Journal of Food Engineering xxx (2017) 1e8

Contents lists available at ScienceDirect

Journal of Food Engineering

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

Applicability of protein and fiber-rich food materials in

extrusion-based 3D printing
€-Kortelainen, Nesli Sozer*
Martina Lille, Asta Nurmela, Emilia Nordlund, Sini Metsa
VTT Technical Research Centre of Finland Ltd., P.O. Box 1000, FI-02044 VTT, Finland

a r t i c l e i n f o a b s t r a c t

Article history: In this study, the applicability of extrusion-based 3D printing technology for food pastes made of protein,
Received 14 October 2016 starch and fiber-rich materials was assessed, as a starting point in the development of healthy,
Received in revised form customized snack products. The printability of starch-, cellulose nanofiber-, milk powder-, oat- and faba
7 February 2017
bean protein-based materials and their mixtures was evaluated by examining the ease and uniformity of
Accepted 26 April 2017
extrusion as well as the precision and stability of the printed pattern. The best printing precision and
Available online xxx
shape stability was obtained with a semi-skimmed milk powder-based paste. Rheological measurements
revealed that the shape stability after printing was linked with the yield stress of the paste. Post-
Keywords:
3D printing
processing by oven drying was most successful at high initial solids contents (<50%) of the printed
Starch samples. Extrusion-based 3D printing is a promising tool for producing healthy, structured foods, but
Protein further research is needed for optimising the mechanical properties of the printed materials.
Cellulose nanofiber © 2017 Elsevier Ltd. All rights reserved.
Rheology
Post-processing

1. Introduction
Digitalization is a rising way of increasing efficiency both in the
supply and manufacturing chain. 3D printing technology is an
example of digitalization of food manufacturing industry. It is
forecasted that 3D printing will be the next disruptive and trans-
formative food technology in the coming decade (Karlgraad, 2011).
3D printing is based on additive manufacturing (AM) that is a layer-
by-layer manufacturing process, which takes advantage of phase
transitions or chemical reactions to fuse material layers together
based on 3D model data (Wegrzyn et al., 2012). The term 3D
printing is defined as fabrication of objects through the deposition
of a material using a print head, nozzle, or another printer tech-
nology, however, it is often used synonymously with additive
manufacturing (SFS-ISO/ASTM 52900:2016; Wohlers, 2014). In this
paper the term 3D printing is used to refer to both additive
manufacturing and 3D printing technologies.
The existing 3D printing technologies are classified into seven
process categories, which are powder bed fusion, directed energy
deposition, material jetting, binder jetting, material extrusion, vat
polymerization and sheet lamination. Each of them have several
* Corresponding author.
E-mail address: nesli.sozer@vtt.fi (N. Sozer).
sub-technologies which all have some similarities but also several
differences related to e.g. form of feedstock, form of energy used in
system, multi-material handling capability and restrictions in
design. (SFS-ISO/ASTM 52900:2016).
The 3D printing technology has generally been used in non-food
sectors such as motor vehicles, aerospace and medical applications.
New 3D printing technologies introduce flexibility and novel design
into industrial mass production and enable the manufacturing of
complex shapes, unique solutions and new kinds of combinations
of materials and components. Growth in the 3D printing industry
has accelerated over the past 5 years as an increasing number of
organizations adopt 3D printed products and services. The
increased production speeds decrease the costs of the manufac-
tured parts which will make the 3D printing technologies relevant
for use also in vehicles, machines as well as in other industry areas.
The trend is that 3D printing is increasingly used for producing
parts or components for final products rather than only for proto-
typing. The selection of 3D printable materials includes plastics,
metals, ceramic, concrete, wood and other bio-based materials
(Wohlers, 2014).
Applying 3D printing in food manufacturing has recently
accelerated. The earliest application of 3D printing in food
manufacturing focused on printing of a “cake mix” by paste
extrusion of a mixture that consists of starch, sugar, corn syrup,
yeast and a cake frosting (Yang et al., 2001). Later on granular bed

http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
0260-8774/© 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
2 M. Lille et al. / Journal of Food Engineering xxx (2017) 1e8

sintering technology which utilizes a laser beam or a binding agent

to melt or fuse granular materials has been adapted to form sucrose
based 3D objects by melting and fusing sugar granules with se-
lective hot air melting technology (Sun et al., 2015) which later on
extended to commercially available 3D printers that enables for-
mation of neat and complex 3D structures based on sugar (http://
www.3dsystems.com/culinary/gallery). Besides sintering, ink jet
printing of sugar by using alcohol as a binder can be applied to
produce 3D candies (Park, 2013). Cornell University, Department of
Mechanical and Aerospace Engineering faculty and students, has
made the first multi-material 3D printer available to public. The
printer was first based on a single head extrusion system which has
been later on extended up to eight extrusion heads enabling
printing of diverse materials from epoxy to silicone to chocolate
and cheese. Later on in collaboration with Cadbury and Dr. Liang
Hao from University of Exeter and his students worked further on
the concept of chocolate printing. The group has focused on the
technological issues of chocolate tempering, deposition precision,
and process control, which led to the development of world's first Fig. 1. Schematic illustration of multi-disciplinary workflow related to the use of 3D
chocolate printer introducing co-creation concept to the consumers printing technology in food manufacturing.

(Causer, 2009). Filled chocolate was later on printed by hot melt

extrusion type 3D printing head (Zoran and Coelho, 2011).
The biggest challenges in 3D food printing are ingredient mix
rheology, structure accuracy and shape-stability, compatibility with
traditional food processing technologies (e.g. baking and drying)
and printing speed. For example, traditional cookie recipes have
been shown to be compatible with 3D printing but due to the
presence of high amount of fats they don't retain their shape and
structure after post-processing (e.g. baking) (Lipton et al., 2010).
Thus, to enable 3D printing of cookie recipes the ingredients and
their proportions need to be modified. For example, shape stability
after printing reduced by increasing the amount of butter in butter-
egg yolk-sugar mixture whereas increasing the egg yolk concen-
tration increased width/length stability but decreased height sta-
bility (Lipton et al., 2010). Pizza recipe is another traditional food
product, which needed recipe adaptation to enable the manufac-
ture of 3D printed pizza. The pizza dough required extra water
addition whereas addition of thickeners to the tomato sauce
needed to enable their printability by a paste extrusion based 3D
printer (Lipton et al., 2015). Prior to extrusion the cheese was
melted by heating at 150
C cooled down on the sauce surface. The
use of 3D printer instead of a mold enabled the manufacture of
custom shaped pizza in a shorter time (30 min vs 24 h including
fermentation) (Lipton et al., 2015). In addition, there are techno-
logical challenges to tackle, for instance printing of batters is
strongly dependent on oscillations in the flow during its pumping
process, which has an impact in the final quality of the printing
outcome (Millen, 2012). 3D printing technology has also been uti-
lized in developing various hydrocolloid based structures. Cohen
et al. (2009) printed various combinations of xanthan gum and
gelatin to create a “mouthfeel matrix” where the material was rated
along two orthogonal axes: 1) weak to firm, and 2) smooth to
granular in comparison to other common food materials. The study
showed that the 3D printed structures made of the two hydrocol-
loids in various percentages can cover a mouthfeel range from
liquids to solid vegetables. Pure xanthan and gelatin followed weak
to firm axis (e.g. 0.5% gelatin/milk like texture, 4% gelatin /
mushroom like texture) as a factor of hydrocolloid concentration
however when the two hydrocolloids were combined they started
to possess granularity (e.g. 1% gelatin & 4% xanthan / risotto like
texture, 1% gelatin & 8% xanthan / tomato like texture). The in-
tensity of granular structure increased with increasing xan-
than:gelatin ratio (Cohen et al., 2009).
Despite the recent studies and steps of progress, the majority of
the technologies for 3D food printing are still under research and
development phase. . The major benefits we foresee for 3D printing
of food materials are flexibility, configurability and high material
use efficiency. The flexibility coming from 3D printing technology
will enable the use of alternative food ingredients which the in-
dustry will face in future to create improved products with respect
to nutritional content, health benefits and shelf-life.
The aim of this study was to utilize 3D printing technology in
design of healthy novel structures, which are high in fibre, protein
and low in fat or sugar. For that purpose various protein, starch and
fiber-rich food ingredients and their mixtures were selected for this
study based on their potential as nutrients and functional in-
gredients in 3D printed foods. . The development of 3D printed
customized structures requires cross disciplinary interaction be-
tween food chemistry (e.g. to define the state/phase transitions of
ingredient mixes), processing (e.g. to optimize the rheological
properties of the ingredient mixes during and post-printing) and
engineering (e.g. 3D printing technologies to create customized
healthy food structures) (Fig. 1). Shape and design elements have
been the major focus within the scarce number of studies done
around 3D food printing. However, our lab aims at utilizing this
“next disruptive food technology” in design of healthy novel
structures.
2. Materials and methods
2.1. Materials
The raw materials used in the pastes prepared for 3D printing
were starch, milk powder, cellulose nanofiber, rye bran, oat protein
concentrate and faba bean protein concentrate. The starch was a
cold water swelling modified food starch derived from waxy maize
starch (Ultra-Sperse® M, Ingredion). Two types of instant milk
powders from Valio Ltd. with a difference in fat content were used
in the pastes. The skimmed milk powder (SMP) consisted of 35%
protein, 53% carbohydrates (all lactose) and 0.6% fat (manufac-
turer's data). The semi-skimmed milk powder (SSMP) was lactose-
free and consisted of 37% protein, 38% carbohydrates and 15% fat
(manufacturer's data).
Cellulose nanofiber (CNF) was prepared from once dried
bleached birch Kraft pulp from Finland. The dry matter composition
of the pulp was 73% cellulose, 26% hemicellulose and 1% lignin. The
pulp suspension (2% dry matter) was pre-refined by passing it once
through an ultra-fine friction grinder (Masuko Supermasscolloider

Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
 M. Lille et al. / Journal of Food Engineering xxx (2017) 1e8
MKZA10-15 J from Masuko Sangyo, Japan). The grinding was per-
formed with standard grinding stones (MKE10-46). The rotation
speed was 1500 rpm and the gap width was 0.18 mm. The gel was
prepared by feeding the pre-refined fiber suspension into a
microfluidizer (M-7115-30, Microfluidics, US) at an operating
pressure of 185 MPa. The first pass was through the chambers
having a diameter of 500 mm and 200 mm. The next five passes were
through the 500 mm and 100 mm chambers. The final fibrillated
product formed a viscous hydrogel at 1.6% dry matter.
The finely ground rye bran was prepared as described by Alam
et al. (2014) and it consisted of 13.1% protein, 44.3% starch, 26.7%
dietary fiber and 1.9% fat on dry weight basis (dw). The oat protein
concentrate (OPC) was prepared as described by Sibakov et al.
(2011) and it consisted of 48.3% protein, 37.0% starch, 6.8% dietary
fiber and 4.0% fat (dw). The faba bean protein concentrate (FBPC)
was prepared from faba beans (cv. Kontu) by stone milling,
dehulling and grinding into flour for further air classification as
described in Coda et al. (2015). From air classification, both protein
and starch rich fractions were obtained. The protein rich fraction,
which had 62.1% protein, 0.9% starch, 10.7% dietary fiber, 6.4% ash
and 3.5% fat (dw), was further used in the printing trials.
2.2. Preparation of pastes for 3D printing
The pastes consisting of starch or milk powder alone were
prepared by mixing the starch or milk powder into deionized water
with a spoon. The starch-milk powder mixtures were prepared by
first dissolving the milk powder into deionized water under mag-
netic stirring. Afterwards all the starch was added and everything
mixed with a spoon. The prepared suspensions were homogenized
by mixing with a Bamix wand mixer (2-speed model M133, ESGE
AG, Switzerland) equipped with a multi-purpose blade. The mixing
was carried out at full speed in three 10 s cycles. A final homoge-
nization step was performed by extruding the paste once through a
50 ml Luer-Lok syringe (BD Plastipak, Becton Dickinson S.A. Spain).
The paste was stored at room temperature for a minimum of 2 h
before 3D printing or rheological characterization, which were
carried out during the same day.
Mixtures of CNF and starch or milk powder were prepared by
adding the starch or milk powder into the CNF gel and mixing with
a spoon. No extra water was added to the mixture. The paste was
homogenized and allowed to settle before printing as described
above.
Plant protein concentrates (OPC, FBPC) and rye bran were sus-
pended in deionized water by mixing with a spoon. The suspension
was heated in a boiling water bath for 10 min under regular spoon
mixing. The suspension was allowed to cool to room temperature
and homogenized by mixing with the Bamix wand mixer for 4e6
times 10 s until a visually homogeneous paste was obtained. The
final homogenization step was performed with the 50 ml syringe as
described above. The paste was stored at 4
C overnight and
allowed to warm to room temperature before it was used for 3D
printing and rheological characterization.
In preliminary experiments, pastes at various concentrations of
the studied materials and their mixtures were prepared. Print-
ability of the pastes was assessed by pressing the paste by hand
through a 10 mL Luer-Lok syringe equipped with the same tip
(0.41 mm) than was used in the 3D printer. Only the best per-
forming pastes, i.e. those which could be passed through the sy-
ringe and showed reasonable shape stability after extrusion, were
chosen for the actual 3D-printing trials.
2.3. Rheological characterization of pastes
The viscoelastic properties of selected pastes prepared for 3D
Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
3
printing were analyzed by stress sweep measurements with a
stress controlled AR-G2 rheometer (TA Instruments, UK). Stainless
steel parallel plates with a diameter of 20 mm and a gap of 1 mm
were used as measuring geometry. Measurements were carried out
at 22
C. After sample loading and trimming, the edge of the sample
was covered with rapeseed oil to prevent evaporation of water and
the sample was allowed to relax for 5 min before starting the
measurement. The stress sweep measurements were carried out
with a logarithmically increasing shear stress at a frequency of
0.1 Hz (start stress 0.001 Pa, 5 points/decade). Each material was
measured in triplicate.
2.4. 3D printing process
VTT's micron scale dispensing environment based on nScrypt
technology (nScrypt, Inc, Orlando, Florida) consists of several
different types of pumps which enable the printing of materials
with versatile rheological properties, such as pastes, molten poly-
mers, dispersions, inks and gels (Metsa €-Kortelainen et al., 2014). A
CAD-controlled xyz-motion control system guides nozzle position
and enables the deposition of 3D structures in a layer-by-layer
approach.
A simplified pump system was used in the 3D food printing
trials and it is based on an air pressure controlled extrusion of the
pastes through a nozzle (tip) of a syringe (Fig. 2). Syringes with a
volume of 3 ml were filled with the pastes prepared for the 3D
printing trials. The filled syringes were processed in a speed mixer
(SpeedMixerTM DAC 150 SP) for 2 min for removing the air bubbles
from the samples before the 3D printing.
The script which includes the movement paths and the speed
for the 3D printing was created using a computer. 25 mm  25 mm
squares filled with diamond-like structures were 3D printed at
room temperature on a plastic substrate so that the tip didn't cross
the formerly printed zigzag-shaped lines. The amount of the layers
was 10 and the printing speed and the diameter of the tip were
2 mm/s and 0.41 mm, respectively. The height of the tip was
increased by 0.3 mm after each layer. The air pressure used for the
extrusion of the paste was chosen experimentally and it normally
varied between 22 and 600 kPa depending on the paste material.
Photographs of the printed structures were taken during and
right after printing and print quality was visually assessed on a
scale from 1 to 5 (1 ¼ very bad, 5 ¼ very good). The precision and
shape-stability of the printed structures were taken into account in
the print quality assessment.
2.5. Post-processing of printed objects
After the 3D printing process, selected samples were post-
processed by oven drying or freeze-drying. The oven drying was
performed in a Heraues UT6060 heating and drying oven (Thermo
Scientific, Germany) at 100
C for 20e30 min. For freeze drying, the
samples were frozen at 18
C and freeze-dried with a CRIST
ALPHA laboratory freeze-drier (B. Braun Biotech International,
Germany).
2.6. Hardness measurement of post-processed samples
The hardness of the printed samples right after oven or freeze
drying was measured with a cutting test, using a Texture Analyser
(TA.XTPlus, Stable Micro Systems Ltd., UK). The samples were
placed on a slotted insert and cut in the middle over their whole
width (~2.5 cm) with a knife blade (HDP/BS) at a speed of 1 mm/s.
The maximum force of the first peak was recorded as hardness. The
dry matter content of the samples was determined right after the
hardness measurement by over night drying at 105
C.
4 M. Lille et al. / Journal of Food Engineering xxx (2017) 1e8

Fig. 2. The material extrusion type of device used for 3D printing of food materials.

2.7. Statistical analysis the SMP concentration to 60% increased the viscosity of the paste so
much that it was impossible to print even if the tip diameter was
Statistical significance of rheological data between selected increased from 0.41 to 1.19 mm (Table 1). An interesting finding was
samples was assessed by independent t-test using a significance the excellent printability of the SSMP at a concentration of 60%. The
level of 0.05 by using SPSS Statistics 22 (SPSS Inc., Chicago, USA) material kept its shape very well after printing which resulted in
software. very good printing precision (Table 1). The protein content of the
SMP and SSMP pastes was at a comparable level (21 vs. 22%), but
the fat content was clearly different (0.4 vs. 9%). The fat present in
3. Results and discussion
the SSMP might have acted as a plasticizer or lubricant in the sys-
tem and made the material more easily flowing (Fu et al., 1997).
3.1. Printability of pastes
There was also a difference in the carbohydrate content between
the two milk powder pastes (32% in SMP vs. 23% in SSMP paste),
The 3D printing trials were conducted with ingredients
which may have influenced the flow behavior of the materials. The
commonly found in foods, which were considered to be useful
carbohydrates in SMP were lactose and lactose hydrolysis products
building blocks in the development of healthy 3D printed foods of
(glucose and galactose) in SSMP. Printing of milk powder-based
diverse textures. The ingredients contained components from all
pastes required relatively high pressures as compared to printing
main groups of nutrients, i.e. carbohydrates, protein, fat and dietary
of pure starch paste (e.g. 110 kPa for 60% SSMP and 37 kPa for 15%
fiber, either alone or in combination with other ingredients.
starch).
Starch was selected to the 3D printing trials as a carbohydrate
As a next step, mixtures of starch and milk powder were stud-
source and its functionality as a thickening/gelling agent. A certain
ied. The printability of the pregelatinized starch was improved by
pregelatinized (cold-swelling) starch type was recommended for
the addition of SMP. A mixture consisting of 10% starch and 15%
this study by the manufacturer, as this type of starch has good
SMP showed better shape stability after printing than 15% starch
shear-stability and a yield stress when used at a concentration
alone (Table 1), but the printing quality was still not as good as with
above close packing of the swollen granules, which helps in
the 60% SSMP alone. The starch-SMP mixtures with up to 15% SMP
retaining the shape after deposition. The starch also shows shear-
could be printed at lower pressure (50 kPa) than the 60% SSMP
thinning behavior in flow, which facilitates the extrusion of the
paste (110 kPa). Increasing the SMP content in the starch-SMP
material through the syringe tip. The pregelatinized starch formed
mixture to 20% resulted in increased stickiness and hampered
a paste suitable for 3D printing at a concentration of 15%. Images of
printability. The printing pressure also increased from 50 to 114 kPa
the material during and right after 3D printing are shown in Table 1
by increasing the SMP content in the mixture from 15 to 20%.
together with an assessment of the print quality. The 15% starch
Cellulose nanofiber (CNF) is considered to have great potential
paste was easy to print, but the material did not fully keep its shape
as a reinforcing ingredient in composite materials (Abitbol et al.,
after printing, which hampered the precision of the printed pattern.
2016; Klemm et al., 2011). The performance of native CNF in 3D
Milk powder was studied as a potential protein source in 3D
printing was therefore also evaluated with the aim to improve the
printed structures. Both skimmed milk powder (SMP) and semi-
shape stability of the printed structures. Images of the CNF con-
skimmed milk powder (SSMP) were studied with the aim to eval-
taining pastes during and right after printing are presented in
uate the role of fat on the performance of the materials in 3D
Table 2. The printability of the 1.6% CNF gel alone was poor (no
printing. SMP formed a high-viscosity paste at a solids content of
image shown). A continuous flow was difficult to maintain and the
50% but it was challenging to print due to its stickiness, which
tip was clogged by the material, and phase separation was observed
resulted in uneven deposition of the material (Table 1). Increasing

Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
 M. Lille et al. / Journal of Food Engineering xxx (2017) 1e8 5

Table 1
Printability of pastes prepared from starch and/or milk powder.

Sample Air pressure (kPa) During printing After printing Print quality Observations

15% starch 37 3 Spreads after printing

50% SMP 145 3 Sticky

60% SMP 600a 1 Very thick and dry, fails to print

60% SSMP 110 5 Keeps shape very well after printing

10% starch þ 15% SMP 50 4 Spreads after printing

10% starch þ 20% SMP 114 3 Sticky

a
Tip size was 1.19 mm.

Table 2
Printability of pastes containing CNF together with starch or milk powder.

Sample Air pressure (kPa) During printing After printing Print quality Observations

1.5% CNF þ 5% starch 37 3 Tip was clogged

1.4% CNF þ 10% SMP 22 2 Tip was clogged

1.3% CNF þ 20% SMP 23 2 Tip was clogged

0.8% CNF þ 50% SSMP 48 5 No clogging of tip

to take place in the material upon extrusion. The clogging might
have been caused by some larger fiber particles remaining in the
material after fibrillation. It is also possible that shear-induced
flocculation of the material took place when forcing it through
the small tip of the syringe, in a similar manner as observed by
Karppinen et al. (2012) in the small gap of a concentric cylinder
geometry in a rheometer. An attempt to reduce the separation of
liquid from the gel was made by adding 5% starch to the CNF gel
(Table 2). Although starch addition reduced water separation it did
not fully prevent clogging of the tip. The printing precision was still
impaired by uneven flow of the material. The addition of 10e20%
SMP to the CNF gel did not make a clear improvement in the
printability of the material and the printing accuracy was even
worse than with starch addition (Table 2). The clogging problem
was, however, overcome by adding 50% SSMP to the CNF gel, which
resulted in excellent shape retention of the printed samples. The

Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
6 M. Lille et al. / Journal of Food Engineering xxx (2017) 1e8
printing of this material could be carried out at relatively low
pressure (48 kPa) as compared to the paste consisting of 60% SSMP
alone (110 kPa), by which similar printing accuracy was obtained.
The printability of plant materials, containing protein, starch
and dietary fiber was analyzed using rye bran, oat protein (OPC)
and faba bean protein (FBPC) concentrate as raw materials
(Table 3). The starch present in the materials was utilized as
thickening agent by gelatinizing the starch by heating the aqueous
suspension of the material. The mixture was cooled to room tem-
perature before 3D printing. The rye bran formed a paste suitable
for printing at a dry matter content of 30%, but the printing was
disturbed by clogging of the syringe tip. The clogging might have
been caused by the presence of some larger rye bran particles in the
milled material. Alam et al. (2014) reported that 26% of the particles
in the material had a particle size between 100 and 1000 mm, so it is
possible that some larger particles were stuck in the 410 mm syringe
tip. The printability of the rye bran paste might be improved by
sifting out the particles larger than the syringe tip diameter from
the milled rye bran. The OPC could be reasonably well printed at a
solids content of 35% although the printed structure was not fully
stable after printing. The print quality of the OPC paste was com-
parable to that obtained with the paste composed of 10% starch and
15% SMP. For FBPC, a higher concentration than for OPC was needed
to create a printable paste. The viscosity of a 35% FBPC paste was too
low for printing. Better printability was obtained by increasing the
dry matter content of the FBPC paste to 45%, even though some
material spreading after printing was still observed. No problems
with tip clogging were encountered with pastes prepared from OPC
or FBPC. There were substantial differences in the concentrations of
the plant materials required to create pastes with a suitable vis-
cosity for printing, i.e. 30% for rye bran, 35% for OPC and 45% for
FBPC. Differences in the composition, size and shape of insoluble
particles and response to heating and shear between the materials
are some of the factors which may have created differences in
rheological properties of the prepared pastes.
3.2. Viscoelastic properties of selected pastes
The printability of soft materials by extrusion is highly governed
by the rheological properties of the materials (Godoi et al., 2016).
Table 3
Printability of pastes prepared from rye bran or plant protein concentrates.
Sample Air pressure (kPa) During printing
30% rye bran 50
35% OPC 38
35% FBPC 31a
45% FBPC 56
a
Tip size was 0.2 mm.
Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
The viscosity of the material need to be low enough at high shear
rates to allow flow through a small nozzle, but the material has to
quickly regain a high viscosity at rest to support the structure after
deposition. Viscosity measurements at high shear rates cannot be
reliably performed for highly viscous materials with a rotational
rheometer due to artefacts caused by e.g. shear fracturing and
sample escaping from the gap. The viscoelastic properties of high-
viscosity materials can, however, be reliably analyzed by oscillatory
measurements at small deformations.
In this study, the viscoelastic properties of selected pastes were
measured by oscillatory stress sweeps, with the aim to find out
whether they could be utilized in predicting the printability of the
materials. Representative storage modulus (G0 ) curves from these
measurements are shown in Fig. 3. In stress sweep measurements,
the G0 typically show constant plateau values at small stress values.
The G0 in this plateau region, i.e. in the linear viscoelastic (LVE)
range, is a measure of the elasticity of a material at very small
deformation and it can be taken as a measure of the structural
strength or mechanical rigidity of a material at rest (Metzger, 2006).
The numerical results of the stress sweep measurements are pre-
sented in Table 4. The results show that all samples had an elasticity
dominating gel-like structure (G’ > G00 , phase angle < 45
). The 60%
SMP paste, which was very difficult to extrude through the tip of
Fig. 3. Oscillatory stress sweeps for selected pastes prepared for 3D printing together
with images of the corresponding materials right after 3D-printing.
After printing Print quality Observations
4 Tip was clogged
4 Spreads after printing
1 Viscosity too low
3 Spreads after printing
 M. Lille et al. / Journal of Food Engineering xxx (2017) 1e8
Table 4
Stress sweep measurement results. The storage modulus (G0 ), loss modulus (G00 ) and phase angle values were calculated as averages (±stdev) of the values in the linear
viscoelastic range.
Sample G' (Pa)
15% starch 261 ± 2a
10% starch þ 15% SMP 280 ± 4b
60% SMP 36000 ± 3400d
60% SSMP 1550 ± 230c
35% OPC 1900 ± 30c
*Independent T-test, Sig. (2-tailed) value less than or equal to 0.05 confidence level. Values marked with same letter within a column were not significantly different.
the syringe showed the highest G’ (~36 000 Pa) of all the measured
materials. The G0 of the sample which performed the best in 3D
printing (amongst the materials tested in this study), i.e. the 60%
SSMP paste, was at an intermediate level (~1500 Pa) as compared to
other measured samples. The G0 of the 35% OPC paste was at a
comparable level (~1900 Pa) with the 60% SSMP paste, although the
printing quality was not as good. The materials which were easy to
extrude but spread to some extent after printing (15% starch and
10% starch þ 15% SMP) showed relatively low storage modulus
values (~300 Pa) (Table 4).
Another interesting parameter extracted from the stress sweep
curves was the yield stress, i.e. the point at which the G’ (and G00 )
started to decrease after the plateau region. This point can be taken
as a measure of the stress at which the network structure of the
material starts to break down. The yield stress is related to the
ability of the material to keep its shape under gravity and under the
stresses generated by material layers deposited on top of it. The
yield stress value (Table 4) of the 60% SSMP paste, which showed
the best shape-stability after printing, was significantly higher than
the yield stress values of the samples which spread to some degree
after printing. It seems that a high structural strength at rest (G0 ) is
not enough to ensure good printability of a material, but that a
certain degree of resistance to external stresses (yield stress) is
required as well. The rate of structure recovery after cessation of
shear also play a role in shape-stability after printing, but that could
not be tackled with the rheometer due to the above described
challenges of measuring high-viscous materials at large de-
formations and high shear rates.
It is worthwhile to mention that none of the rheological pa-
rameters extracted from the stress sweep measurements showed
clear correlation with the pressures used to print the materials
(showed in Tables 1 and 3). The lack of correlation might be
explained by the fact that the stress sweep measurements are
performed at very small deformations, whereas the materials are
subjected to much larger deformations during extrusion through
Fig. 4. The effect of oven and freeze drying on appearance of 3D printed samples: 1) 1.5% CNF þ 5% starch, 2) 15% starch, 3) 30% rye bran, 4) 35% OPC, 5) 45% FBPC, 6) 0.8% CNF þ 50%
SSMP and 7) 60% SSMP.
Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
7
G’’ (Pa) Phase angle (
) Yield stress (Pa)
43.5 ± 0.1a 9.48 ± 0.07a 5.34 ± 0.00a
53.3 ± 0.4b 10.8 ± 0.2b 8.52 ± 0.02b
10600 ± 1300d 16.4 ± 0.4c 99.9 ± 0.0e
290 ± 41c 10.6 ± 0.1b 61.1 ± 0.3d
320 ± 6c 9.6 ± 0.1a 24.3 ± 0.1c
the syringe tip. The material is also sheared in a different manner in
the oscillatory measurements in the rheometer than in the tip of
the syringe.
3.3. Effect of post-processing on appearance of printed samples
Post-processing by drying was studied as a means of increasing
the stiffness of the 3D-printed samples. A comparison between
oven drying and freeze-drying was conducted. Fig. 4 shows images
of a selection of 3D printed materials right after printing and oven
drying at 100
C or freeze-drying. The samples in Fig. 4 are ordered
by increasing initial dry matter content of the paste from left to
right. Fig. 4 clearly demonstrates that the printed structure was
better preserved by freeze-drying than by oven drying for samples
with an initial dry matter content of 35% or lower. Freeze-drying
normally preserves structures better, as the material is dehy-
drated in frozen state and the more solid structure can resist the
stresses caused by the process better (Ratti, 2001). Oven drying
caused considerable shrinkage of the samples (particularly for
samples 1 and 3), which is typical for conventional drying by hot air
(Mayor and Sereno, 2004). With the 15% starch (sample 2) and 35%
OPC also some spreading of the material appears to have occurred
during oven drying, possibly due to a decrease in the viscosity of
the materials during heating. However, for samples with an initial
dry matter content of 45% or higher (samples 5e7), almost com-
parable structures were obtained by both drying methods, due to a
lower amount of water removed and a possible higher initial
structural strength of the materials.
3.4. Effect of post-processing on dry matter content and hardness of
printed samples
The best performing protein-rich samples were characterized by
measuring their hardness and dry matter content (Table 5) right
after post-processing. Oven drying did not result in complete
8 M. Lille et al. / Journal of Food Engineering xxx (2017) 1e8
Table 5
Dry matter content and hardness of selected 3D printed samples after oven- or
freeze-drying.
Sample Dry matter content (%) Hardness (N)
Oven-dried Freeze-dried Oven-dried
60% SSMP 85.2 ± 0.9 89.4 ± 0.3 42.6 ± 3.1
50% SSMP þ 0.8% CNF 74.6 ± 2.6 90.8 ± 0.7 16.3 ± 4.2
35% OPC 79.8 ± 1.9 95.6 ± 0.1 13.0 ± 2.0
45% FBPC 63.1 ± 1.7 95.9 ± 0.1 2.9 ± 0.9
a
Sample very fragile, only one sample could be measured.
drying of the samples as the processing time was limited to
30 min at maximum. The SSMP samples showed the highest
hardness (43 N) after oven drying. Replacing part of the milk
powder with CNF resulted in softening of the samples, possibly due
to the lower dry matter content of the CNF containing sample (75%
with and 85% without CNF). Both SSMP containing samples were
oven dried for 20 min, but the high water binding capacity of CNF
apparently reduced the drying rate. The oven-dried FBPC samples
were much softer than the oven-dried OPC samples. Both of them
were dried for 30 min, but less water was removed from the FBPC-
based samples (63% dm) than from the OPC-based (80% dm) ones.
The freeze-dried samples had a dry matter content of 89e96%
(Table 5). Freeze-drying of the OPC- and FBPC-based samples
appeared somewhat more effective than that of the SSMP-based
ones. The addition of CNF to SSMP reduced the hardness as
compared to using SSMP alone. There was a remarkable difference
in the hardness of the OPC- and FBPC-based samples. The OPC-
based samples were so fragile that most of the samples fell in
pieces already before the hardness measurement (only one sample
could be measured), whereas the FBPC-based samples were even
harder than the SSMP-based samples.
It is apparent that the drying rate and hardness of 3D printed
samples is highly affected by the composition of the printed paste.
Optimization of the raw material mixtures and the drying process is
needed for producing products with desired mechanical properties.
4. Conclusions
This study demonstrated the applicability of diverse food ma-
terials (e.g. cold swelling starch, milk powder, rye bran, oat and faba
bean protein concentrates, cellulose nanofibers) in 3D food print-
ing, which is a starting point for future development of healthy,
customized 3D printed foods. Best printing results were achieved
with pastes containing 10% cold swelling starch þ 15% SMP, 60%
SSMP, 30% rye bran, 35% OPC or 45% FBPC. High yield stress was
required to achieve good shape stability after printing. A high solids
content was shown to be beneficial in post-processing by oven
drying. Further research is needed to understand the effects of 3D
printing on both micro- and macrostructural properties and
texture. The authors predict that 3D printing can be applied to food
processing industry as a rapid prototyping tool to plan, design and
test new ingredients and/or food structures.
Conflicts of interest statement
The authors have declared no conflict of interest.
Acknowledgements
Leila Kostamo and Riitta Pasanen (VTT Technical Research
Centre of Finland Ltd) are acknowledged for their excellent
Please cite this article in press as: Lille, M., et al., Applicability of protein and fiber-rich food materials in extrusion-based 3D printing, Journal of
Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.04.034
assistance in sample preparation. This work was funded by the
Bioeconomy Transformation Spearhead Programme of VTT Tech-
nical Research Centre of Finland.
Freeze-dried Appendix A. Supplementary data
36.7 ± 5.7
9.4 ± 1.4 Supplementary data related to this article can be found at http://
2.2a
dx.doi.org/10.1016/j.jfoodeng.2017.04.034.
59.8 ± 16.3
References
Abitbol, T., Rivkin, A., Cao, Y., Nevo, Y., Abraham, E., Ben-Shalom, T., Lapidot, S.,
Shoseyov, O., 2016. Nanocellulose, a tiny fiber with huge applications. Curr.
Opin. Biotechnol. 39, 76e88.
Alam, S.A., Ja €rvinen, J., Kirjoranta, S., Jouppila, K., Poutanen, K., Sozer, N., 2014. In-
fluence of particle size reduction on structural and mechanical properties of
extruded rye bran. Food Bioprocess Technol. 7 (7), 2121e2133.
Causer, C., 2009. They have got a golden ticket. Potentials IEEE 28 (4), 42e44.
Coda, R., Melama, L., Rizzello, C.G., Curiel, J.A., Sibakov, J., Holopainen, U.,
Pulkkinen, M., Sozer, N., 2015. Effect of air classification and fermentation by
Lactobacillus plantarum VTT E-133328 on faba bean (Vicia faba L.) flour nutri-
tional properties. Int. J. Food Microbiol. 193, 34e42.
Cohen, D.L., Lipton, J.I., Cutler, M., Coulter, D., Vesco, A., Lipson, H., 2009 August.
Hydrocolloid printing: a novel platform for customized food production. In:
Solid Freeform Fabrication Symposium (SFF’09).
Fu, J., Mulvaney, S.J., Cohen, C., 1997. Effect of added fat on the rheological properties
of wheat flour doughs. Cereal Chem. 74 (3), 304e311.
Godoi, F.C., Prakash, S., Bhandari, B.R., 2016. 3d printing technologies applied for
food design: status and prospects. J. Food Eng. 179, 44e54.
Karlgraad, R., 2011. 3D Printing Will Revive American Manufacturing. Forbes. http://
www.forbes.com/sites/richkarlgaard/2011/06/23/3d-printing-will-revive-
american-manufacturing/#6ba9a9a4295c (June 23, 2011) (accessed on 31/01/
2017).
Karppinen, A., Saarinen, T., Salmela, J., Laukkanen, A., Nuopponen, M., Seppa €la
€, J.,
2012. Flocculation of microfibrillated cellulose in shear flow. Cellulose 19 (6),
1807e1819.
Klemm, D., Kramer, F., Moritz, S., Lindstro €m, T., Ankerfors, M., Gray, D., Dorris, A.,
2011. Nanocelluloses: a new family of nature-based materials. Angew. Chem.
Int. Ed. 50 (24), 5438e5466.
Lipton, J., Arnold, D., Nigl, F., Lopez, N., Cohen, D.L., Nore n, N., Lipson, H., 2010,
August. Multi-material food printing with complex internal structure suitable
for conventional post-processing. In: Solid Freeform Fabrication Symposium,
pp. 809e815.
Lipton, J.I., Cutler, M., Nigl, F., Cohen, D., Lipson, H., 2015. Additive manufacturing for
the food industry. Trends Food Sci. Technol. 43 (1), 114e123.
Mayor, L., Sereno, A.M., 2004. Modelling shrinkage during convective drying of food
materials: a review. J. Food Eng. 61 (3), 373e386.
Mets€ a-Kortelainen, S., Ronkainen, H., Varis, T., Ruusuvuori, K., Roozeman, R.,
Vehmas, T., Kupiainen, V., Ahlroos, T., Lagerbom, J., Suhonen, T., 2014. Material
challenges in the manufacturing of tailored structures with direct write tech-
€llinghaus, T., Manonukul, A., Lexow, J.
nologies. In: Udomkichdecha, W., Bo
(Eds.), Materials Challenges and Testing for Manufacturing, Mobility, Biomed-
ical Applications and Climate. Springer, Cham, pp. 57e66.
Metzger, T.G., 2006. The Rheology Handbook, second ed. Vincentz Network,
Hannover, pp. 114e141.
Millen, C.I., 2012. The Development of Colour 3D Food Printing System. Diss. Master
thesis. Massey University, Palmerston North, New Zealand.
Park, R., 2013. Bitter Sweet e 3D Systems Acquires the Sugar Lab (2013, September
12). Retrieved from 3D Printing Industry. https://3dprintingindustry.com/news/
bitter-sweet-3d-systems-acquires-the-sugar-lab-16923/ (accessed on 31/01/
2017).
Ratti, C., 2001. Hot-air and freeze-drying of high value foods: a review. J. Food Eng.
49 (4), 311e319.
SFS-ISO/ASTM 52900, 2016. Additive Manufacturing e General Principles - Termi-
nology. Finnish Standards Association.
Sibakov, J., Myllyma €ki, O., Holopainen, U., Kaukovirta-Norja, A., Hietaniemi, V.,
Pihlava, J.M., Poutanen, K., Lehtinen, P., 2011. Lipid removal enhances separation
of oat grain cell wall material from starch and protein. J. Cereal Sci. 54 (1),
104e109.
Sun, J., Zhou, W., Huang, D., Fuh, J.Y., Hong, G.S., 2015. An overview of 3D printing
technologies for food fabrication. Food bioprocess Technol. 8 (8), 1605e1615.
Wegrzyn, T.F., Golding, M., Archer, R.H., 2012. Food Layered Manufacture: a new
process for constructing solid foods. Trends Food Sci. Technol. 27 (2), 66e72.
Wohlers report 2014. ISBN 978-0-9913332-0-2.
Yang, J., Wu, L., & Liu, J. (2001). Rapid prototyping and fabrication method for 3-D
food objects, U.S. Patent No. 6280785.
Zoran, A., Coelho, M., 2011. Cornucopia: the concept of digital gastronomy. Leonardo
44 (5), 425e431.