A process control and interlayer heating approach to reuse polyamide 12 powders and create parts with improved mechanical properties in selective laser sintering

Abstract

Capable of building high-quality, complex parts directly from digital models, selective laser sintering (SLS) additive manufacture (AM) is a core method of agile manufacturing. Polyamide 12 is the most commonly and successfully used polymer powders to date in SLS due to the conforming thermal behaviors of this thermoplastic polymer. State of the art technology produces a substantial amount of un-sintered powders after the manufacturing process. Failure to recycle and reuse these aged powders not only leads to economic losses but also is environmentally unfriendly. This is particularly problematic for powders close to the heat-affected zones that go through severe thermal degradations during the laser sintering processes. Limited procedures exist for systematically reusing such extremely aged powders. This work proposes a new process control method to maximize reusability of aged and extremely aged polyamide 12 powders. Building on a previously untapped interlayer heating, preprocessing, and mixing of powder materials, we show how reclaimed polyamide 12 powders can be consistently reprinted into functional samples, with mechanical properties even superior to current industrial norms. In particular, the proposed method can yield printed samples with 18.04 % higher tensile strength and 55.29 % larger elongation at break using as much as 30 % of extremely aged powders compared to the benchmark sample.

Full text

A process control and interlayer heating approach to reuse polyamide 12
powders and create parts with improved mechanical properties in selective
laser sintering
Feifei Yang
*
, Tianyu Jiang*, Greg Lalier
†
, John Bartolone†, Xu Chen*
‡
Abstract: Capable of building high-quality, complex parts directly from digital models, selective laser
sintering (SLS) additive manufacture (AM) is a core method of agile manufacturing. Polyamide 12 is the
most commonly and successfully used polymer powders to date in SLS due to the conforming thermal
behaviors of this thermoplastic polymer. State of the art technology produces a substantial amount of un-
sintered powders after the manufacturing process. Failure to recycle and reuse these aged powders not only
leads to economic losses but also is environmentally unfriendly. This is particularly problematic for
powders close to the heat-affected zones that go through severe thermal degradations during the laser
sintering processes. Limited procedures exist for systematically reusing such extremely aged powders. This
work proposes a new process control method to maximize reusability of aged and extremely aged
polyamide 12 powders. Building on a previously untapped interlayer heating, pre-processing, and mixing
of powder materials, we show how reclaimed polyamide 12 powders can be consistently reprinted into
functional samples, with mechanical properties even superior to current industrial norms. In particular, the
proposed method can yield printed samples with 18.04% higher tensile strength and 55.29% larger
elongation at break using as much as 30% of extremely aged powders compared to the benchmark sample.
Key words: Selective laser sintering; Powder reuse; Powder aging and degradation; Interlayer heating;
Sustainability
1. Introduction
Capable of processing almost any laser-absorbent materials including polymers, metals, ceramics and
composites, selective laser sintering (SLS) is one of the most well established and commonly used additive
manufacturing (AM) techniques to rapidly manufacture three-dimensional components [1-4]. Polymeric
powders, semicrystalline or amorphous, are the first and still the most widely applied materials in SLS [1,
*
Department of Mechanical Engineering, University of Washington, Seattle, WA, 98195, USA.
†
Unilever Research & Development, 45 Commerce Drive - Trumbull, CT, 06611, USA.
‡
Corresponding author.
Email addresses: yangff@uw.edu (Feifei Yang), tjiang19@uw.edu (Tianyu Jiang), Greg.Lalier@unilever.com
(Greg Lalier), John.Bartolone@unilever.com (John Bartolone), chx@uw.edu (Xu Chen‡)

5, 6]. Parts printed using amorphous polymer powders are partially consolidated, and consequently can be
useful for applications when the strength and durability of parts are not dominant [7]. On the contrary, parts
printed using semicrystalline polymer powders are fully consolidated with high mechanical strength and
effectively weakened warpage. Among the semicrystalline polymers, polyamide families are most popular
for SLS, and polyamide 12 dominates the market because of the capability to generate strong parts for
common applications [3, 4, 8]. To be more specific, parts printed using polyamide 12 powders have superior
mechanical properties than that of parts from amorphous polymers. The superior mechanical properties
originate from the thermal behaviors of the semicrystalline materials [9-11], including, e.g., a wide
processing window for sintering, high melting enthalpy, and high flowability [3, 12]. The existence of a
wide sintering window between melting onset and crystallization onset temperatures is beneficial for
maximizing part consolidation to get a fully dense part [13]. High melting enthalpy and high flowability
are necessary during laser sintering to get a homogeneous powder layer and to melt the powder locally with
high accuracy and reproducibility [2, 12].
Despite the popular applications of the polyamide 12 powders in SLS, the volume ratio of powders
that translate to parts is small: e.g., commonly 5% - 15% of the total powders in the build chamber. The
85% to 95% residual powders went through deteriorate physical and chemical degradations in the intricate
fabricating processes including preheating, sintering, cooling, and/or post treatment [3, 14], but have the
potential to be recycled and reused for further applications [3, 4, 15, 16]. However, the deteriorated powders
have reduced surface morphologies, larger and more complex molecular chains, decreased flowability, and
deteriorated mechanical and thermal properties, which make it challenging to reuse them directly [3, 4, 13,
15, 16]. Also, polyamide 12 powder is relatively expensive, priced around $150/kg in 2019 [14].
Abandonment of the residual polyamide 12 powders can cause not only economic losses but also
environmental pollution. Thus, reclaim and reuse are difficult yet necessary for a sustainable SLS AM.
Prior to reuse, characterizations, such as powder flowability, MVR and viscosity, are studied to exhibit
the differences between polyamide 12 new powders and reclaimed powders. Sufficient flowability is a
necessity in SLS to enable processing of powders and deposition of powder layers [17]. Many efforts have
been done to understand the static and dynamic flow behaviors of polyamide 12 powders, and to optimize
the flowability [17, 18]. Particle shape significantly impact powder flowability. High sphericity leads to
high flowability, potato-shaped particles the next, and rough-edged irregular particles exhibit the lowest
flowability [19]. MVR index reveals changing flowability of the powders due to degradation [3, 15, 20].
New polyamide 12 powders have higher MVR values, while reclaimed powders have lower MVR values
[20]. Besides, longer aging or multiple times of reusing lead to smaller MVR values. Reclaimed polyamide
12 powders have increased zero-shear viscosity compared to new powders [3, 4]. Moreover, a higher
temperature accelerates aging and increases zero-shear viscosity [4]. The increase of zero-shear viscosity

for reclaimed powders results from the post-condensation phenomena. It also increases molecular weight
of reclaimed polyamide 12 powders, as confirmed by Gel permeation chromatography (GPC) [4].
From there, relevant works on the reuse of the reclaimed polyamide 12 powders have been reported in
recent years. L. Feng et al. [14] reclaimed polyamide 12 from SLS and made the powders into filaments for
fused deposition modeling (FDM). L. Wang et al. [21] demonstrated a closed-loop recycling of
polyamide12 powder from SLS into milled carbon fiber/recycled polyamide12 composite filaments for
extrusion-based AM. The effect of prolonged storage at elevated temperatures on the isothermal
crystallization kinetics of polyamide 12 has been studied using Flash Differential Scanning Calorimetry
(DSC) experiments by F. Paolucci et al. [12]. Clarifying the aging mechanisms on thermal behavior,
coalescence behavior and the resulting crystallinity, microstructure and mechanical properties, and
investigating systematic aging mechanisms and microstructural evolution, P. Chen et al. [3] and S.
Dadbakhsh et al. [4] characterized the aging process of polyamide 12 powders in SLS. Dotchev et al. [15],
Wegner et al. [20] and Josupeit et al. [11] verified the decreased flowability of the reclaimed polyamide 12
powders compared to new ones through a melt volume rate (MVR) index, which was recommended as a
measure of the powder degradation rates. K. Wudy et al. [22, 23] investigated the influences of processing
time and temperature on aging effects of polyamide 12 in SLS. In particular, the references (i) studied
molecular changes and thermal property changes of polyamide 12 partcake material, and (ii) examined the
aging behavior of polyamide 12 in SLS including bulk characteristics and part properties such as porosity
and surface roughness. The effect of powder reuse on mechanical properties has been studied by R.D.
Goodridge et al. [24] and K. Wudy et al. [25], who pointed out the changes of tensile strength as well as
elongation at break in parts built from aged polyamide 12 powders. In addition, the relationships between
preheating temperature [20], energy density [20, 26], combined dwelling time between layers, energy
density [27], and part quality were studied.
The complex aging mechanism of polyamide 12 powders in laser-sintering involves a set of physical
and chemical processes. The reversible physical degradation here mainly leads to changing molecule order
or concentration, particle post-crystallization and agglomeration [20]. The chemical degradation, on the
other hand, is irreversible and predominant in the aging process. The chemical degradation changes such
polymer structures as chain scission, branching and cross-linking caused by oxidation, post-condensation
and hydrolyzes. Previous studies have revealed that thermal oxidation and post-condensation are two main
aging mechanisms in chemical degradation during SLS [4, 25]. The oxidation process is initiated by free
molecule radicals (hydrogen radical) emerging from the decomposition of the polyamide 12. After oxygen
addition and transfer, termination reaction occurs to form stable final products (e.g., hydroperoxide and
imide groups) and to complete the oxidation reaction. Apart from the thermal oxidation reaction, post-
condensation is another irreversible chemical degradation behavior of polyamide 12 that leads to property

changes in reclaimed powders [12, 25]. Here, the lengths of polymer molecular chain increase through
combinations of free radicals at high temperatures. As a result, the molecular weight, flowability and
viscosity of the reclaimed polyamide 12 change.
Although existing efforts have sought to understand the aging mechanisms and reuse of the degraded
polyamide 12 powders, the possibility and feasibility of reusing the residual polyamide 12 powders remain
not fully exploited. In particular, the reuse of the extremely aged polyamide 12 powders closes to the heat-
affected zones (HAZs)
§
has not been reported. This paper seeks to bridge the missing link and proposes a
new method, hereby referred to as active interlayer heating, to build parts with the reclaimed and extremely
aged polyamide 12 powders. We discuss a new process control approach to reuse the polyamide 12 powders
of different degradation levels, different mixing percentages, and different combinations with multiple layer
printing and superior mechanical properties. The result is that reclaimed polyamide 12 powders can be
consistently reprinted into functional samples, with mechanical properties comparable or even superior to
current industrial norms. In particular, the proposed method can yield printed samples with 18.04% higher
tensile strength and 55.29% larger elongation at break using as much as 30% of extremely aged powders.
2. Proposed active interlayer heating for reusing aged and extremely aged polyamide 12
powders
Figure 1 outlines the flowchart of the proposed approach. The main procedures include powder
collection, preprocessing, powder mixing, powder characterizations, parameter control, interlayer heating,
and part characterizations. Below, we discuss each procedure in detail.
§
Areas close to the laser-material interaction during sintering.

Powder samples
Powder mixing
Powder preprocess
Powder characterizations
Parameter control
Part characterizations
Interlayer heating based SLS
New polyamide 12 powder
Aged polyamide 12 powder
Extremely aged polyamide 12 powder
Sieving process
Pure powders
New-aged mixed powders
New-extremely aged mixed powders
New-aged-extremely aged mixed powders
Particle size/shape analysis
Scanning electron microscope (SEM)
Differential scanning calorimetry (DSC)
Fourier-transform infrared spectroscopy (FTIR)
X-Ray Diffraction (XRD)
Preheating temperature
Laser power
Laser speed
Scan spacing
Layer thickness
Layer numbers of samples
Tensile strength
Printing of pure powders
Printing of benchmark sample
Printing of mixed powders without interlayer heating
Printing of mixed powders with interlayer heating
Proposed method for recycling of aged and extremely aged polyamide 12 powders
Elongation at break
Young's modulus
Figure 1 Proposed method for reusing aged and extremely aged polyamide 12 powders
2.1. Materials sample preparation
2.1.1. Powder samples
In this study, we collected polyamide 12 powders purchased from EOS Corp. and reclaimed from
standard SLS processes on an EOS P 390 machine. The powders cover three levels of degradation: (i) new
powders without heat treatment, (ii) aged powders located far away from the HAZs during SLS and
currently reused in the industry, and (iii) extremely aged powders located close to the HAZs and not being

actively reused in SLS. Strong laser-material interaction occurs during the sintering process at or close to
HAZs, leading to complex aging and degradation of the powder materials. The intricate laser-induced
thermophysics change properties of polyamide powders such as particle sizes/shapes, microstructures,
thermal properties and mechanical properties. Consequently, extremely aged powders are more severely
degraded than the aged powders due to its closer location to the HAZs.
2.1.2. Powder preprocess
Different from new powders and aged powders, extremely aged powders went through severe
degradations induced by high temperature and intense laser-material interactions. Consequently, the
collected extremely aged powders clump together and suffer from severe aggregation, which adversely
impact the powder coating process. As shown in Figure 2 (a), new powders and/or aged powders could be
coated well with a smooth surface on the powder bed, which is conducive to part densification and
consolidation. Nevertheless, the extremely aged powders could not be coated smoothly because of the
existence of the aggregated large particles, as shown in Figure 2 (b), as a result of striking drop in flowability.
This was attributed to an increase in the molecular weight, originating from a cross-linking of the polymer
material, and an increase of particles of non-spherical, irregular shapes with featured rough edges in
extremely aged powders [4]. Preprocessing of extremely aged powders is therefore necessary for a smoother
coated surface and improved sintering behaviors.
In this work, a sieving process was applied to the extremely aged powder prior to printing. This process
was done in a fume hood using a sieve with the mesh size of 200 µm to grind the powders. The time for
sieving a batch of 500 ml powders is around 2 hours with preparation and post cleaning procedures. After
the sieving process, about 50 ml~80 ml heavy materials were captured. In all, 84%~90% of the processed
materials can be collected and reused.
(a) New/aged powders (b) Extremely aged powders
Figure 2 Powders coated on the powder bed

2.1.3. Powder mixing
Four different groups of powders were used in experiment (as shown in Figure 1): pure powders, mix
of new and aged powders, mix of new and extremely aged powders, and mix of new, aged and extremely
aged powders, to be specific. For the pure powders, 100% new, 100% aged, and 100% extremely aged
powders were prepared. For the mixed powder groups, powders were prepared with various mixing
percentages (Figure 5). Volume percentages are used in this paper to mix powders.
2.2. Powder characterizations
The particle size/shape distribution were measured using a PartAn 3-D particle size and shape analyzer,
as shown in Figure 3. The measurement range of the analyzer is from 22 microns to 35 millimeters, and it
is suitable for the nylon particles, whose average diameter is around 50 ~ 60 µm [28, 29]. New, aged, and
extremely aged polyamide 12 powder samples were examined using the analyzer. To get reasonable test
results, over 500,000 particles were measured in each experiment.
Figure 3 A 3D particle size and shape analyzer
Scanning electron microscope (SEM) was carried out to evaluate the variations in micro surface
morphologies of the collected powders in different degradation degrees. We used a Teneo LVSEM, and
coated the samples with a thickness of 3 nanometers of Pt/Pd to increase material conductivity. We then
examined these powders under different magnifications, to be specific, 200, 500, 2000 and 10,000.
To measure the thermal properties, including glass transition temperature, melting temperature,
melting enthalpy, crystallization temperature, and crystallization enthalpy, we conducted differential
scanning calorimetry (DSC) on the powder samples of different degradation degrees using a TA
Instruments DSC Q20. Heating and cooling cycles were performed between room temperature and 225 °C
with a rate of 10 °C/min under a nitrogen atmosphere. The DSC results were obtained as guidance for
experimental parameters when sintering aged or extremely aged powders.
To measure the differences of molecular components and the chemical microstructures between new
and reclaimed materials, we applied Fourier-transform infrared spectroscopy (FTIR) to measure infrared
absorption and emission spectra of the powders at different wavelengths. We used a Nicolet Magna-IR 560

FTIR instrument with the wavelength ranges from 6500 cm-1-100 cm-1 and with the spectral resolution of
0.35 cm-1, which is primarily used for polymer identifications. Wavelengths in the different infrared regions
can be absorbed by materials with specific molecular components or microstructures. Thus the ranges of
wavelengths absorbed are used to determine the sample molecular components or microstructures.
In this work, we use X-Ray Diffraction (XRD) to check crystalline properties of different polyamide
12 materials, focusing specifically on changes in polymer crystallinity, crystalline phases, crystallographic
structures, grain sizes and crystal orientations. XRD patterns were obtained on a D2 Phaser diffractometer
(Bruker, German) with CuKα radiation (λ = 1.54060 Å).
2.3. Parameter control
The parameters for new polyamide 12 powders failed to create parts with extremely aged powders. We
tested different printing parameters to obtain optimal results for different powders. Figure 1 shows six
parameters that impact significantly part qualities and were used in this study: preheating temperature, laser
power, laser speed, scan spacing, layer thickness, and layer numbers. These parameters were accurately
controlled by our in-house developed SLS testbed in experiments. In our study, the same six core parameters
are needed for consistent comparison. A variety of parameter settings are suitable for new powders, while
few parameter settings are applicable to the cases when using extremely aged powders due to the degraded
material properties. In particular, excessive laser power and energy density, along with a high preheating
temperature, lead to over melting and deteriorated part properties when sintering extremely aged powders.
The general principles to select the optimal parameter settings in this work are that (i) the parameters are
set as close to OEM recommendations for regular printing as possible, and (ii) the new, aged and extremely
aged polyamide 12 powders can all be successfully printed into parts using the same set of parameters. The
selected optimal parameter settings suitable for powders of different degradation levels are: 160 °C
preheating temperature; 3000 mm/s scan speed; 18 W laser power; 0.3 mm scan spacing; and 150 µm layer
thickness. In our testing, we printed 10-layer tensile bar samples for each powder combination.
2.4. SLS with interlayer heating
2.4.1. SLS machine
The SLS machine used in the paper is an in-house built open-configuration SLS AM machine/research
testbed, as shown in Figure 4. Compared to the black box commercial systems, it enables researchers to
access as well as manipulate key manufacturing process parameters, build online implementable process
models, and discover problems in the complex multi-physical laser sintering process. We designed the
machine using a Coherent GEM100A CO2 laser (the maximum laser power 100 W) and a Scanlab
intelliSCAN 14 scanner. The powder bed dimension designed is 250 mm (L) x 250 mm (W) x 150 mm (H)
with the layer thickness resolution of 20 µm and theoretical XY positioning resolution of 0.24 µm. The
designed typical scan speed is 3.75 m/s and the maximum scan speed is 30 m/s. The powder handling is

carried out by one feed cylinder, one build cylinder, and a recoating arm, which are driven by servo motor
and leadscrew transmission. It is featured with combined heating method: radiation heating with infrared
heaters above the powder bed (200 mm) and conduction heating with a mica heater underneath the powder
bed. The software used to control the machine is an in-house developed LabVIEW program integrated with
Scanlab RTC5 API, which can read G-code and send motion command to galvo scanner.
Figure 4 In-house built and customized SLS testbeds
2.4.2. Proposed SLS with interlayer heating
The SLS with interlayer heating proposed in this paper is a process that interlayer heating is applied to
the specimens being printed during the SLS process. Between the printing of every two layers, the proposed
process maintains the surface temperature of the powder bed the same as the preheating temperature (160 °C)
for a controlled period of time. We tested 0 second (no additional interlayer heating) and 60 seconds
interlayer heating to powders and powder mixtures with different degradation levels to explore the
influences of interlayer heating on part mechanical properties. Through the experiment results, it was
verified that 60 seconds interlayer heating can provide samples with enough heat and energy.
We explored the reusability of aged and extremely aged powders and the influences of interlayer
heating on part properties, as shown in Figure 5. The experiments were separated into six stages, with a
distinguish objective for each stage. In the experiment in stage 1, pure powders were used to verify the
feasibility of multi-layer printing with reclaimed powders. In stage 2, the currently available industrial reuse
combination, 50%-50% new-aged mixed powders were used to print the benchmark samples – samples
whose mechanical properties were regarded as the baselines in this research. The aim of stage 3 was to
check the part mechanical properties when using new-extremely aged mixed powders with the percentages
of extremely aged powders increasing. In stage 4, to examine the influences of interlayer heating on part
mechanical properties, all the mixed powder combinations in stages 2 and 3 were reapplied and were printed
with 60-second interlayer heating. By comparing the mechanical properties of samples in stages 2, 3, and

4, information on influences of interlayer heating on part properties were obtained. In stage 5, different
from the previous stages using the combinations of two kinds of powders, parts were printed with mixed
powders composed of all three types of powders to verify the feasibility and potential benefits of this
practice. To ensure part quality, we selected the combinations of 30% new - 30% aged - 40% extremely
aged powders and 30% new - 40% aged - 30% extremely aged powders. To examine the influences of
interlayer heating, the experiments in stage 5 were repeated with 60-second interlayer heating in stage 6.

Stage 1
Stage 2
Stage 3
100% new powder
100% aged powder
To verify the feasibility of multi-
layer printing with reclaimed
powders
No
No
50% new + 50% agedTo print the benchmark sample No
No
100% extremely aged powder No
To evaluate part properties when
using new- extremely aged mixed
powders
70% new + 30% extremely aged
60% new + 40% extremely aged
50% new + 50% extremely aged
40% new + 60% extremely aged
30% new + 70% extremely aged
20% new + 80% extremely aged
10% new + 90% extremely aged
No
No
No
No
No
No
Stage 4
60 s
To examine the influences of
interlayer heating when using two
kinds of powders
70% new + 30% extremely aged
60% new + 40% extremely aged
50% new + 50% extremely aged
40% new + 60% extremely aged
30% new + 70% extremely aged
20% new + 80% extremely aged
10% new + 90% extremely aged
60 s
60 s
60 s
60 s
60 s
60 s
Stage 5
30% new + 30% aged + 40%
extremely aged
30% new + 40% aged + 30%
extremely aged
To verify the feasibility of printing
with new-aged-extremely aged
powders
No
No
Stage 6
30% new + 30% aged + 40%
extremely aged
30% new + 40% aged + 30%
extremely aged
To examine the influences of
interlayer heating using three kinds
of powders
60 s
60 s
Stages Objectives Combinations Interlayer heating
50% new + 50% aged 60 s
Figure 5 Experiment plans about reusability of aged and extremely aged powders and influences of interlayer heating (The mixed
powders are in volume percentages)

2.4.3. Mechanisms of SLS with interlayer heating
Due to the existence of particles with high melting points, the reclaimed polyamide 12 powders are
more difficult to melt than new polyamide 12 powders. This has been verified through the hot stage
microscopy and the DSC test, where the aggregated spherulite structures and high melting point pieces
cannot melt during the regular sintering process [4]. Lack of coalescing polymer chains, insufficient
consolidation, partial densification and numerous unmolten particles can thus appear, severely degrading
the end mechanical properties. The proposed SLS with interlayer heating aims to provide the needed heating
energy to promote coalescence of polymer powders and improve part densification. In the mean time,
increased diversity of grain sizes in the mixed new-reclaimed powders can improve part tensile strengths
when using reclaimed polyamide 12 powders.
Post-crystallization and recrystallization are different phenomena related to powder aging. Due to post-
crystallization, crystallization ratio increases and spherulites grow in reclaimed powders. Recrystallization
refers to the transfer of crystalline structures. Both post-crystallization and recrystallization can occur
during the SLS process, although it remains unknown which one is dominant. The combined post-
crystallization and recrystallization could affect part microstructures such as crystalline ratios and crystal
sizes. Parts with small crystal sizes have increased flexibility and decreased brittleness, while parts with
larger crystal sizes are easier to break before the separation of crystals. This study will explore the influences
of interlayer heating and powder qualities on part microstructures when using reclaimed powder materials
(e.g. crystalline ratios and crystal sizes, and part elongation at break).
2.5. Part characterizations
Tensile tests were carried out on all the SLS fabricated samples in Figure 5. The samples were designed
based on ASTM standards and were stretched for each tensile measurement using an Instron 5869
Electromechanical testing system with a maximum load frame capacity of 50 kN equipped with Blue Hill
control software, as shown in Figure 6. A polishing treatment was done on both the top and bottom surfaces
of part to remove any skirmish un-sintered particles prior to the tensile testing for accurate measurement.
We used a disc-shaped polishing machine with the diameter of 35 cm and a rubber polishing pad in the
polishing treatment. The machine can rotate with running water above to clean the un-sintered particles on
the surface of the tensile bars by friction. After polishing to clean the samples for proper surface
measurements, tensile tests were conducted using an Instron 5869 Electromechanical testing system at a
testing speed of 5 mm/min. After the tensile test, the stress-strain curves, tensile strength, elongation at
break, and Young’s modulus of various parts were reported.

Figure 6 Tensile bar dimensions and tensile test machine
3. Results and discussions
3.1. Powder test results
3.1.1. Particle size/shape distribution
The particle size distributions are shown in Figure 7. As seen, the percentages of the particles with the
diameter greater than 100 µm are 21%, 22%, and 24% for new powder, aged powder, and extremely aged
powder, respectively, showing a mild increase after the physical/chemical degradations. The most possible
explanation is that the elevated temperature and high-energy laser form an enclosed environment which
leads to edge melting of the un-sintered powders, and even aggravated aggregation, leading to the increase
of particle size.
(a) Tensile bar dimensions (unit: mm)
(b) Tensile test machine - Instron 5869

(a) New polyamide 12 powder
(c) Extremely aged polyamide 12 powder
(b) Aged polyamide 12 powder
Figure 7 Particle size distribution of polyamide 12 powders
The obtained particle shape distributions are shown in Figure 8. Sphericity is the measurement of how
close the shape of an object approaches that of a mathematically perfect sphere, of which the value is 1.00.
In this paper, to analyze the particle shape distribution, the particles with sphericity between 0.9~1.0 are
defined as particles with good sphericity. Figure 8 shows that the percentages of particles with good
sphericity for new, aged, and extremely aged powders are 21%, 16%, and 14%, respectively. The sphericity
of the reclaimed powder decreases mildly, which eventually increases part surface roughness. In other
words, the percentage of particles with non-spherical and irregular shapes or rough edges increases, which
is also a reason for the decreased flowability of the reclaimed powders.
Comparing to the new powders, a mild increase on the particle size distributions and a mild decrease
on the particle sphericity distributions of the reclaimed powders are part of the reasons for the aggregated
large particles and decreased flowability. In particular, a sieving process prior to printing is needed for the
severely degraded extremely aged powders.

(a) New polyamide 12 powders (b) Aged polyamide 12 powders
(c) Extremely aged polyamide 12 powders
Figure 8 Particle shape distribution of polyamide 12 powders
3.1.2. SEM
The SEM results of new, aged and extremely aged polyamide 12 powders are presented in Figure 9.
Each row shows the images of the same sample at different magnifications. Each column shows images of
different samples at the same magnifications. In Figure 9 (a), with the magnification of 200, little
differences were observed for the three samples. In the second column, few aggregated large size particles
were found in new powders, while there were more large size particles found in aged powders and extremely
aged powders as aggregation exacerbated. At a magnification of 2000, more apparent differences emerged
among different samples. There were no cracks in new powders, obvious cracks in aged powders, and more
cracks in extremely aged powder in the SEM scope. These cracks were attributed to the evaporation of
moisture or alcohol during the heating/sintering cycles or repeated expansions/shrinkages during multiple
times recycling. The SEM results in Figure 9 help us to understand the aging mechanisms of polyamide 12
powders and the effects of aging on different surface morphologies between new powders and reclaimed
powders.

Figure 9 SEM results of polyamide 12 powders
3.1.3. DSC
Thermal characteristics of new, aged and extremely aged polyamide 12 powders tested by DSC
instrument are shown in Table 1, (a) Heating cycle and (b) Cooling cycle. Each sample went through two
heating cycles and one cooling cycle. In the first heating cycle, the onset melting temperatures decrease
from new powder to extremely aged powder, which is important knowledge for parameter settings of SLS
printing when the aged or extremely aged powders are used. The setting principle of the preheating
temperature is that the preheating temperature should be set close to the material onset melting temperature
or melting point [3, 20, 30]. The decrease of onset melting temperature of aged and extremely aged powders
gave us significant instructions when printing aged or extremely aged powder because the appropriate
setting of preheating temperature is dominant to keep the sintering processes going and to ensure the part
quality. A higher preheating temperature leads to unnecessary powder melting in the chamber, and a lower
preheating temperature decreases the coalescence behaviors of particles. Furthermore, the decrease of the
onset melting temperatures is consistent with the conclusions that aging can enlarge the melting intervals,
which is caused by post-condensation. In the first heating cycle, the increase of the melting temperatures
New powder
Aged powder
Extremely aged powder
New powder
Aged powder
Extremely aged powder
New powder
Aged powder
Extremely aged powder
(a) 200 magnification (b) 500 magnification (c) 2000 magnification

and melting enthalpies of reclaimed powders originates from the post-crystallization [16]. This is a process
starting with nucleation, followed by spherulite growth. As temperature decreases, spherulites grow and
radiate from the center of the core, forming spherical shapes [31] and consequently leading to the formation
and aggregation of crystallites. These changes affect the melting temperatures and melting enthalpies of
reclaimed powders.
Table 1 Thermal characteristics of new, aged and extremely aged polyamide 12 powders
(a) Heating cycle
Polyamide 12
powder
Heating process
Onset melting
temperature/°C
Peak melting
temperature/°C
Melting enthalpy/J·g-1
New powder
The first heating cycle
183.36
186.38
97.60
The second heating cycle
172.48
177.55
41.25
Aged powder
The first heating cycle
181.81
188.18
104.7
The second heating cycle
171.38
176.97
41.42
Extremely aged
powder
The first heating cycle
180.77
187.71
103.3
The second heating cycle
170.82
176.62
35.69
(b) Cooling cycle
Polyamide 12 powder
Cooling process
Onset crystallization
temperature/°C
Peak crystallization
temperature/°C
Crystallization
enthalpy/J·g-1
New powder
Cooling cycle
152.14
148.23
50.15
Aged powder
150.40
146.48
48.42
Extremely aged powder
150.97
147.02
46.78
3.1.4. FTIR
FTIR test results of polyamide 12 powders are presented in Figure 10, and the infrared bands and the
corresponding assignments are exhibited in Table 2. In polyamides, CH bonds close to nitrogen are the
weakest bonds, thus most oxidation reactions proceed on these carbons [31, 32]. In the FTIR spectra, the
existence of peaks at 1368, 1158, 1062 and 946 cm−1, and the corresponding dramatically diminished
signals after printing reveal the oxidation reactions occurred on carbons close to nitrogen [3]. In Figure 10,
peaks at the vibrational frequencies of 1369.23, 1159.03, 1062.60 and 948.82 cm−1 slightly decreased in
aged powders and extremely aged powders compared to new powders, which was a sign of oxidation
reactions. Therefore, similar to the material aging after the sintering process, the aging mechanisms of the
recycled powders impacted by the existence of laser and high temperature are thermal oxidations.
In the long molecular chains, every 12-carbon can be regarded as a polyamide 12 unit with only one
carbon and oxygen double bond, forming the basic molecular structure of the new polyamide 12 powders.
In each polyamide 12 unit, C-H bonds in the methylene groups adjacent to the nitrogen are the weakest
bonds. Free hydrogen molecule radicals can emerge from these carbons in the presence of laser radiations
to initiate the oxidation process. During the propagation process of oxidation reactions, unavoidable oxygen
residual in the chamber (2% - 5% in state of the art [4]) gets absorbed, forming unstable middle products
or structures such as peroxide radicals and peroxides. After the termination of the oxidation reaction with

stable final products, two carbon and oxygen double bonds exist in every polyamide 12 unit due to the
addition of oxygen in reclaimed powder in contract to the new powders [3].
Due to the oxidation reaction, the molecular structures of reclaimed polyamide 12 powders are more
complex and stable. It is consistent with the conclusion from the DSC test that the melting temperatures
and melting enthalpies of reclaimed powders increase. After the oxidation reaction, or the addition of
oxygen to the molecular chains of polyamide 12 powders, the molecular weight of the materials may
increase. It has been verified that long build time and high build temperature leads to a molecular weight
increase of reclaimed polyamide 12 powders [22].
Table 2 Infrared bands and the assignments
Vibrational frequency (cm−1)
Assignments
3290.02
N-H stretching vibration
3095.24
Fermi resonance of the ν(N-H) stretching
2915.89
CH2 asymmetric stretching
2846.46
CH2 symmetric stretching (ordered)
1637.30
Amide-I (mostly of the ν(C=O) stretches)
1459.87
CH2 reference band
1369.23
CH bend, CH2 twisting
1267.02
Amide III (C-N stretching + C=O in-plane bending)
1193.74
Splitting of amide II (CH2 wagging or CH2 twisting)
1159.03
Skeletal motion involving CONH (am, γ)
1062.60
Skeletal motion involving CONH
948.82
CONH in-plane
709.69
CH2 rocking
620.98
Amide VI (N-H out-of-plane bend)

(a) New polyamide 12 powder
(b) Aged polyamide 12 powder
(c) Extremely aged polyamide 12 powder
Figure 10 FTIR results of polyamide 12 powders
3.1.5. XRD
XRD test results of polyamide 12 powders are shown in Figure 11, indicating crystalline structures of
the new and deteriorated polyamide 12 powders. Polyamide 12 exhibits two crystal structures, α and γ
phases, and usually γ acts as its stable phase. As shown in Figure 11, the XRD test results of new, aged,
and extremely aged polyamide 12 powders presented similar behaviors and are in different colors. The
XRD peak at 10.98° was selected as a reference of crystallization. These curves all exhibited α-structures
with very high crystallinities, unstable structures and oriented in an anti-parallel manner, which were
consistent with reference results [3, 4].

α phase
- New polyamide 12 powder
- Aged polyamide 12 powder
- Extremely aged polyamide 12 powder
Figure 11 XRD test results of polyamide 12 powders
3.2. Printing results
Based on the proposed method, the tensile bars with 10 layers were printed. Using differently degraded
materials, the same series of optimized parameters were applied to all the designed experiments in Figure
5 to ensure the equivalent processing conditions. The printing results are presented in Table 3.
Table 3 Printing results about reusability of extremely aged powders and influences of interlayer heating
Stages
Powder percentages
Interlayer heating
Sample number
Thickness / mm
Stage 1
100% new powder
No
3
0.88; 0.82; 0.81
100% aged powder
No
2
0.91; 0.91
100% extremely aged
No
2
0.65; 0.68
Stage 2
50% new + 50% aged
No
3
1.41; 1.41; 1.40
Stage 3
70% new + 30% extremely aged
No
2
1.46; 1.45
60% new + 40% extremely aged
No
2
1.25; 1.25
50% new + 50% extremely aged
No
3
0.98; 0.97; 0.96
40% new + 60% extremely aged
No
3
1.43; 1.42; 1.41
30% new + 70% extremely aged
No
2
1.38; 1.36
20% new + 80% extremely aged
No
2
1.36; 1.34
10% new + 90% extremely aged
No
3
0.80; 0.78; 0.78
Stage 4
50% new + 50% aged
60 s
3
1.41; 1.41; 1.42
70% new + 30% extremely aged
60 s
2
1.48; 1.47
60% new + 40% extremely aged
60 s
3
0.82; 0.83; 1.12
50% new + 50% extremely aged
60 s
3
1.42; 1.40; 1.40
40% new + 60% extremely aged
60 s
2
1.36; 1.36
30% new + 70% extremely aged
60 s
3
1.32; 1.35; 1.30
20% new + 80% extremely aged
60 s
2
1.00; 1.00
10% new + 90% extremely aged
60 s
3
1.22; 1.26; 1.14
Stage 5
30% new + 30% aged + 40% extremely aged
No
3
1.40; 1.38; 1.40
30% new + 40% aged + 30% extremely aged
No
2
1.34; 1.35
Stage 6
30% new + 30% aged + 40% extremely aged
60 s
3
1.40; 1.39; 1.36
30% new + 40% aged + 30% extremely aged
60 s
2
1.26; 1.33
* Preheating temperature, 160 °C; scan speed, 3000 mm/s; laser power, 18 W; scan spacing, 0.3 mm; layer thickness, 150 µm.

Though we used the same set of parameter settings to print the tensile bar samples, we recoated
different combinations of powders to print different samples. Thus, there are sample variations in, e.g.,
coalescence behaviors, solidification and consolidation effects, yielding variable thicknesses in the
generated tensile bars. Such variations, however, do not affect the measured normalized mechanical
properties. The width of tensile bars is in good consistence with the dimensions designed in Figure 6a. Note
that a few samples (e.g. Figure 12f) contain deformation in pictures. The reason is that we air cooled all
samples outside the chamber after printing, instead of waiting for them to cooldown inside the chamber.
Pictures of tensile bar samples are shown in Figure 12. The samples printed using 50%-50% new-aged
mixed powders are taken as the benchmark parts. Samples printed using powders of different mixing ratios
are listed, with and without interlayer heating. As seen, all samples were successfully 3D printed with no
visible differences on the sample surfaces. More part characterizations will be explained in the following
sections.
(a) Benchmark tensile bars with and without interlayer heating
(b) Tensile bars with 30% and 40% extremely aged powders

(c) Tensile bars with 50% and 60% extremely aged powders
(d) Tensile bars with 70% extremely aged powders
(e) Tensile bars with 80% extremely aged powders

(f) Tensile bars with 90% extremely aged powders
(g) Tensile bars with 30% new + 40% aged + 30% extremely aged powders
Figure 12 Pictures of some tensile bar samples
3.3. Part test results
3.3.1. Stage 1
The stress-strain curves of samples printed using pure powders are exhibited in Figure 13. As shown,
there are several differences between these samples. The average tensile strength of samples printed using
new powders, aged powders and extremely aged powders are respectively 22.96 Mpa, 18.12 Mpa, and
11.08 Mpa, and that of samples using aged and extremely aged decreased by 21.09% and 51.75% compared
to new ones. As explained in Section 2.4.3, parts using more new powders have larger tensile strength under
standard settings. The average Young’s modulus of samples printed using new powders, aged powders, and
extremely aged powders are respectively 503.67 Mpa, 358.50 Mpa, and 177.00 Mpa, and that of samples
using aged and extremely aged decreased by 28.82% and 64.86% compared to new ones. However, the
average elongations at break of samples using new, aged, and extremely aged powders are respectively
5.31%, 12.12%, and 9.56%, and that of samples using aged/extremely aged powders increase by 56.20%
and 44.46% compared to new SLS parts. The reason for this is attributed to that the reclaimed materials
have smaller crystal size, increased flexibility and decreased brittleness. Inversely, parts using new powders

with larger crystal size are easier to break down before the separation of crystals, with decreased flexibility.
Using the Scherrer’s equation and XRD test results, we calculated the crystal sizes of the printed parts in
Figure 13. The crystal sizes of new parts, parts using aged powders and parts using extremely aged powders
are respectively 5.21 nm, 3.39 nm and 4.53 nm, which are consistent with our previous conclusions.
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
11.07 Mpa
11.08 Mpa
17.79 Mpa
18.45 Mpa 21.20 Mpa
23.60 Mpa
Sample 1 of extremely aged powders
Sample 2 of extremely aged powders
Young's modulus: 322 Mpa; elongation: 12.35 %
Young's modulus: 395 Mpa; elongation: 11.88 %
Young's modulus: 502 Mpa; elongation: 5.42 %
Young's modulus: 506 Mpa; elongation: 5.42 %
Stress/Mpa
Strain/%
Sample 1 of new powders
Sample 2 of new powders
Sample 3 of new powders
Sample 1 of aged powders
Sample 2 of aged powders
Young's modulus: 503 Mpa; elongation: 5.08 %
Young's modulus: 178 Mpa; elongation: 9.38 %
Young's modulus: 176 Mpa; elongation: 9.73 %
24.07 Mpa
Figure 13 Stress-strain curves of samples printed using pure powders
3.3.2. Stage 2
As the most popular used mixing percentages for recycling aged powders [4], the samples printed
using the 50%-50% new-aged mixed powders are taken as the benchmark samples in this paper, of which
the mechanical properties are baselines in the proposed method. The stress-strain curves of the benchmark
samples are shown in Figure 14. There are two breakpoints in the stress-strain curves for these samples due
to the layered fabrication process. For sample 1, there is a breakpoint at the strain of 7.82%, and another at
11.57%, indicating that layered fracture occurred during the tensile tests. This is a phenomenon that
multiple layers in one tensile bar break down at different times. We recoated different combinations of
mixed powders to print different samples. The mixed powders affect sample solidification and
consolidation between layers and results in the layered fracture in the parts using mixed powders.
Calculating the average mechanical properties of the benchmark samples, the baselines of tensile strength,
Young’s modulus and elongation at break are respectively 25.80 Mpa, 568 Mpa, and 11.36%. The average
mechanical properties of these samples will be used to evaluate the mechanical properties of the tensile bars
in the following Stages.

0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
23.18 Mpa
26.15 Mpa
Benchmark samples
(9.15, 23.18)
(4.04, 20.66)
(12.93, 25.08)
(6.33, 26.15)
(11.57, 26.44)
Stress/Mpa
Strain/%
Sample 1
Sample 2
Sample 3
50% new powder + 50% aged powder
(7.82, 28.19)
28.19 Mpa
Figure 14 Stress-strain curves of benchmark samples
3.3.3. Stage 3
In stage 3, samples using new-extremely aged mixed powders without interlayer heating were printed,
and the stress-strain curves of these samples are shown in Figure 15 (a) - (g). With the percentages of
extremely aged powders increasing from 30% to 90%, the average tensile strengths are 24.69 Mpa, 29.18
Mpa, 25.32 Mpa, 25.02 Mpa, 20.93 Mpa, 22.65 Mpa and 29.97 Mpa, and the average elongations at break
are 7.10%, 7.76%, 8.60%, 8.43%, 12.33%, 13.38% and 15.36%. The average tensile strength of parts from
the combination of 60%-40% new-extremely aged mixed powders to the combination of 30%-70% new-
extremely aged mixed powders show a decrease. But this did not apply to the remaining mixing percentages,
which can be ascribed to that the mechanical properties of samples are not only related to the properties of
powders but also relevant to the thermal or laser conditions in the sintering chamber. It is worth noting that
the tensile strengths of 10%-90% new-extremely aged samples appear to be better than the benchmark
samples. Due to the severely decreased flowability of the extremely aged powders, the recoating of the
extremely aged powders is difficult. The large percentage of the extremely aged powders in the 10%-90%
new-extremely aged mixed powders accompanied with sustained high temperature in the SLS chamber
failed the recoating process during the printing of the last few layers. And the obtained top layers were
sintered multiple times with no powders recoated, which has the similar effects with printing using
improved energy densities. High energy densities lead to high mechanical properties. This is the reason for
the large tensile strengths of 10%-90% new-extremely aged samples. Overall, when the percentages of new
powders decrease (from 60% to 20%) and the percentages of extremely aged powders increase (from 40%
to 80%), the average tensile strengths of samples decrease. However, the elongations at break increase with
the increasing of extremely aged powders because the microstructure changes.

0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
23.40 Mpa
Stress/Mpa
Strain/%
Sample 1
Sample 2
70 % new + 30 % extremely aged + no interlayer heating
25.99 Mpa
(7.10, 22.74)
(5.49, 23.40)
(7.10, 24.74)
(6.20, 25.99)
0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
27.67 Mpa
(8.37, 28.83)
(6.04, 30.69)
(7.15, 27.28)
60 % new + 40 % extremely aged + no interlayer heating
Stress/Mpa
Strain/%
Sample 1
Sample 2
(4.93, 27.67)
30.69 Mpa
(a) 70% new + 30% extremely aged (b) 60% new + 40% extremely aged
0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
25.27 Mpa
27.06 Mpa
(7.59, 23.63)
(3.49, 20.74)
(8.21, 24.08)
(5.43, 25.27)
(10.01, 25.15)
(6.71, 27.06)
50 % new + 50 % extremely aged + no interlayer heating
Stress/Mpa
Strain/%
Sample 1
Sample 2
Sample 3
23.63 Mpa
0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
25.62 Mpa
25.59 Mpa (8.15, 23.89)
(7.05, 25.62)
(7.99, 22.85)
(5.43, 23.85)
(9.16, 23.80)
(7.10, 25.59)
40 % new + 60 % extremely aged + no interlayer heating
Stress/Mpa
Strain/%
Sample 1
Sample 2
Sample 3
23.85 Mpa
(c) 50% new + 50% extremely aged (d) 40% new + 60% extremely aged
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
Stress/Mpa
Strain/%
Sample 1
Sample 2
30 % new + 70 % extremely aged + no interlayer heating
20.17 Mpa (11.96, 18.22)
(8.49, 20.17)
(12.69, 20.73)
(7.43, 21.69)
21.69 Mpa
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
23.65 Mpa
(14.49, 22.01)
(9.76, 23.65)
(12.27, 20.04)
Stress/Mpa
Strain/%
Sample 1
Sample 2
20 % new + 80 % extremely aged + no interlayer heating
(9.04, 21.66)
21.66 Mpa
(e) 30% new + 70% extremely aged (f) 20% new + 80% extremely aged

0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
31.85 Mpa
30.18 Mpa
Stress/Mpa
Strain/%
Sample 1
Sample 2
Sample 3
10 % new + 90 % extremely aged + no interlayer heating
27.89 Mpa
(g) 10% new + 90% extremely aged
Figure 15 Stress-strain curves of samples printed using new-extremely aged mixed powders without interlayer heating
3.3.4. Stage 4
Figure 16 compares stress-strain curves of benchmark samples and samples using new-extremely aged
mixed powders with and without interlayer heating. As shown, the tensile strengths of some samples remain
no significant changes after interlayer heating, while that of the other samples increase. In Figure 16 (a),
(b), (c), (e), (f), the tensile strengths of samples with 60-second interlayer heating increased by 25.19%,
36.10%, 30.13%, 5.46% and 22.51%. With 60-second interlayer heating, the average tensile strengths of
tensile bars printed using 50%-50% new-aged powder mix, 70%-30% new-extremely aged powder mix,
60%-40% new-extremely aged powder mix, 40%-60% new-extremely aged powder mix and 30%-70%
new-extremely aged powder mix are respectively 32.30 Mpa, 33.61 Mpa, 37.97 Mpa, 26.39 Mpa and 25.64
Mpa. With interlayer heating, tensile bars 3D printed using more new powders tend to have larger tensile
strengths. In total, the tensile strengths of samples using mixed powders can be improved to some extent
after interlayer heating because of the better melting and coalescence behaviors of particles on each layer.
In particular, in addition to the successful reusing of extremely aged powders, the proposed process control
yields parts with tensile strength 25.19% higher than default machine configuration using the standard
material combination (Figure 16a).

0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
25.80 Mpa
23.62 Mpa
25.58 Mpa
28.19 Mpa
31.22 Mpa
31.93 Mpa
Stress/Mpa
Strain/%
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
60 s interlayer heating + sample 3
no interlayer heating + sample 1
no interlayer heating + sample 2
no interlayer heating + sample 3
50 % new + 50 % aged
33.76 Mpa 32.30 Mpa
25.19 %
0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
33.61 Mpa
37.20 Mpa
30.02 Mpa
25.99 Mpa
70 % new + 30 % extremely aged
Stress/%
Strain/%
no interlayer heating + sample 1
no interlayer heating + sample 2
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
23.40 Mpa 24.70 Mpa
36.10 %
(a) 50% new + 50% aged (b) 70% new + 30% extremely aged
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
30.13%
29.18 Mpa
37.97 Mpa
42.26 Mpa
38.56 Mpa
33.10 Mpa
27.67 Mpa
Stress/Mpa
Strain/%
no interlayer heating + sample 1
no interlayer heating + sample 2
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
60 s interlayer heating + sample 3
60% new + 40% extremely aged
30.69 Mpa
0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45 50 % new + 50 % extremely aged
Stress/Mpa
Strain/%
no interlayer heating + sample 1
no interlayer heating + sample 2
no interlayer heating + sample 3
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
60 s interlayer heating + sample 3
(c) 60% new + 40% extremely aged (d) 50% new + 50% extremely aged
0246810 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
5.46%
26.39 Mpa
25.02 Mpa
26.90 Mpa
25.87 Mpa
23.85 Mpa
25.59 Mpa
40% new + 60% extremely aged
Stress/Mpa
Strain/%
no interlayer heating + sample 1
no interlayer heating + sample 2
no interlayer heating + sample 3
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
25.62 Mpa
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
20.93 Mpa
20.16 Mpa
21.71 Mpa 21.17 Mpa
27.10 Mpa
Stress/Mpa
Strain/%
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
60 s interlayer heating + sample 3
no interlayer heating + sample 1
no interlayer heating + sample 2
30 % new + 70 % extremely aged
28.64 Mpa 25.64 Mpa
22.51%
(e) 40% new + 60% extremely aged (f) 30% new + 70% extremely aged

0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
23.61 Mpa
21.66 Mpa 22.19 Mpa
Stress/Mpa
Strain/%
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
no interlayer heating + sample 1
no interlayer heating + sample 2
20 % new + 80 % extremely aged
22.71 Mpa
0 5 10 15 20
0
5
10
15
20
25
30
35
40
45
Stress/Mpa
Strain/%
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
60 s interlayer heating + sample 3
no interlayer heating + sample 1
no interlayer heating + sample 2
no interlayer heating + sample 3
10 % new + 90 % extremely aged
(g) 20% new + 80% extremely aged (h) 10% new + 90% extremely aged
Figure 16 Comparisons of stress-strain curves of benchmark samples and samples using new-extremely aged mixed powders with
and without interlayer heating
3.3.5. Stage 5 and Stage 6
Figure 17 compares stress-strain curves of samples using new-aged-extremely aged mixed powders
with and without interlayer heating. After 60 seconds interlayer heating, the tensile strengths of samples
using 30%-30%-40% new-aged-extremely aged mixed powders increase by 1.19%, and that of samples
using 30%-40%-30% new-aged-extremely aged mixed powders increase by 18.04%. It is concluded that
the tensile strengths of samples using new-aged-extremely aged mixed powders can be improved after
interlayer heating.
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
1.19%
31.77 Mpa
32.15 Mpa
32.35 Mpa
30.59 Mpa
33.50 Mpa
30.79 Mpa
32.70 Mpa
Stress/Mpa
Strain/%
No interlayer heating + sample 1
No interlayer heating + sample 2
No interlayer heating + sample 3
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
60 s interlayer heating + sample 3
31.82 Mpa
30% new + 30% aged + 40% extremely aged
0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
18.04 %
21.89 Mpa
25.84 Mpa
24.02 Mpa
27.66 Mpa
21.58 Mpa
no interlayer heating + sample 1
no interlayer heating + sample 2
60 s interlayer heating + sample 1
60 s interlayer heating + sample 2
Stress/Mpa
Strain/%
30% new + 40% aged + 30% extremely aged
22.20 Mpa
(a) 30% new + 30% aged + 40% extremely aged (b) 30% new + 40% aged + 30% extremely aged
Figure 17 Comparisons of stress-strain curves of benchmark samples and samples using new-aged-extremely aged mixed
powders with and without interlayer heating

3.3.6. Discussions
3.3.6.1. Tensile strength
Figure 18 compares tensile strengths of the samples: (a) without interlayer heating, (b) with 60 seconds
interlayer heating. The baseline here is the samples printed from the 50%-50% new-aged blend without
interlayer heating, i.e., 25.80 Mpa. From Figure 18 (a), the powder blends yielding larger tensile strengths
than the baseline are: 40%-60% new-extremely aged, 30%-30%-40% new-aged-extremely aged, and 10%-
90% new-extremely aged. However, when the percentages of extremely aged powders increased from 50%
to 80%, the average tensile strengths decrease. The reasons are that it is more difficult to move dislocations
through and between the grains when there are more extremely aged powders. Therefore, for the samples
without interlayer heating, the recommended mixing percentages closest to industry current practice is
60%-40% new-extremely aged and 30%-30%-40% new-aged-extremely aged (13.18% and 23.14% better
than baseline).
From Figure 18 (b), there are three blends with 60 seconds interlayer heating of which the tensile
strengths are larger than the baseline (25.80 Mpa): 50%-50% new-aged, 60%-40% new-extremely aged and
30%-30%-40% new-aged-extremely aged (25.19%, 47.17% and 24.69% better than baseline). Thus, the
recommended mixing percentages closest to industry current practice for the samples with interlayer
heating are 50%-50% new-aged, 60%-40% new-extremely aged and 30%-30%-40% new-aged-extremely
aged.
Tensile bars using reclaimed powders cannot coalesce well due to the high melting point particles. As
a result of the insufficient consolidation, partial densification and numerous unmolten particles, tensile
strengths of the 3D printed parts degrade. Thus, tensile bars using more new powders normally have larger
tensile strengths. In Figure 18b, the tensile bars using 60%-40% new-extremely aged mixed powders with
interlayer heating have more new powders and better coalescence behaviors than most of the other parts,
resulting in superior tensile strength. For the tensile bars using 70%-30% new-extremely aged mixed
powders with interlayer heating, the tensile strength is slightly lower than that of the parts using 60%-40%
powder blend. The reason is that the tensile bars using 60%-40% new-extremely aged mixed powders have
better densification than that using 70%-30% new-extremely aged mixed powders, which has been verified
from our SEM test results.

0 2 4 6 8 10 12 14 16 18 20
0
5
10
15
20
25
30
35
40
45
30% new + 30% aged + 40% extremely aged
30% new + 40% aged + 30% extremely aged
23.14%
Average:
29.20 Mpa Average: 31.77 Mpa
Stress/Mpa
Strain/%
50% aged + 50% new
30% extremely aged + 70% new
40% extremely aged + 60% new
50% extremely aged + 50% new
60% extremely aged + 40% new
70% extremely aged + 30% new
80% extremely aged + 20% new
90% extremely aged + 10% new
No interlayer heating
Average:
25.80 Mpa
Base line
13.18%
0 5 10 15 20
0
5
10
15
20
25
30
35
40
45
24.69%
25.19%
Average:
32.17 Mpa
Average: 37.97 Mpa
Stress/Mpa
Strain/%
50% aged + 50% new
30% extremely aged + 70% new
40% extremely aged + 60% new
50% extremely aged + 50% new
60% extremely aged + 40% new
70% extremely aged + 30% new
80% extremely aged + 20% new
90% extremely aged + 10% new
30% new + 30% aged + 40% extremely aged
30% new + 40% aged + 30% extremely aged
60 s interlayer heating
Average:
32.30 Mpa
Baseline
25.80 Mpa
47.17%
(a) Samples without interlayer heating (b) Samples with 60 s interlayer heating
Figure 18 Comparisons of sample tensile strengths
3.3.6.2. Elongation at break
Figure 19 compares sample elongations at break: (a) sample elongations at break, (b) mean values and
standard deviations of sample elongations at break. Similar to the tensile strength test, the baseline
performance is from the samples printed using 50%-50% new-aged blend without interlayer heating, i.e.,
11.36%. From Figure 19b, there are several blends that have similar or larger values of elongations at break
compared to the baseline. For the samples without interlayer heating (Figure 19b (i)), the samples using
more extremely aged powders have larger average elongations. The largest elongation is from the 10%-90%
new-extremely aged blend (the average value is 35.30% better than baseline), because increased extremely
aged powders yields smaller crystal size (verified by XRD) and increased flexibility. The recommended
mixing percentages of samples without interlayer heating are 30%-30%-40% new-aged-extremely aged
and 30%-40%-30% new-aged-extremely aged mixed blends. These combinations are close to industry
current practice, and the elongations are better than baselines. For the samples with interlayer heating
(Figure 19b (ii)), the largest elongation is from the 30%-40%-30% new-aged-extremely aged mixed blend
(the average value is 58.97% better than baseline), because interlayer heating enhances particle bonding
and microstructures. The recommended mixing percentage of samples with interlayer heating is 30%-40%-
30% new-aged-extremely aged mixed blend. The part elongation at break of this combination increases by
55.29% after interlayer heating (from 11.63% to 18.06%).
From Figure 19b, the samples without interlayer heating have relatively uniform standard deviations
on elongations at break. However, the standard deviations on elongations of samples with interlayer heating
are irregular. These samples with 70%, 80% and 90% extremely aged powders have larger standard
deviations than the remaining ones.

In general, parts using more reclaimed powders have increased elongations at break, while parts using
more new powders with larger crystal sizes are easier to break before the separation of crystals. Notice that
10%-90% new-extremely aged powder blends generated competing results regarding elongation at break.
The result is attributed to the fact that the reclaimed powders have smaller crystal size, increased flexibility
and decreased brittleness. However, considering combined mechanical strength and the large deviation
from industrial practice, this powder composition is not recommended for immediate industrial application.
Though parts using more reclaimed powders have smaller crystal size, increased flexibility and elongations
at break, powder mixing increase diversity of grain sizes in the mixed new-reclaimed powders, contributing
to decreased elongations at break.
Sample 1 Sample 2 Sample 3
0
6
8
10
12
14
16
18
20
22
24
26
Elongation at break/%
50% aged + 50% new 30% extremely aged + 70% new
40% extremely aged + 60% new 50% extremely aged + 50% new
60% extremely aged + 40% new 70% extremely aged + 30% new
80% extremely aged + 20% new 90% extremely aged + 10% new
30% new + 30% aged + 40% extremely aged
30% new + 40% aged + 30% extremely aged
(i) No interlayer heating
Sample 1 Sample 2 Sample 3
0
8
10
12
14
16
18
20
22
24
26
Elongation at break/%
50 % aged + 50 % new 30 % extremely aged + 70 % new
40 % extremely aged + 60 % new 50 % extremely aged + 50 % new
60 % extremely aged + 40 % new 70 % extremely aged + 30 % new
80 % extremely aged + 20 % new 90 % extremely aged + 10 % new
30 % new + 30 % aged + 40 % extremely aged
30 % new + 40 % aged + 30 % extremely aged
(ii) 60 s interlayer heating
(a) Sample elongations at break
Average Standard deviation
0
5
10
15
20
25
Elongation at break/%
50% aged + 50% new 30% extremely aged + 70% new
40% extremely aged + 60% new 50% extremely aged + 50% new
60% extremely aged + 40% new 70% extremely aged + 30% new
80% extremely aged + 20% new 90% extremely aged + 10% new
30% new + 30% aged + 40% extremely aged
30% new + 40% aged + 30% extremely aged
(i) No interlayer heating
Average Standard deviation
0
5
10
15
20
25
Elongation at break/%
50 % aged + 50 % new 30 % extremely aged + 70 % new
40 % extremely aged + 60 % new 50 % extremely aged + 50 % new
60 % extremely aged + 40 % new 70 % extremely aged + 30 % new
80 % extremely aged + 20 % new 90 % extremely aged + 10 % new
30 % new + 30 % aged + 40 % extremely aged
30 % new + 40 % aged + 30 % extremely aged
(ii) 60 s interlayer heating
(b) Mean values and standard deviations of sample elongations at break
Figure 19 Comparisons of sample elongations at break

3.3.6.3. Young’s modulus and build time
Figure 20 compares the Young’s modulus of samples printed using powders of different combinations,
(a) sample Young’s modulus, (b) mean values and standard deviations of sample Young’s modulus. The
dotted line in Figure 20a shows the Young’s modulus of the 50%-50% new-aged mixed blend. From Figure
20b, parts with high percentages of extremely aged powders, for instance, 70%, 80% and 90%, have lower
average Young’s modulus. However, there is no distinct rule that the standard deviations of sample Young’s
modulus are affected by powder combinations or interlayer heating. Also, in presence of recycled materials,
the proposed configuration can consistently control the sample Young’s modulus to the same level as the
benchmark samples (50%-50% new-aged mixed blend).
50 aged 30 EX aged 40 EX aged 50 EX aged 60 EX aged 70 EX aegd 80 EX aged 90 EX aged 3:3:4 3:4:3
0
200
400
600
800
1000
Young's modulus/Mpa
Samples
No interlayer heating
60 s interlayer heating
Baseline
(a) Sample Young’s modulus
50 aged 30 EX aged 40 EX aged 50 EX aged 60 EX aged 70 EX aegd 80 EX aged 90 EX aged 3:3:4 3:4:3
0
200
400
600
800
1000 Part without interlayer heating
Young's modulus/Mpa
Samples
Mean data
Standard deviation
Part with interlayer heating
Mean data
Standard deviation
(b) Mean values and standard deviations of sample Young’s modulus
Figure 20 Comparisons of sample Young’s modulus
The total build time for tensile bars without interlayer heating is 35 minutes (25 minutes of preheating
and 10 minutes of printing). As a trade-off of successfully reusing extremely aged powders and improved

mechanical properties, for the tensile bars with interlayer heating, the interlayer heating time is 10 minutes
(60 seconds×10 layers) for a batch. The total build time for tensile bars with interlayer heating is 45 minutes
(25 minutes of preheating, 10 minutes of printing and 10 minutes of interlayer heating). The proposed SLS
with interlayer heating aims to provide the needed heating energy to improve part mechanical properties
through promoting powder coalescence and part densification. In practice, engineering judgements are
recommended to balance material cost, part design, and urgency of the manufacturing task when recycling
and reusing the materials.
4. Conclusions
A new process control method was proposed in this paper to explore the possibility and feasibility of
reusing the differently degraded polyamide 12 powders in different combinations. In particular, the
proposed method successfully reuses extremely aged polyamide 12 powders close to the heat-affected zones.
The proposed method is composed of seven steps, including powders sample collection, powder preprocess,
powder mixing, powder characterizations, parameter control, SLS with interlayer heating and part
characterizations. This method enabled reusing reclaimed polyamide 12 powders in different situations:
only pure powders, new-aged mixed powders, new-extremely aged mixed powders with different mixing
percentages and new-aged-extremely aged mixed powders with different mixing percentages.
The proposed method of SLS created parts with improved mechanical properties: the largest tensile
strength we obtained is 37.97 Mpa from tensile bars 3D printed using 60%-40% new-extremely aged
powders with 60-second interlayer heating, a result 47.17% better than the baseline (25.80 Mpa). The tensile
bars which have stably large elongations at break are from the 10%-90% new-extremely aged blends
without or with interlayer heating (15.37% and 16.59% respectively). Compared to the baseline sample, the
tensile bars 3D printed using 30%-40%-30% new-aged-extremely aged mixed powders with 60-second
interlayer heating yield 18.04% higher tensile strength and 55.29% larger elongation at break.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.
Acknowledgments
The work was supported in part by a research grant from Unilever and by NSF award 1953155.
References
[1] J.-P. Kruth, X. Wang, T. Laoui, L. Froyen, Lasers and materials in selective laser sintering, Assembly
Automation, (2003).
[2] R. Goodridge, C. Tuck, R. Hague, Laser sintering of polyamides and other polymers, Progress in
Materials science, 57 (2012) 229-267.

[3] P. Chen, M. Tang, W. Zhu, L. Yang, S. Wen, C. Yan, Z. Ji, H. Nan, Y. Shi, Systematical mechanism
of Polyamide-12 aging and its micro-structural evolution during laser sintering, Polymer Testing, 67
(2018) 370-379.
[4] S. Dadbakhsh, L. Verbelen, O. Verkinderen, D. Strobbe, P. Van Puyvelde, J.-P. Kruth, Effect of PA12
powder reuse on coalescence behaviour and microstructure of SLS parts, European Polymer Journal, 92
(2017) 250-262.
[5] B. Chen, Y. Wang, S. Berretta, O. Ghita, Poly Aryl Ether Ketones (PAEKs) and carbon-reinforced
PAEK powders for laser sintering, Journal of Materials Science, 52 (2017) 6004-6019.
[6] S. Berretta, K.E. Evans, O. Ghita, Processability of PEEK, a new polymer for high temperature laser
sintering (HT-LS), European Polymer Journal, 68 (2015) 243-266.
[7] K. McAlea, Materials and applications for the Selective Laser Sintering, in: Process, Proc. 7^< th>
Int. Conf. on Rapid Prototyping, 1997, pp. 23-33.
[8] S. Ziegelmeier, P. Christou, F. Wöllecke, C. Tuck, R. Goodridge, R. Hague, E. Krampe, E.
Wintermantel, An experimental study into the effects of bulk and flow behaviour of laser sintering
polymer powders on resulting part properties, Journal of Materials Processing Technology, 215 (2015)
239-250.
[9] H. Zarringhalam, N. Hopkinson, N. Kamperman, J. De Vlieger, Effects of processing on
microstructure and properties of SLS Nylon 12, Materials Science and Engineering: A, 435 (2006) 172-
180.
[10] G. Wang, P. Wang, Z. Zhen, W. Zhang, J. Ji, Preparation of PA12 microspheres with tunable
morphology and size for use in SLS processing, Materials & Design, 87 (2015) 656-662.
[11] S. Arai, S. Tsunoda, R. Kawamura, K. Kuboyama, T. Ougizawa, Comparison of crystallization
characteristics and mechanical properties of poly (butylene terephthalate) processed by laser sintering and
injection molding, Materials & Design, 113 (2017) 214-222.
[12] F. Paolucci, M. van Mook, L. Govaert, G. Peters, Influence of post-condensation on the
crystallization kinetics of PA12: From virgin to reused powder, Polymer, 175 (2019) 161-170.
[13] M. Schmid, A. Amado, K. Wegener, Polymer powders for selective laser sintering (SLS), in: AIP
Conference proceedings, AIP Publishing LLC, 2015, pp. 160009.
[14] L. Feng, Y. Wang, Q. Wei, PA12 Powder Recycled from SLS for FDM, Polymers, 11 (2019) 727.
[15] K. Dotchev, W. Yusoff, Recycling of polyamide 12 based powders in the laser sintering process,
Rapid Prototyping Journal, (2009).
[16] D.T. Pham, K. Dotchev, W. Yusoff, Deterioration of polyamide powder properties in the laser
sintering process, Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical
Engineering Science, 222 (2008) 2163-2176.
[17] S. Ziegelmeier, F. Wöllecke, C. Tuck, R. Goodridge, R. Hague, Characterizing the bulk & flow
behaviour of LS polymer powders, in: Proceedings SFF Symposium, Austin (TX), USA, 2013.
[18] M. Krantz, H. Zhang, J. Zhu, Characterization of powder flow: Static and dynamic testing, Powder
Technology, 194 (2009) 239-245.
[19] S.C. Ligon, R. Liska, J.r. Stampfl, M. Gurr, R. Mlhaupt, Polymers for 3D printing and customized
additive manufacturing, Chemical reviews, 117 (2017) 10212-10290.
[20] A. Wegner, C. Mielicki, T. Grimm, B. Gronhoff, G. Witt, J. Wortberg, Determination of robust
material qualities and processing conditions for laser sintering of polyamide 12, Polymer Engineering &
Science, 54 (2014) 1540-1554.

[21] L. Wang, A. Kiziltas, D.F. Mielewski, E.C. Lee, D.J. Gardner, Closed-loop recycling of
polyamide12 powder from selective laser sintering into sustainable composites, Journal of Cleaner
Production, 195 (2018) 765-772.
[22] K. Wudy, D. Drummer, Aging effects of polyamide 12 in selective laser sintering: Molecular weight
distribution and thermal properties, Additive Manufacturing, 25 (2019) 1-9.
[23] K. Wudy, D. Drummer, Aging behavior of polyamide 12: interrelation between bulk characteristics
and part properties, in: Solid Freeform Fabrication Symposium—an Additive Manufacturing Conference,
2016, pp. 770-781.
[24] R. Goodridge, R. Hague, C. Tuck, Effect of long-term ageing on the tensile properties of a
polyamide 12 laser sintering material, Polymer Testing, 29 (2010) 483-493.
[25] K. Wudy, D. Drummer, F. Kühnlein, M. Drexler, Influence of degradation behavior of polyamide 12
powders in laser sintering process on produced parts, in: AIP Conference Proceedings, American
Institute of Physics, 2014, pp. 691-695.
[26] A. Wegner, G. Witt, Correlation of process parameters and part properties in laser sintering using
response surface modeling, Physics Procedia, 39 (2012) 480-490.
[27] M. Pavan, M. Faes, D. Strobbe, B. Van Hooreweder, T. Craeghs, D. Moens, W. Dewulf, On the
influence of inter-layer time and energy density on selected critical-to-quality properties of PA12 parts
produced via laser sintering, Polymer testing, 61 (2017) 386-395.
[28] E.C. Hofland, I. Baran, D.A. Wismeijer, Correlation of process parameters with mechanical
properties of laser sintered PA12 parts, Advances in materials science and engineering, 2017 (2017).
[29] G. Salmoria, V. Lauth, M. Cardenuto, R. Magnago, Characterization of PA12/PBT specimens
prepared by selective laser sintering, Optics & Laser Technology, 98 (2018) 92-96.
[30] M. Schmid, R. Kleijnen, M. Vetterli, K. Wegener, Influence of the origin of polyamide 12 powder on
the laser sintering process and laser sintered parts, Applied Sciences, 7 (2017) 462.
[31] R. Li, X. Hu, Study on discoloration mechanism of polyamide 6 during thermo-oxidative
degradation, Polymer degradation and Stability, 62 (1998) 523-528.
[32] P. Gijsman, D. Tummers, K. Janssen, Differences and similarities in the thermooxidative degradation
of polyamide 46 and 66, Polymer degradation and stability, 49 (1995) 121-125.