Journal of Cleaner Production 197 (2018) 1046e1055

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Close-looped recycling of polylactic acid used in 3D printing: An

experimental investigation and life cycle assessment
Peng Zhao a, b, Chengchen Rao a, b, Fu Gu a, c, d, *, Nusrat Sharmin e, Jianzhong Fu a, b
a
State Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027, China
b
Key Laboratory of 3D Printing Process and Equipment of Zhejiang Province, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027,
China
c
Department of Industrial Engineering, Zhejiang University, Hangzhou 310027, China
d
National Institute of Innovation Management, Zhejiang University, Hangzhou 310027, China
e
Department of Chemical and Environmental Engineering, Nottingham University, Ningbo 315100, China

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

Article history: The objective of this work is to investigate the potential of close-looped recycling of polylactic acid (PLA)
Received 17 April 2018 that used in 3D printing, from the perspectives of material properties and environmental performance. A
Received in revised form commercial grade of PLA was extruded into filament, then was subjected to fused deposition modelling
25 June 2018
(FDM) 3D printing process. The printed products were later shredded and re-extruded for repeating 3D
Accepted 26 June 2018
Available online 27 June 2018
printing cycle. Samples were taken from each repeating cycle, were characterised in terms of mechanical,
rheological, molecular, thermal and morphological properties. In the experiment, the material can only
^ as de
Handling Editor: Cecilia Maria Villas Bo be reprocessed for two 3D printing cycles, as the reprocessed material can no longer be further pro-
Almeida cessing. Although little changes are observed in the mechanical properties, the viscosity measurements
suggest that PLA deteriorates significantly due to the repeated FDM 3D printing cycles. The reduced
Keywords: viscosity values are compatible with the decrements in average molecular weights and thermal stability,
3D printing the increments of carbonyl bonds, crystallinity and pinholes in the materials' surfaces. Virgin PLA was
Close-looped recycling blended into the recycled material to improve the viscosities, which enables the blends 3D printable. The
Polylactic acid
results of life cycle assessment approach show that the environmental burdens associated with close-
Viscosity
looped recycling are lower than those of placing the 3D printed products in incineration or in landfill.
Life cycle assessment
End-of-life option Incineration might be another environmental alternative, due to the material embodies a relatively high
calorific value; however, the material is not recuperated, thus incineration cannot satisfy the more
stringent environmental legislations, in which recovery and reuse of materials are rigorously required.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction final products will be manufactured in 3D printing processes

Three-dimensional (3D) printing is a typical additive
manufacturing (AM) technique for fabricating products of various
structures and geometries (Ngo et al., 2018). Owning to its capacity
in highly customisable products, 3D printing has been widely
applied in many industries, including construction (Wu et al.,
2016), energy (Zhang et al., 2017), food (Godoi et al., 2016), medi-
cal (Jang et al., 2018), prototyping (Ngo et al., 2018) and
manufacturing (Upadhyay et al., 2017). It is predicted that 50% of
* Corresponding author. State Key Laboratory of Fluid Power and Mechatronic
Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027,
China.
E-mail address: gufu@zju.edu.cn (F. Gu).
(Berman, 2012). Proliferation of 3D printing brings emerging
environmental problems such as emissions and waste disposal.
Compared to extensive efforts that paid on characterising emis-
sions from 3D printing (see Stephens et al., 2013; Kim et al., 2015;
Azimi et al., 2016; Steinle, 2016; Mendes et al., 2017; Stabile et al.,
2017) and mitigating this problem (Kwon et al., 2017; Rao et al.,
2017), little research concerns about the sustainability issue of 3D
printing materials, that is, disposal of waste 3D printed products.
Although 3D printing possesses higher material efficiency (Gebler
et al., 2014), material sustainability is an issue that can no longer
be ignored due to wide adoption of 3D printing. Close-looped
recycling can be a promising measure in tackling this particular
problem due to the following four reasons. First, despite intensive
efforts have been paid on developing new materials for 3D printing,

https://doi.org/10.1016/j.jclepro.2018.06.275
0959-6526/© 2018 Elsevier Ltd. All rights reserved.
 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055
only a limited number of materials are commercially viable (Gebler
et al., 2014; Ngo et al., 2018). Second, the global production of 3D
printing materials like polylactic acid (PLA) is still limited, though
PLA production has enjoyed a substantial growth during the past
decade (Beltra n et al., 2018). Third, mechanical recycling of waste
plastics is becoming a common and environmental practise in
mitigating solid plastic waste problems (Andreoni et al., 2015; Gu
et al., 2017), and recycled plastics have comparable performance
to their virgin counterparts (Gu et al., 2016a, 2016b) and have even
been applied on manufacturing high-value products such as auto-
mobiles (Gu et al., 2017). Fourth, high-quality recycling includes
close-looped recycling is more sustainable than downcycling (Di
Maria et al., 2018) and other alternatives like landfill (Bensadoun
et al., 2016).
Yet, to the best of our knowledge, there is little literature that
investigates this particular topic of close-looped recycling of 3D
printing material, and the extant research concentrates on exam-
ining changes in mechanical properties that associated with
repeated 3D printing cycles (Sanchez et al., 2017; Tian et al., 2017).
Sanchez et al. (2017) studied the mechanical performance of PLA
during repeated 3D printing processes and found that there were
significant deteriorations in most mechanical properties after five
recycling cycles. Tian et al. (2017) achieved material recovery rates
of 100% for continuous carbon fibre and 73% for PLA matrix in
recycling 3D printed continuous carbon fibre reinforced PLA com-
posites. The recycling potential of 3D printing materials is only
investigated from the perspective of mechanical performance.
Although mechanical performance is crucial in the applications of
3D printing, there are also other properties that should be
considered during recycling. For example, viscosity is a particularly
important characteristic for fused deposition modelling (FDM), the
most common and frequently studied 3D printing process. FDM 3D
printers can only adopt materials which have viscosity in a limited
range, as the viscosity should be high enough to provide structural
support and low enough to enable extrusion (Wang et al., 2017).
From material sustainability perspective, applying life cycle ac-
counting techniques like life cycle assessment (LCA) to calculate the
overall environmental impacts of the whole recycling process,
would provide quantitative evidence that ensures the best route of
disposing the plastic waste and demonstrates the gain of using the
recycled polymers (Ragaert et al., 2017). However, the data on
environmental impact of disposing 3D printed waste is in scarcity
or even non-existed, as the studies on recycling 3D printing ma-
terials are also extremely limited.
To address this knowledge gap, the main objective of this work
is to investigate the potential of close-looped recycling PLA that
used in 3D printing, in terms of changes in properties and related
environmental impacts. PLA is studied due to its limited global
production (around 0.21 Mt in 2016) (Beltr an et al., 2018) and
extensive application in 3D printing (Rao et al., 2017). As a type of
biopolymers, PLA is made from crops and increasing its production
could jeopardise the food security of poorer countries (Mülhaupt,
2013). Moreover, PLA is also used in other applications like pack-
ages (Beltran et al., 2018). These characteristics justify the choice of
material, as prolonging the life of PLA used in 3D printing is
essential in reducing material consumption and waste generation.
In this study, a commercial grade of PLA was extruded into filament
and fed into an FDM 3D printer, and the printed products were
shredded, extruded and then subjected to the 3D printing process
again. The cycle was repeated till the material cannot be subjected
to further 3D printing. During each cycle, a certain amount of
sample was extracted for detailed characterisation with the pur-
pose of identifying possible properties changes. The properties
tested include: (1) mechanical performance in terms of tensile and
1047
flexural properties; (2) rheological properties which refer to the
viscosities of PLA in different cycles; (3) molecular structures,
including number average molecular weights (Mn) and weight
average molecular weight (Mw) that measured by gel permeation
chromatography (GPC), and functional groups that detected by
Fourier transform infrared spectroscopy (FTIR); (4) thermal prop-
erties that tested via differential scanning calorimetry (DSC) and
thermogravimetric analysis (TGA); (5) morphological features that
investigated via scanning electron microscopy (SEM). The selection
of properties was determined in accordance with the conceptual
framework for evaluating the quality of recycled plastics proposed
by Vilaplana and Karlsson (2008). Virgin PLA was then blended
with recycled PLA in different proportions, since compounding
recycled plastics with virgin counterparts is a simple and effective
route for improving desirable properties (Gu et al., 2014; Hopmann
et al., 2015). Apart from this, blending recycled PLA with virgin
materials for upgrading the performance has been pointed out as
one of the future research directions in mechanical recycling of PLA
(Badia and Ribes-Greus, 2016). The environmental performance
was assessed using an LCA approach for comparing the close-
looped recycling with other end-of-life alternatives such as land-
fill and incineration, since these three routes are most common in
disposing plastic waste and are frequently compared in the previ-
ous LCA studies (Bensadoun et al., 2016; Gu et al., 2017, 2018).
The contribution of this study is threefold. First, the materials
recycled from FDM 3D printed products of different cycles have
been characterised in a comprehensive manner, while the extant
literature tends to concentrate solely on mechanical properties
(Sanchez et al., 2017; Tian et al., 2017). Second, the key performance
issue, i.e., viscosity, in achieving close-looped recycling of 3D
printing material was identified, and an attempt has been carried
out to mitigate the problem via blending with virgin material.
Third, environmental impacts of close-looped recycling of 3D
printing material were first modelled and compared to other
disposal routes, with the purpose to identify the most sustainable
disposal route for 3D printed products.
The rest of this paper is organised as follows. The experimental
sector in this study are illustrated in the next section. The experi-
mental results are presented and discussed in Section 3, and Sec-
tion 4 concludes this study.
2. Materials and methods
2.1. Materials
A commercial grade of PLA with a trademark of Ingeo™ 4032D,
produced by NatureWorks Co., Ltd., Guangdong, China was used in
this work. The specific density of this PLA material is 1.24 g cm
3.
This type of material is frequently applied as raw material of 3D
printing filament, see online 3D printing sites such as Stro3D (2018)
and 3DprintWorks (2018).
2.2. Processing
Based on the selection of properties and simulated PLA recycling
processes in the previous literature (Badia and Ribes-Greus, 2016;
Sanchez et al., 2017), a scheme for close-looped recycling of PLA in
3D printing was adapted as illustrated in Fig. 1. In this experimental
work, virgin PLA pellets were extruded into filament and then
subjected to FDM 3D printing, and 3D printed products were
shredded and re-extruded into filament; the properties of the
materials were characterised during each cycle, as the samples
were taken from the filaments, while only mechanical tests use 3D
printed specimens. The experimental scheme enables the com-
parison of materials of different cycles and proportions, as specified
1048 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055
Fig. 1. Scheme of close-looped recycling of PLA that used in 3D printing.
in the characterisation Section 2.3. The samples that taken from the
nth cycles are denoted as PLA (n-1).
2.2.1. Drying and extrusion
Prior to processing, PLA (virgin and/or recycled) was dried at
80  C for 4 h to remove most moisture content, as suggested by
NatureWorks (2018). Required amount of PLA was weighed and
made into filament through extrusion using a Friend™ SJ25/25 co-
rotating twin-screw extruder (screw diameter ¼ 25 mm and L/
D ¼ 25). The temperature profile was set at 180 C-200 C-200 C-
200 C-200  C with a screw speed of 50 rpm. The parameters were
determined in accordance with the recommendations of
NatureWorks (2018), and the die with diameter of 1.75 mm was
used.
2.2.2. 3D printing
The 3D printing process was carried out using an FDM 3D
printer, model Delta-pro, Suzhou Zijingang Intelligent Manufacturing
Equipment Co., Ltd, China. Referred to the work of Tian et al. (2017),
the processing parameters in the 3D printing process were deter-
mined and shown in Table 1. These parameters remained constant
during this experiment for yielding comparable samples. The
shapes of 3D printed products were set according to the specifi-
cations of mechanical tests, that is, dumbbell-shaped specimens
required by ISO527:2012 and rectangle specimens required by
ISO178:2010.
Table 1
Process parameters for 3D printing.
Parameters
Toolpath*
Filling density (F, vol.%)
Layer thickness (L, mm)
Bottom layers
Sealing layers
Feed rate of filament (E, mm min
1)

1
Printing speed (V, mm min )
Diameter of nozzle (mm)
Temperature of liquefier (Tf,  C)
Temperature of bed (Tb,  C)
*A parameter that describes the superposition of the melt strands in a
3D printing product, values refer to the angles between the strands,
see Sanchez et al. (2017).
2.2.3. Shredding
Apart from some samples that extracted for characterisation, the
rest of 3D printed specimens were shredded by a Friend™ plastic
mill. Washing and sorting were not included in this experiment,
since the purpose of this work is to simulate close-looped recycling
of 3D printed products; the source and the end-use of PLA were
deterministic.
2.3. Characterisation
2.3.1. Mechanical
The tensile properties tested were tensile modulus, tensile
strength, yield strength and elongation at break, which were
assessed in accordance with ISO527e2:2012. Tensile tests were
performed on a Gotech TCS-2000NE universal testing machine
with a full-scale load range of 20 kN and a crosshead speed of
1 mm min
1. Flexural properties refer to flexural modulus and
flexural strength, which were assessed in accordance with
ISO178:2010. Flexural tests were performed on a Gotech TCS-
2000NE universal testing machine with a full-scale load range of
20 kN with a crosshead speed of 1 mm min
1, and a span of 64 mm.
Referring to our previous studies on investigating mechanical
performance of polymers (Gu et al., 2016a, 2016b), values reported
are the average of eight measurements for each fraction to ensure
their representativeness, and all specimens were conditioned at
23  C and 50% R.H. for over 88 h before tests, in accordance with ISO
291:2008.
2.3.2. Rheological
The viscosity was measured by a rotary rheometer, model DHR-
1, TA, USA. The experimental parameters were: frequency of 1 Hz,
temperature range was set between 160  C and 220  C, a heating
rate at 5  C min
1 and a strain of 5%. Each fraction was tested three
times, and the average values were recorded. The number of
measurements was determined in accordance with the previous
study on PLA (Tiwary et al., 2017).
2.3.3. Molecular
Mn was measured by GPC using an Agilent PL-GPC220 machine
with PL gel Olexis 300  7.5 mm chromatographic columns, in
accordance with ASTM D6474 standard. The specimens were dis-
solved in trichlorobenzene at 150  C, and the concentration of PLA
was kept around 5.0 mg ml
1 and the sampling size was kept 10 ml.
The FTIR spectra of PLA was recorded using a Vector 22 FTIR
equipment, Bruker, Switzerland, equipped with an attenuated total
reflectance (ATR) accessory. Each spectrum was recorded at a res-
olution of 2 cm
1, with a total of eight scans. The FTIR spectra were
corrected and normalised.
2.3.4. Thermal
DSC analysis was performed on samples of about 10 mg, in
standard aluminium pans, using a TA Q-20 calorimeter, under ni-
Values trogen atmosphere. The samples were heated from 25  C to 220  C
at a heating rate of 10  C min
1, and then kept at 220  C for 3 min to
45/45
20
elucidate the effects of previous thermal treatments. Afterward, the
0.2 samples were cooled to 0  C and finally a second heating scan was
3 performed until 220  C. Glass transition temperature (Tg), crystal-
3 lisation temperature (Tc) and melting temperature (Tm) were ob-
100
tained, and the crystallinity (c) of PLA was determined by Eq. (1)
100
0.4 (Beltran et al., 2016):
210
60 DHm
DHcc
cð%Þ ¼ (1)
DH∞
where DHm and DHcc (J g
1) are the melting enthalpy and the cold
 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055
crystallisation enthalpy respectively. DH∞ is set at 93.1 J g
1, which
 et al.,
is the melting enthalpy for 100% crystalline PLA (Deroine
2014).
TGA was carried out using a TA TGA-Q50 analyser under a room
condition. Samples around 20 mg were weighted and heated from
ambient temperature to 700  C at a heating rate of 10  C min
1.
2.3.5. Morphological
The specimens were scanned using a HITACHI SU8000 (Japan)
scanning electron microscope at an acceleration voltage of 3.00 kV.
2.4. Life cycle assessment
The life cycle impact assessments on comparing the end-of-life
alternatives of 3D printed products were performed in accordance
to ISO 14040 and ISO 14044. The GaBi ts software was used to
construct the LCA study, and the ReCiPe method (www.lcia-recipe.
net, version 1.08) was employed to assess the impacts on the mid-
point level. The choice of modelling software and method was
supported by previous studies in modelling the lifecycles of poly-
mers or composites (see Hervy et al., 2015; Gu et al., 2017; Gu et al.,
2018). On the ReCiPe midpoint level, the following main impact
categories were included using a top-down approach: global
warming potential (GWP, in kg CO2 eq), ozone depletion potential
(ODP, in CFC-11 eq), fossil depletion potential (FDP, in kg oil eq),
human toxicity potential (HTP, in kg 1,4-DB eq), particulate matter
formation potential (PMFP, in kg PM10 eq), photochemical oxidant
formation potential (POFP, in kg NMVOC), freshwater eutrophica-
tion potential (FEP, in kg P eq), freshwater ecotoxicity potential
(FETP, in kg 1,4-DB eq), marine eutrophication potential (MEP, in kg
N eq), marine ecotoxicity potential (METP, in kg 1,4-DB eq),
terrestrial acidification potential (TAP, in kg SO2 eq), and terrestrial
ecotoxicity potential (TETP, in kg 1,4-DB eq). To conduct a reliable,
quantified comparison, normalisation was carried out (Pizzol et al.,
2017), and Europe was selected as the reference system due to data
availability.
In this study, close-looped (mechanical) recycling was
compared to landfill and incineration, as these routes are common
routes of disposing plastic waste (see Bensadoun et al., 2016; Gu
et al., 2017; Gu et al., 2018). To measure the energy that could be
recuperated from incineration, a Parr 1261 bomb calorimeter with a
Parr 1108 oxygen combustion bomb was employed to measure the
heat (calorific value, in terms of M J kg
1) from combustion of the
composite constituents include all carbon hydrogen and oxygen
compounds.
3. Results and discussion
3.1. Effects of repeated three-dimensional printing
3.1.1. Changes in mechanical properties
_
Contrary to the finding of Zenkiewicz et al. (2009), the 3D
printing only repeated for two cycles, as the filament (PLA (2))
cannot be printed into specimens during its third cycle. This could
be attributed to the 3D printing process, which introduces more
significant deteriorations into the material. Therefore, the me-
chanical properties of PLA (0) and PLA (1) were measured due to
limited sample availability, as shown in Table 2. As shown in
Table 2, both tensile and flexural properties only exhibit slight
deteriorations after one processing cycle, in which two extrusion
processes and one 3D printing process were took places. This
finding is comparable to the previous literature on PLA recycling, as
_
Zenkiewicz et al. (2009) observed a modest reduction in tensile
strength of PLA that has been extruded up ten cycles, and recycled
1049
Table 2
Mechanicals properties of PLA in different cycles.
PLA (0) PLA (1)
Tensile modulus (MPa) 1572.43 ± 27.16 1566.54 ± 45.61
Tensile strength (MPa) 30.21 ± 0.89 29.47 ± 1.21
Yield strength (MPa) 27.69 ± 0.77 27.65 ± 1.09
Elongation at break (%) 2.74 ± 0.53 2.45 ± 0.47
Flexural modulus (MPa) 2423.73 ± 56.42 2234.70 ± 70.68
Flexural strength (MPa) 64.48 ± 2.49 57.55 ± 1.97
PLA is becoming stiffer and more brittle (Scaffaro et al., 2011). In
general, the changes in tested mechanical properties are alike to the
results of Sanchez et al. (2017), but this experiment only lasted for
two 3D printing cycles. To investigate the possible deteriorations
that prevent the materials from further 3D printing cycles, PLA (0),
PLA (1) and PLA (2) were studied in all the other characterisation
tests.
3.1.2. Changes in rheological properties
Fig. 2 shows the viscosity values of PLA (0), PLA (1) and PLA (2).
The initial viscosities at 160  C are significantly differentiated, and
this observation is distinctive different from the extant literature,
where only marginal decreases in viscosity values have been re-
ported during mechanical recycling of PLA (Pillin et al., 2008; Badia
n et al., 2018). It is well-known that PLA is sus-
et al., 2012; Beltra
ceptible to thermo-mechanical and hydrolytic degradation during
melt processing (Scaffaro et al., 2014). Beltra n et al. (2018) found
that the reduction in viscosity of recycled PLA is minor without
washing the material, that is, hydrolytic degradation is more sig-
nificant than thermo-mechanical impacts. Since the washing step is
not included in this experimental setting, see Fig. 1, the observation
shows that without introduction of hydrolytic degradation,
thermo-mechanical degradation in 3D printing is quite significant,
could seriously limit the recyclability of PLA. The rapid declination
in the value of viscosity (see Fig. 1) well explains why the PLA
material cannot be subjected to further FDM 3D printing cycles,
therefore the rheological performance might be able to serve as an
indicator of 3D printability.
3.1.3. Changes in molecular structures
The diminution of mechanical and rheological performance
could be attributed to chain scission during recycling, and the GPC
results shown in Table 3 confirm the chain scission had been taken
places in the recycled samples. Further, the value of polydispersity
coefficient (d, defined as Mw/Mn) is increasing with the number of
3D printing cycles, which indicates the intensified diffusion and
randomisation of the polymer chains during reprocessing cycles.
Fig. 2. Viscosity results of PLA that processed for different cycles.
1050 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055

Table 3 Table 4
GPC results of PLA that processed for different cycles. Thermal results of PLA that processed for different cycles.

PLA (0) PLA (1) PLA (2) PLA (0) PLA (1) PLA (2)

Mn 50,906 42,668 15,901 DSC Tg ( C) 59.86 59.47 59.11

Mw 106,963 89,460 42,037 Tc ( C) 123.69 104.34 98.32
d 2.10 2.10 2.64 Tm ( C) 165.64 168.39 167.55
DHm (J g
1) 36.69 39.79 50.71
DHcc (J g
1) 37.78 19.95 11.89
Х (%) 1.17 19.87 38.06
The FTIR-ATR spectra of PLA (0), PLA (1) and PLA (2) are shown TGA 10 wt.% loss ( C) 341 339 330
in Fig. 3, in the region between 400 and 4000 cm
1. The band 50 wt.% loss ( C) 362 360 356
assignment for PLA is shown in the work of Kister et al. (1998), 90 wt.% loss ( C) 376 374 371

which is not significantly different from Fig. 3. Compatible to the

observations of Scaffaro et al. (2014) and Beltra n et al. (2018), the
region of 1750 cm
1 that corresponds to the stretching of the CO
bonds of the ester groups (Kister et al., 1998) is deviated from
noticeably. The bond has been recognised as an indicator of the
degradation of PLA, as this implies the appearance of new carbonyl-
linked species and could be the sign of reduction in molecular
weight (Badia et al., 2012). Fig. 3 shows that there are some
noticeable differences between virgin material (PLA (0)) and
reprocessed ones (PLA (1) and PLA (2)), indicating that certain
amounts of new carbonyl compounds have been generated during
reprocessing, and the volumes are proportional to the number of
reprocessing cycles. Apart from the region of 1750 cm
1, the vari-
ation of the absorption bands from 700 to 2000 cm
1 can also be
regarded as a sign of degradation of PLA during repeated 3D
printing. These changes could be attributed to the presence of
shorter polymer chains in the recycled materials, which may rear-
range into lamellar structures (Beltran et al., 2016).
3.1.4. Changes in thermal properties
Thermal properties that measured by DSC and TGA are sum-
marised in Table 4. It can be seen that all three samples present a
glass transition (Tg) around 60  C, similar to the previous report
(Beltran et al., 2018). The reduction in Tc can be attributed to a
higher mobility of the polymer chains, as a consequence of reduced
molecular weight during reprocessing cycles (see Pillin et al., 2008;
Scaffaro et al., 2011; Beltr an et al., 2018). Chain scission and
reduction in molar mass have been demonstrated by the results of
viscosity and molecular weight tests, as discussed previously.
Table 4 shows that there are no major differences in the melting
temperatures (Tm) of the three samples. Owning to higher molec-
ular weight, the mobility of polymer chains in PLA (0) is limited,
which prevents them to rearrange during the melting analytical
process, therefore the sample has the lowest Tm.
Concerning the crystallinity (Х ), the value tends to increase
significantly with the reprocessing cycles, as shown in Table 4.
Fig. 3. FTIR-ATR spectra of PLA that processed for different cycles.
Compared to the previous observations on repeated extrusion (see
_
Zenkiewicz n et al., 2016;
et al., 2009; Sikorska et al., 2012; Beltra
Beltran et al., 2018) or injection moulding (see Badia et al., 2012),
the degradation during 3D printing is much more significant in
fewer cycles, as suggested by the increasing crystallinity (Х ) and
melting enthalpy (DHm) and the decreasing cold crystallisation
enthalpy (DHcc). Since a single cycle of extrusion-based recycling
would not significantly affect the thermal properties (Nascimento
et al., 2010), this thermal degradation can be attributed to the 3D
printing process. 3D printing resulted in shorter polymers chains,
which are able to crystallise more easily and forms more crystals
during thermal processing. These thermal results show a good
agreement with the results of viscosity and molecular weight tests.
Table 4 reports the temperatures of weight losses of 10 wt.%,
50 wt.% and 90 wt.%. These results indicate that reprocessing
lowers the thermal stability of materials, which is also attributed to
the reduction in molecular weight, as shorter polymer chains are
responsible for this thermal degradation (Carrasco et al., 2010).
3.1.5. Changes in surface morphology
Fig. 4 shows the SEM micrographs of the fractured surfaces of
the three samples. The formation of small pinholes can be observed
on the surfaces of the reprocessed samples, and the number and
average size of pinholes in PLA (2) seemed larger than those of PLA
(1). The observation is similar to the findings of Sikorska et al.
(2012), yet the impacts of 3D printing seemed more significant.
3.2. Recycling of reprocessed material
Blending virgin PLA with recycled PLA is now becoming a
common measure for realising close-looped recycling. Hopmann
et al. (2015) found that addition of up to 50 wt.% recycled PLA
would not affect the mechanical and optical properties of PLA
packaging films. Tian et al. (2017) recovered 73 wt.% of PLA in PLA/
continuous-carbon-fibre composites that used in 3D printing.
Adopting the findings of this literature and our previous experience
in blending virgin-recycled plastics (Gu et al., 2014, 2016a), virgin
PLA was blended with reprocessed and shredded materials in
proportions of 20 wt.%, 50 wt.% and 80 wt.%, denoted as PLA (2)-
20 V, PLA (2)-50 V and PLA (2)-80 V, respectively, to investigate the
potential of close-looped recycling of 3D printed products.
Based on the results presented in the last subsection, it can be
found that most properties are somehow correlated, for example,
reduced molecular weight is responsible for changes in rheological,
mechanical and thermal properties. With the objective to select the
representative property as a performance indicator, that is,
threshold for acceptance of materials, cluster analysis was applied
in the similar manner of the previous works (Gu et al., 2016b). The
results of the selected performance indicator, the viscosities of the
materials are plotted in Fig. 5.
Compared to the results of PLA (2) that shown in Table 2, the
 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055
(a)
Fig. 4. SEM images of fractured surfaces of filaments at magnification of 10 k: (a) PLA (0), (b) PLA (1), and (c) PLA (2).
Fig. 5. Viscosity results of PLA that processed for different cycles.
addition of virgin PLA has significantly improved the viscosities of
the blends, as all the three blends exhibit higher viscosity values
than that of PLA (2). The viscosity results of PLA (2)-50 V and PLA
(2)-80 V are even higher than that of PLA (1), indicating both the
materials can be used in 3D printing. Adding virgin polymer is
essential for close-looped recycling of PLA that repeatedly used in
3D printing.
3.3. Assessments of disposal routes
3.3.1. Definition of goal and scope
The goal of this sector was to quantify and compare the envi-
ronmental impacts associated with different routes of disposing 3D
printed products, which, in this case, close-looped recycling,
incineration and landfill. The selection of disposal routes was in
accordance with the previous literature (Gu et al., 2017, 2018), as
these treatments are typical end-of-life options for polymeric
materials.
In this study, unlike the previous research that used mechanical
performance to determine materials of equivalent mass
(Bensadoun et al., 2016; Gu et al., 2018), the function unit (FU) was
set as 1 kg of 3D printable PLA, regardless virgin or recycled. The
choice was made on the basis of the experimental results, as the
rheological property (i.e., viscosity) is selected as the representative
performance indicator, as other properties show a good agreement
with the declining value of viscosity, excepting for mechanical
performance which shows little differences during reprocessing.
Besides, the FDM 3D printing is frequently applied in some less-
demanding fields such as prototyping (Ngo et al., 2018), where
simple tests are more preferable than sophisticated ones. Since
rheological tests are more convenient than molecular, thermal and
morphological characterisations, selecting viscosity as the repre-
sentative performance indicator is more suitable for FDM 3D
printing. Based on the experimental results, it was assumed that
1051
(b) (c)
the 3D printed PLA material can only be recycled once. To further
simply the comparative LCA, the virgin-recycled PLA blends were
not included, as well as transportation, since the goal of this work is
to quantitatively compare the environmental impacts of three end-
of-life options for one FU of PLA. Adopting the LCA settings in the
previous literature (Bensadoun et al., 2016; Gu et al., 2018), the goal
and scope of this assessment on environmental performance of the
three disposal procedures are defined as follows:
(1) Close-looped Recycling: the environmental savings (or
burdens) in this end-of-life route are defined as the negative
(or positive) values of the categorised impacts associated
with substituting virgin PLA filament with recycled PLA
filament. The processes of fabricating virgin PLA filament
include PLA polymerisation, drying and extrusion, and the
processes of manufacturing recycled PLA filament are
shredding, drying and extrusion. Focusing on the material
substitution, the environmental impacts of 3D printing itself
are not included.
(2) Incineration: the environmental savings achieved in incin-
eration are defined as the negative values of the environ-
mental burdens associated with generating certain amount
of electric power in the current power mix that equals to the
energy recuperated from combustion of one FU of PLA. The
measurement of energy recuperation from PLA incineration
is described in the end of Section 2.4.
(3) Landfill: the environmental impacts of disposing PLA in
landfill.
3.3.2. Inventory for disposal options
Both primary and secondary data from different sources have
been used in this assessment. Primary data, including energy
consumption and mass losses, that related to 3D printing pro-
cessing and reprocessing (see Section 2.2) were measured in the
experiment. Secondary data is referred to the life cycle inventory
(LCI) datasets that provided by the Gabi software, which cover most
relevant processes, including PLA polymerisation, energy supply
and landfill. Table 5 compiles the data and assumptions used to
model the material disposal routes.
3.3.3. Environmental impact assessment
Based on the datasets and assumptions in Table 5, the normal-
ised scores of the ReCiPe midpoint indicators of disposing 1 kg of
the PLA material that used in 3D printing were calculated and
plotted in Fig. 6.
As shown in Fig. 6, close-looped recycling and incineration can
achieve significant environmental savings in the selected impact
categories, whereas landfill only results in environmental burdens.
The landfill’s scores are noticeable in the categories of HTP, PMFP,
FEP, FETP and TETP, which could possibly be the consequence of the
1052 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055

Table 5
LCI for disposing 1 kg PLA.

Disposal route Process Description of data and assumptions used

Close-looped PLA Dataset “DE: Polylactic acid” was used to denote the fabrication of virgin PLA pellets. The environmental savings of close-looped
recycling polymerisation recycling are equivalent to the values in this dataset subtracted by the summed environmental impacts of the whole reprocessing
procedure. The environmental benefits of material substitution are calculated in the same manner as the works of Gu et al. (2017) and
(2018). According to the dataset, virgin PLA is synthesised according to over-ring opening polymerisation that take places in deionised
water.
Drying Dataset “GLO: Air drying” was used, and the actual electric energy consumption was employed to replace the corresponding figures in
the dataset.
Extrusion Datasets “Plastic Extrusion” was used, and the actual production data of the extruding system (model Friend™ SJ25/25) was employed
to replace the corresponding figures.
Shredding Dataset “Shredding” was used, with the actual production data of the shredder (Friend™ plastic mill) was employed to replace the
corresponding figures.
Incineration Combustion Dataset “Waste incineration (plastics)” was used, and the actual calorific value (18.8 MJ kg
1) of PLA were employed to replace the
corresponding figure in the dataset.
Energy Dataset “CN: Electricity grid mix” was used to calculate the environmental savings based on the assumption that the thermal energy
generation released from incineration can be exploited for power generation, thermal power generation is a dominant route of power generation
in China and the efficiency is set at 45% based on the work of Meng et al. (2016).
Landfill Landfill Dataset “Plastic waste on landfill” was used.

(a)

(b)
Fig. 6. Environmental impacts associated with disposal of 1 kg PLA in three end-of-life options, that is, landfill, recycling and incineration; the positive values denote environmental
burdens, while the negative values indicate environmental savings. (a) Impact scores presented in relative to the highest values of the corresponding indicator, and (b) normalised
impact scores of the selected categories.

release of persistent organic pollutants (POPs) and particles avoided. This observation is comparable with the work of Gu et al.
(Melnyk et al., 2015). Specifically, close-looped recycling achieves (2018), in which landfill was identified as the least preferable route
the most significant environmental savings in the category of FEP, of disposing solid plastic wastes.
as the huge amount of water consumption in PLA polymerisation is Close-looped recycling leads in environmental savings of the
 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055 1053

Repeated 3D printing process can only last for two cycles, as sig-
nificant deteriorations were detected in the viscosity values, which
made the material unsuitable for further reprocessing. Though
mechanical performance deteriorated to a limited extent, signifi-
cant decreases in rheological property conformed with losses in
molecular weights. Chain scissions have triggered increases of
crystallinity and compromised thermal stability. Noticeable de-
teriorations have also been observed in the FTIR spectra and SEM
micrographs; the presence of carbonyl groups and surface pinholes,
indicating that thermomechanical degradation has taken places
during melt processing and it is responsible for decrements in the
molecular weight of PLA. Blending recycled PLA with virgin mate-
rial improved the viscosity, which facilitates the close-looped
recycling process. The environmental performance of close-
looped recycling was assessed using the ReCiPe LCA approach,
was then compared to the environmental impacts associated with
incineration and landfill. The normalised LCA results indicate that
the most significant environmental benefits can be achieved in the
close-looped recycling, while landfill is the end-of-life option with
the highest environmental burdens. According to the values of the
selected impact indicators, incineration can also be regarded an
environmentally friendly end-of-life alternative, however, it does
not allow recuperation of the materials and therefore makes no
contribution to the legal requirement of using recycled materials. In
sum, close-looped recycling can be a feasible solution of disposing
wastes generated from FDM 3D printing processes, a practise which
Fig. 7. Normalised single-score values of environmental impacts associated with particularly sound in the environmental aspect. Yet, the number of
disposal of 1 kg PLA in the three end-of-life options; the positive values denote envi- reprocessing cycles is limited due to significant thermomechanical
ronmental burdens, while the negative values indicate environmental savings.
deteriorations, but it can be remediated via adding virgin PLA. In
future research, more comprehensive and in-depth study of the
indicators of GWP, FDP, FEP, FETP, MEP, while incineration achieves thermomechanical degradation mechanism of 3D printing should
the highest environmental benefits in categories of ODP, HTP, PMFP, be considered.
POFP, METP, TAP, and TETP. The reduction of the impacts of ODP and
TAP is probably attributed to the electric power mix of China, where Acknowledgements
coal-fired power generation still holds a large share (Wang et al.,
2018). Comparatively speaking, combustion of PLA is more envi- This work is financially supported by National Natural Science
ronmentally friendly, yet economically unviable. The emissions Foundation of China (no. 51475420 & 51635006), the Fundamental
from the mechanical recycling process could lower its environ- Research Funds for the Central Universities of China (no.
mental performance in these categorised indicators. Still, the nor- 2018QNA4006), Industrial Technology Innovation and Industriali-
malised single-score values (see Fig. 7) confirm that the close- sation of Science and Technology Project (no. 2014A35001-2), and
looped recycling is the most sustainable alternative among the Zhejiang Provincial Natural Science Foundation of China (no.
three options, showing a good agreement with the conclusion of LZ18E050002).
Hahladakis et al. (2018). In addition, material recuperation is
becoming more desirable with the advent of more stringent envi- Nomenclature
ronmental legislations, for example, Extended Producer Re-
sponsibility in China requires the use of recycled materials in
manufacturing new products and the stipulated proportion of Abbreviations
recycled materials is set at 20 wt.% (Gu et al., 2017). On the basis of AM Additive manufacturing
above discussion, it can be concluded that close-looped recycling is ATR Attenuated total reflectance
the most suitable technique for disposing 3D printed products from DSC Differential scanning calorimetry
an environmental point of view, as supported by both the LCA re- FDP Fossil depletion potential
sults and the legal requirements. FDM Fused deposition modelling
From the economic perspective, close-looped recycling only FEP Freshwater eutrophication potential
requires energy cost of shredding, drying and extrusion (less than 1 FTIR Fourier transform infrared spectroscopy
USD for reprocessing 1 kg PLA), which are far less than the cost of FU Function unit
PLA pellets (18 USD per 1 kg virgin PLA, according to 3DprintWorks, GPC Gel permeation chromatography
2018). Besides, the fast processing speed can be another advantage GWP Global warming potential
of choosing close-looped recycling as primary disposal route of HTP Human toxicity potential
polymeric materials (Bensadoun et al., 2016). LCA Life cycle assessment
LCI Life cycle inventory
MEP Marine eutrophication potential
4. Conclusion METP Marine ecotoxicity potential
ODP Ozone depletion potential
In this study, the potential of close-looped mechanical recycling PLA Polylactic acid
the PLA material that used in 3D printing was investigated. PMFP Particulate matter formation potential
1054 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055

POFP Photochemical oxidant formation potential Gu, F., Zhang, W., Guo, J., Summers, P., Hall, P., 2017. From waste plastics to industrial
raw materials: a life cycle assessment of mechanical plastic recycling practice
POP Persistent organic pollutant
based on a real-world case study. Sci. Total Environ. 601e602, 1192e1207.
SEM Scanning electron microscopy Gu, F., Zheng, Y., Zhang, W., Yao, X., Pan, D., Wong, A.S.M., Guo, J., Hall, P.,
TAP Terrestrial acidification potential Sharmin, N., 2018. Can bamboo fibres be an alternative to flax fibres as mate-
TETP Terrestrial ecotoxicity potential rials for plastic reinforcement? A comparative life cycle study on poly-
propylene/flax/bamboo laminates. Ind. Crop. Prod. 121, 372e387.
TGA Thermogravimetric analysis Hahladakis, J.N., Velis, C.A., Weber, R., Iacovidou, E., Purnell, P., 2018. An overview of
chemical additives present in plastics: migration, release, fate and environ-
mental impact during their use, disposal and recycling. J. Hazard Mater. 344,
Symbols 179e199.
d Polydispersity coefficient Hervy, M., Evangelisti, S., Lettieri, P., Lee, K.-Y., 2015. Life cycle assessment of
nanocellulose-reinforced advanced fibre composites. Compos. Sci. Technol. 118,
F Filling density
154e162.
L Layer thickness Hopmann, C., Schippers, S., Ho €fs, C., 2015. Influence of recycling of poly(lactic acid)
P Printing speed on packaging relevant properties. J. Appl. Polym. Sci. 132 (9), 41532e41537.
Mn Number average molecular weight Jang, J., Park, J.Y., Gao, G., Cho, D.-W., 2018. Biomaterials-based 3D cell printing for
next-generation therapeutics and diagnostics. Biomaterials 156, 88e106.
Mw Weight average molecular weight Kim, Y., Yoon, C., Ham, S., Park, J., Kim, S., Kwon, O., Tsai, P.-J., 2015. Emissions of
n Number of cycles nanoparticles and gaseous material from 3D printer operation. Environ. Sci.
Tb Temperature of bed Technol. 49 (20), 12044e12053, 2015.
Kister, G., Cassanas, G., Vert, M., 1998. Effects of morphology, conformation and
Tc Crystallisation temperature configuration on the IR and Raman spectra of various poly(lactic acid)s. Poly-
Tf Temperature of liquefier mers 39 (2), 267e273.
Tg Glass transition temperature Kwon, O., Yoon, C., Ham, S., Park, J., Lee, J., Yoo, D., Kim, Y., 2017. Characterization
and control of nanoparticle emission during 3D printing. Environ. Sci. Technol.
Tm Melting temperature 51 (18), 10357e10368.
c Crystallinity Melnyk, A., Dettlaff, A., Kuklin  ska, K., Namiesnik, J., Wolska, L., 2015. Concentration
DHcc Cold crystallisation enthalpy and sources of polycyclic aromatic hydrocarbons (PAHs) and polychlorinated
biphenyls (PCBs) in surface soil near a municipal solid waste (MSW) landfill. Sci.
DHm Melting enthalpy
Total Environ. 530e531, 18e27.
DH∞ Melting enthalpy for 100% crystalline PLA Mendes, L., Kangas, A., Kukko, K., Mølgaard, B., Sa €€
am€anen, A., Kanerva, T.,
Ituarte, I.F., Huhtiniemi, M., Stockmann-Juvala, H., Partanen, J., Ha €meri, K.,
Eleftheriadis, K., Viitanen, A.-K., 2017. Characterization of emissions from a
References desktop 3D printer. J. Ind. Ecol. 21 (S1), S94eS106.
Meng, M., Mander, S., Zhao, X., Niu, D., 2016. Have market-oriented reforms
Andreoni, V., Saveyn, H.G.M., Eder, P., 2015. Polyethylene recycling: waste policy improved the electricity generation efficiency of China's thermal power in-
scenario analysis for the EU-27. J. Environ. Manag. 158, 103e110. dustry? An empirical analysis. At. Energ. 114, 734e741.
Azimi, P., Zhao, D., Pouzet, C., Crain, N.E., Stephens, B., 2016. Emissions of ultrafine Mülhaupt, R., 2013. Green polymer chemistry and bio-based plastics: dreams and
particles and volatile organic compounds from commercially available desktop reality. Macromol. Chem. Phys. 214, 159e174.
three-dimensional printers with multiple filaments. Environ. Sci. Technol. 50 Nascimento, L., Gamez-Perez, J., Santana, O.O., Velasco, J.I., Maspoch, M.L., Franco-
(3), 1260e1268. Urquiza, E., 2010. Effect of the recycling and annealing on the mechanical and
Badia, J.D., Ribes-Greus, A., 2016. Mechanical recycling of polylactide, upgrading fracture properties of poly(Lactic Acid). J. Polym. Environ. 18, 654e660.
trends and combination of valorization techniques. Eur. Polym. J. 84, 22e39. NatureWorks, 2018. Ingeo™ Biopolymer 4032D Technical Data Sheet. Available
Badia, J.D., Stromberg, E., Karlsson, S., Ribes-Greus, A., 2012. Material valorisation of from: https://www.natureworksllc.com/~/media/Technical_Resources/
amorphous polylactide. Influence of thermo-mechanical degradation on the Technical_Data_Sheets/TechnicalDataSheet_4032D_films_pdf. accessed on
morphology, segmental dynamics, thermal and mechanical performance. 13th March, 2018.
Polym. Degrad. Stabil. 97 (4), 670e678. Ngo, T.D., Kashani, A., Imbalzano, G., Nguyen, K.T.Q., Hui, D., 2018. Additive
n, F.R., Lorenzo, V., de la Orden, M.U., Martínez-Urreaga, J., 2016. Effect of
Beltra manufacturing (3D printing): a review of materials, methods, applications and
different mechanical recycling processes on the hydrolytic degradation of challenges. Compos. B Eng. 143, 172e196.
poly(l-lactic acid). Polym. Degrad. Stabil. 133, 339e348. Pillin, I., Montrelay, N., Bourmaud, A., Grohens, Y., 2008. Effect of thermo-
n, F.R., Lorenzo, V., Acosta, J., de la Orden, M.U., Urreaga, J.M., 2018. Effect of
Beltra mechanical cycles on the physico-chemical properties of poly(lactic acid).
simulated mechanical recycling processes on the structure and properties of Polym. Degrad. Stabil. 93 (2), 321e328.
poly(lactic acid). J. Environ. Manag. 216, 25e31. Pizzol, M., Laurent, A., Sala, S., Weidema, B.P., Verones, F., Koffler, C., 2017.
Bensadoun, F., Vanderfeesten, B., Verpoest, I., Van Vuure, A.W., van Acker, K., 2016. C. Normalisation and weighting in life cycle assessment: quo vadis? Int. J. Life
Environmental impact assessment of end of life options for flax-MAPP com- Cycle Assess. 22 (6), 853e866.
posites. Ind. Crop. Prod. 94, 327e341. Ragaert, K., Delva, L., Van Geem, K., 2017. Mechanical and chemical recycling of solid
Berman, B., 2012. 3-D printing: the new industrial revolution. Bus. Horiz. 55 (2), plastic waste. Waste Manag. 69, 24e58.
155e162. Rao, C., Gu, F., Zhao, P., Sharmin, N., GU, H., FU, J., 2017. Capturing PM2.5 emissions
Carrasco, F., Page s, P., Gamez-Perez, J., Santana, O.O., Maspoch, M.L., 2010. Pro- from 3D printing via nanofiber-based air filter. Sci. Rep. 7, 10366.
cessing of poly(lactic acid): characterization of chemical structure, thermal Sanchez, F.A.C., Boudaoud, H., Hoppe, S., Camargo, M., 2017. Polymer recycling in an
stability and mechanical properties. Polym. Degrad. Stabil. 95 (2), 116e125. open-source additive manufacturing context: mechanical issues. Add. Manuf.
Deroine , M., Le Duigou, A., Corre, Y., Le Gac, P., Davies, P., Ce
sar, G., Bruzaud, S., 2014. 17, 87e105.
Accelerated ageing of polylactide in aqueous environments: comparative study Scaffaro, R., Morreale, M., Mirabella, F., La Mantia, F.P., 2011. Preparation and recy-
between distilled water and seawater. Polym. Degrad. Stabil. 108, 319e329. cling of plasticized PLA. Macromol. Mater. Eng. 296 (2), 141e150.
Di Maria, A., Eyckmans, J., Van Acker, K., 2018. Downcycling versus recycling of Scaffaro, R., Botta, L., Passaglia, E., Oberhauser, W., Frediani, M., Di Landro, L., 2014.
construction and demolition waste: combining LCA and LCC to support sus- Comparison of different processing methods to prepare poly(lactid acid)e
tainable policy making. Waste Manag. (in press). hydrotalcite composites. Polym. Eng. Sci. 54 (8), 1804e1810.
3D printWorks, 2018. PLA Pellets for 3D Printing. Available from: https://3d-print- Sikorska, W., Richert, J., Rydz, J., Musioł, M., Adamus, G., Janeczek, H., Kowalczuk, M.,
works.com/products/pla-pellets-for-3d-printing. accessed on 13th March, 2018. 2012. Degradability studies of poly(l-lactide) after multi-reprocessing experi-
Gebler, M., Uiterkamp, A.J.M.S., Visser, C., 2014. A global sustainability perspective ments in extruder. Polym. Degrad. Stabil. 97 (10), 1891e1897.
on 3D printing technologies. Energy Pol. 74, 158e167. Stabile, L., Scungio, M., Buonanno, G., Arpino, F., Ficco, G., 2017. Airborne particle
Godoi, F.C., Prakash, S., Bhandari, B.R., 2016. 3d printing technologies applied for emission of a commercial 3D printer: the effect of filament material and
food design: status and prospects. J. Food Eng. 179, 44e54. printing temperature. Indoor Air 27 (2), 398e408.
Gu, F., Hall, P., Miles, M.J., Ding, Q., Wu, T., 2014. Improvement of mechanical Steinle, P., 2016. Characterization of emissions from a desktop 3D printer and indoor
properties of recycled plastic blends via optimising processing parameters us- air measurements in office settings. J. Occup. Environ. Hyg. 13 (2), 121e132.
ing the Taguchi method and principal component analysis. Mater. Des. 62, Stephens, B., Azimi, P., El Orch, Z., Ramos, T., 2013. Ultrafine particle emissions from
189e198. desktop 3D printers. Atmos. Environ. 79, 334e339.
Gu, F., Hall, P., Miles, M.J., 2016a. Performance evaluation for composites based on Stro3D, 2018. Stronghero3D 3D Printer Pla Filaments Splendid Rainbow Multicolor
recycled polypropylene using principal component analysis and cluster anal- Gradient 1.75mm 1kg(2.2lb). Available from: http://www.stronghero3d.com/
ysis. J. Clean. Prod. 115, 343e353. stronghero3d-3d-printer-pla-filaments-splendid-rainbow-multicolor-gradient-
Gu, F., Hall, P., Miles, N.J., 2016b. Development of composites based on recycled 1-75mm-1kg-2-2lb. accessed on 13th March, 2018.
polypropylene for injection moulding automobile parts using hierarchical Tian, X., Liu, T., Wang, Q., Dilmurat, A., Li, D., Ziegmann, G., 2017. Recycling and
clustering analysis and principal component estimate. J. Clean. Prod. 137, remanufacturing of 3D printed continuous carbon fiber reinforced PLA com-
632e643. posites. J. Clean. Prod. 142, 1609e1618.
 P. Zhao et al. / Journal of Cleaner Production 197 (2018) 1046e1055
Tiwary, P., Park, C.B., Kontopoulou, M., 2017. Transition from microcellular to
nanocellular PLA foams by controlling viscosity, branching and crystallization.
Eur. Polym. J. 91, 283e296.
Upadhyay, M., Sivarupan, T., El Mansori, M., 2017. 3D printing for rapid sand
castingda review. J. Manuf. Process. 29, 211e220.
Vilaplana, F., Karlsson, S., 2008. Quality concepts for the improved use of recycled
polymeric materials: a review. Macromol. Mater. Eng. 293 (4), 274e297.
Wang, X., Jiang, M., Zhou, Z., Gou, J., Hui, D., 2017. 3D printing of polymer matrix
composites: a review and prospective. Compos. B Eng. 110, 442e458.
Wang, J., Wang, R., Zhu, Y., Li, J., 2018. Life cycle assessment and environmental cost
1055
accounting of coal-fired power generation in China. Energy Pol. 115, 374e384.
Wu, P., Wang, J., Wang, X., 2016. A critical review of the use of 3-D printing in the
construction industry. Autom. ConStruct. 68, 21e31.
_
Zenkiewicz, M., Richert, J., Rytlewski, P., Moraczewski, K., Stepczynska, M.,
Karasiewicz, T., 2009. Characterisation of multi-extruded poly(lactic acid).
Polym. Test. 28 (4), 412e418.
Zhang, F., Wei, M., Viswanathan, V.V., Swart, B., Shao, Y., Wu, G., Zhou, C., 2017. 3D
printing technologies for electrochemical energy storage. Nano Energy 40,
418e431.