Industrial Crops & Products 115 (2018) 306–314

Contents lists available at ScienceDirect

Industrial Crops & Products

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

Effect of formulation and process on the extrudability of starch-based foam T

cushions
⁎
Abinader Georges, Catherine Lacoste , Erre Damien
Laboratory of Science Engineering (GRESPI), Department of Material Process and Packaging System (MPSE), Université de Reims Champagne-Ardenne (URCA), Reims
Engineering School (ESIReims), 3 Esplanade Roland Garros, F 51100, Reims, France

A R T I C L E I N F O A B S T R A C T

Keywords: Solid foams produced from petroleum sources have a number of environmental and sustainability issues. Using
Starch foams renewable resources to produce foams has great perspectives but those kinds of foams are technically difficult to
Formulation process. Starch foams from wheat, pea and potato starches were prepared using reactive extrusion with different
Reactive extrusion concentrations of starch/glycerol/gluten/sodium bicarbonate. The density, the expansion ratio, the elasticity
Cushion foams
and the dynamic damping properties of such foams were determined. Foams made from wheat starch exhibited
the best mechanical properties, especially during ageing. Using the ratios of wheat starch/glycerol/gluten/so-
dium bicarbonate that gave the best results (100/46/25/1), extrusion parameters were adjusted and higher and
more stable expansion (Expansion ratio = 9.1) was obtained when the screw speed was as low (300 rpm) as the
highest input rate could afford (21 kg/h).

1. Introduction Kanlaya Pornsuksomboon et al., 2016; Sanhawong et al., 2017; Shey

Synthetic plastic cushioning foams are usually used as foamed
particles. They come from oil, a non-renewable resource and are non-
biodegradable. These foams are obtained from the injection or the in
situ creation of gas into the macromolecular matrix of a polymer. They
are generally used as protective material for fragile products during
transport and storage. The most commonly used cushioning foams in
the market are expanded polystyrene foams to protect large and light
products; expanded polyethylene foams for less bulky, heavy and fra-
gile products; expanded polypropylene foams to provide additional
benefits such as high heat resistance and good insulation properties;
and polyurethane foams for fragile and light products. Cushion foams
coming from agricultural resources are part of the family of foams
produced from bioplastics. The development of extrusion of starch-
based foams is the most commonly used substitute the traditional pet-
rochemical foams, which are mainly manufactured by extrusion. The
expansion created by injected gas or in situ formed gas during extrusion
causes high pressure at the exit of the die. In contact with the ambient
air, the pressure suddenly drops and this results in the foaming of the
material. The gas bubbles formed in the starch paste are generally
molecules of carbon dioxide (CO2), but in some cases they are nitrogen
molecules (N2). In order to obtain trays or solid block in starch foams,
the most common process is the baking process using compressed molds
(Bergel et al., 2017; Kaisangsri et al., 2012; Machado et al., 2017;
et al., 2006). Starch foams could also be produced by a microwave
heating process. It leads to rigid foams for both structural or packaging
applications (Lopez-Gil et al., 2015). Soykeabkaew et al. (2015) pre-
sented a large overview of starch-based composite foams. They com-
pared different starch foaming process (extrusion, baking/compression,
microwave heating, freeze-drying/solvent exchange and supercritical
fluid extrusion). They concluded that extrusion process led to large cell
size foams with low density so this process was well suitable for
packaging transportation. The development of extrusion starch-based
foam was studied by other authors. Indeed, Bhatnagar and Hanna
(1995) studied the properties of extruded starch-based foams.
Chinnaswamy and Hanna (1987), Lee et al. (2009) and Robin et al.
(2010) added a blowing agent such as water or sodium bicarbonate in
the formulation that decomposes at high temperature and under pres-
sure to form CO2 gas bubbles thus expanding foams. While other au-
thors, like Alavi et al. (1999, 2003), Cho and Rizvi (2009), Goel and
Beckman (1995), Jeong and Toledo (2004) and Muljana et al. (2009)
preferred to inject the carbon dioxide or the nitrogen into the form of
supercritical fluid at low temperature directly during the extrusion,
thus causing the formation of gas bubbles in the material. Della Valle
et al. (1997) determined the relationship between the rheological
properties of molten starches and their expansion upon extrusion. They
found that the expansion was dependent on the extrusion temperature,
the water content of the starch and its concentration in amylose.

⁎
Corresponding author.
E-mail address: catherine.lacoste@univ-reims.fr (C. Lacoste).

https://doi.org/10.1016/j.indcrop.2018.02.001
Received 21 August 2017; Received in revised form 29 January 2018; Accepted 1 February 2018
Available online 22 February 2018
0926-6690/ © 2018 Elsevier B.V. All rights reserved.
A. Georges et al.
Although the method of manufacturing starch-based foams appears
simple, the complexity lies in the stability of the process and the
properties of the resulting foams, especially for extrusion process.
Concerning the ingredients to add to starch to improve the prop-
erties of the foam, several studies have been performed to find the best
formulation and the ideal process for developing starch-based foams
having mechanical properties comparable to those of petrochemical
foams. Among these works, Lin et al. (1995) studied the mechanical
properties of extruded starch foam cushioning particles that contains
high level of amylose. Pushpadass et al. (2008) examined the effect of
temperature, moisture and talc on the physical properties of the ex-
truded foam particles based on starch. However, the cushion foams
based on extruded starch still present technological barriers in terms of
processability and final properties. Most previous works have studied
loose fill chips for protective packaging (Pushpadass et al., 2008;
Heartwin et al., 2010). In order to enlarge the range of starch foams in
packaging, this paper is devoted to the development of starch-based
foams as strips rather than loose-fill particles (chips) to overcome the
lack of control during processing.
The main objective of this paper is to understand and optimize the
process of manufacturing starch foam plates by extrusion to ensure
stability and reproducibility of the process. In the first part, we compare
the properties of starch foams with different additives (plasticizer,
gluten, yeast) and from different sources of starch (wheat, pea and
potato) and the impact of the formulation on the expansion ratio,
rheological properties. In the second part, extrusion parameters are
studied in order to obtain a stable starch foam with best expansion ratio
and dynamic damping properties.
2. Materials and methods
Wheat starch (25–28% amylose content) and wheat gluten were
obtained from Chamtor, France. Pea starch (35% amylose content) was
obtained from Cosucra, Belgium and potato starch (25% amylose con-
tent) was from Haussimont, France. Glycerol, used as a plasticizer, was
provided from by A.R.D society, France. Sodium bicarbonate, used as a
blowing agent, was coming Ecodis, France. Petrochemical foams were
used to compare their behavior with ours: one polyethylene foam
(Etafoam 220, 35 kg/m3, from Dow Chemical) and a polyether foam
and an expanded polystyrene cushion foam (from our laboratory).
The foams were compounded using a Clextral BC21 co-rotating
twin-screw extruder machine, with an optimized screw configuration
and temperature profile and with a ribbon die (20 mm width, 2 mm
thickness). Plasticizer was first introduced by a liquid pump, just before
the kneading zone of the barrel. After extrusion, samples were cut into
15 cm strips, identified and conditioned for one week at 23 ± 2 °C and
50 ± 10% relative humidity according to NF EN ISO 7214 prior
characterization.
Each formulation is denoted by a nomenclature which refers to its
constituents by abbreviations followed by their percentage rate relative
to the starch. These classifications are obtained as follows:
The first letter of the nomenclature defines the nature of the starch
used, for example the letter “B” refers to wheat starch, “P” to pea starch
and “PDT” to potato starch. The second letter defines the plasticizer
used. In our case a big “G” refers to glycerol. Then the third letter uses a
small “g” to refer to the wheat gluten and the fourth letter defines the
blowing agent (in our case sodium bicarbonate) and is designated by a
large “L”. For example, the formulation |BG46g25L1| contains wheat
starch as the main resin, with 46 parts of glycerol as a plasticizer, 25
parts wheat gluten and 1 sodium bicarbonate, per 100 parts of starch
(pcs).
Table A (in Appendix A) shows the different formulations used in
this paper.
307
Industrial Crops & Products 115 (2018) 306–314
2.1. Foam density
In accordance with the NF ISO 845 standard, foam density was
calculated in kg/m3 by dividing the weight of the sample in kg into its
volume in m3. Five samples from each formulation were tailored into a
rectangular shape and were weighed. The length, width and thickness
of each sample were measured three times with a graduated Vernier
caliper.
2.2. Expansion ratio (ER)
The expansion ratio of the foam was calculated by dividing the
cross-sectional area of the foam by the cross-sectional area of the die in
mm2. The result was the mean of five samples of each formulation.
2.3. Foam microstructure
The structural properties were determined by using a dnt® Digital
Micro Capture Pro microscope. 5 mm thick slices were cut using a
double blade cutter. The surface of each sample was colored using blue
ink in order to show the contrast of the cellular structure. The cell size,
which is the cross-sectional area of each cell in mm2, and the cell wall
thickness, which is the cell wall thickness of the cells in mm, were
calculated using ImageJ 1.43 software. The cell density of each spe-
cimen in cell/mm2 was then determined by dividing the numbers of
cells of the cross-sectional area of the foam by the area of this foam
section. The average values and standard deviation of all the mea-
surements were calculated and sorted in the results and discussion
section.
2.4. Differential thermal analysis (DSC)
The melting properties of the starch based materials were de-
termined using differential scanning calorimetry. The raw materials and
the mixtures were analyzed by a measuring apparatus DSC “NETZSCH
Thermal Analysis”, after 48 h conditioning at a temperature of 23 °C
and 65% relative humidity. Samples of 10–20 mg are placed and sealed
in aluminum seals to prevent loss of water and are heated under pure
nitrogen and controlled by a thermocouple at temperatures ranging
from 25 to 200 °C with a heating rate of 2 °C/min.
2.5. Moisture content (MC)
The moisture content of the different foams was determined by
using an “OHAUS moisture analyzer”. This thermogravimetric analyzer
calculates the weight change between the value of the foam dry sample
and the wet weight of that sample (20 × 20 × 15 mm3) at 80 °C.
2.6. Viscoelastic properties
The viscoelastic properties of our foams were examined on the re-
sulting foams using a stress-controlled rheometer (AR2000Ex, TA
Instruments), with torsion geometry. The samples were cut into a
12 mm wide, 50 mm long rectangular solid. The thickness varied with
the sample but was around 10 mm. The tests were performed at 25 °C
and at a strain of 0.005 Pa (such a strain was chosen within the linear
viscoelasticity range).
2.7. Creep tests
The compressive creep tests were performed on the resulted foams
using a dynamic mechanical analysis instrument (DMA) designed by TA
Instruments, with a thermal relative humidity accessory. This allowed
mechanical properties of the foam sample to be determined under
standard conditioning atmosphere of 23 ± 1 °C and 50 ± 2% relative
humidity. The creep/recovery test was performed using parallel
A. Georges et al.
rounded plate geometry of 40 mm diameter for compression use.
Samples were cut into 4 cm2 square slices. A preload force of 0.01 N was
assigned and a constant stress of 10 kPa was performed on the foam’s
surface for five minutes. This creep test was followed by ten minutes of
relaxation. The resulting foam’s creep strain γ(t) and the recoverable
compliance J(t) were registered as a function of time.
2.8. Dynamic cushion curves
According to NF ISO 4651, the dynamic cushion curves were es-
tablished on a Lansmont free fall drop machine monitored by an ac-
celerometer connected to a computer. The curves were generated by
dropping a series of known steel weights (1–5 kg) of 465 cm2 surface
onto a series of 10 mm thick cushion sample (30–300 cm2), from a
46 cm height and by measuring the shock impact (number of decel-
erations G’s, with G = 9.81 m/s2) on the foam. The number of the
samples and the platen mass were modified to vary the static load range
from 0.5 to 12 kPa. The drop height was identified as the height from
which a package was likely to fall, and is related to the weight of the
product being handled. According to ASTM D-3332, fragile products
have a fragility level around 25 G, while rugged products have a fra-
gility level greater than 120 G.
3. Results and discussion
3.1. Formulation
Water is usually considered to be the natural starch plasticizer.
Indeed, its strong polar nature allows it to spread and to easily break the
hydrogen bond interchains of the starch. But because of its moisture,
water has a negative effect on the recrystallization of the starch-based
foams during extrusion and during storage. Because of its non-volatile
viscous nature (T boiling,glycerol = 290 °C), glycerol was chosen as a
plasticizer. To understand the thermal transformations of the starch in
the mixture, DSC analyses have been performed on hydrated wheat
starch (12% of moisture) and on hydrated wheat starch (12% of
moisture) mixed with 25% glycerol. Melting temperature of starch
mixture has decreased from 153.6 °C without glycerol to 127.5 °C with
25% glycerol. These results show that the melting temperature of wheat
starch decreases when it is combined with glycerol and reaches a value
lower than its degradation temperature which is of the order of 275 °C.
The glycerol will facilitate the processability and destructuring of ex-
truded starch by protecting it from degradation. Thus, it allowed to
impart more interesting viscoelastic properties to the resulting foams.
Glycerol was introduced along the barrel thanks to a pump and the
starch was put through the main feeder.
Nevertheless, acting as a simple plasticizer, glycerol did not confer
enough flexibility and softness to the starch foam for cushioning ap-
plications. The cell nucleation was low, and the generated cells were
very large and irregular due to the lack of stability of the water vapor
released during extrusion. The foam obtained is then very dry and
brittle after cooling, which does not allow the protection of products
during transport. Taking inspiration from the baking process, gluten
and sodium bicarbonate were finally added to the formulation in order
to improve the elasticity and reduce the retraction after the expansion
at the die exit. Wheat gluten is a viscoelastic compound with long
chains and water insoluble molecules. It consists in a heterogeneous
mixture of proteins (75–85%) linked by covalent bonds (disulfides)
giving it elasticity and toughness, lipids (5–7%) bound by hydrophobic
interactions and starch (5–10%) and water (5–8%) bound by non-
covalent bonds (hydrogen, ionic) that confer viscosity (Feillet, 2000).
Wheat gluten is often used in the food industry because it allows, by its
elastic structure, to trap the carbon dioxide secreted by the yeast and
increase the behavior of the dough during the production of bread. To
understand the thermal transformations of gluten in the mixture, DSC
analyses were performed for gluten and the mixtures associated with it.
308
Industrial Crops & Products 115 (2018) 306–314
Fig. 1. Elastic and viscous moduli versus frequency of starch foams made with different
ratios of glycerol. full marks: elastic modulus G’; empty marks: viscous modulus G”.
(measured with the rheometer AR2000) with their expansion ratios.
The results show that the glass transition and melting temperatures of
wheat gluten (containing 10% water) are respectively 51 °C and
124.4 °C. With 35 pcs of gluten (Bg35), these temperatures increased
slightly to 63.1 °C and 136.8 °C, respectively. But in the case of the
starch-gluten-glycerol mixture (BG46g35), the glass transition and
melting temperatures decreased significantly to 33.4 °C and 120 °C,
respectively. These results show that the combination of glycerol-gluten
with wheat starch has a positive effect on the decrease of the glass
transition and melting temperatures of wheat starch, which thus allows
the starch to be more easily plasticized and processed by extrusion.
In order to have a more uniform nucleation and a better expansion
ratio, the concentration of sodium bicarbonate used to expand the
foams was 1 pcs. Indeed, under 1 pcs, the expansion was not regular and
stable and above 2 pcs, it created too many cells and the strip shredded.
Thus the ratio expansion of such foams was lower than the one obtained
with 1 pcs of sodium bicarbonate. Consequently, a rate of 1 pcs was
chosen in order to have the more stable expansion.
The influence of glycerol ratio on rheological properties are pre-
sented in Fig. 1. The amount of glycerol had to be above 21 pcs,
otherwise the extruder had trouble swallowing matter and the resulting
foams leaving the die were dry and hard. Increasing the glycerol ratio
decreased the elastic modulus of the foam. Indeed, glycerol was needed
to plasticize starch but, acting as a plasticizer, it facilitated the mobility
of the starch chains and thus affected the elastic modulus. The Fig. 1
also shows a difference of behavior between the foam containing 26 pcs
of glycerol and the one having more than 30 pcs of glycerol. Above
30 pcs, rheological properties of starch foams were not affected by in-
creasing glycerol concentration, meaning that for high glycerol con-
centration, plasticized starch was supersaturated which induced a very
low expansion at the die exit (ER = 1.2). Therefore, a concentration of
25 pcs of glycerol is a mean ratio that makes it easier to process and
destructure the starch during the extrusion. The expansion ratio of such
a foam was up to 3.8. And unlike water, it made the foam more resistant
to crystallization.
For cushioning applications, starch foams require good elastic and
dynamic dumping properties. Fig. 2 showed that adding a small amount
of gluten (10 pcs) in the formulation, made the expansion of the starch
foam easier. Gluten facilitated the movement and the incorporation of
the plasticizer on the starch. By increasing the levels of gluten, glycerol
content also needed to be increased needed to be also increased in order
to keep its constant fraction (25%) compared to the powdery system
[+gluten starch]. The resulting material mixture | BG31g25| has con-
sistently expanded at the die exit, and the retraction rate was reduced
by 50% compared to the gluten-free formulation. At high levels of
gluten (35 pcs) (Fig. 2d) and still at a concentration of 25% of glycerol
with respect to the powdery gluten-starch system (34 pcs compared to
starch), the expansion of the die exit was very high, which resulted in a
moderately stable flow, with a low retraction rate of about 43%. With
A. Georges et al. Industrial Crops & Products 115 (2018) 306–314

Fig. 2. Formulation from |BG25| (a), from |BG27g10| (b,) from |BG31g25| (c) and from |BG34g35| (d) for different gluten ratios; after one week of storage at 20 °C and 50% HR.

these results, we concluded that gluten had a very positive effect on the
formulation. It facilitated plasticization of starch increasing the solu-
bilization of the plasticizer (glycerol) in the amorphous phase of the
starch. This allowed the addition of more glycerol in the matrix. In
addition, the gluten reduced foam shrinkage phenomenon after ex-
pansion at the die exit. The resulting foams were more elastic with
better expansion.
Fig. 3 shows viscous and elastic properties of starch foams in pre-
sence of 10 pcs of gluten with different ratios of glycerol. The foam with
30 pcs of glycerol exhibits the most viscous behavior (higher G” slope)
with high rigidity (the higher elastic modulus). Indeed, the critical rate
of glycerol was exceeded and the material did not expand well after the
die because of its heaviness. The sample containing 27 pcs of glycerol
has the weakest slope for the viscous and elastic moduli with higher
values which means the foam should have good elastic. The fact that Fig. 4. Creep compliance during a five-minute test at a stress of 10−2 MPa and strain
glycerol hinders cell formation and thus foam expansion is corroborated recovery of wheat starch foams for different glycerol rates.

by Fig. 4 in which it can be seen that the more glycerol the foam
contains the lower compliance it has meaning that the foam is less
deformable. For very low content of glycerol (20 pcs), the foam has lost
some stiffness. The foam has become more brittle due to a lack of
plasticizer. The foam |BG24g10| is more deformable but takes more
time to recover its strain. |BG27g10| seems to be a good compromise for
a packaging application: it has a moderate deformation under stress and
recovers its dimension fast.
Adding more gluten brought more stiffness to the foam by de-
creasing its elastic modulus (Fig. 5) and by allowing more deformation
under stress (Fig. 6). Thanks to its network, the gluten improved the
elasticity of the starch foam. The gluten promoted better expansion.
Fig. 7 shows the microscopic pictures of the foam containing 10 pcs
(Fig. 7a) and 30 pcs (Fig. 7b) of gluten. Cells are bigger and more
homogeneous for 30 pcs of gluten. Its intrinsic elasticity induced higher
stable expansion and lower retraction after cooling. Fig. 5. Comparison of behavior between cushion foam with 10 and 30 pcs of gluten in
wheat starch. Viscoelastic properties: full marks: elastic modulus G’; empty marks: viscous
modulus G”.

Fig. 8 shows the cushion curves of 15-mm-thick foams with two

Fig. 3. Elastic and viscous moduli versus frequency of starch foams made with different
rates of glycerol with 10 pcs of gluten. full marks: elastic modulus G’; empty marks:
viscous modulus G”.
different levels of gluten (10 and 30 pcs). The foam |BG27g10L1| has a
typical u-curve shape. Its deceleration peak is high and this foam would
only protect rugged products. However, the foam with 30 pcs of gluten
has a lower deceleration peak meaning that it would better protect
fragile product. This can be explained by the high elasticity of this
material |BG24g30L1| (higher creep compliance) which is probably
influenced by the large cell size of the foams (Fig. 6). Finally, the most
suitable foam for cushion application in packaging was obtained with
high levels of gluten with an average rate of glycerol (around 25 pcs)
Considering the base resin, three types of starch: wheat starch, pea
starch and potato starch were tested to optimize the properties of the
foams. The thermal transitions of wheat starch, pea and potato starch in
terms of melting temperatures are determined and listed in Table 1. The
DSC results showed that the melting temperatures of the three types of

309
A. Georges et al. Industrial Crops & Products 115 (2018) 306–314

Fig. 6. Comparison of behavior between cushion foam with 10 and 30 pcs of gluten in Fig. 8. Dynamic cushion curves of 15 mm thick starch foam for two gluten contents
wheat starch. Creep compliance during a five-minute test at a stress of 10−2 MPa (full (10 pcs and 30 pcs) at a 46 cm drop height.
marks) and strain recovery (empty marks).

Table 1
starch were close to each other, although their melting enthalpies were Thermal transitions of the different starches obtained by DSC.
different because of their difference of crystallinity. Indeed, the potato
Characteristic points obtained by DSC Wheat Pea Potato
starch, with a size of large grain structure (100 μm), is more crystalline
than pea starch and wheat medium grain (2–38 μm) (Buléon et al., Incipient melting temperature of the starch (°C) 152.0 138.4 146.4
1998). The melting temperatures of those three types of starch are in Melting peak of the starch (°C) 153.6 151.0 149.4
accordance with the literature values (Colonna and Mercier, 1985; End temperature of melting of the starch (°C) 159.4 161.5 156.3
Melting enthalpy (J/g) 214.8 237.0 304.3
Homer et al., 2014; Steeneken and Woortman, 2009)
The results showed that foams made from wheat, potato and pea
starch had comparable expansion ratios after leaving the die (Fig. 9). and pea starch should be better than the ones made from potato starch.
Concerning the morphology of those blends, cells were much more This was confirmed by the creep tests made on those three foams
regular for wheat starch foams (Fig. 9a.) and it could be compared to (Fig. 10). Indeed, the foam made from pea starch (triangle marks) was
the morphology of the polyether foams (Fig. 9c.). For foams made from more resistant than the foam made from wheat (diamond-shaped
pea starch, the thickness of the cell walls was thinner but regular. The marks) and potato (circle marks) starch. Those three foams had good
morphology of potato starch foams was much less well defined. Indeed, recovery properties but when we did the creep test after five-month
the cell walls were not homogeneous (see Fig. 9b.). Concerning the storage again, one can see on Table 2 that potato starch foams had very
foams from polyethylene (Fig. 9d.), the cell walls did not resist when poor recovery. It could be explained by the recrystallization of this foam
cut. To conclude, those pictures showed that foams made from wheat and it became thus brittle compared to the foams made from wheat and

Fig. 7. Microscopic pictures of extruded foams for two ratio of gluten (10 pcs and 30 pcs).

310
A. Georges et al. Industrial Crops & Products 115 (2018) 306–314

Fig. 9. Microscopic pictures of extruded foams from different natures of starch and petroleum foams (magnifying: X100).

Fig. 10. Strain recovery after a five-minute test at a stress of 10−2 MPa of starch made
foams (wheat, pea and potato). Fig. 11. Viscoelastic properties of foam from wheat starch, pea starch and polyether. Full
marks: elastic modulus G’; empty marks: viscous modulus G”.

Table 2
Ageing of foams made from different sources of starch (wheat, pea and potato). Strain
recovery (%) of one week after extrusion and 5 months later.

Strain recovery (%)

t = 0 (one week after extrusion) After 5 months

Wheat BG36g30L1 99 79
Pea PG31g30L1 90 60
Potato PDTG47g25L1 90 20

pea starch.
Considering that pea and wheat starch led to the best cushion foams
compared to those using potato starch, we focused then only on pea and
wheat starch foams. In order to verify that cushion foams coming from
Fig. 12. Cushion curves of foam from wheat starch, pea starch, PSE, Etafoam 220 and
starch could be competitive with petroleum cushion foams, the visc- polyether.
osity properties (Fig. 11) and dynamic cushioning properties (Fig. 12)
of cushion foams made from starch, PE, PS and polyether were estab-
lished. In Fig. 11, we compared the three foams having almost the same higher with foams made from pea starch (ER = 4.6 for pea starch
morphology (see Fig. 9a, c, e). Starch foams were more rigid but ex- compared to ER = 4 for wheat starch).
hibited quite good elastic properties anyway. Foams coming from pea Considering the cushion curves (Fig. 12), the behavior of the
starch had slightly greater moduli than the foams coming from wheat cushion foams made from starch was comparable to the polyethylene
starch and that concorded with the fact that the expansion was a little foam (Etafoam 220): they could undergo a wide range of static

311
A. Georges et al.
pressures, whether for light or heavy products. Nevertheless, our
cushion foams are more dedicated for rugged products whereas etafoam
220 protects fragile product better. Thus, at this early stages of devel-
opment, cushion foams from pea and wheat starch showed good
cushioning properties. One can notice that polyether foam would be
used with light and fragile products and the PSE would be used for
products with a fragility comparable to those protected by etafoam 220
but for lighter weight.
With these tests and because of a better ageing (Table 2), wheat
starch has been selected for the rest of this paper, especially because of
its better behavior after ageing. Indeed, wheat starch has a semi-crys-
talline structure medium grain (2–38 μm) and low water content (10%)
compared to potato starch (100 μm and 13%) and a low concentration
of amylose (25%) and low gel temperature (59 °C) compared to pea
starch (35% and 71 °C). These properties allows the wheat starch,
during extrusion, to mix more easily with the plasticizer (for rapid in-
flation of the grains) and to deconstruct more easily (by fusion of the
crystalline chains high shear) than the pea starch and potato starch.
These properties allows the starch-based mixture of wheat to expand
with the lowest withdrawal and that makes wheat starch the most
suitable material base for the development of cushion foams. To enlarge
the range of starch foam properties, biopolymers and natural fibers
could be added to the formulation. This is the subject of a further paper
of our laboratory.
The next part of this paper is dedicated to the choice of foam de-
velopment process to define the conditions for optimum implementa-
tion of the extrusion of starch-based foams.
3.2. Processability
In view of those previous results, formulations |BG46g25L1| based
on wheat starch, glycerol, gluten and sodium bicarbonate were chosen.
The extrusion technique we chose was the cooking-extrusion with a
twin screw extruder and a strip die. But several other points could be
improved to optimize the development of the foams. Indeed, the point
that limited the foam development process gave a non-uniform shape to
starch foam which was due to the difficulty in obtaining a stable ratio
between starch and glycerol that was directly injected into the extruder.
Thus, the concept of pre-mixture was adapted from coated tablets to
prepare granules from the compound system “resin + plasti-
cizer + additives”, which were then directly incorporated into the ex-
truder. This concept allowed the preparation of various formulations
without having to inject the plasticizer by an external feeder during
extrusion, which facilitated the production of a homogeneous blending.
And above all, it ensured constant glycerol content throughout the
extrusion of a given foam.
For the implementation of the concept of pre-mixing
“starch + glycerol + sodium bicarbonate”, a climatic chamber was
used for conditioning ingredients, a turbo mixer was used for mixing
the various components, and a screen was used to form fine and uni-
form granules. The implementation of the pre-mixture was divided into
four steps.
The first step, “drying”, consisted in preparing the base resin, made
of the powder mixture of starch and wheat gluten, which was stored in
a climatic chamber at 23 °C and 65% relative humidity, for at least 24 h.
In these climatic conditions, the water content of the mixture reached
10%, amount needed to facilitate the swelling and the solubilization of
the plasticizer in the starch. The second step, “granulation”, was to
place the conditioned powdery mixture [starch + gluten] in the
SCAMEX turbo mixer, then spraying it with the liquid plasticizer (gly-
cerol) and mixing it in the mixer for a few minutes. The third step,
“drying”, was to spread the pellets in aluminum pans and to store them
again for at least 24 h at 23 °C and 65 % relative humidity. This step of
drying the granules led to a water content of 13%. The fourth step,
“sieving”, consisted in sieving the granules with a sieve of 2.33 mm to
obtain a fine and uniform powder. The granules bigger than the sieve
312
Industrial Crops & Products 115 (2018) 306–314
diameter were returned to the turbo mixer and were blended again at
ambient temperature before sifting. The final mixture was placed for
the last time in an oven for 48 h at 23 °C and 65% relative humidity, to
achieve a water content of 13%. The fifth and last step, “finishing”,
consisted in adding, just prior to extrusion, sodium bicarbonate, used as
a blowing agent to the mixture of sieved powders, and mixing them at
ambient temperature for a few minutes (1–2 min) in the turbo mixer. It
was important, prior to extruding, that the granules have a minimum
water content of 13%, as the water vapor promoted the expansion of
the material. The technique of compounding solved the problem of
heterogeneous plasticization of the extruded material. This technique
also facilitated the extrusion process by stabilizing the flow of material
at the extruder inlet and at the outlet of the die, ensuring the re-
producibility of the tests with a constant glycerol content throughout
the extrusion. This is considered to be the key to scaling up to industrial
production.
In order to optimize the parameters of the extrusion process, the
effects and the interactions between the two most influential factors on
the machinability and expanding foams, the speed of the screw and the
incoming flow material were determined. A machinability study was
performed by varying the rotational speeds of the screws from 100 to
600 rev/min and the feed rate of the material from 10 to 50 kg/h during
the extrusion of the foams. One can thus find the extrema that should
not be exceeded in order to have a stable process with acceptable ex-
pansion while remaining within the power and torque extruder capa-
city.
The results are summarized in a “runnability” plan in Fig. 13. By
conducting machinability tests, we noticed that the rotation speed of
the screw and the input material flow have a significant impact on the
macroscopic appearance of the foams, especially in terms of regularity
and foam expansion rate. Following these results, the parameters
leading to a stable process were selected to set the new study plan
entitled “process” (Table B). For this study, we varied the screw rotation
speed from 300 to 600 rpm/min in order to keep a safety margin be-
cause below 300 rpm/min, the technical feasibility of the foams was not
fully assured. As for the feed rate of the material, we varied from 11 to
21 kg/h, and this by keeping the EMS between 300 and 500 kJ/kg,
which was the energy needed to completely deconstruct the starch
without degrading it.
Similarly, for each experimental point, the specific mechanical en-
ergy (EMS) and the expansion ratio of the resulting foams (ER) have
been determined. The EMS was calculated according to the equation 1,
taking into account the maximum power of our twin-screw extruder,
which was 8.5 kW, the maximum rotation speed of the screws was 680
rev/min and the maximum torque was 172 Nm.
Torque × RPM P N C
EMS = = motor × × × 3600
Q Q Nmax Cmax (1)
Where P motor and Q were the maximum power of the motor of the
extruder in kW and the material input in kg/h respectively, N was the
Fig. 13. Runnability plan for the formulation |BG46g25L1|.
A. Georges et al.
Fig. 14. Expansion ratio (full marks) of wheat starch foam and EMS (empty marks) versus
screw speed and input rate for the formulation |BG46g25L1|.
rotation speed of the extruder screw (in rpm) that was measured and
Nmax was the maximum rotation speed of the extruder screw (in rpm)
that the extruder could afford, C and Cmax were respectively the
measured and maximum torque of the motor in Nm.
The results are described in Table B. Finally, Fig. 14 illustrates the
expansion ratio of the foam versus the screw speed and the input rate.
The greater expansion was obtained for the highest possible input rate
with the lowest screw speed available in those conditions.
At a constant feed rate, by increasing the speed of rotation of the
screws (300–600 rpm) the engine torque, residence time, die pressure
and foam expansion ratio decreased. However, EMS increased. In fact,
increasing the speed of rotation of the screws caused an increase in
shear, which reduced the viscosity of the material and the engine
torque, and increased the EMS. Since the material was more easily
swallowed when the screw speed increased, the residence time and the
pressure at the die outlet decreased (Fig. 15) also causing a decrease in
the expansion of the foam after the sector.
The best expansion ratio was obtained at a speed of screw rotation
of 300 rev/min and at a flow rate of material 21 kg/h. This expansion
was due to the high shear observed by a very high torque (102 Nm),
which caused a high pressure in the die (34 bar) with a mean residence
time of 27 s and an average EMS of 381 kJ/kg. Indeed, when the feed
rate of the material increases in the extruder at a medium screw speed,
the motor torque which is connected to the die pressure, increased, thus
generating such great expansion that the material was pushed into the
die with great force, then generating a greater variation of pressure
with the outside pressure. Foams were visually the best when it ap-
proached the boundary conditions of machinability. Their expansion
rate was thus maximum and they showed less irregularity in appear-
ance.
Contrary to the rotational speed of the screw, increasing the mate-
rial throughput (11–21 kg/h) to the rotational speeds of the screw
Fig. 15. Variation of the pressure at die (full marks) and of the residence time (empty
marks) during the extrusion of the formulation |BG46g25L1| at different screw speeds.
313
Industrial Crops & Products 115 (2018) 306–314
Fig. 16. Effect of the pressure at die (diamond marks), of the residence time (triangle
marks) and of the motor torque (square marks) on the expansion ratio (empty marks)
during the extrusion of the formulation |BG46g25L1|.
constant, the engine torque, die pressure and expansion of the material
increases. However, the material residence time decreases. In fact, at a
set speed, increasing the flow of material results in a higher filling
material in the extruder thereby reducing the residence time, and the
viscosity of the material and causes an increase in torque (Fig. 16).
Because the motor torque increases, the pressure at the outlet of the die
increases causing an increase in the expansion of the foam after the
process (Fig. 16).
4. Conclusion and perspectives
The development of starch-based foam sheets with a twin-screw
extrusion was successfully achieved. Wheat starch had shows better
suitability for cushioning, especially during ageing than pea and potato
starch. Sodium bicarbonate increases the nucleation of the cells and
stabilizes the expansion of the foam. Gluten improves the elasticity of
the foam and makes the incorporation of glycerol easier during extru-
sion. The formulation that exhibits best expansion ratio and rheological
properties is the one containing wheat starch with 46 pcs of glycerol,
25 pcs of gluten and 1 pcs of sodium bicarbonate. Nevertheless, in order
to store these cushioning foams over time for commercial uses, the
formulation must be adapted by using additives, including pest re-
pellents. A new way of preparing formulation was explained. It consists
in pre-mixing all the ingredients (starch, gluten, glycerol and sodium
bicarbonate) at room temperature as a dry blend and then putting all
this sifted powdery mixture in the main feeder of the extruder. Finally,
the machinability analysis performed allowed the selection of the levels
of the screw speed and of the material feed rate to be set. The effect of
extrusion parameters, such as the screw speed, the motor torque, the
die pressure, the residence time and the input rate, on the expansion
ratio was studied. The best expansion ratios are obtained at high input
rate (21 kg/h) and at a screw speed of 300 rpm which was the lowest
screw speed the extruder could afford with the highest input rate. In
those conditions, very stable and homogeneous foams are obtained.
And the recovery properties of such foams are suitable for cushioning in
packaging application. All the originality of this work has been to
propose a technique of twin-screw extrusion making it possible to ob-
tain starch foam strips in stable conditions that could allow a change of
scale. In fact, an industrial extruder, of greater power, will allow the
extrusion of foam plates that can then be thermoformed into cushioning
packaging. Such cushioning packaging would enlarge the range of ap-
plications that already exists for loose fill cushioning materials and
would be competitive with polyethylene, polypropylene, polyether or
polystyrene petrochemical setting foams. Further research will study
the influence of the amount of each ingredient, the influence of ad-
ditives such as PLA, CaCO3, linen on the mechanical properties and the
biodegradability of such foams.
A. Georges et al.
Acknowledgements
This research paper was a contribution of Reims engineering school,
ESIReims. The study was partially financed by the INTERREG IV
European program.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.indcrop.2018.02.001.
References
Alavi, S.H., Gogoi, B.K., Khan, M., Bowman, B.J., Rizvi, S.S.H., 1999. Structural properties
of protein-stabilized starch-based supercritical fluid extrudates. Food Res. Int. 32,
107–118.
Alavi, S.H., Rizvi, S.S.H., Harriott, P., 2003. Process dynamics of starch-based micro-
cellular foams produced by supercritical fluid extrusion. I: model development. Food
Res. Int. 36, 309–319.
Bergel, B.F., Da Luz, L.M., Santana, R.M.C., 2017. Comparative study of the influence of
chitosan as coating of thermoplastic starch foam from potato, cassava and corn
starch. Prog. Org. Coat. 106, 27–32.
Bhatnagar, S., Hanna, M.A., 1995. Properties of extruded starch-based plastic foam. Ind.
Crops Prod. 4, 71–77.
Buléon, A., Colonna, P., Planchot, V., Ball, S., 1998. Starch granules: structure and bio-
synthesis. Int. J. Biol. Macromol. 23, 85–112.
Chinnaswamy, R., Hanna, M.A., 1987. Nozzle dimension effects on the expansion of ex-
trusion cooked corn starch. J. Food Sci. 52, 1746–1747.
Cho, K.Y., Rizvi, S.S.H., 2009. 3D microstructure of supercritical fluid extrudates I: melt
rheology and microstructure formation. Food Res. Int. 42, 595–602.
Colonna, P., Mercier, C., 1985. Gelatinization and melting of maize and pea starches with
normal and high-amylose genotypes. Phytochemistry 24, 1667–1674.
Della Valle, G., Vergnes, B., Colonna, P., Patria, A., 1997. Relations between rheological
properties of molten starches and their expansion behaviour in extrusion. J. Food
Eng. 31, 277–295.
Feillet, P., 2000. Le Grain De Blé: Composition Et Utilisation, first ed. INRA, France.
Goel, S.K., Beckman, E.J., 1995. Nucleation and growth in microcellular materials: su-
percritical CO2 as foaming agent. AIChE J. 41, 357–367.
314
Industrial Crops & Products 115 (2018) 306–314
Heartwin, A., Pushpadass, H.A., Weber, R.W., Dumais, J.J., Hanna, M.A., 2010.
Biodegradation characteristics of starch–polystyrene loose-fill foams in a composting
medium. Bioresour. Technol. 101, 7258–7264.
Homer, S., Kelly, M., Day, L., 2014. Determination of the thermo-mechanical properties in
starch and starch/gluten systems at low moisture content –a comparison of DSC and
TMA. Carbohydr. Polym. 108, 1–9.
Jeong, H., Toledo, R., 2004. Twin-screw extrusion at low temperature with carbon di-
oxide injection to assist expansion: extrudate characteristics. J. Food Eng. 63,
425–432.
Kaisangsri, N., Kerdchoechuen, O., Laohakunjit, N., 2012. Biodegradable foam tray from
cassava starch blended with natural fiber and chitosan. Ind. Crops Prod. 37, 542–546.
Kanlaya Pornsuksomboon, K., Holló, B.B., Szécsényi, K.M., Kaewtatip, K., 2016.
Properties of baked foams from citric acid modified cassava starch and native cassava
starch blends. Carbohydr. Polym. 136, 107–112.
Lee, S.Y., Eskridge, K.M., Koh, W.Y., Hanna, M.A., 2009. Evaluation of ingredient effects
on extruded starch-based foams using a supersaturated split-plot design. Ind. Crops
Prod. 29, 427–436.
Lin, Y., Huff, H.E., Parsons, M.H., Iannotti, E., Hsieh, F., 1995. Mechanical properties of
extruded high amylose starch for loose-fill packaging material. LWT – Food Sci.
Technol. 28, 163–168.
Lopez-Gil, A., Silva-Bellucci, F., Velasco, D., Ardanuy, M., Rodriguez-Perez, M.A., 2015.
Cellular structure and mechanical properties of starch-based foamed blocks re-
inforced with natural fibers and produced by microwave heating. Ind. Crops Prod. 66,
194–205.
Machado, C.M., Benelli, P., Tessaro, I.C., 2017. Sesame cake incorporation on cassava
starch foams for packaging use. Ind. Crops Prod. 102, 115–121.
Muljana, H., Picchioni, F., Heeres, H.J., Janssen, L.P.B.M., 2009. Supercritical carbon
dioxide (scCO2) induced gelatinization of potato starch. Carbohydr. Polym. 78,
511–519.
Pushpadass, H.A., Babu, G.S., Weber, R.W., Hanna, M.A., 2008. Extrusion of starch-based
loose-fill packaging foams: effects of temperature, moisture and talc on physical
properties. Packag. Technol. Sci. 21, 171–183.
Robin, F., Engmann, J., Pineau, N., Chanvrier, H., Bovet, N., Della Valle, G., 2010.
Extrusion, structure and mechanical properties of complex starchy foams. J. Food
Eng. 98, 19–27.
Sanhawong, W., Banhalee, P., Boonsang, S., Kaewpirom, S., 2017. Effect of concentrated
natural rubber latex on the properties and degradation behavior of cotton-fiber-re-
inforced cassava starch biofoam. Ind. Crops Prod. 108, 756–766.
Shey, J., Iman, S.H., Glenn, G.M., Orts, W.J., 2006. Properties of baked starch foam with
natural rubber latex. Ind. Crops Prod. 24, 34–40.
Soykeabkaew, N., Thanomsilp, C., Suwantong, O., 2015. A review: starch-based compo-
site foams. Composites Part A: Appl. Sci. Manuf. 78, 246–263.