10th International Conference on Fracture Mechanics of Concrete and Concrete Structures

FraMCoS-X
G. Pijaudier-Cabot, P. Grassl and C. La Borderie (Eds)
https://doi.org/10.21012/FC10.235356

3D PRINTED CAPSULES FOR SELF-HEALING CONCRETE APPLICATIONS

* † ††
GIOVANNI ANGLANI , PAOLA ANTONACI , SUSANA I. CARILLO GONZALES ,
†† ††
GIORGIA PAGANELLI AND JEAN-MARC TULLIANI
*
Department of Structural, Geotechnical and Building Engineering (DISEG), Politecnico di Torino
Corso Duca degli Abruzzi 24, 10129 Turin, Italy
e-mail: giovanni.anglani@polito.it, www.diseg.polito.it
†
Department of Structural, Geotechnical and Building Engineering (DISEG), Politecnico di Torino
Corso Duca degli Abruzzi 24, 10129 Turin, Italy
e-mail: paola.antonaci@polito.it, www.diseg.polito.it
†† Department of Applied Science and Technology (DISAT), Politecnico di Torino
Corso Duca degli Abruzzi 24, 10129 Turin, Italy
e-mail: jeanmarc.tulliani@polito.it, www.disat.polito.it

Key words: Self-healing concrete, Additive Manufacturing, 3D printed capsules, Crack repair,
Mechanical recovery, Durability

Abstract: This study focuses on the use of additive manufacturing techniques as a possible way to
produce novel types of macro-capsules, suitable to be filled with different healing agents and
subsequently embedded in cementitious materials to introduce self-healing properties and deal with
the sensitivity of these materials to crack formation. Specifically, tubular macro-capsules were
produced by 3D printing and filled either with an expansive polyurethane resin or with a liquid
sodium silicate. The 3D-printed capsules were first characterized in terms of water tightness and
shell continuity, to evaluate the eventual need for additional protective layers. Then, they were
inserted into cement mortar prisms and subjected to controlled cracking by three-point-bending to
experimentally investigate their bonding with the cement matrix and the ability to release their
content at the crack site to promote self-healing. The actual self-healing effect was expressed in
terms of recovery of load-carrying capacity under mechanical tests. Positive results were achieved,
showing a good potential of the proposed capsule-based system for structural applications. Further
improvements can be envisaged through the optimization of the capsule shape, for which the 3D
printing technology offers virtually unlimited possibilities.

1 INTRODUCTION be one of the most promising due to its

Recently, a growing attention has been versatility. This technological approach is
devoted by the international scientific based on the principle that when a crack comes
community to the development of across a capsule, the capsule shell is broken
technological solutions able to introduce self- and thus, the healing agent contained in it can
healing properties in cementitious materials, in be released into the crack, restoring the
such a way to deal with the intrinsic sensitivity damage [3]. With regard to their dimension,
of these materials to crack formation. Different capsules can be divided into micro- and
strategies have been proposed so far, as macro-capsules, respectively presenting a
extensively reviewed in [1,2]. Among all of maximum size smaller or larger than 1 mm.
them, the capsule-based approach appears to Whereas the micro-capsules [4–6] can be well

1
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

dispersed into the matrix while introducing almost immediately and cold weld to the
smaller pores, the volume of carried healing previous extruded layers [22]. The turning
agent could be too small to fill completely point for the diffusion of the FDM was the
large cracks. Moreover, also due to their expiration in 2009 of the abovementioned
typical spherical shape, their adhesion with the patent and the widespread open-source
surrounding matrix could be not enough to movement which created significant cost
allow the shell breakage upon cracking, thus reductions for these 3D printers, which are
not triggering the self-healing effect. nowadays available on the market also with
A possible solution to these issues can be low costs (< € 1000) as not-assembled DIY
offered by tubular macro-capsules, which (Do It Yourself) open-source printers, fully
allow higher storage capacity and likelihood of assembled open-source printers or commercial
shell breakage upon crack interception. systems with proprietary software [24].
Different encapsulation techniques have been In this work, macro-capsules with two
extensively investigated and successfully used different sizes were produced by FDM using
by different research groups, whether in terms an open-source 3D printer and different types
of size, materials or production techniques. of thermoplastic filaments, then filled either
Examples include hollow glass tubes [7–9], with an expansive polyurethane resin or with a
ceramic tubes produced by phase-inversion liquid sodium silicate. Their shape was
spinning [7,10], cementitious capsules tubular, with wall thickness of 200 µm,
produced by extrusion [3,11], polymeric external diameter of 7.9 or 8.5 mm, and total
capsules produced by extrusion [12–15]. length of 40 or 50 mm. Such simple shape
This study focuses on exploiting the settings were selected for the sake of
potential offered by Additive Manufacturing comparison with similar macro-capsules
(AM), also popularly known as 3D printing, as developed and thoroughly investigated by the
a possible way to produce novel types of authors, which are made of a different material
macro-capsules, suitable to be filled with (cement paste) and produced by extrusion,
different healing agents and subsequently filled with the same healing agents and used in
embedded in cementitious materials. In the last the same cementitious systems [3,11,25]. The
decade, the rapid development of AM 3D-printed capsules were first characterized in
technology allowed the fabrication of terms of water tightness and shell continuity,
materials with increasing complex geometries to evaluate the eventual need for additional
[16–18] and its progressive introduction to the protective layers. Then, they were inserted into
construction industry [19–21]. Unlike cement mortar prisms and subjected to
conventional manufacturing techniques that cracking by three-point-bending, to evaluate
fabricate products by removing materials from experimentally their ability to create a stable
a larger stock of bulk material, AM processes bond with the cement matrix, to be ruptured
allow the creation of the final shape by adding concurrently with the formation of the fracture
materials under the control of an automated in the matrix and to release their content at the
system, usually layer by layer that are built crack site to promote self-healing. The actual
and consolidated using different techniques self-healing effect was expressed in terms of
[22], for example the common processes of recovery of load-carrying capacity under
Fused Deposition Modelling (FDM), Selective mechanical tests, after reloading of the healed
Laser Sintering (SLS) or stereolithography specimen by three-point-bending.
(SLA). The technique used in this study was The purpose of this study was to assess the
the FDM, which was developed and patented feasibility of this proof-of-concept self-healing
by S. Scott Crump in 1992 [23]. It is based on system, in view of further improvements that
the heating and successively extrusion of can be envisaged through the optimization of
thermoplastic filaments from a movable hot the capsule shape, for which the 3D printing
head that deposit the melted material in thin technology offers virtually unlimited
layers onto a substrate, which then solidify possibilities.

2
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

2 3D PRINTED CAPSULES renewable resources such as corn or

sugarcane and it is a biodegradable
2.1 Production of the 3D printed capsules thermoplastic aliphatic polyester. PLA
The open-source Prusa I3 Hephestos has a low Tg (around 60 °C) and limited
printer, equipped with a 0.40 mm diameter elongation at break (<10%) [12]. It is
nozzle, was used to fabricate the capsules. The widely used in the packaging and
highest achievable print resolution by the medical industry and it is one of the
printer is 60 µm and it can reach a maximum most common material for 3D printing.
print speed of 80 mm/s. Figure 1 shows the • Polyethylene terephthalate (PET) is the
printer and its main components. most common thermoplastic polymer
resin of the polyester family, largely
used to produce beverage and food
packaging, clothing and engineering
resins. As a semi-crystalline polymer,
PET properties depend heavily on
processing conditions and its physical
and chemical properties are a
combination of different variables such
as composition, crystallinity,
orientation, etc. [26]. Due to the urgent
issues connected to PET waste disposal,
the use of recycled PET filaments to
produce 3D printed capsules for self-
healing applications could represent a
possible disposal solution.
• Polyethylene terephthalate glycol-
Figure 1: Prusa I3 Hephestos printer. modified (PETG) is a copolyester of the
pure PET prepared by partially
Nowadays, several materials are replacing the ethylene glycol (EG)
commercially available in the form of filament groups with 1,4-cyclohexanedimethanol
spools for FDM. In order to be eligible for (CHDM) groups in order to modify the
being used as autonomous self-healing polymer properties, such as reducing its
encapsulating material, several aspects should melting temperature.
be considered, such as their melting
• Poly(methyl methacrylate) (PMMA),
temperature (Tm), their glass transition
the synthetic polymer of methyl
temperature (Tg) and their chemical resistance
methacrylate, is a transparent
both to the coating and healing agent
thermoplastic and is a strong, tough and
substances and towards the alkaline
lightweight material, commonly used
environment of the concrete. Tg of the selected
due to its properties, its ease of handling
material is an important feature as a relatively
and processing and its low cost. This
brittle behavior is required to trigger the
material was selected especially due to
encapsulated healing agent release, namely
its brittle behavior, for which it was also
low elongation at rupture and relatively high
compared with concrete [27].
tensile strength. Finally, low cost materials are
preferable in order to produce a cost-effective Preliminary testing on samples of the
self-healing system. Consequently, four different filaments were performed in order to
different materials were selected to produce assess the accuracy of the technical sheets’
the capsules: declared values, namely Fourier-transform
infrared spectroscopy (FTIR) and tensile
• Poly(lactic acid) (PLA) is derived from testing. The testing confirmed the purity of the

3
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

filaments’ chemical compounds and the cementitious capsules. For the same purpose,
mechanical properties typical of the selected length was set to 40 mm, obtaining a
materials. comparable volume capacity of healing agent.
The models for printing the capsules were A positive aspect offered by manufacturing
realized using the open-source Ultimaker Cura the capsules using FDM was the possibility of
software. The main printing parameters used producing the sealing of the capsule ends and
for each material are listed in Table 1. filling the capsules during the printing
execution. The seal thickness was set to 3 mm.
Table 1: Configurations of printing parameters (when The printing procedures were carried out as
configurations differ among polymers, they are reported
in the following order: PLA / PET / PETG / PMMA)
follows:
1. the bottom capsule end was printed,
Printing parameters Configurations
with a thickness of 3 mm;
Layer height (mm) 0.06 2. the shell of the capsule was printed until
Printing temperature (°C) 211 / 235 / 235 / 240 reaching 40 mm height, then printing
Bed temperature (°C) 60 / 30 / 60 / 90 was arrested;
3. while printing procedure was on stand-
Filament diameter (mm) 2.85
by, the capsule was filled using a
Nozzle diameter (mm) 0.40 syringe;
Printing speed (mm/s) 60 4. printing was resumed and the final
Travel speed (mm/s) 120 capsule end was sealed, with a thickness
of 3 mm.
Two models of capsules were designed and After preliminary testing aimed to assess
printed. The first one (Figure 2), hereafter the water tightness and the ability of the
referred to as Model 1, were designed for the capsules to correctly store and protect the
sake of comparison after the cementitious healing agents (see Section 2.2), it was
capsules developed and thoroughly considered necessary to add an external
investigated by the authors [3,11,25]. The shell coating. The coating was made by applying an
thickness was set to 0.2 mm in order to obtain external layer of a two-component liquid
a good compromise between ease of printing, epoxy resin (Plastigel, API SpA). Before
holding capacity and to guarantee the relative coating drying, the capsules were rolled into
brittle behavior that is essential for triggering sand in order to improve the adhesion with the
the healing agent release. Consequently, the surrounding mortar matrix. Figure 2 shows the
external diameter of the capsule was set to manufacturing of the Model 1 capsules.
7.9 mm, in order to obtain a comparable However, the preliminary testing showed
internal diameter with the aforementioned that, despite their simple manufacturing and

Figure 2: Model 1 capsules: a) printing of the first capsule end sealing and its shell; b) print is interrupted,
capsule is filled and then print is resumed in order to realize the end sealing; c) 3D printed capsules, filled
and with both ends sealed (from left to right PMMA, PET, PLA, PETG); d) coating application and e) capsule
are rolled in the sand in order to improve their bonding with the cementitious matrix.

4
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

Figure 3: Model 2 capsules: a) 3D printed capsules, without capsule ends sealing (from left to right
PMMA, PLA, PETG, PET); b) epoxy primer applied by immersion; c) top view of the capsule after epoxy
film applied by immersion; d) lateral view of the capsule with dried epoxy coating and e) finished
capsule with sealed ends, filled and rolled in the sand while epoxy was still fresh.

filling procedures and the ability to correctly SpA), one before and one after the injection of
encapsulate and release agents with a slow the healing agent. A last layer of Plastigel was
reacting mechanism, the Model 1 capsules, applied and, while it was still fresh, the
even after coating, were not effective in capsules were rolled into sand to improve the
encapsulating highly reactive healing agents bonding with the surrounding mortar matrix.
that are triggered for example by the contact Figure 3 shows the manufacturing of the
with moisture, which require an almost perfect Model 2 capsules.
water/air tight condition in order to be Table 2 summarizes the specific features of
correctly encapsulated and protected by the each 3D printed capsule model configuration.
harsh and humid cementitious environment.
Therefore, a second model was designed, Table 2: Configurations of 3D printed capsules models
hereafter referred to as Model 2 (Figure 3). Features Model 1 Model 2
In order to provide a better protection of the
External diameter (mm) 7.9 8.5
healing agents, it was deemed necessary to
apply a double coating, applied by immersion Length (mm) 40 50
to have a complete coverage of the shell. Thus, Shell thickness (mm) 0.2 0.2
to allow a correct application by immersion, Type of sealing 3D printed Stucco K
only the lateral shell was printed, without end Primer AQ
sealings that were realized later. To take into Plastigel
Type of coating Plastigel
account the loss of internal volume caused by (ext.)
(int.+ext.)
the internal coating and the application of the Finishing Sand Sand
sealings, the capsule dimensions were slightly
modified in order to obtain a comparable 2.2 Water tightness and shell continuity
internal volume (i.e. volume of healing agent
stored) with the Model 1. Consequently, the An aspect of paramount importance in the
external diameter of the capsule was set to development of an effective encapsulated self-
8.5 mm and the length to 50 mm. healing is the ability of the capsules to protect
After printing, the first coating was applied the healing agent from the cementitious
by immerging the tubes in an epoxy-based environment and not allowing loss of agent
two-component primer (Primer AQ, API from the capsule before rupture. In order to
SpA). This operation was repeated two times test this ability and assess the need of an
within 24 hours. Then, the tubes were further external coating, Model 1 capsules were filled
dip coated with a two-component liquid epoxy with water during printing, as explained in
film (Plastigel, API SpA). The capsule ends Section 2.1. For each material used for
were sealed with an epoxy-based two- printing, two capsules were filled and not
component thixotropic plaster (Stucco K, API coated (as in Figure 2.c) and two capsules
were filled and coated (as in Figure 2.e), for a

5
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

total of 16 capsules. The capsules weight was Model 1 capsules with the external coating, as
then monitored in time after exposition to air, in Section 2.1.
in order to assess the mass of water loss. The After testing the ability of the capsules of
permeation of water through the capsules correctly retaining the healing agents inside
polymeric shell and coating could be described the capsules, it was investigated their ability of
as a Fickian diffusion process, hence linearly protecting them by the harsh cementitious
dependent with the square root of time [28]. environment. Therefore, Model 1 capsules
Consequently, it is possible to define: produced with each plastic filament were filled
either with the liquid sodium silicate or the
Mw = p∙t0.5 (1)
moisture-reactive polyurethane precursor that
where Mw (mg) is the mass loss in time would have been used as healing agent for
attributed to the water loss, p (mg/h0.5) is the testing the mechanical recovery after cracking
water permeation rate and t (h) is the time of (see Section 3.1) and embedded inside cement
permeation. Hence, it is possible to assess the mortar prisms during casting. After 28 days of
ability of protecting the healing agent by curing in water, the specimens were split in
examining the reduction of the water half to check the viability of the healing
permeation rate. Table 3 shows the water agents. Regardless of the plastic used for
permeation rate for each tested capsule. printing, the sodium silicate was still liquid
and available for enabling the self-healing
Table 3: Water permeation rate of the Model 1
effect, while the polyurethane was already
capsules, either with or without coating
hardened inside the capsules (Figure 4).
Water permeation rate (mg/h0.5)
Model 1 Model 1
Material #
(no coating) (with coating)
1 5.13 11.38
PLA
2 5.32 9.62
1 23.07 3.28
PET
2 7.07 4.84
1 1.30 3.99
PETG
2 23.13 3.25
1 4.43 3.44
PMMA
2 5.66 3.05

Except in the case of PLA, the application Figure 4: Polyurethane hardened inside Model 1
of the external coating resulted in an overall capsules embedded in the cement mortar matrix.
reduction of the water absorption rate and an
increased repeatability in the results, as in the In light of these findings, a more protective
case of the PET and PETG where the results coating system was deemed necessary.
were significantly different and most likely Consequently, Model 2 capsules were
governed by the printing and layer adhesion designed in order to face these further
defects. Consequently, even if the small requirements. The application of the double
number of specimens does not allow a protective coating on the capsule shell before
statistically significant analysis of the results, filling it, allows to reduce the possibility of
this test provides a good indication of the moisture ingress during mortar samples curing.
importance of the external coating in providing The retaining ability of the Model 2
a good protection of the healing agent and capsules was tested in the same way as
compensating the defects caused by the Model 1, by monitoring the encapsulated
production of the capsules with FDM such as water loss over time. For each material used
the interlayer adhesion. As a result of these for printing, two Model 2 capsules (as in
findings, it was decided to always produce the Figure 3.e) were produced and filled, for a

6
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

Figure 5: Water permeation rate of Model 1 capsules (with or without coating) and Model 2 capsules.

total of 8 capsules. between the printed filaments. On the contrary,

Figure 5 shows the comparison between the it is possible to observe the effectiveness of the
water permeation rates obtained for the double coating, which allows to obtain a shell
Model 1 capsules (with or without coating) free of defects that could cause leakage of the
and the Model 2 capsules, in which it is healing agents and vulnerability against the
possible to see the water loss reduction, hence surrounding cementitious matrix.
a better retaining of the healing agent, when
Model 2 is used. These findings about the
better protection provided by the Model 2
were confirmed by filling these capsules with
the polyurethane precursor and embedding it
in the mortar matrix. Unlike Model 1, the
Model 2 capsules were able to correctly
protect the reactive healing agent during
mortar samples curing and releasing it upon
cracking.
The shell continuity of the capsules plays a
leading role on the tightness of the capsules. In
order to investigate and compare the shell
continuity of the capsules printed with the four
different polymers, without coating and with
the double coating used in the Model 2
configuration, they were sputter coated with
gold and their surfaces were analyzed using a
Field Emission Scanning Electron Microscope
(FE-SEM Hitachi S-4000). Figure 6 shows the
60x magnification images of the capsules
printed with each material, with and without
coating. It is possible to observe how the
printed polymeric capsules in the absence of
coatings are not able to effectively isolate the
healing agent from the external environment, Figure 6: FE-SEM images of the capsules printed
as they present defects due to the cohesion with each material, with and without coatings.

7
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

3 SELF-HEALING EFFECT developed properties and allow a high degree

of hydration of the matrix, hence limiting the
3.1 Self-healing mortar specimens effect of autogenous healing due to unhydrated
In order to realize an autonomous self- cement particle and thus being able to attribute
healing system, the 3D printed capsules were the healing efficiency to the contribution of the
filled either with a liquid sodium silicate capsule system.
solution (Na2O 10.6 wt%, SiO2 26.5 wt%, and
3.2 Self-healing effect evaluation
H2O 62.9 wt%, provided by Sigma Aldrich) or
an expansive polyurethane (PU) resin At an age of 28 days from casting, the
(CarboStop U, provided by Orica). These specimens were pre-cracked via a crack-width
healing agents were selected due to the good controlled three-point bending test with a span
results obtained in previous studies [3,11,25]. of 10 cm, using a 100 kN closed-loop servo-
Since sodium silicate encapsulation does not controlled MTS hydraulic press. The target
require high protection standards due to its maximum crack mouth opening displacement
reaction mechanism, it was encapsulated using (CMOD) during loading was set at 800 µm, in
Model 1 capsules. Conversely, the PU order to test the ability of the system to heal
precursor is very moisture-reactive and require large cracks and in order to reduce the
to be encapsulated with a sufficiently water/air influence of the autogenous healing. A U-
tight system, hence Model 2 was selected for shaped notch was created by saw cutting a day
its encapsulation. before the pre-cracking procedure, measuring
Mortar prisms (40 mm x 40 mm x 160 mm) 4 mm in width and 5 mm in height. Figure 7
were made by using ordinary Portland cement shows the setup for the pre-cracking phase.
CEM I 52.5 N, tap water (water to cement
ratio of 0.5) and normalized sand (sand to
cement ratio of 3), in accordance with the
standard UNI EN 196-1. Mortar is often
chosen instead of concrete as cementitious
material reference for testing the efficiency of
the self-healing systems. In addition to the fact
that mortar is more homogenous than concrete,
the main reason for using mortar is the often
high production cost of the self-healing
systems in the prototype phase, thus it is very
useful to do a fast screening of their
effectiveness in mortar before upscaling them
for the use in concrete [29].
The capsules were manually embedded in
the middle of the mortar specimens during Figure 7: Specimen during the pre-cracking phase.
casting. For each plastic filament, two
specimens containing one Model 1 capsule During the pre-cracking phase, it was
filled with sodium silicate and two specimens possible to observe the leakage of the healing
containing one Model 2 capsule filled with PU agent up to the external surfaces of the
resin were produced for testing the correct specimens, which denoted the correct breakage
capsule breakage and consequent mechanical of the capsules, thus the triggering of the
recovery, for a total of 16 specimens. autonomous healing mechanism. The leakage
After 24 hours covered with plastic foils, of the sodium silicate was visible due to the
the specimens were demolded and then stored wetting of the crack sides, while for the PU
in water for 28 days. This curing condition resin it was possible to clearly observe the
was chosen in order to obtain sufficiently expansion of the yellow closed-cell foam.

8
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

The specimens were stored in air in the during the pre-cracking stage and Pu is the
laboratory environment for a time well beyond residual load obtained at the moment of
the curing time of both the healing agents, so unloading preceding the re-loading stage.
as to allow the full curing of the healing agents Table 4 summarizes the LRI obtained for each
and avoid that water could promote tested specimen.
autogenous healing or cause the foaming of
the unpolymerized polyurethane that could Table 4: Load Recovery Indexes of the self-healing
system realized with 3D printed capsules
have remained in the capsules. In order to
maintain the condition as close as possible to Load Recovery Index (%)
the crack formation phase, thus better Sodium silicate PU resin
simulating the real healing condition of a Material #
(Model 1) (Model 2)
damaged structure, the specimens were placed 1 33 22
PLA
on two supports with a span of 10 cm with the 2 57 49
crack mouth pointing downward. 1 27 36
PET
After complete curing of the healing agent 2 18 65
and before evaluating the strength regain, a 1 31 35
PETG
visual inspection of the crack filling was 2 45 12
1 16 46
performed. While the PU resin filling was PMMA
2 24 18
visible to the naked eyes due to its yellow
color and the presence of foam above the
On average, both the healing agents showed
crack, in order to observe the sodium silicate, a
a good recovery of mechanical properties, with
40x magnification microscope was used.
an average LRI of (31 ± 14) % for the sodium
Figure 8 shows the crack on the lateral face of
silicate and of (35 ± 18) % for the PU resin.
one specimen, filled with the sodium silicate.
The best results for the sodium silicate were
obtained when PLA was used as printing
material, while in the case of the PU resin the
best printing material was PET.
After completion of the mechanical
recovery testing, the samples were split at the
location of the crack in order to visually
evaluate the spreading of the healing agent on
the crack faces, in order to gain a qualitative
information about the filling of the crack.
Figure 9 shows an example of the spreading of
the sodium silicate and the PU resin that
highlights in both cases a large covered area,
hence a good bridging of the crack faces.
Figure 8: Filling of the crack by the sodium silicate
(lateral face of the specimen).

The strength regain was evaluated, after re-

loading the healed specimens via a crack-
width controlled three-point bending test as in
the crack creation phase, through a load
recovery index (LRI) [2,3,25]. The load
recovery index was defined as:
LRI(%) = (Pr - Pu)/(Pp - Pu)∙100 (2) Figure 9: Spreading of the liquid sodium silicate
(left) and the PU foam (right) on the crack faces. The
where Pr is the peak load obtained during the area covered by the sodium silicate was colored in
re-loading stage, Pp is the peak load reached cyan in order to highlight its spreading region.

9
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

4 CONCLUSIONS AND FUTURE encapsulated system, by using simple tubular

PERSPECTIVES capsules for the sake of comparison of
In this work, Fusion Deposition Modelling extensively investigated system. Further
was used to produce novel macro-capsules to improvements can be envisaged through the
be used in self-healing cementitious systems. optimization of the capsule shape, for which
Four different types of thermoplastic filaments the 3D printing technology offers virtually
were selected, based on characteristics that unlimited tailoring possibilities.
make them suitable for the requirements of a
functioning autonomous self-healing system. ACKNOWLEDGEMENT
Two types of tubular capsules were The authors would like to acknowledge the
produced, namely Model 1 and Model 2 contribution of the COST Action CA15202
capsules, with slightly different shapes and (http://www.sarcos.enq.cam.ac.uk). The
different sealing and coating configurations. financial support from the Interdepartmental
The two models were characterized in terms of Center “Responsible Risk Resilience Centre”
water tightness and shell continuity, which (R3C) in funding part of the equipment is
proved them effective to be used to gratefully acknowledged. Buzzi Unicem SpA
encapsulate liquid healing agents and is thanked for providing cement. Mr. Stefano
highlighted their unique features. The Model 1 Airoldi from Minova CarboTech GmbH is also
capsule presents simple manufacturing, filling thanked for Orica CarboStop U resin
and coating procedures and was proved providing. Dr. Beniamino Magnaghi from API
suitable for the encapsulation of healing agents S.p.A. is greatly thanked for waterproof epoxy
that does not require a perfect air and water resins providing (Primer AQ, Plastigel and
tight encapsulation system, such as liquid Stucco K). The technical support of Mr.
silicates. However, it is not suitable for Vincenzo Di Vasto from Department of
reactive healing agents, which require on the Structural, Geotechnical and Building
contrary a better protection from the humid Engineering of Politecnico di Torino is
cementitious environment, such as gratefully acknowledged.
polyurethane precursors. To overcome this
limit, the Model 2 capsule was designed in REFERENCES
order to guarantee high performances in terms [1] De Belie, N., Gruyaert, E., Al-Tabbaa, A.,
of tightness at the cost of a slightly more Antonaci, P., Baera, C., Bajare, D., et al.,
complex sealing, coating and filling procedure. 2018. A Review of Self-Healing Concrete
Both models, when filled with the healing for Damage Management of Structures.
agents suitable for their characteristic features, Adv. Mater. Interfaces. 5:1800074.
showed satisfactory results in terms of
mechanical recovery and spreading of the [2] Ferrara, L., Van Mullem, T., Alonso,
healing agent on the crack faces. This last M.C., Antonaci, P., Borg, R.P., Cuenca,
feature suggests a good self-sealing ability, E., et al., 2018. Experimental
which is very important for preventing the characterization of the self-healing
ingress of harmful substances inside the capacity of cement based materials and its
fractured matrix. This aspect will be further effects on the material performance: A
investigated through tests aimed to assess self- state of the art report by COST Action
sealing and durability-related features, such as SARCOS WG2. Constr. Build. Mater.
water permeability and absorption tests. 167:115-42.
These positive results show a good
potential of the proposed capsule-based system [3] Formia, A., Irico, S., Bertola, F.,
for structural applications. This work served as Canonico, F., Antonaci, P., Pugno, N.M.,
proof of concept of the use of Additive et al., 2016. Experimental analysis of self-
Manufacturing to produce self-healing healing cement-based materials

10
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

incorporating extruded cementitious [11] Formia, A., Terranova, S., Antonaci, P.,
hollow tubes. J. Intell. Mater. Syst. Struct. Pugno, N.M. and Tulliani J.-M., 2015.
27:2633-52. Setup of extruded cementitious hollow
tubes as containing/releasing devices in
[4] Yang, Z., Hollar, J., He, X. and Shi, X., self-healing systems. Materials (Basel)
2011. A self-healing cementitious 8:1897-923.
composite using oil core/silica gel shell
microcapsules. Cem. Concr. Compos. [12] Hilloulin, B., Van Tittelboom, K.,
33:506-12. Gruyaert, E., De Belie, N. and Loukili, A.,
2015. Design of polymeric capsules for
[5] Kanellopoulos, A., Giannaros, P. and Al- self-healing concrete. Cem. Concr.
Tabbaa, A., 2016. The effect of varying Compos. 55:298-307.
volume fraction of microcapsules on
fresh, mechanical and self-healing [13] Gruyaert, E., Van Tittelboom, K., Sucaet,
properties of mortars. Constr. Build. J., Anrijs, J., Van Vlierberghe, S.,
Mater. 122:577-93. Dubruel, P., et al., 2016. Capsules with
evolving brittleness to resist the
[6] Wang, J.Y., Soens, H., Verstraete, W. and preparation of self-healing concrete.
De Belie, N., 2014. Self-healing concrete Mater. Construcción 66:e092.
by use of microencapsulated bacterial
spores. Cem. Concr. Res. 56:139-52. [14] Šavija, B., Feiteira, J., Araújo, M.,
Chatrabhuti, S., Raquez, J.-M., Van
[7] Van Tittelboom, K., De Belie, N., Van Tittelboom, K., et al., 2016. Simulation-
Loo, D. and Jacobs, P., 2011. Self-healing Aided Design of Tubular Polymeric
efficiency of cementitious materials Capsules for Self-Healing Concrete.
containing tubular capsules filled with Materials (Basel) 10:10.
healing agent. Cem. Concr. Compos.
33:497-505. [15] Araújo, M., Chatrabhuti, S., Gurdebeke,
S., Alderete, N., Van Tittelboom, K.,
[8] Kanellopoulos, A., Qureshi, T.S. and Al- Raquez, J.-M., et al., 2018. Poly(methyl
Tabbaa, A., 2015. Glass encapsulated methacrylate) capsules as an alternative to
minerals for self-healing in cement based the ‘’proof-of-concept’’ glass capsules
composites. Constr. Build. Mater. 98:780- used in self-healing concrete. Cem. Concr.
91. Compos. 89:260-71.

[9] Qureshi, T.S., Kanellopoulos, A. and Al- [16] Bückmann, T., Stenger, N., Kadic, M.,
Tabbaa, A., 2016. Encapsulation of Kaschke, J., Frölich, A., Kennerknecht,
expansive powder minerals within a T., et al., 2012. Tailored 3D Mechanical
concentric glass capsule system for self- Metamaterials Made by Dip-in Direct-
healing concrete. Constr. Build. Mater. Laser-Writing Optical Lithography. Adv.
121:629-43. Mater. 24:2710-4.

[10] Buysse, C., Kovalevsky, A., Snijkers, F., [17] Taheri Andani, M., Karamooz-Ravari,
Buekenhoudt, A., Mullens, S., Luyten, J., M.R., Mirzaeifar, R. and Ni, J., 2018.
et al., 2010. Fabrication and oxygen Micromechanics modeling of metallic
permeability of gastight, macrovoid-free alloys 3D printed by selective laser
Ba0.5Sr0.5Co0.8Fe0.2O3−δ capillaries melting. Mater. Des. 137:204-13.
for high temperature gas separation. J.
Memb. Sci. 359:86-92. [18] Xu, Y., Zhang, H., Šavija, B., Chaves
Figueiredo, S. and Schlangen, E. 2019.

11
 Giovanni Anglani, Paola Antonaci, Susana I. Carillo Gonzales, Giorgia Paganelli, Jean-Marc Tulliani

Deformation and fracture of 3D printed study of PP/PET composites: Factorial

disordered lattice materials: Experiments design, mechanical and thermal
and modeling. Mater. Des. 162:143-53. properties. Polym. Test. 56:167-73.

[19] Chaves Figueiredo, S., Romero [27] Melin, S. 1989. Why are crack paths in
Rodríguez, C., Ahmed, Z.Y., Bos, D.H., concrete and mortar different from those
Xu, Y., Salet, T.M., et al., 2019. An in PMMA? Mater. Struct. 22:23-7.
approach to develop printable strain
hardening cementitious composites. [28] Thomas, S., Wilson, R., Anil Kumar, S.
Mater. Des. 169:107651. and George, S.C., 2018. Transport
Properties of Polymeric Membranes.
[20] Van Der Putten, J., De Schutter, G. and Elsevier.
Van Tittelboom, K., 2018. The Effect of
Print Parameters on the (Micro)structure [29] Van Mullem, T., Gruyaert, E., Debbaut,
of 3D Printed Cementitious Materials. In B., Caspeele, R. and De Belie, N., 2019.
Wangler and Flatt (eds), First RILEM Int. Novel active crack width control
Conf. on Concr. Digit. Fabr. - Digit. technique to reduce the variation on water
Concr. 2018, September 10-12, 2018, permeability results for self-healing
Zürich, Switzerland: Springer; pp.234-44. concrete. Constr. Build. Mater. 203:541-
51.
[21] Xia, M. and Sanjayan, J., 2016. Method of
formulating geopolymer for 3D printing
for construction applications. Mater. Des.
110:382-90.

[22] Huang, S.H., Liu, P., Mokasdar, A. and

Hou, L. 2013. Additive manufacturing and
its societal impact: a literature review. Int.
J. Adv. Manuf. Technol. 67:1191-203.

[23] Crump, S.S., 1992. Apparatus and method

for creating three-dimensional objects. US
Patent 5,121,329.

[24] Lanzotti, A., Grasso, M., Staiano, G. and

Martorelli, M. 2008. Effect of processing
conditions on the bonding quality of FDM
polymer filaments. Rapid. Prototyp. J.
21:252-64.

[25] Anglani, G., Antonaci, P., Idone, G. and

Tulliani, J.-M., 2018. Self-healing of
cementitious materials via embedded
macro-capsules. In Ye et al (eds), Proc. of
the 4th Inter. Conf. on Serv. Life Des. for
Infrastructures (SLD4), August 27-30,
2018, Delft, the Netherlands: RILEM
Publications S.A.R.L.; pp.385-8.

[26] Nonato, R.C. and Bonse, B.C., 2016. A

12