See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/305729134

Method of formulating geopolymer for 3D printing for construction

applications

Article  in  Materials & design · July 2016

DOI: 10.1016/j.matdes.2016.07.136

CITATIONS READS

57 2,373

2 authors:

Ming Xia Jay G. Sanjayan

Swinburne University of Technology Swinburne University of Technology
28 PUBLICATIONS   168 CITATIONS    289 PUBLICATIONS   7,011 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

3D printing of structures using fibre-reinforced Portland cement/geopolymer concrete View project

Retrofitting Public Buildings for Energy and Water Efficiency View project

All content following this page was uploaded by Ming Xia on 29 March 2018.

The user has requested enhancement of the downloaded file.

 Materials and Design 110 (2016) 382–390

Contents lists available at ScienceDirect

Materials and Design

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

Method of formulating geopolymer for 3D printing for

construction applications
Ming Xia, Jay Sanjayan ⁎
Department of Civil and Construction Engineering, Centre for Sustainable Infrastructure, Faculty of Science, Engineering and Technology, Swinburne University of Technology, Hawthorn, Victoria
3122, Australia

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• A new methodology is presented for

formulating geopolymer for 3D printing
process.
• The study identified the essential char-
acteristics for printability and dimen-
sional accuracy.
• Identified parameters for assessing the
quality of the geopolymer printed
shapes.
• A new post-processing method for
geopolymer to gain strength for con-
struction application.

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

Article history: An innovative methodology is presented for formulating geopolymer-based material for the requirements and
Received 8 July 2016 demands of commercially available powder-based 3D printers. Different key powder parameters such as particle
Received in revised form 15 July 2016 size distribution, powder bed surface quality, powder true/bulk densities, powder bed porosity, and binder drop-
Accepted 28 July 2016
let penetration behavior were used to quantitatively evaluate the printability of prepared geopolymer-based ma-
Available online 29 July 2016
terial. Results indicated the prepared geopolymer-based material achieved sufficient printability to be used in a
Keywords:
powder-based 3D printer. 3D printed specimens from the geopolymer were assessed for printing accuracy, ap-
Additive manufacturing parent porosity and mechanical property. All printed cubic specimens exhibited an anisotropic phenomenon in
Powder-based 3D printing dimensional accuracies and mechanical properties. Before further post-processing, the printed cubic structure
Construction gained compressive strength of up to 0.9 MPa with an acceptable low dimensional expansion of b4%. Post pro-
Geopolymer cessing by immersion in saturated anhydrous sodium metasilicate solution at 60 °C gained compressive strength
of up to 16.5 MPa.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction processes build a finished structure in successive layers with less

Additive manufacturing (AM) is a group of emerging techniques for
fabricating three-dimensional (3D) structures directly from a digital
model. Unlike conventional industrial manufacturing processes, AM
⁎ Corresponding author.
E-mail address: jsanjayan@swin.edu.au (J. Sanjayan).
waste material. Since its emergence 30 years ago, AM technologies
have been successfully applied in a wide range of industries including
aerospace, automotive, biomedical, consumer and food [1].
AM is also now gaining popularity in construction industry. When
compared with conventional construction processes, the application of
AM techniques in construction industry may (1) reduce the labor re-
quirements which would result in a decreased construction cost and

http://dx.doi.org/10.1016/j.matdes.2016.07.136
0264-1275/© 2016 Elsevier Ltd. All rights reserved.
 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
Fig. 1. Schematic illustrations of the 3DP process: (a) 3DP inkjet printing system, (b) Enlargement of the area in red rectangle: powder/binder interaction between adjacent layers.
an increased level of safety, (2) reduce on-site construction time by op-
erating at a constant rate, (3) minimize the chance of errors by highly
precise material deposition, and (4) increase architectural freedom
which would enable more sophisticated designs for structural and aes-
thetic purposes [2,3].
In the last few years, a variety of technologies has been developed to
adopt AM in construction industry. These techniques can be divided
into two forms.
1.1. Extrusion printing
The extrusion printing technique is analogous to the Fused Deposi-
tion Modelling (FDM) method by extruding cementitious material
from a nozzle mounted on a gantry to print a structure layer by layer.
Examples such as Contour Crafting, developed by Khoshnevis [4,5]
and Concrete Printing, designed by Lim et al. [6,7].
1.2. Powder printing
The powder printing technique, also named powder-based three di-
mensional printing (3DP), is another typical AM process that creates ac-
curate structures with complex geometries by depositing binder liquid
(or "Ink") selectively into to powder bed to bind powder where it im-
pacts the bed [8]. Two typical examples are the D-shape technique de-
veloped by Cesaretti et al. [9] and Emerging Objects [10].
Both of the two techniques are designed for construction purpose
and have many similarities. However, each technique has distinct fea-
tures in construction industry. The extrusion printing technique has
been aimed at on-site construction applications such as large-scale
building components with complex geometries, while the powder
printing technique is an off-site process designed to manufacturing pre-
cast components. It is the authors' belief that powder-based 3DP tech-
nique is highly suitable for small-scale building components such as
Table 1
The chemical composition of slag (wt%).
Chemical Al2O3 SiO2 CaO
Component 12.37 32.76 44.64
a
Loss on ignition (unburnt carbon content).
383
panels, permanent formworks and interior structures that then can be
assembled on site.
The work presented in this paper focuses on powder-based 3DP
technique as this method is capable of producing building components
with fine details and intricate shapes. There is a demand in construction
industry for such components which can only be made with expensive
formwork systems with the currently available construction systems.
Powder-based 3DP technique needs to be developed so that they can
produce robust and durable components at a reasonable speed to satisfy
this industrial demand. Hence the urgency for research in this area.
A schematic of the powder-based 3DP process is illustrated in Fig. 1.
At the start, a roller, mounted together with a print head, spreads a layer
of powder to cover the base of the build plate. The base layer is about
3 mm in thickness. Then, according to the layer thickness setting of
the 3D printer, a thin layer of powder (approximately 0.1 mm) is spread
and smoothed by the roller over the powder bed surface. Once a layer is
completed, the binder solution is delivered from binder feeder to the
print head and jetted by nozzles mounted in the print head. The mech-
anism of controlling the binder solution is a non-continuous approach
called drop-on-demand (DoD) technique [11] which has been widely
used in contemporary desktop printing systems. In the print head, the
binder solution is pushed out through nozzles either thermally by a
thermal bubble or mechanically by a piezoelectric actuator. Then binder
droplets are formed and are selectively applied on powder layer, caus-
ing powder particles to bind each other (Fig. 1b). Repeating the de-
scribed steps, the built part is completed and removed after a
particular drying time and unbound powder is removed by using an
air blower.
Powder-based 3DP is a versatile AM technique characterized by a
wide range of potentially suitable materials in powder form, including
ceramics, polymers, metals and composites [12,13]. Powder-based
3DP technique has a great potential to make a significant and positive
contribution to the construction industry. There are mainly two
Fe2O3 K2O MgO Na2O TiO2 SO3 L.O.I.a
0.54 0.33 5.15 0.22 0.51 4.26 0.09
384 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
Fig. 2. The process of preparing Geopolymer powder.
limitations to prevent this technique performing at its maximum poten-
tial for application in construction industry: the printer and the printing
material.
As most powder-based 3D printers are in small or medium sizes,
consideration needs to be given for the feasibility of making large size
printers for the large-scale manufacturing requirement in construction
industry. In recent years, several large-scale powder-based 3D printers
have been developed. For example, the one designed by D-shape has a
dimension of 15 × 4 × 1 m, and the VX4000 printer from Voxeljet
which is capable of making sand casts as large as 4 × 2 × 1 m.
In addition to the printer size, printing speed is another important
consideration. In powder-based 3DP process, several factors will affect
printing speed, such as the movement on X, Y and Z axes of the gantry
system, velocity of the binder jetting in the print head and velocity of
the powder spreading. As shown in Fig. 1, the binder jetting process
and powder spreading process are not in parallel and this nonsynchro-
nous process has become the most critical issue that affects the printing
speed in powder-based 3DP process. Designing a system which can
make these two processes operate in parallel will improve the printing
speed. Future research works in this area need to focus on such im-
provements to increase printing efficiency.
Another limitation lies in the low availability of printing materials.
Until recently, only a small range of cement-based materials have
been explored such as rapid hardening Portland cement (RHPC) [14],
calcium aluminate cement (CAC) [15], magnesium oxychloride cement
(also known as Sorel cement) [9], fiber reinforced cement polymer [10]
and ultra-high performance concrete (UHPC) [16]. As such, it is urgently
needed to expand the current severely limited scope of materials that
can be used in a powder-based 3D printer.
Geopolymer [17], also known as an inorganic polymer, is an alterna-
tive material that acts as the binding agent in concrete. Geopolymer is a
Fig. 3. Schematic illustrations of drop penetration behavior test: (a) Drop penetration apparatus, (b) Drop penetration parameters.
type of alkali aluminosilicate cement which can have superior mechan-
ical, chemical and thermal properties compared to Ordinary Portland
Cement (OPC), and with significantly lower CO2 emission. Geopolymers
are synthesized by activation of an aluminosilicate source (metakaolin,
fly ash and slag) with alkaline activators.
In recent years, geopolymer has attracted considerable attention be-
cause of its high compressive strength, excellent resistance to sulfate at-
tack, good acid resistance, low creep and minimal drying shrinkage
behavior. The utilization of industrial by-products such as slag and fly
ash in geopolymer is considered particular beneficial, as the disposition
of industrial by-products has always been a global issue.
However, the applications of geopolymer have been restricted to
small areas using conventional manufacturing techniques. A novel
method of manufacturing geopolymer component, such as powder-
based 3DP technique, should be taken into consideration to expand
the application of this eco-friendly material.
This paper presents an innovative methodology to adopt
geopolymer-based material for the requirement and demand of com-
mercially available powder-based 3D printers, intended for the expan-
sion of the limited scope of materials that can be used in a powder-
based 3D printer in construction industry. In this study, printable
geopolymer-based material was prepared by blending slag, anhydrous
sodium metasilicate and fine sand. Different key parameters such as
particle size distribution, powder bed surface quality, powder true/
bulk densities, powder bed porosity, and binder droplet penetration be-
havior were used to evaluate the printability of geopolymer-based ma-
terial quantitatively and compare with the commercially available
plaster-based material. The printing accuracy, apparent porosity and
mechanical property of the printed structures were investigated and
benchmarked against with those printed from a commercially available
plaster-based powder.
 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
Fig. 4. CAD-designed 3D structure to be printed (Dimensions in mm).
2. Materials and methods
2.1. Materials
2.1.1. Powders
Two powder systems, slag-based geopolymer powder (marked as
GP) and commercially available plaster-based powder (marked as ZP),
were used in this study. A mixture of slag (Building Products Supplies
Pty Ltd., Australia), silicate-based activator (Redox Pty Ltd., Australia)
and fine sand (TGS Industrial Sands Pty. Ltd., Australia) was used to pre-
pare the printable geopolymer-based material. The chemical composi-
tion of the slag, determined by X-ray fluorescence (XRF) analysis, is
shown in Table 1. The total does not sum up to 100% due to rounding off.
The silicate-based activator used in this study was composed of
beaded anhydrous sodium metasilicate with a chemical composition
of 50.66 wt% Na2O, 47.00 wt% SiO2 and 2.34 wt% H2O. TGS Industrial
70 Grade Sand was classified as semi-sharp sand with a median size of
184 μm. The incorporation of TGS sand was aimed to function as an
inert filler in the GP powder to improve its printability.
The GP powder was prepared as follows: Firstly, anhydrous sodium
metasilicate beads were dry milled for 5 min in a planetary ball mill
with ceramic balls. Then slag, ball milled anhydrous sodium metasilicate
and fine sand were dry mixed in a Hobart mixer until a homogeneous
mixture was obtained. The process of preparing GP powder is shown
in Fig. 2.
For comparison, a commercial plaster-based powder (Zp® 150, Z-
Corp, USA) was used in this study without further treatment. The ZP
powder was supplied by the same company manufacturing the 3D
printer used in this study. The composition of the ZP powder was mainly
calcium sulfate hemihydrate (CaSO4·1/2H2O) by X-ray diffraction
(XRD) analysis [18] and energy dispersive spectroscopy (EDS) analysis
[19].
2.1.2. Binder
An aqueous solvent (Zb® 63, Z-Corp, USA) was used as binder during
the 3DP process. The binder was an aqueous commercial clear solution
with the viscosity similar to pure water. The composition of the binder
was mainly water with 2-Pyrrolidone as identified by Fourier transform
infrared (FTIR) spectroscopy analysis [20].
2.2. Powder characterizations
The printability of powder, defined by Butscher et al. [21], is a funda-
mental requirement for the 3DP process, such as depositability and
printability. It is influenced by several parameters of the powder such
as particle size distribution, powder true/bulk densities, powder bed po-
rosity, and binder droplet penetration behavior. Before adopting a new
powder to a commercially available 3D printer, printability is an impor-
tant characteristic that should be assessed [21].
Various characterization techniques have been applied in this study,
including laser diffraction particle analyzer to determine the particle
size and particle size distribution (PSD) of the powders, digital camera
385
for inspecting powder bed surface quality, true/bulk and in-process
bed densities, and a drop penetration measurement apparatus for char-
acterizing wettability of the powders.
2.2.1. Particle size distribution
The particle size and particle size distribution of the powders were
measured by using a laser diffraction particle analyzer (Cilas 1190,
CILAS, France). Each powder was measured three times to find out the
D10, D50 and D90 values, which indicate 10%, 50% and 90% of the particles
are below those sizes.
2.2.2. Powder bed surface quality
After preparing the powder bed in the printer, the build plate with
powder bed was removed from the 3D printer. The surface qualities of
the two powder beds were inspected by using a Nikon D810 camera.
The inspection direction was 45° to the top of the powder bed.
2.2.3. True, bulk and in-process bed densities
True (ρtrue.power) and bulk (ρbulk.powder) densities of the powders
were measured according to Australian Standard AS 1774.6-2001
(R2013) for determination of true density and Australian Standard AS
1774.2-2001 (R2013) for determination of bulk density. In-process
bed density (ρbed.powder), first proposed by Zhou et al. [22], is the density
measured after the powder being spread to the build plate. It is an im-
portant characteristic to ensure high-quality print quality in powder-
based 3DP process.
Prior to 3DP process, the powder was spread by a roller from the
powder feeder to the build plate. Then the build plate was removed
from the 3D printer. The volume of the powder bed and the mass of
powder in the build plate were measured to determine the in-process
bed density. The measurement was conducted three times for each
powder. Using the in-process bed density and true powder density,
Fig. 5. Particle size distributions of ZP and GP powder.
386 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
Table 2
Properties of ZP and GP powders.
Powder properties ZP
True density (g·cm−3) 2.60
Bulk density (g·cm−3) 0.91
In-process bed density (g·cm−3) 0.94
Powder bed porosity (%) 64%
powder bed porosity (Pbed) can be calculated from Eq. (1).
ρbed:powder
P bed ¼ 1−  100%
ρtrue:powder
2.2.4. Drop penetration behavior and wettability
An apparatus was developed for visualizing binder droplet impac-
tion on powder bed, penetration into powder bed and granule forma-
tion within the powder bed. A schematic diagram of the drop
penetration behavior measurement apparatus is shown in Fig. 3a.
After the powder bed being prepared on the build plate, the build
plate with powder bed was removed from the 3D printer. Then a binder
(Zb®63, Z-Corp, USA) droplet with a volume of approximately 60 μL was
deposited by a 1-mL medical syringe (25-gauge needle) which was po-
sitioned above powder bed surface. The droplet formed at the tip of the
needle and grown until its weight exceeded the surface tension force.
In this study, the wettability of powder was quantified by three pa-
rameters (Fig. 3b): drop penetration time (tp), drop penetrating depth
(dp) and drop spreading diameter (φp).
Drop penetration time (tp) is defined as the time taken for the drop-
let to penetrate completely into the powder with no liquid remaining on
the powder bed surface. A high-speed video recorder operating at 240
frames/s was applied to track a single binder droplet penetrating into
the powder bed. Ten replicates were performed for each powder. The
drop penetration time was calculated from the video frame by frame.
The starting point was the time when binder droplet firstly contacted
with the powder bed. The end point was the time when light reflection
on the droplet disappeared.
Drop penetrating depth (dp) and spreading diameter (φp) were
measured by using a digital caliper with an accuracy of 0.01 mm when
the granule was solidified. A sieve and an air blower were used to
Fig. 6. Surface quality inspection of (a) ZP and (b) GP powder bed.
remove the unbound powder. These two parameters were measured
after each drop penetration time measurement.
GP
2.81
2.3. 3D printing process
0.78
0.83 Fig. 4 demonstrates two digital models designed by using a comput-
70% er-aided design (CAD) software (SolidWorks® 2014, SolidWorks Corp.,
USA). A plate structure (40.4 × 40.4 × 4 mm) with different pore sizes
and shapes was designed to identify the printing resolution. This plate
consists of 3 shapes of pores, varying the side dimension ranging from
0.5 to 5 mm. A cubic structure with a dimension of 20 × 20 × 20 mm
was designed for quality assessing purpose.
3D printing process was carried out on a commercial 3D printer
ð1Þ (ZPrinter® 150, Z-Corp, USA) with an HP11 print head (C4810A). The
3D printer has a specified resolution of 300 × 450 dpi and an
185 × 236 × 132 mm build volume. For simplicity, default printing pa-
rameters setting was used with a layer thickness of 0.1016 mm and a
binder/volume ratio of 0.24 for the shell section and 0.14 for the core
section of the printed structure. The printed cubic structures were
dried within the powder bed at room temperature for 2 h and then un-
bound powder was removed using compressed air. After depowering,
the printed cubic structures were divided into two batches. In the first
batch, no further post-processing procedures have been applied on
both the ZP and GP cubic structures. In another batch, the GP cubic
structures were immersed in saturated anhydrous sodium metasilicate
solution until 1 and 7 days at a temperature of 60 °C.
2.4. 3D printed parts characterizations
2.4.1. Dimensional accuracy
Dimensions (the length of cubic structure sides) of printed cubic
structures were measured using a digital caliper with an accuracy of
0.01 mm and then compared with those of the CAD designed model. Di-
mensional deviation ratio (DDR) was used to quantify the dimensional
accuracy of printed structures. It can be calculated from Eq. (2).
L−LCAD
DDR ¼  100% ð2Þ
LCAD
where L is the measured value of length, LCAD is the length of CAD de-
signed model.
In this study, DDR value was measured in three orientations, which
are defined as follows: X-orientation is the direction of binder jetting,
 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
Fig. 7. Sequential images of binder droplet impacting on surfaces of (a) ZP and (b) GP powder bed. The number at the bottom of each frame indicates the time (ms).
Y-orientation is the direction of new layer spreading and Z-orientation
is the direction of layer stacking (Fig. 1a).
2.4.2. Apparent porosity
The apparent porosity of printed cubic structures was measured
based on Australian Standard AS 1774.5:2014. First of all, samples
were weighted in dry state (Mdry) using a precision balance with an ac-
curacy of 0.001 g, and then submerged in immersion liquid for 30 min.
Then, the weight of the sample suspended in an immersion liquid
(Msusp) was measured. Subsequently, each sample was taken out from
the immersion liquid and dabbed with a wet cloth to remove excess im-
mersion liquid and weighted to determine the wet weight (Mwet). The
bulk density (ρbulk.cube) and apparent porosity (Pcube) was calculated
by Eq. (3) and Eq. (4) respectively.
Mdry
ρbulk:cube ¼ ρliquid 
Mwet −Msusp
ρbulk:cube
P cube ¼ 1−  100%
ρtrue:powder
2.4.3. Mechanical property
To evaluate the mechanical property of the printed cubic structure,
uniaxial compressive strength was carried out at a loading rate of
0.03 kN/s. In this study, uniaxial compressive strength was measured
in both X-orientation and Z-orientation.
3. Results and discussions
3.1. Powder characterizations
The particle size distributions of the ZP and GP powders are present-
ed in Fig. 5.
The ZP and GP powders exhibit similar multimodal distribution, cov-
ering approximately the same size range of 0.1 to 100 μm. However, the
GP powder curve is slightly shifted to the left, indicating that the pres-
ence of particles with smaller diameters than the ZP powder.
The PSD analysis result showed the average particle size of ZP and GP
powders were 37.85 μm and 24.60 μm respectively. The ZP powder had
a particle size distribution as D10 = 1.50, D50 = 31.64 and D90 =
69.51 μm, while the GP powder had D10 = 1.23, D50 = 17.24 and
D90 = 52.75 μm.
True, bulk and in-process bed densities are listed in Table 2.
Results showed the GP powder exhibited lower in-process bed den-
sity and bulk density but higher powder bed porosity, compared to the
ZP powder. The PSD analysis result (Fig. 5) shows that the GP powder
387
contained more fine particles than the ZP powder. These fine particles
have a tendency to agglomerate during the deposition process, causing
macro voids or "bubble voids" within the powder bed [23], and thus in-
creasing the powder bed porosity. The slight difference between in-pro-
cess bed density and bulk density indicated both powders had not been
compressed significantly by the roller during the printing process. Pow-
der particles were deposited in a close condition to the free-flowing
state [22].
In 3DP process, one crucial requirement is the powder depositability,
which depends heavily on powder particle size [12,13]. For dry deposi-
tion, the suggested powder particle size is N20 μm [12]. Due to the pre-
dominance of inter-particle attraction force, fine powder particle
(b20 μm) tends to adhere to one another and form agglomerates,
which will result in insufficient depositability [24].
The surface quality of ZP and GP powder bed after deposition are
shown in Fig. 6. It was noted that the GP powder bed exhibited rough
ð3Þ
but even surface, while the ZP powder bed surface was smooth and ho-
mogenous. In this study, the surface quality of the GP powder bed was
already sufficient for printing [25].
ð4Þ
In 3DP process, another crucial requirement is the wettability of
powder by the binder solution. Fig. 7 shows a series of images taken
from the footage of typical drop penetration behavior of Zb® 63 binder
on ZP and GP powder.
As is shown in Fig. 7, the impaction and penetration behavior of the
binder droplet on both powder bed surfaces were quite similar. Upon
impact the powder surface, the binder droplet elastically deformed
and made a shallow crater. Solids spreading around the droplet surface
did not occur. Then the droplet reached its maximum spreading diame-
ter and started to penetrate into the powder bed driven by capillary
forces. At the end of penetration, a flat granule with a slightly higher
rim was formed within the powder bed because the rate of penetrating
was slower than that of spreading. In this measurement, the penetration
time of GP powder was 350 ms, which was longer than that of the ZP
powder.
Table 3
Drop penetration test results of ZP and GP powders.
Powder Binder Drop penetration parameters
tp (ms)a dp (mm)b φp (mm)c
ZP ®
Zb 63 316 ± 14 1.23 ± 0.05 2.42 ± 0.12
GP 344 ± 20 2.10 ± 0.11 1.95 ± 0.16
a
Drop penetration time.
b
Drop penetrating depth.
c
Drop spreading diameter.
388 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
Fig. 8. 3D printed structures using geopolymer-based material. (a) Plate structure, (b) Cubic structure.
A summary of the drop penetration test results is given in Table 3.
Ten replicates were performed for each powder. Each value is presented
as average ± standard deviation.
Results showed the GP powder had longer drop penetration time of
344 ± 20 ms than the ZP powder. In general, the drop penetration time
is sensitive to the powder bed structure. The drop penetration time will
be longer for powder formulation with a finer average particle size than
coarser one [22,26,27]. The presence of macro voids formed by fine par-
ticles agglomeration within the powder bed will increase the drop pen-
etration time. This is because binder liquid tends to flow through
capillary pores rather than macro voids [26].
For the GP powder, the spreading diameter (1.95 ± 0.16 mm) was
smaller than that of the ZP powder (2.42 ± 0.12 mm). Spreading is
the first stage of drop penetration. The droplet is flattened and spread
out horizontally into a pancake shape after impacting. In a very short
time, the droplet reaches to its maximum spreading diameter, which
is sensitive to the physical properties of the binder liquid and wettability
of the powder [28].
Compared with the ZP powder, the GP powder had a higher pene-
trating depth (2.10 ± 0.11 mm), indicating that the binder droplet dem-
onstrated more vertical penetrating behavior in the GP powder bed. For
a single fixed-volume binder droplet, the smaller spreading diameter,
the higher penetrating depth. Meanwhile, the capillary paths in the
finer particle powder bed for binder droplet will become more irregular
due to the presence of macro voids. The binder droplet has to penetrate
Table 4
Dimensional deviation ratio of 3D printed cubic structures.
Dimensional deviation ratio (DDR) (%)
Powders X-orientation Y-orientation Z-orientation
ZP 1.05 ± 0.55 1.75 ± 0.60 3.65 ± 0.95
GP 0.85 ± 0.25 1.65 ± 0.35 2.65 ± 0.85
deeper in the powder bed to complete the whole penetration process.
This also explains the longer penetration time on GP powder bed.
3.2. Characterizations of 3D printed structures
The 3D printed structures using geopolymer-based material are
shown in Fig. 8. The plate structure presented in Fig. 8a shows that the
prepared geopolymer-based material allowed the production of struc-
ture with a minimum rectangle-shape pore size of 1 mm via the pow-
der-based 3D process. The quality of 3D printed cubic structure was
determined by three characteristic features: printing accuracy, apparent
porosity and mechanical property. Table 4 summarizes the DDR mea-
surement results of 3D printed cubic structures in three orientations.
Thirty replicates were performed for each powder. Each value is pre-
sented as average ± standard deviation.
According to the results listed in Table 4, the lengths of all printed
cubic structures were longer than CAD model (20 mm). The dimension-
al deviation ratios in three orientations were b4%. Moreover, Z-orienta-
tion had the highest deviation among three orientations. The positive
and different deviation values indicated all cubic structures exhibited
anisotropic expansion in all orientations. This anisotropic phenomenon
might be associated with the bleeding mechanism in printing process
[29] and preferential orientation of the powder particles [30] in the de-
position process.
Table 5
Apparent porosity of 3D printed cubic structures.
Cubic structure Bulk density (g·cm–3) Apparent porosity (%)
ZP 0.99 ± 0.03 61.9 ± 1.8
GP 1.20 ± 0.04 57.1 ± 1.4
 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
Table 6
Uniaxial compressive strength of cubic structures (Without post-processing).
Powders Uniaxial compressive strength (MPa)
X-orientation Z-orientation
ZP 0.69 ± 0.12 0.61 ± 0.11
GP 0.91 ± 0.03 0.76 ± 0.10
For GP cubic structure, DDR values in all orientations were lower
than those printed from ZP powder. It was also noted that DDR value
in Z-orientation was significantly lower compared with the ZP cubic
structure. This might be explained by the difference in powder bed
structure and powder/binder reactivity [31,32].
The bulk density and apparent porosity of the printed cubic struc-
tures are presented in Table 5. Fifteen printed cubic samples were mea-
sured and each value is presented as average ± standard deviation.
As is shown in Table 5, the GP cubic structure had a lower apparent
porosity. The apparent porosities of both cubic structures were lower
than the powder bed porosity (Table 2) respectively. By comparing
Table 2 and Table 5, it was noted that ZP powder had a bed porosity of
64%, and its average cube apparent porosity decreased by 2.1% after print-
ing. For the GP powder, the difference between bed porosity and average
cube apparent porosity was 12.9%. This could also be explained by the for-
mation of different reaction product between powder particles.
The uniaxial compressive strength measurement results without ap-
plying any post-processing procedures are listed in Table 6. Fifteen
printed cubic samples were measured and each value is presented as
average ± standard deviation.
The strengths presented in Table 6 are too low for construction ap-
plications. However, the strengths are sufficiently high to withstand
the pressure from the compressed air during depowering. The post pro-
cessing is an essential process so that sufficient strengths can be
achieved.
As is shown in Table 6, GP cubic structure exhibited higher compres-
sive strength in both orientations than the ones made from the com-
mercially available ZP powder. This could be attributed to the different
powder/binder reaction product. For the ZP powder, particles were
bonded by interlocking CaSO4·2H2O crystals in the form of plates and
needles. Those CaSO4·2H2O crystals were formed by the hydration of
CaSO4·1/2H2O through a solution mechanism [33]. For the GP powder,
particles were bonded by calcium silicate hydrate (C-S-H) gel and sodi-
um aluminosilicate hydrate (N-A-S-H) gel [34], which were formed by
slag and anhydrous sodium metasilicate through geopolymerization
process [35] in the presence of water. Mechanical properties in porous
ceramic powder system are determined by the bonding force at the par-
ticle neck [36]. The bonding force caused by plate/needle like crystals
between ZP powder particles was lower than that resulting from
dense gels between GP powder particles.
It was also noted that the compressive strength in X-orientation was
higher than that in Z-orientation. For GP cubic structure, the compres-
sive strength in X-orientation was 19.7% higher than that in Z-orienta-
tion, while it was 13.1% higher for ZP cubic structure. Similar results
have been found in other printable ceramic powder system [30,37–
40]. This indicates that printing direction has a significant effect on the
mechanical property of the printed structure. The anisotropic
Table 7
Uniaxial compressive strength of GP cubic structures (With post-processing).
Uniaxial compressive strength (MPa)
Immersed time X-orientation Z-orientation
1 day 10.3 ± 0.2 9.2 ± 0.2
7 days 16.5 ± 0.4 15.7 ± 0.2
389
mechanical property might be related to the week inter-layer bond
strength [38] and the preferential orientation of the powder particles
[30] in the deposition process.
The uniaxial compressive strength of GP cubic samples immersed in
saturated anhydrous sodium metasilicate solution at 60 °C for different
times are listed in Table 7. Fifteen printed cubic samples were measured
and each value is presented as average ± standard deviation.
As is shown in Table 7, the compressive strength of GP cubic struc-
tures measured in X-orientation has been significantly improved by
1031% from 0.91 ± 0.03 MPa to 10.3 ± 0.2 MPa after 1-day alkaline so-
lution immersion and increased by 1713% after 7-day immersion. This
sharp increase could be regarded as a consequence of the continued
geopolymerization process in the presence of alkaline activator
solution.
It was also noted that the compressive strength in X-orientation was
still higher than that in Z-orientation, but the gap was reduced com-
pared with the compressive strength listed in Table 6, indicating that
post-processing procedure will reduce the anisotropic phenomenon.
4. Conclusions
An innovative methodology is developed in this study for formulat-
ing geopolymer-based material for the requirements and demands of
commercially available powder-based 3D printers. Different key pow-
der parameters were used to quantitatively evaluate the printability of
prepared geopolymer-based material. Cubic structure and plate struc-
ture were manufactured by the prepared geopolymer-based material
using a commercially available 3D printer. The following conclusions
can be drawn from this study:
(1) The prepared geopolymer-based material had gained sufficient
depositability and wettability and it is applicable to replace com-
mercially available material in a powder-based 3D printer.
(2) Geopolymer-based cubic structure can be made using a powder-
based 3DP process that is sufficiently strong enough to handle for
de-powdering and post-processing and shown to have good di-
mensional accuracy.
(3) All printed cubic structures exhibited an anisotropic phenome-
non in dimensional accuracy and mechanical property. X-orien-
tation (binder jetting direction) has better accuracy and higher
strength than other directions.
(4) A new post-processing method introduced in this study in-
creased the strengths many times from the original strength.
Strengths about 16 MPa can be achieved with this post-process-
ing method. While this is sufficient for some construction appli-
cations, further improvements to strength will be necessary for
many construction applications.
(5) The new post-processing method reduced the anisotropy of the
printed material.
The method developed in this study is readily scalable to produce
large structural components.
References
[1] T. Wohlers, 3D Printing and Additive Manufacturing State of the Industry, Wohlers
Associates Inc., Colorado, 2014.
[2] W. Gao, Y. Zhang, D. Ramanujan, K. Ramani, Y. Chen, C.B. Williams, C.C. Wang, Y.C.
Shin, S. Zhang, P.D. Zavattieri, The status, challenges, and future of additive
manufacturing in engineering, Comput. Aided Des. 69 (2015) 65–89.
[3] R.A. Buswell, R. Soar, A.G. Gibb, A. Thorpe, Freeform construction: mega-scale rapid
manufacturing for construction, Autom. Constr. 16 (2007) 224–231.
[4] B. Khoshnevis, S. Bukkapatnam, H. Kwon, J. Saito, Experimental investigation of con-
tour crafting using ceramics materials, Rapid Prototyp. J. 7 (2001) 32–42.
[5] B. Khoshnevis, D. Hwang, K.-T. Yao, Z. Yeh, Mega-scale fabrication by contour
crafting, Int. J. Ind. Syst. Eng. 1 (2006) 301–320.
[6] S. Lim, T. Le, J. Webster, R. Buswell, S. Austin, A. Gibb, T. Thorpe, Fabricating construc-
tion components using layer manufacturing technology (GICC'09) Paper presented
390 M. Xia, J. Sanjayan / Materials and Design 110 (2016) 382–390
at the Global Innovation in Construction Conference, Loughborough University,
Leicestershire, UK, 2009 13–16.
[7] S. Lim, R.A. Buswell, T.T. Le, R. Wackrow, S.A. Austin, A.G. Gibb, T. Thorpe, Develop-
ment of a viable concrete printing process, Proceedings of the 28th International
Symposium on Automation and Robotics in Construction, (ISARC2011), Seoul,
South Korea 2011, pp. 665–670.
[8] E.M. Sachs, J.S. Haggerty, M.J. Cima, P.A. Williams, Three-dimensional printing
techniques(U.S. Patent No. 5,204,055. Washington, DC: U.S. Patent and Trademark
Office) 1993.
[9] G. Cesaretti, E. Dini, X. De Kestelier, V. Colla, L. Pambaguian, Building components for
an outpost on the lunar soil by means of a novel 3D printing technology, Acta Astro-
naut. 93 (2014) 430–450.
[10] R. Rael, V. San Fratello, Developing concrete polymer building components for 3D
printing, Proceedings of the 31st Annual Conference of the Association for Computer
Aided Design in Architecture, (ACADIA 11), Banff, Canada 2011, pp. 152–157.
[11] H.P. Le, Progress and trends in ink-jet printing technology, J. Imaging Sci. Techn. 42
(1998) 49–62.
[12] B.R. Utela, D. Storti, R.L. Anderson, M. Ganter, A review of process development steps
for new material systems in three dimensional printing (3DP), J. Manuf. Process. 10
(2008) 96–104.
[13] B.R. Utela, D. Storti, R.L. Anderson, M. Ganter, Development process for custom
three-dimensional printing (3DP) material systems, J. Manuf. Sci. Eng. 132 (2010)
0110081–0110089 (Transactions of the ASME).
[14] G.J. Gibbons, R. Williams, P. Purnell, E. Farahi, 3D printing of cement composites,
Adv. Appl. Ceram. 109 (2010) 287–290.
[15] A.-K. Maier, L. Dezmirean, J. Will, P. Greil, Three-dimensional printing of flash-set-
ting calcium aluminate cement, J. Mater. Sci. 46 (2010) 2947–2954.
[16] C. Gosselin, R. Duballet, P. Roux, N. Gaudillière, J. Dirrenberger, P. Morel, Large-scale
3D printing of ultra-high performance concrete – a new processing route for archi-
tects and builders, Mater. Design 100 (2016) 102–109.
[17] J. Davidovits, Geopolymers and geopolymeric materials, J. Therm. Anal. Calorim. 35
(1989) 429–441.
[18] A. Farzadi, M. Solati-Hashjin, M. Asadi-Eydivand, N.A.A. Osman, Effect of layer thick-
ness and printing orientation on mechanical properties and dimensional accuracy of
3D printed porous samples for bone tissue engineering, PLoS One 9 (2014),
e108252.
[19] N. Afshar-Mohajer, C.-Y. Wu, T. Ladun, D.A. Rajon, Y. Huang, Characterization of par-
ticulate matters and total VOC emissions from a binder jetting 3D printer, Build. En-
viron. 93 (2015) 293–301.
[20] M. Asadi-Eydivand, M. Solati-Hashjin, S.S. Shafiei, S. Mohammadi, M. Hafezi, N.A.A.
Osman, Structure, properties, and in vitro behavior of heat-treated calcium sulfate
scaffolds fabricated by 3D printing, PLoS One 11 (2016), e0151216.
[21] A. Butscher, M. Bohner, C. Roth, A. Ernstberger, R. Heuberger, N. Doebelin, P.R. Von
Rohr, R. Müller, Printability of calcium phosphate powders for three-dimensional
printing of tissue engineering scaffolds, Acta Biomater. 8 (2012) 373–385.
[22] Z. Zhou, F. Buchanan, C. Mitchell, N. Dunne, Printability of calcium phosphate: calci-
um sulfate powders for the application of tissue engineered bone scaffolds using the
3D printing technique, Mater. Sci. Eng. C 38 (2014) 1–10.
View publication stats
[23] M. Fayed, L. Otten, Handbook of Powder Science & Technology, Springer Science &
Business Media, UK, 2013.
[24] Q. Li, V. Rudolph, B. Weigl, A. Earl, Interparticle van der Waals force in powder
flowability and compactibility, Int. J. Pharm. 280 (2004) 77–93.
[25] S. Spath, H. Seitz, Influence of grain size and grain-size distribution on workability of
granules with 3D printing, Int. J. Adv. Manuf. Technol. 70 (2014) 135–144.
[26] K.P. Hapgood, J.D. Litster, S.R. Biggs, T. Howes, Drop penetration into porous powder
beds, J. Colloid Interface Sci. 253 (2002) 353–366.
[27] T. Nguyen, W. Shen, K. Hapgood, Drop penetration time in heterogeneous powder
beds, Chem. Eng. Sci. 64 (2009) 5210–5221.
[28] E. Nefzaoui, O. Skurtys, Impact of a liquid drop on a granular medium: inertia, vis-
cosity and surface tension effects on the drop deformation, Exp. Thermal Fluid Sci.
41 (2012) 43–50.
[29] M. Castilho, B. Gouveia, I. Pires, J. Rodrigues, M. Pereira, The role of shell/core satu-
ration level on the accuracy and mechanical characteristics of porous calcium phos-
phate models produced by 3Dprinting, Rapid Prototyp. J. 21 (2015) 43–55.
[30] Y. Shanjani, Y. Hu, R.M. Pilliar, E. Toyserkani, Mechanical characteristics of solid-
freeform-fabricated porous calcium polyphosphate structures with oriented stacked
layers, Acta Biomater. 7 (2011) 1788–1796.
[31] E. Sachs, M. Cima, P. Williams, D. Brancazio, J. Cornie, Three dimensional printing:
rapid tooling and prototypes directly from a CAD model, J. Manuf. Sci. Eng. 114
(1992) 481–488.
[32] A. Khalyfa, S. Vogt, J. Weisser, G. Grimm, A. Rechtenbach, W. Meyer, M.
Schnabelrauch, Development of a new calcium phosphate powder-binder system
for the 3D printing of patient specific implants, J Mater. Sci.-Mater. M. 18 (2007)
909–916.
[33] E. Gartner, J. Gaidis, J. Skalny, in: J.P. Skalny (Ed.), Materials Science of Concrete, The
American Ceramic Society, Westerville, OH 1989, pp. 95–125.
[34] C.K. Yip, G.C. Lukey, J.S.J. van Deventer, The coexistence of geopolymeric gel and cal-
cium silicate hydrate at the early stage of alkaline activation, Cem. Concr. Res. 35
(2005) 1688–1697.
[35] P. Duxson, A. Fernández-Jiménez, J.L. Provis, G.C. Lukey, A. Palomo, J.S.J. Deventer,
Geopolymer technology: the current state of the art, J. Mater. Sci. 42 (2006)
2917–2933.
[36] S.A. Uhland, R.K. Holman, S. Morissette, M.J. Cima, E.M. Sachs, Strength of green ce-
ramics with low binder content, J. Am. Ceram. Soc. 84 (2001) 2809–2818.
[37] M. Vlasea, R. Pilliar, E. Toyserkani, Control of structural and mechanical properties in
bioceramic bone substitutes via additive manufacturing layer stacking orientation,
Addit. Manuf. 6 (2015) 30–38.
[38] W. Zhang, R. Melcher, N. Travitzky, R.K. Bordia, P. Greil, Three-dimensional printing
of complex-shaped alumina/glass composites, Adv. Eng. Mater. 11 (2009)
1039–1043.
[39] M. Asadi-Eydivand, M. Solati-Hashjin, A. Farzad, N.A.A. Osman, Effect of technical
parameters on porous structure and strength of 3D printed calcium sulfate proto-
types, Robot. Cim.-Int. Manuf. 37 (2016) 57–67.
[40] P. Feng, X. Meng, J.-F. Chen, L. Ye, Mechanical properties of structures 3D printed
with cementitious powders, Constr. Build. Mater. 93 (2015) 486–497.