield,
because the presence of surfactant supports the deformation and the stretching of the
interface, from where the brittleness of formed interfacial film [70-73]. During the second
phase of emulsification, the concentration of Tween 80 is important, it shouldn’t be too weak
to avoid the fast exhaustion of its presence in the interfacial film; thus, the stretching of the
film lead a reduction in the excess of surface, then the film can be very unstable. On the other
 ACCEPTED MANUSCRIPT
hand, with a sufficient quantity of surfactant, the part of the thinnest film having the most
raised elasticity, has a great resistance to the stretching [74, 75].
Fig.6.f shows response surfaces of the combined effects of stirring speed and duration
on the encapsulation yield. The encapsulation yield variation ranges according to low and
high level of stirring time; for 30 min stirring time the yield decreases when stirring speed
increases from 400 rpm to 1000 rpm, then, it changes from 74% to 72%, whereas for 120 min
stirring time, the encapsulation yield increases with the increase in stirring speed (from 79%
to 81,5%).

PT
The stirring speed increasing from 400 rpm to 700 rpm allow to reach a slight
increasing of encapsulation yield due to the suspension homogenization resulting in protein

RI
charges meeting with polysaccharides ones. From this stirring speed (700rpm) we notice a
decrease of encapsulation yield; this can be explained by the fragmentation of the formed
particles and the release of the active ingredient initially encapsulated [74]. At the operational

SC
temperature to which polymers are exposed (40°C) the reactive sites of proteins are revealed,
therefore, more protein is exposed more the hydrostatic bonds are formed. On another side,
the hydrophobic bonds are enhanced at this temperature [76-81].

U
By analyzing the various response surfaces, the data presented in the Pareto chart
AN
(fig.5) was confirmed. The various response surfaces let us to select the optimum conditions,
which are presented in table 5 in coded and real values. For the optimized values of the four
parameters, the calculated yield was 84 %.
M

Table 5: Optimized operational parameters values concerning the microencapsulation

yield.
D

Factors Coded variables Process variables

X1 (mmole/ L) -0.691 0.157
TE

X2 (mmole/ L) 0.873 23×10-4

X3 (rpm) -0.286 614
X4 (min) 0.946 117.558
EP

3.5.2. Optimizing the size distribution

The particles size distribution (span) obtained by complex coacervation during
C

different tests is reported by the second degree polynomial equation; thus, it is expressed
AC

according to the estimated effects of each coefficient as represented in table 6.

Span = 2.327-0.112 X1-0.175 X2-0.231 X3-0.05 X4-0.002 X1X2-0.12 X1X3 + 0.027
X1X4+0.166 X2X3+0.099 X2X4+0.15 X3X4-0.292 X12+0.017 X22-0.366 X32
+ 0.04 X42

The value of the determination coefficient R2 and calculated Fisher value are
respectively 0.86 and 3.94. These tests allow to judge relevance of the variables and to model
the studied responses through a second degree polynomial model (table 6).
For each parameter, the effects significance of linear, quadratic and interactions
between the variables were tested and reported on a Pareto chart represented in fig.7 which
reveals five significant effects ordered according to their t-value and p-value as the stirring
 ACCEPTED MANUSCRIPT
speed followed by the Tween 80 concentration, the interaction of the tween 80 concentration
and the stirring speed and the interaction between the stirring speed and the stirring duration
and finally the quadratic effect of stirring speed. The increase of the stirring speed and the
tween 80 concentration allow the decrease of the size distribution (desired effect), the
quadratic effect of the stirring speed affects the response in the same way. In the other way,
the interaction between the tween 80 concentration * the stirring speed and the interaction
between the stirring speed * the stirring duration rise the size distribution (undesired effect).
We deduce that the size distribution is affected by the process factors (stirring speed and its

PT
duration) more than the chemical factors (surfactants concentrations).
Neglecting insignificants effects, the mathematical model is simplified as:
Span = 2.327-0.175 X2-0.231 X3 +0.166 X2X3+0.150 X3X4-0.366 X32

RI
Table 6: Coefficients Significance of the span quadratic equation.

SC
Variables Regression Standard t-value Significance
coefficient error (p-value)
Constant 2.327 0.093 24.97 <0.0001*

U
X1 -0.112 0.0596 -1.88 0.0851
X2 -0.175 0.0596 -2.94 0.0124*
AN
X3 -0.231 0.0596 -3.88 0.0022*
X4 -0.05 0.0596 -0.84 0.4195
X1X2 -0.002 0.063 -0.03 0.9745
M

X1X3 -0.12 0.063 -1.89 0.0830

X1X4 0.027 0.063 0.43 0.6734
X2X3 0.166 0.063 2.63 0.0220*
D

X2X4 0.099 0.063 1.57 0.1422

TE

X3X4 0.150 0.063 2.37 0.0354*

2
X1 -0.292 0.158 -1.85 0.0892
2
X2 0.017 0.158 0.11 0.9167
2
EP

X3 -0.366 0.158 -2.32 0.0389*

2
X4 0.040 0.158 0.26 0.8024
* Values in boldface represent significant factors (p < 0.05).
C

The 3D response surface graphs for the size distribution as a function of the four
AC

selected parameters (X1, X2, X3, and X4) were shown in fig. 8.
Fig.8.a illustrates the response surface of combined effects of soy lecithin
concentration with Tween 80 concentration on the size distribution, whereas the stirring speed
and duration are fixed at their central levels (respectively 700 rpm and 75 min). From these
charts we can conclude that a span of 1.8 is obtained when the maximum concentrations of
surfactants are used (0,8 mmole/ L of soy lecithin and 25×10-4 mmole/ L of Tween 80).
Fig.8.b illustrates the combined effects linking the first surfactant concentration (soy
lecithin) to the stirring speed on the span; Combining the high levels of the two variables, a
size distribution result on 1,2.
The response surfaces of the combined effect of soy lecithin concentration with
stirring duration are reproduced on the fig.8.c. A combination of maximum soy lecithin
 ACCEPTED MANUSCRIPT
concentration (0,8 mmole/ L) including stirring time in the interval of 50-90 min leads to
minimal span.
Fig.8.d describes the response surface of the combined effects of Tween 80
concentration with stirring speed on the obtained span value. Size distribution decreases when
the surfactant concentration moves to its higher level. Stirring speed also presents an
influence on the span which decreases when the speed tends towards his high level. The span
is weaker when the Tween 80 concentration and the stirring speed tend towards their high
levels.

PT
Analysis of fig.8.e indicates a span response surface depending on Tween 80
concentrations combined with stirring duration. Thus, the weakest span is obtained when the
maximum Tween 80 concentration and a weak time of agitation were used.

RI
Size distribution response surface of the combined effects between stirring speed and
stirring duration reveals that the span reaches its minimal value for a maximum speed and a

SC
minimal stirring time (1000 rpm and 30 min respectively) (fig.8.f). Indeed, as previously
indicated, during the analysis of the Pareto chart, the linear and quadratic stirring speed
effects present a more marked influence. For the weakest size distribution, stirring speed must
tend towards its high level whereas the stirring time must be on its low level.

U
The suitable homogenization in terms of stirring speed and time of the emulsions
AN
confers a narrow distribution of the obtained product. Also, the coalescence of the droplets is
slowed down due to the increase of the repulsion between these droplets, and that because of
the increase in the surfactant concentration [82-84]. Thus, according to our results the size
M

distribution of the microparticles is more affected by the stirring speed and its duration,
conversely, the encapsulation yield is more effected by the surfactants concentration.
D

Optimized operational parameters values in coded and real values are presented on
table 7. With these optimized values of the studied factors, we obtain a value of the size
TE

distribution bordering the unit (1).

Table 7: Values of the optimized operational parameters concerning the span.

Factors Coded variables Process variables
EP

X1(mmole/ L) +1 0.8
X2 (mmole/ L) +1 25×10-4
C

X3 (rpm) +1 1000
X4 (min) -1 30
AC

3.6 Optimization and verification of the model

In order to obtain the maximum encapsulation yield and the minimum size
distribution, the predicted combination of parameters was summarized in table 9. Under these
conditions, the model predicted an encapsulation yield of 84% with a size distribution of 1.3.
To validate the optimum conditions predicted by this model, triplicate experiments were
conducted using the optimized process parameters and mean encapsulation yield value of 83.7
± 0.45 % and a size distribution of 1.26 ± 0.07 was obtained. The results are closely related
with the data obtained from the predicted model.
Table 9 : Values of the optimized operational parameters for both responses.
Factors Coded variales Process variables
 ACCEPTED MANUSCRIPT
X1 (mmole/ L) +1 0.8
X2 (mmole/ L) -1 10-4
X3 (rpm) +1 1000
X4 (min) +1 120

PT
RI
4. Conclusion
In this study, Metformin microcapsules with SPI/Pectin as wall materials were

SC
successfully prepared by double emulsion followed by complex coacervation technique. Thus,
27 experiments were necessary to elaborate the central composite design using response
surfaces methodology to determine the most significant studied factors that affect both MEY

U
and span of obtained suspension of microparticles.
Literature results show that microencapsulation of some hydrophilic actives other than
AN
metformin by double emulsion followed by complex coacervation allows to reach 42 -71 % of
encapsulation for aspartam using arabic gum and gelatin as wall materials with a particles
diameter varying between 84 and 100 µm. In addition, an encapsulation yield of 91% for
M

casein hydrolizate by SPI and pectin with lower diameters varying from 16 to 25 µm are
observed. Some studies carried out to encapsulate metformin hydrochloride using thin layer
evaporation report a particles size of 0.7 to 1.7 µm but with a low encapsulation efficiency
D

varying from 8% to 17%. However, hydrogel beads technic allow a higher encapsulation
efficiency of metformin reaching 32%-73% with an important particle size of 1.56 mm to
TE

1.95 mm. In addition, ionotropic gelation method, like Hydrogel beads technic; allow a higher
encapsulation efficiency of metformin with an average of 30 to 94% and a diameter of 1.24
mm to 1.61 mm.
EP

Our results manage with the objectives of high encapsulation yield and low particle
size offering a metformin encapsulation yield from 70 % to 85% with a particle size ranging
between 0.37µm and 28 µm. The optimized MEY values of 84 % corresponding respectively
C

to 0.157 (mmole/L) and 23×10-4 (mmole/L) as concentration of soy lecithin and Tween 80,
AC

614 rmp as stirring speed and 117 min as stirring duration.

Metformine microencapsulation with the SPI/pectin ratio of 1:1 have diameter varying
from 0.37 µm to 28 µm; those microparticles show a size distribution from 0.73 to 2.67,
whereas, optimum span bordering the unit (1) consequent respectively to 0.8 (mmole/L) and
25×10-4 (mmole/L) as concentration of soy lecithin and Tween 80, 1000 rmp as stirring speed
and 30 min as stirring duration. The optimal conditions for microencapsulation of Metformin
hydrochloride were respectively 0.8 (mmole/L) and 10-4 (mmole/L) as concentration of soy
lecithin and Tween 80, 1000 rpm as stirring speed and 120 min as stirring duration. In
conclusion, the microencapsulation of Metformin hydrochloride under the optimized
conditions by RSM ensures the mean size (16µm) of microcapsules with the tight size
distribution (span) of 1.3 and highest yield reaching to 84 %.
 ACCEPTED MANUSCRIPT
Finally, the new proposed double emulsion followed by complex coacervation technic
for the preparation of polymeric microparticles of metformin was found to be simple, easily
controllable and economical by using low coast excipient polymers (SPI and pectin).

PT
RI
References

SC
[1] British Pharmacopoeia, 2009.
[2] K. Sundaramoorthy, S. Kavimani, Formulation and evaluation of extended release

U
dosage form of metformin hydrochloride using a combined hydrophobic and hydrophilic
AN
matrix: rate of in-vitro and in-vivo release studies, Int. J Pharm Sci Rev Res, (2013).
[3] R. Aswal, A. Semalty, Chitosan microspheres of metformin hydrochloride and the
effect of using different concentrations of crosslinking agent, Int. Res. J. Invent. Pharm. Sci, 2
(2013) 22-27
M

[4] P. Banerjee, J. Deb, A. Roy, A. Ghosh, P. Chakraborty, Fabrication and

Development of Pectin Microsphere of Metformin Hydrochloride, ISRN Pharm, (2012).
D

http://dx.doi.org/10.5402/2012/230621
[5] I. Hasan, S. Paul, S. Akhter, N. J. Ayon, M. S. Reza, Evaluation and optimization of
TE

influence of permeability property and concentration of polymethacrylic polymers on

microspheres of metformin HCl, Dhaka Univ. J. Pharm. Sci., 12 (2013) 93-
103. http://dx.doi.org/10.3329/dujps.v12i2.17611
EP

[6] R. Maji, S. Ray, B. Das, A.K. Nayak, Ethyl cellulose microparticles containing
metformin HCl by emulsification-solvent evaporation technique : effect of formulation
variables, ISRN Polymer Science, 2012. http://dx.doi.org/10.5402/2012/801827
C

[7] P.K. Choudhury, M. Kar, Controlled release metformin hydrochloride microspheres

of ethyl cellulose prepared by different methods and study on the polymer affected
AC

parameters, J Microencapsul, 26 (2009) 46-53. doi: 10.1080/02652040802130503

[8] M. Ahuja, M. Yadav, S. Kumar, Application of response surface methodology to
formulation of ionotropically gelled gum cordia/gellan beads, Carbohydr Polym, 80 (2010)
161-167. https://doi.org/10.1016/j.carbpol.2009.11.005.
[9] G. Navneet, G. Akanksha, J. Neetesh, Formulation design and in-vitro evaluation of
metformin microspheres using ionotropic gelation technique, J Pharm Res, 4 (2011) 2103-
2106.
[10] F.M. Patel, A.N. Patel, K.S. Rathore, Release of metformin hydrochloride from
ispaghula sodium alginate beads adhered cock intestinal mucosa, Int J Curr Pharm Res, 3
(2011) 52-55.
 ACCEPTED MANUSCRIPT
[11] A.A. Kumar, T. Balakrishna, R. Jash, T.E.G.K. Murthy, A.A Kumar and B.
Sudheer, Formulation and evaluation of mucoadhesive microcapsules of metformin HCl with
gum karaya, Int J Pharm Pharm Sci, 3 (2011) 150-155.
[12] H.G. Bungenberg de Jong, Crystallisation-coacervation-flocculation. In H. G. Kruyt
(Ed.), Colloid Science, vol. 2, pp. 232-258, Amsterdam: Elsevier, 1949.
[13] E. Kizilay, A. B. Kayitmazer, P. L. Dubin, Complexation and coacervation of
polyelectrolytes with oppositely charged colloid, Adv Colloid Interface Sci, 167 (2011) 24-
37. https://doi.org/10.1016/j.cis.2011.06.006.

PT
[14] Z. H. Ma, D. G. Yu, C. J. Branford-White, H. L. Nie, Z. X. Fan, L. M. Zhu,
Microencapsulation of tamoxifen: Application to cotton fabric, Colloids Surf B Biointerfaces,
69 (2009) 85-90. https://doi.org/10.1016/j.colsurfb.2008.11.005.

RI
[15] Z. J. Dong, S. Q. Xi, S. Huab, K. Hayat, X. M. Zhang, S. Y. Xua, Optimisation of
cross-linking parameters during production of transglutaminase-hardened spherical

SC
multinuclear microcapsules by complex coacervation, Colloids Surf B Biointerfaces, 63
(2008) 41-47. https://doi.org/10.1016/j.colsurfb.2007.11.007.
[16] X. Jun-Xia., Y. Hai-Yan, Y. Jian, Microencapsulation of sweet orange oil by
complex coacervation with soybean protein isolate/gum Arabic, Food Chem, 125 (2011)

U
1267-1272. https://doi.org/10.1016/j.foodchem.2010.10.063.
AN
[17] I.O. Adeleke, I.S. Okafor, G. Alebiowu, Studies on the micromeritic properties of
ibuprofen microcapsules, The Pacific Journal of Science and Technology, 13 (2012) 469-472.
[18] V.B. Junyaprasert, A. Mitrevej, N. Sinchaipanid, P. Broome, D.E. Wurster,
M

Effect of process variables on the microencapsulation of vitamin A palmitate by gelatin-

acacia coacervation, Drug Dev. Ind. Pharm, 27 (2001) 561-566.
[19] X-Y. Qv, Z-P. Zeng, J-G. Jiang, Preparation of lutein microencapsulation by
D

complex coacervation method and its physicochemical properties and stability, Food
Hydrocolloid, 25 (2011)1596-1603. https://doi.org/10.1016/j.foodhyd.2011.01.006.
TE

[20] Y.T. Yu, L.N. Liu, X. L. Zhu, X.Z. Kong, Microencapsulation of dodecyl acetate by
complex coacervation of whey protein with acacia gum and its release behavior, Chin. Chem.
Lett, 23 (2012) 847-850. https://doi.org/10.1016/j.cclet.2012.05.006.
EP

[21] N. Devi, M. Sarmah, B. Khatun, T.K. Maji, Encapsulation of active ingredients in

polysaccharide–protein complex coacervates, Adv Colloid Interfac, 239 (2017) 136-145.
https://doi.org/10.1016/j.cis.2016.05.009.
C

[22] V. Ducel, J. Richard, Y. Popineau, F. Boury, Adsorption kinetics and rheological

AC

interfacial properties of plants at the oil-water interface, Biomacromolecules, 5 (2004) 2088-

2093. DOI: 10.1021/bm049739h.
[23] V. Ducel, J. Richard, J. Saulnier, Y. Popineau, F. Boury, Evidence and
characterization of complex coacervates containing plants protein: Application to
microencapsulation of oil droplets, Colloids Surf. A, 232 (2004) 239-
247.https://doi.org/10.1016/j.colsurfa.2003.11.001.
[24] D.V. Mendanha, S.E.M. Ortiz, C.S. Favaro-Trindade, A. Mauri, E.S. Monterrey-
Quintero, M. Thomazini, Microencapsulation of casein hydrolysate by complex
coacervation with SPI/pectin, Food Res Int., 42 (2009) 1099-1104.
https://doi.org/10.1016/j.foodres.2009.05.007.
 ACCEPTED MANUSCRIPT
[25] M.P. Nori, C.S. Favaro-Trindade, S.M. de Alencar, M. Thomazini, J.C. de
Camargo Balieiro, C.J. Contreras Castillo, Microencapsulation of propolis extract by
complex coacervation, LWT Food Sci Technol, 44 (2011) 429-435.
https://doi.org/10.1016/j.lwt.2010.09.010.
[26] Malvern Instruments, Application note: Using laser diffraction to support early stage
product development, 2014.
[27] A. F.T. Silva, A. Burggraeve, Q. Denon, P. Van der Meeren, N. Sandler, T. Van
Den Kerkhof, M. Hellings, C. Vervaet, J. P. Remon, J. Almeida Lopes, T. De Beer,

PT
Particle sizing measurements in pharmaceutical applications: Comparison of in-process
methods versus off-line methods, Eur. J. Pharm. Biopharm, 85(2013) 1006–1018.
https://doi.org/10.1016/j.ejpb.2013.03.032.

RI
[28] HORIBA Instruments, A guidebook to particle size analysis, HORIBA Instruments,
Inc. USA, (2012).

SC
[29] A.G.J. Voragen, W. Pilnik, J-F. Thibault, M.A.V. Axelos, C.M.G.C. Renard,.
Pectins. In: A. M. Stephen, Food polysaccharides and their applications. CRC Press Inc, 1995.
[30] V.B. Tolstoguzov, Functional properties of food proteins and role of protein-
polysaccharide interaction, Food Hydrocolloid, 4 (1991) 429-468.
https://doi.org/10.1016/S0268-005X(09)80196-3.
U
AN
[31] G. Dongowski, A. Lorenz, J. Proll, The degree of methylation influences the
degradation of pectin in the intestinal tract of rats and in vitro, J Nutr., 132 (2002) 1935-1944.
[32] D. Saito, S. Nakaji, S. Fukuda, T. Shimoyama, J. Sakamoto, K. Sugurawa,
M

Comparison of the amount of pectin in the human terminal ileum with the amount of orally
administration pectin, Nutrition, 21 (2005) 914-919.
https://doi.org/10.1016/j.nut.2005.01.005.
D

[33] R. Awasthi, G. T. Kulkarni, M. V. Ramana, T. de Jesus Andreoli Pinto, I. S.

Kikuchi, D. D. Molim Ghisleni, M.de Souza Braga, P. De Bank, K. Dua, Dual crosslinked
TE

pectin–alginate network as sustained release hydrophilic matrix for repaglinide, Int J Biol
Macromol, 97 ( 2017) 721-732. http://dx.doi.org/10.1016/j.ijbiomac.2017.01.050.
[34] S.E. Molina Ortiz, A. Mauri, E.S. Monterrey-Quintero, M.A. Trindade, A.S.
EP

Santana, C.S. Favaro-Trindade, Production and properties of casein hydrolysate

microencapsulated by spray drying with soybean protein isolate, LWT Food Sci Technol, 42
(2009) 919–923. https://doi.org/10.1016/j.lwt.2008.12.004.
C

[35] C.E. Stauffer, La protéine de soja en boulangerie, published and translated by ASA
AC

Europe & Maghreb, Cincinnati, Ohio, USA, 2002.

[36] A. Bhattacharjee, S. Verma, P.R.P. Verma, S.K. Singh, A. Chakraborty,
Fabrication of liquid and solid self-double emulsifying drug delivery system of atenolol by
response surface methodology, J Drug Deliv Sci 41 (2017) 45-57. doi: 10.1016/
j.jddst.2017.06.014.
[37] J. Goupy, Plans d’expériences pour surfaces de réponse, DUNOB Edition, 1999.
[38] L. Pachuau, B.Mazumder, A study on the effects of different surfactants on
Ethylcellulose microspheres, Int J PharmTech Res, 1 (2009) 966-971.
[39] K.S. Mayya, A. Bhattacharyya, J-F. Argillier, Microencapsulation by complex
coacervation: influence of surfactant, Polymer International, 52 (2003) 644-647.
DOI: 10.1002/pi.1125.
 ACCEPTED MANUSCRIPT

[40] A. Bhattacharyya, J-F. Argillier, Microencapsulation by complex coacervation:

Effect of cationic surfactants, J. Surface Sci. Technol, 21 (2005)161-168.
[41] N. Wilson, N.P. Shah, Microencapsulation of vitamins, ASEAN Food J, 14 (2007), 1-
14.
[42] Malvern Instruments, Protocole : Distribution de tailles de particules par diffraction
laser (Malvern Master Sizer), Malvern instruments (www.malvern.com), 2008.
[43] Y. F. Maa, H. R. Constantino, P. A. Nguyen, C. C. Hsu, The effect of operating and

PT
formulation variables on the morphology of spray dried protein particles. Pharm. Dev.
Technol. 2 (1997) 213-223. DOI:10.3109/10837459709031441.
[44] M.I. Amaro, L. Tajber, O.I. Corrigan, A.M. Healy, Optimisation of spray drying

RI
process conditions for sugar nanoporous microparticles (Npmps) intended for inhalation, Int.
J. Pharm, 421 (2010) 99-109. https://doi.org/10.1016/j.ijpharm.2011.09.021.

SC
[45] F. Qi, J. Wu, Q.Z. Fan, F. He, G.F. Tian, T.Y. Yang, G.H. Ma, Z.G. Su,
Preparation of uniform-sized exenatide-loaded PLGA microspheres as long-effective release
system with high encapsulation efficiency and bio-stability, Colloids Surf B Biointerfaces,
112 (2013) 492-498. https://doi.org/10.1016/j.colsurfb.2013.08.048.

U
[46] V. Kunasekaran, K. Krishnamoorthy, Experimental Design for the Optimization of
AN
Nanoscale Solid Lipid Particles Containing Rasagiline Mesylate, J Young Pharm, 7 (2015)
285-295. doi:10.5530/jyp.2015.4.2.
[47] R.P. Raffin, D.S. Jornada, M.I. Ré, A.R. Pohlmann, S.S. Guterres, Sodium
M

pantoprazole-loaded enteric microparticles prepared by spray drying: effect of the scale of

production and process validation, Int. J. Pharm, 324 (2006) 10-18. DOI:
10.1016/j.ijpharm.2006.06.045.
D

[48] Yunqing Kang, Jiang Wu, Guangfu Yin, Zhongbing Huang, Yadong Yao,
Xiaoming Liao, Aizheng Chen, Ximing Pu, Li Liao, Preparation, characterization and in
TE

vitro cytotoxicity of indomethacin-loaded PLLA/PLGA microparticles using supercritical

CO2 technique, Eur J Pharm Biopharm, 70 (2008) 85-97.
https://doi.org/10.1016/j.ejpb.2008.03.011.
EP

[49] L. A. Caetano, A. J. Almeida, L. M .D. Gonçalves, Effect of Experimental

Parameters on Alginate/Chitosan Microparticles for BCG Encapsulation, Mar Drugs. 14
(2016) 90. doi: 10.3390/md14050090
C

[50] I. Fuminori, M. Kimiko, Preparation and properties of monodispersed rifampicin-

AC

loaded poly (lactide-co-glycolide) microspheres. Colloids Surf B Biointerfaces, 39 (2004) 17-

21. https://doi.org/10.1016/j.colsurfb.2004.08.016.
[51] K. Arash, K. Rassoul, Effect of Alyssum homolocarpum seed gum, Tween 80 and
NaCl on droplets characteristics, flow properties and physical stability of ultrasonically
prepared corn oil-in-water emulsions, Food Hydrocolloid, 25 (2011) 1149-1157.
https://doi.org/10.1016/j.foodhyd.2010.10.012.
[52] En-Ping Lai, Yu-Xia Wang, Yi Wei, Guang Li, Preparation of uniform-sized and
dual stimuli-responsive microspheres of poly(N Isopropylacrylamide)/Poly(Acrylic acid) with
semi-IPN structure by one-step method, Polymers, (2016) 8, 90. doi:10.3390/polym8030090
 ACCEPTED MANUSCRIPT
[53] D. Patel, N. Patel, V. Thakkar, A. Modi, T. Gandhi, Development and
characterization of mucoadhesive microspheres of levosalbutamol sulphate, Int J Pharm Sci
Res, 4 (2013) 1838-1851. http://dx.doi.org/10.13040/IJPSR.0975-8232.4(5).1838-51.
[54] O.R. Fennema, Water and Ice, Food Chemistry, 3rd Edition, published by O.R.
Fennema, Marcel Dekker, New York, USA, 1996.
[55] J-L. Salager, Surfactifs en solution aqueuse, University of Los Andes, Mérida.
Venezuela, 1993.
[56] G.A. Rocha-Selmi, F.T. Bozza , M. Thomazini, H.M.A. Bolini, C.S.Fávaro-

PT
Trindade, Microencapsulation of aspartame by double emulsion followed by complex
coacervation to provide protection and prolong sweetness, Food Chem, 139 (2013) 72-78.
https://doi.org/10.1016/j.foodchem.2013.01.114.

RI
[57] F. Maestrelli, P. Mura, M.L. Gonzalez-Rodriguez, M.J. Cozar-Bernal, A.M.
Rabasco, L.D.C. Mannelli, C. Ghelardini, Calcium alginate microspheres containing

SC
metformin hydrochloride niosomes and chitosomes aimed for oral therapy of type 2 diabetes
mellitus, Int. J. Pharm, 15(2017)430-439. http://dx.doi.org/10.1016/j.ijpharm.2017.07.083.
[58] B. Y. Swamy, Y-S. Yun, In vitro release of metformin from iron (III) cross-linked

U
alginate–carboxymethyl cellulose hydrogel beads, Int J Biol Macromol, 77 (2015) 114–119.
https://doi.org/10.1016/j.ijbiomac.2015.03.019.
AN
[59] A. K. Nayak, D. Pal, K. Santra, Ispaghula mucilage-gellan mucoadhesive beads of
metformin HCl: Development by response surface methodology, Carbohydr Polym, 107
(2014) 41–50. https://doi.org/10.1016/j.carbpol.2014.02.022.
M

[60] A. K. Nayak, D. Pal, K. Santra, Swelling and drug release behavior of metformin
HCl -loaded tamarind seed polysaccharide-alginate beads, Int J Biol Macromol, 82 (2015)
1023-1027. https://doi.org/10.1016/j.ijbiomac.2015.10.027.
D

[61] Y.K. Lee, S.I. Ahn, H.S. Kwak, Optimizing microencapsulation of peanut sprout
extract by response surface methodology, Food Hydrocolloid, 30 (2013) 307-314.
TE

https://doi.org/10.1016/j.foodhyd.2012.06.006.
[62] M. Chacón, L. Berges, J. Molpeceres, M.R. Aberturas, M. Guzman, Optimized
preparation of poly D,L (lactic-glycolic) microspheres and nanoparticles for oral
EP

administration, Int. J. Pharm, 141 (1996) 81-91. https://doi.org/10.1016/0378-5173(96)04618-2.

[63] S. Honary, P. Ebrahimi, R. Hadianamrei, Optimization of particle size and
encapsulation efficiency of vancomycin nanoparticles by response surface methodology,
C

Pharm. Dev. Technol, 19 (2014) 987-998. doi: 10.3109/10837450.2013.846375.

AC

[64] R. Jangde, D. Singh, Preparation and optimization of quercetin-loaded liposomes for

wound healing, using response surface methodology, Artif Cells Nanomed Biotechnol 44
(2016) 635-641. doi: 10.3109/21691401.2014.975238.
[65] S. Aziz, J. Gill, P. Dutilleul, R. Neufeld, S. Kermasha, Microencapsulation of krill
oil using complex coacervation, J Microencapsul, 31(8) (2014) 774–784. doi:
10.3109/02652048.2014.932028.
[66] Y. Hu, T. Wu, C. Wu, S. Fu, C. Yuan, S. Chen, Formation and optimization of
chitosan-nisin microcapsules and its characterization for antibacterial activity, Food Control,
72 (2017) 43-52. https://doi.org/10.1016/j.foodcont.2016.06.013.
 ACCEPTED MANUSCRIPT
[67] L. Hu, J. Zhang, Q. Hu, N. Gao, S. Wang, Y. Sun, X. Yang, Microencapsulation of
brucea javanica oil: Characterization, stability and optimization of spray drying conditions, J
Drug Deliv Sci Technol, 36 (2016) 46-54. https://doi.org/10.1016/j.jddst.2016.09.008.
[68] M. Abul Kalam, A.A. Khan, S. Khan, A. Almalik, A. Alshamsan, Optimizing
indomethacin-loaded chitosan nanoparticle size,encapsulation, and release using Box–
Behnken experimental design, Int J Biol Macromol, 87 (2016) 329–340. doi:
10.1016/j.ijbiomac.2016.02.033.
[69] A.A. D’Archivio, M.A. Maggi, Investigation by response surface methodology of the

PT
combined effect of pH and composition of water-methanol mixtures on the stability of
curcuminoids, Food Chem, 219 (2017) 414–418.
https://doi.org/10.1016/j.foodchem.2016.09.167.

RI
[70] J.G. Delgado-Linares, J.C. Pereira, M. Rondón, J. Bullón, J.L Salager, Breaking
of Water-in-Crude Oil Emulsions. 6. Estimating the Demulsifier Performance at Optimum

SC
Formulation from both the required dose and the attained instability, Energy Fuels,
30 (2016) 5483–5491. DOI: 10.1021/acs.energyfuels.6b00666.
[71] B. Borges, M. Rondon, O. Sereno, J. Asuaje, Breaking of Water-in-Crude-Oil

U
Emulsions. 3. Influence of Salinity and Water-Oil Ratio on Demulsifier Action, Energy Fuels,
23 (2009) 1568–1574. DOI: 10.1021/ef8008822.
AN
[72] P. Bouriat, M. Rondon, J. Lachaise, J.L Salager, Correlation between interfacial
tension bump and optimal crude oil dehydration, Energy Fuels, 23 (2009) 3998–4002.
DOI: 10.1021/ef900221z.
M

[73] E. Del Carpio, S.Rodríguez, M. Rondón, B. Borges, Stability of water–Boscan

crude oil emulsions: Effect of salts, alcohols and glycols, Journal of Petroleum Science and
Engineering, 122 (2014) 542–550. https://doi.org/10.1016/j.petrol.2014.08.022.
D

[74] I. Souilem, C. A. Serra, R. Muller, Y. Holl, M. Bouquey, C. Sutter, Dimensional

Analysis of a Novel Low-Pressure Device for the Production of Size-Tunable Nanoemulsions,
TE

AIChE J, 61 (2015) 23-30. DOI: 10.1002/aic.14690.

[75] I. Souilem, R. Muller, Y. Holl, M. Bouquey, C. A. Serra, T. Vandamme, N.
Anton, Anovel low-pressure device for production of nanoemulsions, Chem. Eng. Technol.,
EP

35 (2012)1692-1698. DOI: 10.1002/ceat.201100676.

[76] G.K. Bédié, S.L. Turgeon, J. Makhlouf, Formation of native whey protein isolate–
low methoxyl pectin complexes as a matrix for hydro-soluble food ingredient entrapment in
C

acidic foods, Food Hydrocolloid, 22 (2008) 836–844.

AC

https://doi.org/10.1016/j.foodhyd.2007.03.010
[77] C. Schmitt, C. Sanchez, F. Thomas, J. Hardy, Complex coacervation between
β-lactoglobulin and acacia gum in aqueous medium, Food Hydrocolloid, 13 (1999) 483–496.
https://doi.org/10.1016/S0268-005X(99)00032-6
[78] C. Schmitt, C. Sanchez, S. Despond, D. Renard, F. Thomas, J. Hardy, Effect
of protein aggregates on the complex coacervation between β-lactoglobulin and acacia gum at
pH 4.2, Food Hydrocolloid, 14 (2000) 403–413. https://doi.org/10.1016/S0268-
005X(00)00022-9
[79] C. Schmitt, L. Aberkane, C. Sanchez, Protein-polysaccharide complexes and
coacervates, Handbook of hydrocolloids, pp. 420-476, 2009.
 ACCEPTED MANUSCRIPT
[80] S.L. Turgeon, M. Beaulieu, C. Schmitt, C. Sanchez, Protein–polysaccharide
interactions: phase-ordering kinetics, thermodynamic and structural aspects, Curr. Opin.
Colloid Interface Sci, 8 (2003) 401–414. https://doi.org/10.1016/S1359-0294(03)00093-1
[81] S.L. Turgeon, C. Schmitt, C. Sanchez, Protein–polysaccharide complexes and
coacervates, Curr. Opin. Colloid Interface Sci, 12 (2007) 166–178.
https://doi.org/10.1016/j.cocis.2007.07.007
[82] A. Khalil, F. Puel, Y. Chevalier, J.M. Galvan, A. Rivoire, J.P. Klein, Study of
droplet size distribution during an emulsification process using in situ video probe coupled

PT
with an automatic image analysis, Chemical Engineering Journal, 165 (2010) 946–957.
https://doi.org/10.1016/j.cej.2010.10.031
[83] A. Khalil, F. Puel, X. Cosson, O. Gorbatchev, Y. Chevalier, J.M. Galvan, A.

RI
Rivoire, J.P. Klein, Crystallization-in-emulsion process of a melted organic compound: In
situ optical monitoring and simultaneous droplet and particle size measurements, Journal of

SC
Crystal Growth, 342 (2012) 99–109. https://doi.org/10.1016/j.jcrysgro.2011.06.005
[84] V. Schmitt, S. Arditty, F. Leal-Calderon, Stability of concentrated emulsions, D.N.
Petsev (eds.), Emulsions: Structure Stability and Interactions, pp. 607-639, 2004.

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

Fig.1: Method of preparation of the microcapsules.

D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
Fig.2: Study field of Tween 80 concentration (Pr: Soy porein isolate, T80: Tween 80).

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
a b

RI
U SC
AN
c d
M

Fig.3: Micrographs of optimal formulation. a. primary W/O emulsions (40X

magnification), b. double W/O/W emulsions (40X magnification), c. microcapsules of
D

metformin with SPI/Pectin in coacervate suspension (40X magnification), d.

microcapsules of metformin with SPI/Pectin in coacervate suspension (100X
TE

magnification).
C EP
AC
 ACCEPTED MANUSCRIPT

21 : +++− 10 : −++−
(a) (b)

PT
RI
SC
20 : +−−− 3 : −−−+
(c) (d)
Fig. 4: The granulometric particles size distribution profiles, (a) unimodal distribution

U
of the particles, (b) imperfect unimodal distribution, (c) bimodal distribution with two
AN
distinct peaks, (d) imperfect bimodal distribution.
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
Fig.5: Pareto chart for microencapsulation yield;
X1: soy lecithin concentration; X2: Tween 80 concentration;
AN
X3: stirring speed; X4: stirring duration.
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
(a) (b)

U SC
AN
M
D

(c) (d)
TE
C EP
AC

(e) (f)
Fig.6 : Response surface plots showing combined effect of (a) concentration of soy
lecithin (X1) and concentration of Tween 80 (X2), (b) concentration of soy lecithin (X1)
and stirring speed (X3), (c) concentration of soy lecithin(X1) and stirring duration (X4),
(d) concentration of Tween 80 (X2) and stirring speed (X3), (e) concentration of Tween
80 (X2) and stirring duration (X4), (f) stirring speed (X3) and stirring duration (X4) on
microencapsulation yield (MEY).
 ACCEPTED MANUSCRIPT

PT
RI
SC
Fig.7: Pareto chart for the size distribution;

U
X1: soy lecithin concentration; X2: Tween 80 concentration;
AN
X3: stirring speed; X4: stirring duration.
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
(a) (b)

U SC
AN
M
D

(c) (d)
TE
C EP
AC

(e) (f)
Fig.8. : Response surface plots showing combined effect of (a) concentration of soy
lecithin (X1) and concentration of Tween 80 (X2), (b) concentration of soy lecithin (X1)
and stirring speed (X3), (c) concentration of soy lecithin(X1) and stirring duration (X4),
(d) concentration of Tween 80 (X2) and stirring speed (X3), (e) concentration of Tween
 ACCEPTED MANUSCRIPT
80 (X2) and stirring duration (X4), (f) stirring speed (X3) and stirring duration (X4) on
size distribution (span).

PT
RI
U SC
AN
M
D
TE
C EP
AC