Additive Manufacturing 43 (2021) 102017

Contents lists available at ScienceDirect

Additive Manufacturing
journal homepage: [Link]/locate/addma

Review

Metal powders in additive manufacturing: A review on reusability and

recyclability of common titanium, nickel and aluminum alloys
Pouya Moghimian a, *, Thomas Poirié b, Mahdi Habibnejad-Korayem b, Javier Arreguin Zavala b,
Jens Kroeger b, Frédéric Marion b, Frédéric Larouche b
a
GE Additive Germany GmbH, Freisinger Landstrasse 50, 85748 Garching bei München, Germany
b
AP&C Advanced Powders and Coatings Inc., 241 Rue Allée du Golf, Saint-Eustache, Québec J7R 0L5, Canada

A R T I C L E I N F O A B S T R A C T

Keywords: Metal additive manufacturing (AM) has received a lot of attention since its introduction, owing to its cost effi­
Additive manufacturing ciency and freedom in design. For a successful and repeatable AM process, metal powders conform to strict
Powder reusability chemical and physical properties requirements. There are numerous factors affecting these characteristics in both
Powder recyclability
virgin and used powders. This article presents a comprehensive overview on manufacturing methods of AM metal
Powder bed fusion
Selective laser melting
powders and the reusability of titanium, nickel and aluminum alloys (mainly Ti-6Al-4V, nickel alloy 718 and Al-
Electron beam melting Si10-Mg) as well as criteria for such reuse. Moreover, the feasibility of recycling AM metal powders is discussed.
Plasma atomization This article also provides perspectives and recommendations for AM users to develop a well-defined and standard
powder reuse process for consistent powder characteristics and built parts.

1. Introduction accepted or defined process for AM metal powder reuse. We would like
to first differentiate between the terms powder reuse and powder recycling.
The origins of metal additive manufacturing (AM) dates back to Recycling relates to powder or scrap material (in-process recovery metal
1920, when an electric arc was used by Ralph Baker to fuse superposed or post-consumer scrap) remelting or grinding in order to form it into a
layers of metallic deposits [1,2]. This process was further developed in new atomization feedstock [3]. Reuse relates to repetitive use of a single
the 1980–90s through extensive collaboration between research in­ (initially virgin and then used [7]) powder batch during multiple cycles
stitutes and research and development laboratories [3]. in an AM machine, usually until it falls out of pre-defined limits (i.e. out
Metal powders used in a powder bed fusion AM machine must have of specification (OOS)) [3]. Reuse usually requires little or no process­
distinct physical and chemical characteristics for a reliable and repro­ ing. These two terms have been used interchangeably in the literature.
ducible printing outcome. These characteristics include, but are not However, it is important to differentiate between them. These two
limited to, particle morphology, particle size distribution (PSD), density, processes, both currently a major attraction in the AM industry and
porosity, flowability, occurrence of satellites, agglomeration, humidity, which have been a subject of several studies to date, will be discussed in
tribocharging properties and chemical composition. These characteris­ detail in this review. Several studies have been conducted on the reus­
tics need to be tested and optimized for each AM process or technology ability of different kinds of stainless steels such as 316L and 17–4PH
to produce high quality and reliable parts through metal AM. A large [8–11] where similar changes in physical properties such as morphology
number of metal AM research articles focus on the following three types and PSD were reported, comparing with other metal alloy powders.
of alloys: titanium, nickel and aluminum alloys [4]. Nickel and We focus here mainly on three metal alloy powders: 1) Ti-6Al-4V is
aluminum alloys find applications in different industries such as aero­ an α-β titanium alloy, very well-known for its high strength-to-weight
space and automotive. Additionally, medical and dental industries also ratio, excellent mechanical properties at low or moderate tempera­
tailor titanium alloys [5]. Adoption of additive technology is driven by tures (< 350 ◦ C) and corrosion resistance. This alloy is used for the
freedom of design and the increasing demand for lighter and more fabrication of aerospace components such as jet engines, for marine/
efficient materials with enhanced final part functionality [3,6]. offshore components in sea water like propellers, and for biomedical
A thorough literature review reveals that there is still not a generally implants [12,13]. 2) Nickel alloy 718 is a precipitation nickel-base

* Corresponding author.
E-mail address: [Link]@[Link] (P. Moghimian).

[Link]
Received 12 January 2021; Received in revised form 11 March 2021; Accepted 16 April 2021
Available online 6 May 2021
2214-8604/© 2021 Elsevier B.V. All rights reserved.
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

superalloy and is a critical component of aerospace technologies such as material (see Fig. 1c). The energy source used to melt the raw material
compressor blades, combustion chambers and gas turbine discs [4,14, can be an induction coil, a plasma arc or plasma torches. The molten
15]. 3) Al-Si10-Mg is a hypoeutectic Al–Si casting alloy, which has stream will be struck with a high-velocity jet of gas, plasma or water to
recently received attention for the ability of AM fabrication of light­ pulverize it into a cloud of fine particles.
weight and intricate structures, especially in automotive and aerospace
industries [16–18]. The attention that Al-Si10-Mg alloy continues to
attract in the AM industry is also due to the fact that the high cooling 2.1. Gas and water atomization
rates in metal AM machines (up to ~ 108 K/s) can be much higher than
in conventional casting processes (~ 102 K/s) [19,20]. This rapid so­ Gas Atomization (GA) is a traditional manufacturing process for
lidification has a positive influence on the microstructure and the me­ obtaining close-to-spherical powders, which was first patented in 1872
chanical properties of Al-Si10-Mg [21,22]. by Marriott of Huddersfield who used steam as the atomizing gas of the
We will discuss the effect of the reuse of these powders on their molten material [26–28]. Hall [29] filed a patent in 1919, where he
properties after being processed in two widely adopted metal AM constructed a nozzle such that the molten material could be suctioned
powder bed processes, namely laser powder bed fusion (L-PBF, or se­ vertically from its tip by the force of the steam jet. In 1973 Hellman and
lective laser melting) and electron beam powder bed fusion (E-PBF, or Josefsson [30] patented a new gas atomization apparatus, where the
electron beam melting (EBM)). The current review gives insight into tapping stream of molten material from a tundish (i.e. a vessel with an
different metal powder manufacturing methods, the essential properties opening at the bottom through which the molten material is dispensed)
of metal AM powders, how the AM process can change and influence is redirected by the first gas jet and is then pulverized by the second one,
these powder properties, and finally the reusability and recyclability of in order to speed up the solidification process. Crucible Materials Cor­
these powders after several printing cycles. Finally, it will compare poration further developed the gas atomization technique in the 1980s
different powder reuse strategies and provide recommendations in order [25].
to optimize the powder lifecycle. As a first step of this method, the raw material, elemental or pre-
alloyed ingot or a bar is melted using an electric arc or induction coils
2. Powder production methods in a ceramic crucible or a water-cooled copper crucible (skull melting)
and is kept for some time in this state to become chemically homoge­
There are currently several metal powder atomization methods nized. Secondly, it will be either freely poured down (i.e. free fall) or
tailored for additive manufacturing including water atomization, gas through a confined refractory nozzle with a high pressure difference into
atomization, centrifugal atomization and plasma atomization. In met­ an atomization chamber (i.e. a close-coupled system). The melt will then
allurgy, atomization is known as the process for breaking liquid into fine be pulverized into droplets using a high-velocity jet of noble gas, typi­
droplets [23]. It provides a strong control over the production rate and cally argon or nitrogen [31]. Air or helium can also be used as the
desired physical characteristics of the resulting powders when compared atomizing gas. The choice of the atomizing gas depends on its cost, its
to other powder production methods such as mechanical or chemical thermal conductivity and reactivity with the alloy [32]. As a final step,
methods. Atomization accounts for a large fraction of manufactured the metal droplets cool down to form powder particles. The character­
metal powders. In 2015, over 60% of all metal powders produced in istics of the resulting atomized powder depend on several parameters
North America were produced by atomization [24]. All atomization such as the diameter and the velocity of the melt stream (mass flow rate)
methods use a source of high energy to melt raw material into a molten as well as the mass-flow and shear rates of the atomizing gas. This
stream within an adjusted environment to solidify in a controlled method can atomize a broad range of powder size distributions (PSDs)
manner [25]. Different methods use different raw materials. Fig. 1 below 500 µm economically with a high yield of fine powder [24,33].
shows four different commercial atomization methods for metal pow­ For achieving a high yield of fine particles, close coupled systems are
ders used in additive manufacturing. For instance, gas and water at­ mainly used in gas atomization [34,35]. For an optimized process, the
omization use elemental or pre-alloyed feedstock (see Fig. 1a and b). On powder can be nearly spherical. However, disadvantages to this method
the other hand, plasma atomization uses a pre-alloyed wire as its raw include the presence of a large to moderate fraction of particles with
irregular shapes, a high number of satellite particles as well as entrapped

Fig. 1. Four different commercial atomization methods [25]. Schematic illustrations of (a) gas atomization, (b) water atomization (c) plasma atomization with a
snapshot of a pre-alloyed wire getting atomized at the apex of plasma torches and (d) the plasma rotating electrode process (PREP) which is a centrifugal atomi­
zation process.

2
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

gas pockets within individual particles, especially in coarser ones [24]. microscope (SEM) images of gas-atomized Ti-6Al-4V, nickel alloy 718
Satellites can form inside the atomization chamber when producing and Al-Si10-Mg powders. As seen in these images, the degree of particle
large amount of fine particles, where re-circulating gas brings finer sphericity is relatively low, and a high number of satellite particles can
particles in contact with the coarser heated ones. be observed.
The circularity of virgin gas-atomized Ti-6Al-4V, nickel alloy 718 Water atomization technology is similar to the gas atomization
and Al-Si10-Mg powders was reported to be between 0.7 and 0.9 using process as shown in Fig. 1b. However, a high-pressure water stream is
the equation fcircularity = 4π A/P2 where A is the cross-sectional area and P used as the atomizing fluid to produce spattered droplets. Since the heat
is the perimeter of the particle [36]. transfer and momentum is much higher in water atomization (compared
With gas atomization, there is a risk of contamination associated to gas atomization where the atomization jet has a much lower mass
with crucible and nozzle material because the molten material that is flow rate and heat capacity), it can increase the shape irregularity and
poured through a nozzle into the atomization chamber is kept within a the porosity in the powder. Water atomization benefits from a relatively
ceramic or ceramic-grouted crucible. Reactive materials with high low production cost and high productivity rate (~ 1 ton/min). It is the
melting points can react with the crucible walls and traces of non- most widely used method for the production of ferrous metal powders,
metallic inclusions may be found in the resulting powder. To avoid but it has also been used to produce nickel alloys or reactive materials
these aggressive crucible-melt interactions, mainly physical erosion and like copper and copper alloys [24]. With reactive materials there is a
chemical dissolution, a crucible-free atomization can be used [37,38]. limitation for the powder purity [39]. Drawbacks of this method include
The gas jet hitting the molten stream, unless at extremely elevated low sphericity of powders with high oxygen content and surface
temperature, can result in a rapid cooling and solidification of powder oxidation due to the reaction with water molecules [32].
particles. In these conditions, entrapped gas pockets lack the time to
leave powder particles and powder droplets lack the time to minimize 2.2. Plasma atomization and plasma rotating electrode process
their surface area by adopting a highly spherical shape. This can lead to
the formation of irregular particle shapes with entrapped pores, Plasma atomization (PA) was developed in 1996 by Entezarian et al.
adversely affecting powder flowability. These are some of the issues that [40] to produce highly spherical powders of reactive metals with an
AM users face when using gas-atomized powders. On the other hand, gas average particle size of 40 µm. The method was patented by Pegasus
atomization is a well-established and mature technology. The process is Refractory Materials and Hydro-Québec in Canada in 1998 [41]. This
economical, fast and can produce a broad range of PSDs that can be method was similar to the PREP method which was developed by Nu­
tailored to different applications. Fig. 2a, b and c show scanning electron clear Metals Inc. in 1980. It enables a one-step operation, combining the

Fig. 2. SEM secondary-electron images of virgin gas

atomized (a) Nickel 718, (b) Ti-6Al-4V and (c) Al-Si10-Mg
powders. Images were cropped and resized, and scale bars
were added. Virgin plasma atomized (d) Nickel 718, (e) Ti-
6Al-4V and (f) Al-Si10-Mg powders produced by Advanced
Powders and Coatings Inc. (AP&C). Gas atomized powders
have a fraction of irregular shapes and/or a high number of
satellite particles compared to plasma atomized powders.
All scalebars represent 100 µm.
(a,b,c) Adapted from the Ref. [36] ([Link]
[Link]/licenses/by/4.0/; study by Cordova et al.)

3
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

melting and atomization steps using a pre-alloyed wire as the raw ma­ the form of an electrode bar which will be rotated at approximately
terial, as shown in Fig. 1c. Therefore, the risk of ceramic contamination 15,000 rpm while being melted by a plasma arc. Centrifugal forces eject
from the melting of a reactive metal in a ceramic crucible was elimi­ the molten material from the bar and the particles solidify before they
nated, ensuring a high purity in the final powder product [42]. In this hit the chamber wall. This method was invented as a response to the
method, the molten material is superheated using an extremely hot (up high consumption volume of gas per unit weight of atomized powder
to 11,000 K) argon plasma produced by converging plasma torches. This and a generally very low yield of fine powder particles by traditional gas
inhibits the molten metal particles from solidifying rapidly into irregular atomization [46]. In the PREP process, this yield is dependent on the
shapes [25,43]. The molten droplets spend sufficient time in the su­ centrifugal force thus rotation speed. As with plasma atomization, this
perheated state, allowing them to adopt an equilibrium shape driven by method is also a ceramic-free process, since there is no contact between
the surface tension forces: ideal spheres [42,43]. molten metal and container. However, PREP produces tighter PSD
There are several parameters in the plasma atomization process that ranges vis-à-vis other atomization methods [24].
affect the powder characteristics such as morphology and the PSD. These
include torch power, the wire feed speed, height and width of the at­ 2.3. Post-processing of powders
omization chamber, cooling rate, production rate and velocity of the
plasma gas. In addition to the atomization process, another key process in the
This method has been successfully used for reactive and non-reactive determination of powder characteristics is the post-processing of raw
metals such as titanium, nickel, aluminum and copper but also for ce­ metal powders. These include passivation of powder particles against
ramics such as carbides. It offers several advantages compared to gas oxidation, classification (i.e. sieving or air classification), and blending
atomization: low capital expenditure (CapEx) technology with low gas to obtain the final homogenous powder lot. There are also various
consumption, high degree of particle sphericity, high density, low technologies in order to improve the flowability of powders post at­
porosity (typically below 0.05% by metallographic analysis), excellent omization. There are currently multiple methods to quantify powder
flowability, low contamination, low number of satellite particles, and rheology (flowability and spreadability), but correlation between them
low agglomeration of powder particles. Several plasma atomization is not fully established. No method is widely used except for the simplest
methods have been proposed and patented during recent decades. one that is funnel flow rate (ASTM B213/ISO 4490/ASTM B964). A good
However, most of them do not offer a high yield of fine powders understanding of what is an acceptable flowability value for a given
(<106 µm). Others suffer from high production costs or low production metal alloy is lacking in the AM community. There is a need for more
rates [44]. quantitative studies on powder flowability for additive manufacturing.
As seen in Fig. 2d, e and f, virgin plasma atomized nickel 718, Ti-6Al-
4V and Al-Si10-Mg powders show a high degree of sphericity and a low 3. Powder reusability
amount of satellite particles compared to gas atomized powders.
A schematic of the PSD of atomized powder with a high yield of fine Using AM metal powders is rapidly being adopted by the industry,
powder is shown in Fig. 3. Fine powder finds applications in L-PBF and led primarily by the aerospace and orthopedic sectors. Since AM pow­
E-PBF printing, binder jet, direct energy deposition (DED), metal in­ ders constitute one of the main cost drivers, especially in high-
jection molding (MIM), (spray) coating, and hot isostatic pressing (HIP) performance applications, there has been a great deal of interest in
which is the consolidation and densification of metal powder under high reusing them. Since only a small portion of this powder melts and fuses
pressure and temperature [45]. The latter three were historically those into a part, the rest of the powder can be used several times until it
benefiting from the high-quality powder produced by atomization reaches a condition when the powder cannot be used anymore for a
before metal additive manufacturing became widely adopted. The given application [47,48].
benefits of plasma atomized powder in coating, HIP and MIM resulted Several factors can influence the reusability of a batch of powder.
respectively in an up to 10% more efficiency and a lower product These differ from application to application and user to user. Factors
shrinkage. It has been reported that in the MIM process, a good flow­ include: 1) The chemical composition of the used powder. For instance,
ability of plasma atomized powder can lead to less required additives for ASTM F3001–14 and ASTM F2924–14 dictate that a used Ti-6Al-4V
the powder to be processed [43]. powder which is not conforming to a specified chemical composition
In the plasma rotating electrode process (PREP) which is a centrif­ shall not be further used in an AM machine [49,50]. However, the
ugal atomization process, shown in Fig. 1d, feedstock material comes in number of reuse cycles has not been limited. 2) Contamination of the
used powder. When the powder is being reused in an AM machine it can
collect several contaminants. The contamination source can be the AM
chamber (wear debris) or the tools used in the machine, the gaseous
elements present during the AM process, and finally humidity that the
powder is exposed to during storage, conveying, sieving or inside the
powder removal systems. 3) Physical characteristics such as PSD,
flowability, tap/apparent density or the appearance of a high portion of
angular or interfused particles. If the powder loses its flowability, it can
lead to inconsistency or even a failure during the printing process. If the
density decreases, packing density of powder in layers will decrease as
well, potentially affecting the printed part density due the increased
presence of porosity by lack of melting or the keyhole effect [1,51,52]. If
a drastic change occurs in the size distribution, the user might not be
able to print with the previous machine settings and parameters. 4) As a
consequence, the mechanical properties of the parts being printed with
those powders, such as tensile and fatigue properties, may be affected.
Fig. 3. Schematic of the powder size distribution (PSD) for as-produced Therefore, it is important to understand which decisive criteria are to
atomized powder (black curve) which shows volume distribution percentage determine whether a powder is reusable or not. In principle, each AM
against the particle diameter. Optimal PSD ranges for different applications user can define their own criteria and acceptance ranges for used pow­
such as L-PBF, E-PBF, MIM, HIP, binder jet, spray coating and direct energy der. However, this process needs to be validated by a production phase,
deposition (DED) are shown as red arrows. especially in highly regulated industries such as aerospace and

4
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

biomedical [51]. Some AM or powder manufacturers offer guidance and cycles in an E-PBF machine, without adding additional virgin powder
support for defining these criteria. In the next two sections, we discuss after each build job. A very slight variation in the iron and vanadium
powder reusability studies in 2 AM processes: E-PBF and L-PBF. Of all contents (beta-stabilizers) within fluctuation of analysis were observed.
powder bed fusion AM machines, only the E-PBF process works under Aluminum content decreased on average by approximately 3% after 12
vacuum; all major L-PBF systems include vacuum pumps to purge prior reuse cycles. Hydrogen content, one of the most undesirable impurities
to printing but then do so at or close to atmospheric pressure [53]. They in titanium, showed no significant variation. It is important to note that
print under inert gas (mostly argon or nitrogen) but partial pressures of hydrogen can cause brittle fracture in titanium. The initial oxygen
oxygen and humidity remain, depending on purge cycle, powder con­ content for the virgin powder lot was 0.14 wt%. This oxygen level
dition, etc. increased linearly and exceeded the maximum acceptable Ti-6Al-4V
grade 5 level (i.e. 0.20 wt%) after the twelfth reuse cycle. The powder
3.1. Reusability of titanium alloys in E-PBF particles became partially deformed after several build cycles, which
may be attributed to the blasting of the powder in the powder recovery
In the E-PBF process, it is important to bear in mind that the whole system (PRS) to break off pre-sintered particles, or to exposure to
powder bed is kept at elevated temperatures (usually above 650 ◦ C) elevated temperature in the E-PBF chamber.
during the whole process and that the melting occurs where a high- Shanbhag et al. [61] observed a rapid increase in the oxygen level of
intensity convergent electron beam with a small probe size bombards a Ti-6Al-4V grade 5 powder from 0.13 to 0.17 wt% after two reuse cycles
the powder at its focus point [54]. The E-PBF building chamber is kept in an E-PBF machine. Authors stated that blasting in the PRS could speed
under vacuum at 10-3-10-4 mbar using turbomolecular pumps [55]. The up the oxidation process by the generation of new oxidation-prone
reason is that a coherent electron beam cannot be transmitted through surfaces on the powder particles. Furthermore, the sieving and the
dense gas and strong scattering of electrons by gas atoms can occur transfer of the powder under ambient atmosphere was suggested as
along its path [56]. Therefore, the presence of foreign impurities is another factor leading to the abnormally rapid increase in the oxygen
drastically reduced by the high vacuum compared to the L-PBF process. level of the powder. The latter hypothesis does not seem plausible as
A possible contamination stems from residual hydrocarbons in the titanium powder, especially with a PSD for the E-PBF process, is known
chamber. These carbonaceous materials can crack under the electron to be chemically stable when stored at room temperature.
beam, and deposit on or react with the hot layer of the powder bed [57]. Ghods et al. [62] conducted research on the powder reusability of a
Another contamination can come from the permeation of water mole­ plasma-atomized Ti-6Al-4V grade 5 powder during 30 build cycles in an
cules into the electron beam chamber [58]. It has been reported that one E-PBF machine, over a period of six months. As with other studies, a
main source of oxygen pickup outside of the E-PBF chamber is the water linear increase in the oxygen level was observed over the 30 build cycles
adsorption on the powder surface during exposure to air during sieving totaling approximately 480 h. An average increase of ~ 0.008 wt% in
or storage [59]. It is worth mentioning that since an E-PBF machine the oxygen level per 16 h of build time could be observed. The oxygen
operates under vacuum and high electron beam energy, it can cause a level exceeded the grade 5 limit of 0.20 wt% after the 11th reuse cycle,
loss of some alloy elements such as aluminum in Ti-6Al-4V. Therefore, given a starting oxygen level in the virgin powder of 0.14 wt%. Other
an E-PBF-dedicated Ti-6Al-4V powder feedstock usually contains an elements showed a very slight variation over the 30 build cycles and
additional amount of aluminum to compensate for this loss [1]. The stayed below the limits defined by ASTM F2924–14. These results sup­
amount of aluminum is critical in this alloy, since it can significantly port the conclusions of Petrovic et al. [60]. In a following study by Ghods
alter the material’s strength [60]. and coworkers [63] it was shown that the volumetric porosity in the
Tang et al. [54] studied a Ti-6Al-4V grade 23 (or ELI (extra low in­ Ti-6Al-4V grade 5 built parts remained nearly constant during 30 build
terstitials)) powder through 21 reuse cycles in an E-PBF system and cycles. The origin of porosity in the built parts was mostly associated
characterized the powder physically and chemically after each few cy­ with the porosity in the powder particles and with the incomplete or lack
cles. It was shown that no significant change happened in the elemental of fusion of the melt pool in the E-PBF machine.
composition except for the oxygen content. The oxygen level increased It was shown that virgin Ti-6Al-4V powder particles from an E-PBF
continuously with each reuse cycle and the level fell out of ELI specifi­ size-cut have an oxide layer of about 6.5 nm on their surface, rising to
cation (maximum oxygen level of 0.13 wt%) after the fourth reuse cycle. around 7.5 nm after 10 reuse cycles in the E-PBF process [64]. This in­
This increase was due to the oxygen pickup in the powder depending on crease in the surface oxide layer leads to a local hardening of this surface
the humidity and the exposure time. The PSD of the reused powder layer compared to the core of the particle, as confirmed by nano­
became narrower after each cycle, suggesting the occurrence of fusion of indentation testing. However, this oxidation makes the metal surface
fine particles and their coalescence during the melting process. The brittle and can lead to microcracking [65]. The diffusion of oxygen and
number of satellite particles did not increase dramatically. However, the the accompanying structural changes will be described in the next
surface of the particles became rougher especially after the sixteenth sub-section. The average aluminum content decreased due to the its high
reuse cycle. This was due to the melting process as the particles were vapor pressure. Furthermore, on the roughened surface of reused pow­
heated in the vicinity of the melt pool (i.e. powder spatter). The flow, tap der particles, local oxides include titanium oxide, aluminum oxide and
and apparent density of the reused powders remained virtually un­ iron oxide [64] (see Fig. 4).
changed, despite the change in the PSD. In the as-built tensile samples It was microstructurally shown by Sun et al. [66] that picked-up
where a heat-treatment was not reported, the yield strength and tensile oxygen is not uniformly distributed in the recycled powder particles.
strength both increased with increasing powder reuse cycles, consistent The surface oxide thickness remains below 10 nm even after 30 reuse
with the oxygen pickup after each cycle [54]. cycles, despite an increase of approximately 35% in the average powder
In the study of Nandwana et al. [59], a virgin gas atomized Ti-6Al-4V oxygen content. However, oxygen diffuses and incorporates preferen­
powder was used five times in an E-PBF machine and the oxygen content tially in the β phase of the recycled powder particles, given that the
was raised almost linearly from 0.14 to 0.18 wt%. Observing these data, metastable martensitic α′ microstructure in the virgin Ti-6Al-4V powder
we might be able to speculate that after the seventh reuse, the Ti-6Al-4V transforms into equilibrium α + β two-phase microstructure during
powder would fall out of grade 5 specifications, i.e. the maximum oxy­ powder annealing in an E-PBF machine.
gen content of 0.20 wt%. Moreover, no noticeable change in the
morphology (confirmed by SEM imaging), flowability and PSD curves 3.2. Reusability of titanium alloys in L-PBF
was observed.
Petrovic et al. [60] investigated the chemical and physical properties Oxygen pickup is also a well-known issue for titanium powders in the
of a plasma-atomized Ti-6Al-4V grade 5 powder lot during 16 build L-PBF process. Titanium has a high reactivity with oxygen [67,68].

5
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

Fig. 4. Schematic illustration of some possible chemical

and physical changes after a virgin Ti-6Al-4V powder goes
through the E-PBF process. (a) A virgin Ti-6Al-4V particle
having a native oxide layer, as shown with red arrows. (b)
The particle that was reused several times in an E-PBF
process but has not been molten and integrated into a
printed part. The oxide layer on the surface becomes
thicker, the particle surface roughens, local regions of ti­
tanium oxide, aluminum oxide and iron oxide can form on
the surface. Heated particles can partially sinter/melt and
attach to the particle surface. Aluminum (shown as blue
circles) content decreases due to the evaporation under the
electron beam [64]. (For interpretation of the references to
colour in this figure, the reader is referred to the web
version of this article)

Because of their atomic radii, oxygen and nitrogen, which are strong In another study carried out by Quintana et al. [73], a gas atomized
alpha stabilizers, occupy the octahedral interstices in the structure [69]. Ti-6Al-4V ELI powder certified to ASTM F3001 (with an initial oxygen
The increase in number of these interstitials will increase the lattice level of 0.09 wt%) was used 31 times to build test parts. The study
strain in the c axis. Consequently, the c/a lattice parameter will increase, investigated the effect of reuse on the chemical and physical properties
leading to an inhibition of slip planes in the HCP structure and changing of the powders. Authors sampled 200 g of powder from the dispenser
the material strength and hardness [70]. each time prior to the beginning of the printing process and sieved it to
L-PBF machines work under inert gas atmosphere, usually argon or remove the agglomerates resulting from a build cycle. Surprisingly, the
nitrogen. It should be noted under nitrogen gas atmosphere (N2), the powder stayed within the acceptance ranges prescribed by the ASTM
high reactivity of titanium with nitrogen can lead to the formation of F3001 chemical specification, even after the thirty first reuse cycle. It is
brittle TiN phase [71,72]. There are oxygen sensors in the machine important to note that the L-PBF chamber oxygen content was kept
chamber that limit the oxygen content to usually a maximum of 1000 below 100 ppm during each cycle. The powder oxygen content varied
ppm [68]. Unfortunately, this oxygen concentration is sufficient for an between 0.09% and 0.13 wt% during the reuse cycles but did not show a
oxygen pickup by the powder. Since oxygen is an α-stabilizer, the in­ constant upward trend. A similar behavior was shown in another report
crease in oxygen influences the material’s strength and hardness. This during 38 build cycles [76]. This is in contrast to the findings of Tang et.
comes at the cost of a reduction in ductility and fracture toughness [66, al [54] which showed a constant oxygen pickup of Ti-6Al-4 V ELI
68]. However, it has been reported that a relatively sharp decrease in powder after each build cycle in the E-PBF machine. Tang and coworkers
ductility occurs if the oxygen content exceeds ~ 0.33 wt% [67]. used a virgin powder that contained 0.08 wt% oxygen in an E-PBF
The oxygen pick-up from cycling through the L-PBF process, and in machine versus the 0.09 wt% oxygen content powder used in a L-PBF
general the change in powder characteristics from reuse, depends on the machine (in Quintana’s study). The elevated temperature and the inci­
location of the particle with respect to the laser path. Particles which are dent high-energy electron beam in an E-PBF machine could explain this
molten by the laser beam will be incorporated into the build, but those in apparent discrepancy (this topic will be elaborated in Section 5). The
the vicinity are most likely to be displaced. These particles, labelled slight changes of the aluminum and vanadium contents in the reused
droplet spatter and powder spatter, have the highest absorption of ox­ powder throughout 31 reuse cycles were lower than those reported by
ygen. Droplet spatter is formed by the oxidized molten particles while Tang and coworkers [54]. Indeed, the higher overall temperature, vac­
the powder spatter is formed by the heated, but not molten particles uum level, and energy of the electron beam cause more evaporation of
[73]. The oxygen will bond mostly in the form of TiO2 or TiO on the these elements compared to the laser system. In addition, variations
surface of the powder particles [33,68]. Powder particles which are on between laboratory to laboratory results in the chemical analysis must
the powder bed but not exposed to, or close to, the melt pool are unlikely be considered. For instance, product analysis tolerances of 0.02 wt%,
to be to transferred or ejected by the recirculating gas. 0.40 wt% and 0.15 wt% for oxygen, aluminum and vanadium, respec­
Among the earliest studies on the effect of reuse cycles on Ti-6Al-4V tively, have been covered by ASTM F3001. The PSD curve became
was the study in 2012 by Seyda et al. [74], where a gas atomized narrower with reuse cycles and the particle morphology remained
Ti-6Al-4V powder with particle size of D90 ~ 50 µm was used 12 times spherical but became rougher, which is in agreement with other reports.
in a L-PBF machine. The PSD, morphology, apparent density and visual The flowability improved after the reuse of powders. Tang et al. [54]
appearance were examined on the reused powder. The built parts were reported this as an unexpected result, and attributed it to two possible
also analyzed. It was shown that after 12 cycles of reuse, the average factors. One was the reduction of satellite particles which are very
particle size increased from 37.4 to 51.2 µm. The number of fine parti­ difficult to quantify in a powder sample. The other reason was the
cles was reduced, the apparent density increased by 9% and the flow­ reduction of humidity because of long powder exposure to the vacuum
ability was enhanced. The flowability enhancement was explained by in an E-PBF machine. The latter is not relevant to Quintana’s study. It
the reduction of the number of fine powder particles (<20 µm), hence a should be pointed out that humidity level in the powder can have an
decrease in average specific surface area and thus friction between adverse influence on the flowability of a powder lot due to capillary
particles. It was speculated that fine particles melt and sinter to form action, which is the flow of the liquid in narrow channels by surface
agglomerates that are removed by sieving after each reuse cycle. How­ tension forces [1,47,77]. Moisture or water vapor can be reduced during
ever, if the agglomerates are smaller than the mesh openings, they could vacuum pumping and during the preheating process. The ultimate ten­
still pass through the sieve. The parts printed with reused powder sile strength (UTS) and yield strength (YS) were observed to increase
showed a slight increase in density of ~ 0.3%, which was attributed to slightly with reuse cycles in the HIPed tensile specimens. This is
the increased flowability of reused powder. It has been observed that a attributed to the strengthening effect of oxygen, a potent alloying
higher flowability usually translates into a higher apparent density and element in titanium, consistent with other studies [54,74]. Again, such
therefore a better spreading behavior of the powder [75]. an increase in strength is often accompanied by a loss of ductility and

6
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

significant embrittlement [67,78]. reached the maximum 0.13 wt% level after 22 cycles) and consequently
One of the few works focusing on the effect of powder reuse on the an increasing trend in the tensile strength of the built parts in the
final properties of the printed parts (mechanical, fatigue and thermal) heat-treated condition.
was conducted by Carrion et al. [47] Fatigue is a critical factor in metals, In conclusion, oxygen pickup is a major issue for the reuse of Ti-6Al-
since it causes approximately 90% of all metallic failures [57]. A gas 4V powders in L-PBF and E-PBF machines. The rate of oxygen pickup
atomized Ti-6Al-4V ELI powder with a PSD of 15–45 µm was used 15 was shown to be more significant in the E-PBF process. Fig. 5 provides an
times and after each cycle the powder was sieved with a 80 µm mesh approximate rate of increase in the oxygen content in Ti-6Al-4V powder
sieve. The test parts were fabricated with the virgin powder or the per reuse cycle in both L-PBF and E-PBF machines. In these cited studies,
15-times reused powder using a L-PBF machine. The parts printed dur­ except the ones marked in blue, no rejuvenation of the reused powder lot
ing the other 14 cycles were not characterized. A narrowing of the PSD was carried out during multiple printing cycles. Although several factors
curve, a decrease in the number of satellite particles, an increase in the and machine parameters differ in these studies, the rates given in Fig. 5
number of coarser particles and an enhanced flowability were observed, can form a reference point for AM users to develop a defined reuse
similar to results reported earlier in this section. The thermal conduc­ process for Ti-6Al-4V powder. The average increase of oxygen content in
tivity slightly increased in the parts fabricated with the used powder. a powder lot was estimated to be ~ 0.001 wt% (~ 10 ppm) per reuse
The reuse of powder did not influence the grain morphology of the cycle in the L-PBF process. This is an average increase of oxygen content
fabricated parts. After 15 reuse cycles, the powder showed a slight in­ per ~ 22 h of laser active time for a 100 kg powder batch, based on
crease in oxygen content from ~ 0.10 wt% (the virgin powder’s oxygen Quintana’s study [73]. The average increase of oxygen content per reuse
content) to ~0.12 wt%, which did not exceed the ELI specification. The cycle is shown to be ~ 0.007 wt% (~ 70 ppm) in the E-PBF process. This
authors did not report the powder oxygen content variation after each is an average increase of oxygen content per ~ 14 h of build time for a
build cycle. If this increase is constant for each cycle, then it can be 50 kg powder batch based on the Ghods’ study [62].
extrapolated that the powder would have been out of ELI specification Several factors can influence the oxygen pickup results among
after the 23rd reuse cycle, contrasting with Quintana et al. Tensile different studies. For instance, the active laser or electron beam time, the
stress-strain curves of two as-built annealed fabricated parts with virgin power of energy source, the build volume and geometry and the powder
and used powder were almost identical and had similar YS and UTS bed volume need be measured and reported. Residual oxygen content in
values, indicating a similar fracture toughness [57]. Interestingly, an AM machine also differs from machine to machine and can eventually
non-machined (i.e. as built) parts fabricated using these two powders affect the oxygen pickup results. Therefore, not considering all other
had also a similar fatigue behavior. The stress amplitude values did not controlling factors and merely reporting the numbers of reuse cycles
show significant variation. However, the authors observed that the fa­ cannot provide a reasonable basis of comparison for future reusability
tigue behavior in high cycle regimes might be improved in the machined studies.
parts fabricated with reused powder. Another important finding of this
study was that the aging of the powder does not have an influence on the 3.3. Reusability of nickel alloys in E-PBF
fractographic results. The factors suggested to play a role were the defect
size and their location which are different in machined and Nickel alloy 718 with excellent mechanical properties and oxida­
non-machined parts. For the non-machined parts, the failures initiated tion/corrosion resistance was first introduced to overcome the low
from the micro-notches on the specimen surface, whereas for the weldability of superalloys in the 1960s [82].
machined parts, cracks initiated from internal close-to-surface pores, Nandwana et al. [59] reused 100 kg of nickel alloy 718 (PSD:
regardless of using virgin or used powder in both cases. Masuo et al. [79] 45–106 µm) six times to build parts in an E-PBF machine. The powder
showed that the surface roughness of parts built by E-PBF and L-PBF has was produced with electrode induction melting inert gas atomization
a strong detrimental influence on the fatigue strength, more than other process (EIGA). The powder was not rejuvenated with virgin powder
defects (e.g. gas porosity). The fatigue strength was shown to be during the reuse cycles. The elemental composition did not significantly
enhanced drastically by surface polishing and eventual HIP. change in the reused powder compared to the virgin powder except for
Recently, Denti et al. [80] investigated the effect of 100 reuse cycles oxygen which showed an increase from 0.014 to 0.022 wt%. It was re­
on the characteristics of a Ti-6Al-4V powder with D90 = 59.9 µm. In this ported that the chemistry of powder particles collected from different
study, the authors focused on the duration of powder handling, rather locations after build number six was also unchanged. The elemental
than the number of reuse cycles. The number of powder reuse was the composition was still within the alloy-specific requirements after six
subject of several studies so far, whereas the time duration of powder times of reuse. There was no noticeable change in the flowability and no
handling received much less attention. The powder was used in an L-PBF clear trend in the PSD data after the reuse cycles. Nevertheless, the PSD
machine and sieved with an 80 µm mesh sieve under inert gas before data after the sixth build cycle showed a slight increase in the mean
reuse. The powder was used in total for 1750 h. Virgin powder was diameter of particles. Authors concluded that nickel alloy 718 powder is
added to the used powder every time the powder feeder could not be stable under electron beam and could go through many reuse cycles
filled sufficiently for a complete build job. This repeated mixing shifted without a noticeable change in powder and build chemistry.
the powder to a finer PSD after 100 reuse cycles, in contrast with results Gruber et al. [14] also reported a slight oxygen increase in the reused
of previous studies. Authors did not measure the oxygen content in the nickel 718 powder (from 146 ppm in the virgin powder to 266 ppm after
reused powder, but they did mention a negligible oxidation of the 14 reuse cycles). It was observed that nanoparticles rich in aluminum
finished parts. The UTS of the final parts as a function of reuse time and oxygen (possibly Al2O3) formed on the surface of the powder par­
showed scattered results but with a slight upward average trend. We ticles after 70 h of use in the E-PBF machine. This could be due to the
postulate that an increase in the powder oxygen content during reuse selective oxidation of aluminum on the powder particle surface. The
cycles explains this behavior. nitrogen content in the powder was stable throughout the build cycles.
Kakko et al. [81] used a so-called sub-lot approach to retain the Samples printed using this reused powder showed a higher number of
proportion of Ti-6Al-4V ELI powder and the built parts in a L-PBF ma­ oxide-rich regions, which could adversely affect the structural integrity
chine during reuse cycles. Three powder sublots, each having 35 kg of of the parts [83].
powder, were reused for 6 cycles separately. Thereafter, the sublots were The elements such as Ni, Cr, Al, Fe and Ti in Nickel 718 alloy have a
mixed in a certain proportion and reused in the machine resulting in 22 high tendency to form oxides in the presence of oxygen, which could
reuse cycles in total. During reuse cycles, the powder stayed within the lead to a possible formation of NiO, Cr2O3, Al2O3, FeO/Fe2O3 or TiO/
acceptance ranges prescribed by the ASTM F3001 chemical specifica­ TiO2. According to the Ellingham diagram [84,85], the Gibbs free en­
tion. There was a slight increase in the oxygen content (oxygen content ergy of formation of Al2O3 is the lowest among these oxides, giving it the

7
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

Fig. 5. Average oxygen content increase per

reuse cycle in Ti-6Al-4V powders reused in L-
PBF and E-PBF machines. The average rates are
derived and estimated from the data given in
the cited studies. The orange lines show the
average of all the studies cited herein without
rejuvenation (for the L-PBF and E-PBF pro­
cesses). The blue line shows the average for the
single reuse study with rejuvenation (for the E-
PBF process). Powder rejuvenation lowers the
average oxygen pickup in a powder batch. (For
interpretation of the references to colour in this
figure, the reader is referred to the web version
of this article)

highest stability. Therefore, based on entirely thermodynamic grounds, parameter fcircularity can drop by about 7% after 38 reuse cycles [36]. It
the formation of Al2O3 is favored compared to other oxides in Nickel 718 should be mentioned that there have been limited data available in the
alloy. The oxidation of Al, Ti, Fe or Cr in the E-PBF process seems to be literature on the reusability studies on nickel super alloys.
inevitable. Extremely low oxygen concentration can suppress these In a reuse study on nickel alloy 625 with 2% ceramic reinforcement
oxide formations on the powder surface, although it is not feasible in an in a L-PBF machine, it was shown that the tensile bars printed with 8-
AM machine [86]. This is because the equilibrium partial pressure of times reused powder had a consistent strength with bars printed from
oxygen at the melting point of these elements must be lower than ~ virgin powder [89].
10-7–10-9 mbar to avoid these oxidations. On the other hand, high pro­
cessing temperature in an E-PBF machine leads to a high diffusion rate of 3.5. Reusability of aluminum alloys in L-PBF
oxygen. Therefore, there is a driving force for the reaction to take place.
Cr2O3 and Al2O3 were observed to be formed on powder surface after the Until 2015, limited research had been published on the L-PBF pro­
L-PBF and E-PBF process. Although Cr2O3 has a lower affinity for oxygen cess of aluminum alloys [90]. However, aluminum L-PBF printing has
compared to Al2O3, its formation can be kinetically favorable, especially recently received a lot of attention, despite the processability challenges
considering also its higher content in Nickel 718 alloy [84]. During reuse associated with these alloys. These challenges include the high reflec­
cycles, one can expect that existing oxides decompose gradually into the tivity at 1 µm wavelength of solid state lasers and the thermal conduc­
most stable oxide (i.e. Al2O3) when exposed to high temperatures in the tivity [90–92]. Aluminum powders are relatively light and normally
AM process [87]. have a poor flowability. Therefore, a homogenous spread on the powder
Formation of satellite particles has been also reported to be one of the bed is challenging. One method to tackle this issue could be the utili­
major surface defects in the E-PBF process. This is caused by the elevated zation of a broader PSD range in a L-PBF machine. Aluminum powders
temperatures during pre-heating of the powder bed, which sinters par­ have a high affinity for surface oxidization to form Al2O3, which in turn
ticles together [88]. can increase the porosity [36,93,94]. Oxygen is first chemisorbed to
form an oxide layer and then it thickens further [95]. Therefore Al-based
3.4. Reusability of nickel alloys in L-PBF powder handling is critical to avoid such oxidation. Riener et al. [96]
observed a native oxide layer of approximately 5 nm in a fresh
Ardila et al. [51] studied the effect of nickel alloy 718 powder reuse Al-Si10-Mg powder as well as in the same powder stored in a humid
up to 14 cycles of (PSD: 15–45 µm) using a L-PBF machine with an environment. The authors concluded that humid environment leads to
operational oxygen level of maximum 2000 ppm. Before printing a new the physical adsorption of water molecules onto the particles surface
test part, the reused powder was sieved and dried without being reju­ which can be removed by a drying process. Raza [97] and coworkers
venated by virgin powder. The sieving process has removed only 0.6 kg also measured a similar native oxide thickness in a virgin Al-Si10-Mg
of powder from the initial 25 kg of virgin powder after the end of all the powder of approximately 4 nm. The oxide thickness grew to ~ 38 nm
reuse cycles. The particle size distribution of the reused powder showed after reusing the powder in a L-PBF machine for 30 months (approxi­
a slight coarsening after the seventh reuse cycle, which might be due to mately 5000 h of system operation time). It should be noted that the
the aggregation and fusion of particles in the vicinity of the melt pool. powder was rejuvenated with virgin powder after each reuse cycle. After
No significant change in the chemical composition was reported except 30 months of powder reuse, a volume fraction of ~ 3% of heavily
for possibly a slight oxidation of nickel. The average volume porosity oxidized spatter particles with a non-uniform oxide layer of 50–125 nm
showed a marginal increase after the third reuse cycle. No apparent was observed in the powder batch.
correlation could be found between the reuse of powder and the Another issue in the L-PBF process is that the oxide surface on the
toughness of the material. Authors concluded that the mechanical and molten pool introduces an obtuse contact angle and therefore it hinders
chemical properties of nickel 718 do not significantly change during the the wettability of aluminum [95]. Read et al. [16] observed unmelted
reuse cycles in the L-PBF process. powder particles with a diameter of approximately 6–30 µm on the
In the study of Cordova et al. [36], nickel 718 was reused 38 times in surface of the Al-Si10-Mg tensile samples, which could be due to the
a L-PBF machine. However, the powder was rejuvenated with virgin presence of oxide layers on these particles. These local and thick oxide
powder. Therefore, conclusions on the effect of reuse on the powder layers could inhibit the local consolidation on these sites. These loca­
properties cannot be drawn. Particle elongation and formation of sat­ tions could initiate crack sites leading to large failures in the built parts.
ellites were observed after 38 reuse cycles. It was reported that the One of the few studies on the reusability of Al-Si10-Mg and

8
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

Scalmalloy® (an aluminum-magnesium-scandium alloy) in a L-PBF schematically shown in Fig. 7. The first strategy is called a single batch
machine has been carried out by Cordova and coworkers [36]. The au­ strategy, in which a single batch of powder (i.e. virgin powder quantity
thors investigated gas atomized Al-Si10-Mg and Scalmalloy® powders in produced under traceable conditions from one or different powder lots)
a L-PBF machine after six and three reuses respectively. In their study, will be used in several build cycles and after each cycle the powder is
Al-Si10-Mg was shown to be highly influenced by the reuse cycles sieved to remove sintered agglomerates or deformed angular particles.
regarding PSD, morphology and the increase in the oxygen content. The The sieving is performed with a mesh size opening larger than the upper
oxygen content doubled after six reuse cycles from 0.005 to 0.01 wt%. In cut of the PSD used. The powder batch could be reused until it becomes
contrast, Scalmalloy® showed even a slight decrease in the oxygen degraded/OOS (out of specification) or until the remaining powder is
content after three reuse cycles. It is important to note that since the not sufficient for the completion of the next build cycle. The degradation
number of reuse cycles were not the same for these two powders, a level of powder includes changes in chemical composition of the powder,
comparison between the two is not feasible. density, porosity, morphology, flowability and narrowing of PSD. When
In an earlier study using X-ray photoelectron spectroscopy (XPS), the powder is degraded to a specific level, its use would be detrimental
Maamoun et al. [98] showed that the surface of the gas-atomized to the properties and quality of the printed parts.
Al-Si10-Mg powder particles does not display a significant change in Powder manufacturers might offer a so-called reconditioning of
oxygen content after 18 reuse cycles. They suggested that Al-Si10-Mg powders, meaning that the reused and OOS powder can be recondi­
powder could be reused for part fabrication assuming that a proper tioned and brought back to the initial specification. This process usually
sieving is in place. includes characterizing the chemical properties of the OOS powder to
Asgari et al. [21] showed that the Al-Si10-Mg powder used in an AM determine the elemental composition, sieving to the PSD upper size limit
machine has an aluminum matrix (α-Al) with a distributed fibrous and blending the powder with a determined amount of virgin powder.
network of the silicon phase [21]. Interestingly, the mechanical prop­ This process is illustrated in Fig. 8.
erties (YS, TS and tensile elongation) of the as-built and machined The second strategy is called the frequent refreshing (or Top Up)
Al-Si10-Mg fabricated by reused powder was comparable to the parts strategy, in which the used powder is rejuvenated and blended with an
fabricated by virgin powder. These properties were much better amount of a virgin powder after a defined number of build cycles or in
compared to cast aluminum counterparts. between each cycle. This strategy enables AM users to save powder costs
Fig. 6 summarizes UTS and YS changes for Ti-6Al-4V and Al-Si10-Mg and to reuse the powder for a higher number of build cycles compared to
powders in the E-PBF and L-PBF process from different reuse studies. the single batch strategy. This second strategy also reduces the powder
Generally, there is a slight increase in the UTS and YS of Ti-6Al-4V parts waste (see Fig. 7), but only partially addresses the inhomogeneity of the
printed with reused powder, due to the increase of oxygen interstitials in powder if no proper blending is performed. However, when using the
the material’s structure. However, there was an exception in Carrion’s frequent refreshing strategy, a fraction of the powder can reach a very
study, where the UTS declined in the as-built annealed specimens (see high degradation level. One disadvantage would be the loss of trace­
the blue dashed line in Fig. 6). Carrion and coauthors [99] observed this ability since the powder is topped up with virgin powder of different
contrast and stated that powder reuse has a slight effect on the tensile batches [100,101]. This is not desired especially in industries like
and fatigue properties of L-PBF Ti-6Al-4V parts. Al-Si10-Mg parts did not aerospace and biomedical where traceability is critical.
show a significant change in the UTS and YS results by printing with Lutter‑Günther et al. [100] proposed another strategy for powder
15-times used powder. No data could be found for Nickel 718 alloys. reuse, that is to mix used powders with the same reuse age. This allows
to have sufficient powder for a build job and to have a lower powder
variation within the batches of reused powder. But it demands storage
3.6. Powder reusability strategies and control on powder batches.
Another strategy is powder spheroidization which transforms the
There are generally two strategies for powder reuse, which are

Fig. 6. The effect of multiple reuse cycles on UTS (ultimate tensile strength) and YS (yield strength) of printed Ti-6Al-4V and Al-Si10-Mg parts in the E-PBF and L-PBF
process. The results have been produced from references [21,47,54,73]. No rejuvenation of the reused powder lot was conducted during these studies, and all the
tensile specimens were printed vertically. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article)

9
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

Fig. 7. Two different powder reuse strategies

shown schematically: (i) The single batch
strategy, in which a single batch of powder is
used several times in an AM machine without
being rejuvenated with virgin powder. (ii) The
frequent refreshing (or Top Up) strategy, in
which the powder is rejuvenated with virgin
powder (white bars) frequently in order to
reduce the material waste, the costs and to in­
crease the number of possible reuse cycles.
Darker bars represent a powder which has been
degraded more due to the reuse cycles.

Fig. 8. The reconditioning process including

blending of a used powder batch with a deter­
mined amount of virgin powder, in order to
bring the used powder back to an acceptable
condition (e.g. within-specification oxygen
content). The thin red layers on the powder
surface represent a native oxide layer on virgin
powder. The thicker red layers represent
heavily oxidized used powder. (For interpreta­
tion of the references to colour in this figure,
the reader is referred to the web version of this
article)

used and degraded feedstock into spherical and dense powder particles goes through multiple printing cycles inside a machine chamber, gets
[25,34,102]. It can be applied to various melting point materials from sieved in different chambers, comes in direct contact with machine
pure metals (e.g. W, Mo and Ta) to ceramics (e.g. SiO2 and ZrO2) at tools, oil, reactive gases, moisture, build chamber erosion debris, storage
different particle sizes. Powder is spheroidized by a plasma torch chambers, etc., which can adversely affect the performance of the
through in-flight melting and consequent solidification by giving suffi­ powder, and/or change its chemical composition. In addition, some
cient flight time [102]. Powder spheroidization leads to an enhanced impurities are very difficult to remove. There should be a possibility to
density, less friction between particles and thus an improved flowability. recycle powders or parts by including them into the atomization feed­
This method could also improve the purity of the feed material by stock fabrication process or by directly melting them in a crucible prior
vaporizing possible impurities [102]. However, it can come at the cost of to atomization. In this latter case, the scrap parts or used powder can be
partial vaporization of low melting point elements such as aluminum melted as is or mixed with master alloy or elemental stock, depending on
[25]. compositional requirements.
Each of these reuse strategies has an impact on the costs, quality There are ongoing tests for such recycling methods [22]. Economi­
(variations in the powder and part quality), and ease of implementation cally speaking, recycling would be highly appealing for high-priced
[103]. powders such as titanium alloys. Improvements and developments are
continuing in order to enable AM users to recycle metal waste in a
4. Powder recycling cost-effective way that is compatible with metal powder AM operation.
An important aspect that should be ensured during the recycling of
Recently, the recycling of AM alloys into feedstock that will be powders is traceability, especially in industries like aerospace, where
turned into powder for AM processing has attracted the attention of the quality control and standardization are crucial [21].
industry. Material recycling has a number of environmental and eco­
nomic benefits including significant energy and cost savings [3104]. AM 5. Summary and perspectives
usually generates less waste material compared to other conventional
fabrication methods. The main concern is the feasibility of the direct use A current challenge for AM is to develop standards for the powder
of these materials in additive manufacturing. There are two categories of feedstock characteristics that ensure printed parts with consistent me­
materials that are considered for recycling: 1) used and out of specifi­ chanical and physical properties [98]. This can be difficult, due to
cation AM powder and 2) scrap metals or built parts. variation in virgin powder characteristics across powder suppliers, AM
Accumulation of impurities in these materials is a concern and may machine/parameters, exposure duration of powders to the energy
disqualify the recycling of these powders or parts for direct use. Powder source in an AM machine, sieving machines and processes, etc. In

10
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

Table 1, we have depicted a summary of most of the studies done on the batch and equal distribution of the powder particles. On the other
reuse of Ti-6Al-4V, nickel 718 and Al-Si10-Mg powders in L-PBF and hand, one should take into account that sieving and blending of
E-PBF machines. We observe that oxygen pickup is an issue for metal powders (such as titanium and aluminum powders) can in­
Ti-6Al-4V powders owing to the high reactivity with oxygen. crease the level of static electricity, which leads to a decline in
Oxygen pickup for the Ti-6Al-4V powder was significantly higher in flowability [75].
an E-PBF process than in a L-PBF process (see Fig. 5 and Table 1). • Moisture was not considered as an important factor for a comparison
Derived from the reviewed studies, the average increase of oxygen of powders after reuse. Moisture can reduce the flowability due to
content in a powder lot was estimated to be ~ 0.001 wt% (~ 10 ppm) capillary effects [1,47]. Powder drying before use and storage of
per reuse cycle (or per 22 h of build time) in the L-PBF process versus ~ powder under inert gas is highly recommended. Drying powder in an
0.007 wt% (~ 70 ppm) per reuse cycle (or per 14 h of build time) in the oven for several hours at non-elevated temperatures (<80 ◦ C) could
E-PBF process. The rejuvenation of powder was shown to decrease the drastically increase the flowability. Keeping desiccant bags in the
apparent or net rate of oxygen pickup when accounting is done over powder containers is another way to decrease the humidity level.
many reuse cycles. Rejuvenation, i.e. the addition of virgin powder in However, there is a lack of consensus in the AM industry regarding
between reuse cycles, decreases the oxygen content in a powder lot, thus the use of desiccants. Besides, it is prohibited as per ISO/ASTM
lowering the net or apparent increase in oxygen content across many 52904.
reuse cycles. To clarify further, we can still anticipate an oxygen pick-up • Manufacturing methods and thus characteristics of virgin powder
for ~ 10 ppm (L-PBF) or ~ 70 ppm (E-PBF) if the powder lots are should be considered vital in reusability studies. Critical factors
inspected immediately prior and after the printing process (i.e. before include manufacturing method (e.g. gas atomized or plasma atom­
the rejuvenation process is carried out in between reuse cycles). In an E- ized), powder sphericity, starting PSD and flowability, starting oxy­
PBF machine, it is assumed that the preheating of the powder which is gen level, porosity, etc. For example, if there exists a noticeable
around 650 ◦ C under vacuum can drastically reduce the moisture con­ fraction of powder particles with angular shapes, this could increase
tent in the powder which is a source of oxygen pickup in titanium the surface oxidation because of the increased total surface area of
powders [105]. The quality of the vacuum is also assumed to be critical these particles [111]. As a second example, the presence of entrap­
for oxygen pickup. On the other hand, the reaction of titanium with ped gas pockets in Al-Si10-Mg parts was claimed as a possible
oxygen increases by increasing the temperature [106,107]. Oxygen consequence of improper spreading of powder, as the shape irregu­
pickup in an E-PBF machine is a gradual process caused by the reaction larity of the particles was high [16].
of titanium with water molecules (H2O) present both on the chamber • Powders from different powder manufacturers may behave differ­
walls and on the powder itself at elevated temperatures [60,107]. ently in an AM machine. For example, plasma atomization produces
Additionally, since an electron beam is an ionizing radiation, it causes powders with a generally higher sphericity and lower amount of
radiolysis of water molecules within the chamber. This leads to creation porosity compared to gas atomization. Garboczi and Hrabe [112]
of radical and oxidizing OH groups [108]. showed that a reused plasma-atomized Ti-6Al-4V powder is more
As seen in Table 1, agreeing on a general and accepted reuse process spherical and has lower average porosity and lower fraction of par­
or a number of allowable cycles for each alloy is not feasible using the ticles with pores compared to a reused gas-atomized powder. Using a
existing and available data. Below, we elaborate on the reasons thereof, plasma-atomized Al-Si10-Mg powder (compared to three
and we suggest several measures to be considered for future powder gas-atomized powders), Riener and coworkers [113] observed
reuse investigations. improved mechanical properties and surface quality of the built
parts, especially at high scan speeds above 2000 mm/sec. Higher
• The reviewed articles reported that the sphericity of powder particles sphericity and hence a lower surface area together with a minimized
remains nearly the same after being used. However, a coarsening/ surface tension of PA powders could lead to lower oxidation rate in a
roughening of particle surface was reported. A clear measure for L-PBF or E-PBF process compared to gas-atomized powders. The
sphericity calculation would be advantageous in order to have a characteristics of powders differ from one manufacturing method to
numerical comparison for this metric. Cordova [36] used the circu­ another and can affect the reusability of a powder lot. For instance,
larity index (fcircularity = 4πA/P2 ) to measure the average circularity the difference of the native oxide layer on the surface of Ti-6Al-4V
in SEM images. It is important to note that several factors cause powders can lead to variations in their electrical and thermal
variation in the circularity or sphericity index including the resolu­ behavior in an E-PBF machine [33]. Excellent powder quality leads
tion and magnification of analyzed images as well as the analysis to a higher degree of reproducibility and process stability [42].
software and its algorithm used to calculate circularity of particles. • There have been reports mentioning the effect of the AM machine
Therefore, circularity indices in different studies can vary. chamber itself on the used powder characteristics. For instance, in an
• In order to analyze the morphology of powder particles, typically E-PBF machine, a metallization process occurs on the heat shields
SEM images were visually analyzed to compare the irregularity and which can influence the local powder chemical homogeneity. Since
roughening of the particles in virgin and used powders. SEM images high vapor pressure elements can evaporate and condense on the
provide high spatial resolution. However, this comes at the expense heat shield, these can drop onto the powder bed. Therefore, under­
of analysis volume, since these images typically cover an area of standing the metallized layer becomes important [59]. In the case of
500 × 500 µm for each powder sample [109]. Complementary Ti-6Al-4V and nickel alloy 718, this metallized layer is rich in
methods are needed to have a better statistical understanding of the aluminum and chromium respectively. During reuse cycles of
morphological differences in powders. Dynamic image analysis is Ti-6Al-4V, AM users may observe a decrease in the aluminum con­
one method that can provide accurate information with high statis­ tent (due to evaporation) and after a few reuse cycles, a jump in the
tical confidence on particle size and sphericity. This method can aluminum content due to the metallized layer. The same holds for the
analyze millions of particles, but it captures only a reconstruction of chromium content in a nickel alloy 718 powder. Hence, cleaning
the particle shape [78,110]. heat shields between each build is recommended. Another example
• In almost all the studies reviewed in this article, blending of powder is the presence of water molecules on the E-PBF chamber walls which
did not take place (or has not been reported) before sampling from a can lead to oxygen pickup during elevated E-PBF processes.
used powder. Therefore, the used sample under investigation might • The active time duration of build jobs, build volume and geometry,
not be representative of the powder lot. There is variation in the and powder bed volume need to be measured and reported in powder
chemical composition of powder sampled from different locations in reusability studies. For instance, in an E-PBF machine, the building
the powder bed [59]. Blending ensures homogeneity of the powder time can range from about 10–100 h [33]. Measuring the build and

11
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

Table 1
An overview on the reusability investigations on Ti-6Al-4V, nickel 718 and Al-Si10-Mg carried out by different researchers in an E-PBF or L-PBF machine. 0 cycle refers
to a virgin powder. Types of powder include gas atomized (GA), plasma atomized (PA), or not defined (-). The horizontal bars show the number of reuse cycles that
were investigated by the authors. The green bars represent the successful cycles of powder reuse, i.e. with no significant impairment in the powder or built part
characteristics. A red cross represents an out-of-specification powder and/or part failure. A yellow cross represents our estimation for the cycle number at which the
powder will fall out of specification. (For interpretation of the references to colour in this table, the reader is referred to the web version of this article)

Number of Reuse Cycles

Type

Ref.
PSD

0 10 20 30 40

Nickel Alloy 718

45-106
E-PBF

100 kg inial powder batch [59]

-

15-45

~ 768 cm3 built volume 25 kg inial powder batch

GA

[51]
L-PBF

~15-63
GA

7 kg total printed parts out of 80 kg inial loaded powder [36]

Ti-6Al-4V
45-106

90 kg inial powder batch Grade [54]

-

23
45-106

480 hours of total build me Grade

PA

[62]
5
E-PBF

45-106

Grade
PA

[60]
5
45-106

100 kg inial powder batch Grade [59]

-

5
~15-63

Grade
GA

[74]
-
~25-53
L-PBF

Grade
GA

653 hours cumulated laser acve me [73]

23
15-45

Grade
GA

[47]
23

Al-Si10-Mg
~15-63
GA

[98]
L-PBF

~20-63

5 kg total printed parts out of 20 kg inial loaded powder

GA

[36]

12
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

powder volumes and the real usage time of a powder lot under laser [20] K. Dietrich, J. Diller, S. Dubiez-Le Goff, D. Bauer, P. Forêt, G. Witt, The influence
of oxygen on the chemical composition and mechanical properties of Ti-6Al-4V
or electron beam radiation, rather than reporting only the numbers
during laser powder bed fusion (L-PBF), Addit. Manuf. 32 (2020), 100980.
of reuse cycles, can help to set the stage for a realistic comparison. It [21] H. Asgari, C. Baxter, K. Hosseinkhani, M. Mohammadi, On microstructure and
makes a difference in the reusability of a powder if for example only mechanical properties of additively manufactured AlSi10Mg_200C using recycled
a small volume of powder is exposed to the energy source during powder, Mater. Sci. Eng.: A 707 (2017) 148–158.
[22] D.J. Fisher, Additive Manufacturing of Metals. Materials Research Forum LLC,
each printing cycle. In other words, if a small volume is printed in 2020.
each cycle, the powder might be used for a higher number of cycles. [23] C.F. Dixon, Atomizing molten metals — a review, Can. Metall. Q. 12 (3) (1973)
Conversely, if a large component is built, the powder would start to 309–322.
[24] C. Schade, J.J. Dunkley, Atomization[1], in: P. Samal, J. Newkirk (Eds.), Powder
degrade after a few cycles. Another important aspect is the part Metallurgy, 7, ASM International, 2015, p. 0.
surface area. A porous part brings on more powder exposure to the [25] P. Sun, Z.Z. Fang, Y. Zhang, Y. Xia, Review of the methods for production of
energy source (heat) compared to a dense part. Therefore, it is spherical Ti and Ti alloy powder, JOM Rev. 69 (10) (2017) 1853–1860.
[26] L. Yang, et al., Additive Manufacturing of Metals: The Technology, Materials,
essential to normalize the criteria values in order to obtain compa­ Design and Production, Springer International Publishing, 2017.
rable results in reuse studies. These criteria include, but are not [27] M. Kamal, U.S. Mohammad, A Review: Chill-Block Melt Spin Technique, Theories
limited to, build volume/powder bed volume fraction, part surface & Applications, Bentham Science Publishers, 2013.
[28] N.A. Yefimov, et al., Handbook of Non-Ferrous Metal Powders: Technologies and
area, the active laser or electron beam time, the power of energy Applications, Elsevier Science, 2009.
source and the number of laser sources in a L-PBF machine. [29] E.J. Hall, Nozzle intended for use in disintegrating apparatus, US Patent U.S.
Patent No. 1,545,253, 1925.
[30] P. Hellman, E. Josefsson, Means for continuously cooling powder produced by
granulating a molten material, US Patent U.S. Patent No. 3,771, 929, 1973.
Declaration of Competing Interest
[31] A. Lawley, Atomization of specialty alloy powders, JOM Orig. 33 (1) (2014)
13–18. -11-27 2014, doi: doi:10.1007/BF03354395.
P.M. is an employee of GE Additive Germany GmbH. T.P., [Link]., J. [32] A. Lawley, Preparation of metal powders, Annu. Rev. Mater. Sci. 8 (1) (1978)
AZ., J.K, F.M. and F.L. are employees of AP&C (a GE Additive Company). 49–71.
[33] S. Axelsson, Surface Characterization of Titanium Powders with X-ray
Photoelectron Spectroscopy, 2012.
References [34] Z.Z. Fang, J.D. Paramore, P. Sun, K. Chandran, Y. Zhang, Y. Xia, F. Cao,
M. Koopman, M. Free, Powder metallurgy of titanium – past, present, and future,
Int. Mater. Rev. 63 (7) (2018) 407–459.
[1] B. Dutta, S. Babu, B.H. Jared, Science, Technology and Applications of Metals in
[35] F. Henrik, L. Plöchl, M. Hohmann, Modern methods for the production of metal
Additive Manufacturing, Elsevier Science, 2019.
powders by gas atomisation, Millenn. Steel 2007 (2007) (2007) 103–108.
[2] R. Baker, Method of making decorative articles, US Patent Appl. US1533300,
[36] L. Cordova, M. Campos, T. Tinga, Revealing the effects of powder reuse for
1925.
selective laser melting by powder characterization, JOM Orig. 71 (3) (2019)
[3] J.O. Milewski, Additive Manufacturing of Metals: From Fundamental Technology
1062–1072.
to Rocket Nozzles, Medical Implants, and Custom Jewelry, Springer International
[37] S. Fashu, M. Lototskyy, M.W. Davids, L. Pickering, V. Linkov, S. Tai, T. Renheng,
Publishing, 2017.
X. Fangming, P.V. Fursikov, B.P. Tarasov, A review on crucibles for induction
[4] L. Bian, N. Shamsaei, J. Usher, Laser-Based Additive Manufacturing of Metal
melting of titanium alloys, Mater. Des. 186 (2020), 108295.
Parts: Modeling, Optimization, and Control of Mechanical Properties, Taylor &
[38] R. Gerling, M. Hohmann, F.P. Schimansky, Gas atomization of high melting
Francis, 2017.
reactive metals by a crucible- and ceramic-free technique, Mater. Sci. Forum
[5] A. Nouri, A. Sola, 8 - electron beam melting in biomedical manufacturing, in:
539–543 (2007) 2693–2698.
C. Wen (Ed.), Metallic Biomaterials Processing and Medical Device
[39] A. Lawley, Atomization, in: K.H.J. Buschow (Ed.), Encyclopedia of Materials:
Manufacturing, Woodhead Publishing, 2020, pp. 271–314.
Science and Technology, Elsevier, Oxford, 2001, pp. 387–392.
[6] B. AlMangour, Additive Manufacturing of Emerging Materials, Springer
[40] M. Entezarian, F. Allaire, P. Tsantrizos, R.A.L. Drew, Plasma atomization: a new
International Publishing, 2018.
process for the production of fine, spherical powders, JOM Orig. 48 (6) (1996)
[7] Standard terminology for additive manufacturing – general principles –
53–55.
terminology, West Conshohocken, PA, 2015.
[41] F.A. P.G. Tsantrizos, M. Entezarian, Method of production of metal and ceramic
[8] U. Ali, R. Esmaeilizadeh, F. Ahmed, D. Sarker, W. Muhammad,
powders by plasma atomization, U.S., 1998.
A. Keshavarzkermani, Y. Mahmoodkhani, E. Marzbanrad, E. Toyserkani,
[42] A. Alagheband, C. Brown, Plasma Atomization goes commercial, Met. Powder
Identification and characterization of spatter particles and their effect on surface
Rep. 53 (11) (1998) 26–28.
roughness, density and mechanical response of 17-4 pHstainless steel laser
[43] Plasma Atomization gives unique spherical powders, Metal Powder Report, 52,
powder-bed fusion parts, Mater. Sci. Eng. A 756 (2019) 98–107.
11, pp. 34–37, 1997.
[9] F. Ahmed, U. Ali, D. Sarker, E. Marzbanrad, K. Choi, Y. Mahmoodkhani,
[44] F. Larouche, M. Balmayer, F. Trudeau-lalonde, Plasma atomization metal powder
E. Toyserkani, Study of powder recycling and its effect on printed parts during
manufacturing processes and system therefor, Patent Appl. 2018.
laser powder-bed fusion of 17-4 pHstainless steel, J. Mater. Process. Technol. 278
[45] N. Town, S. Lawler, 12 - construction methods for small modular reactors (SMRs),
(2020), 116522.
in: M.D. Carelli, D.T. Ingersoll (Eds.), Handbook of Small Modular Nuclear
[10] P. Quinn, S. O’Halloran, J. Lawlor, R. Raghavendra, The effect of metal EOS 316L
Reactors, Woodhead Publishing, 2015, pp. 293–317.
stainless steel additive manufacturing powder recycling on part characteristics
[46] A. Kaufmann, Rotating gas jet apparatus for atomization of metal stream, Patent
and powder reusability, Adv. Mater. Process. Technol. 5 (2) (2019) 348–359.
Appl. 323,347, 1974.
[11] N.E. Gorji, P. Saxena, M. Corfield, A. Clare, J.P. Rueff, J. Bogan, P. González,
[47] P.E. Carrion, A. Soltani-Tehrani, N. Phan, N. Shamsaei, Powder recycling effects
M. Snelgrove, G. Hughes, R. O’Connor, R. Raghavendra, D. Brabazon, A new
on the tensile and fatigue behavior of additively manufactured Ti-6Al-4V parts,
method for assessing the recyclability of powders within Powder Bed Fusion
JOM, Orig. Pap. 71 (3) (2018) 963–973.
process, Mater. Charact. 161 (2020), 110167.
[48] E. Santecchia, S. Spigarelli, M. Cabibbo, Material reuse in laser powder bed
[12] D. Destefani, Introduction to titanium and titanium alloy, Met. Handb. (1990)
fusion: side effects of the laser—metal powder interaction, Metals 10 (3) (2020)
586–591.
341.
[13] J.C. M. Li, Microstructure and Properties of Materials (no. v. 2). World Scientific,
[49] ASTM F2924-14, Standard specification for additive manufacturing titanium-6
1996.
aluminum-4 vanadium with powder bed fusion, ASTM,, West Conshohocken, PA,
[14] H. Gruber, M. Henriksson, E. Hryha, L. Nyborg, Effect of Powder Recycling in
2014.
Electron Beam Melting on the Surface Chemistry of Alloy 718 Powder, Metall.
[50] Standard Specification for Additive Manufacturing Titanium-6 Aluminum-4
Mater. Trans. A Orig. 50 (9) (2019) 4410–4422.
Vanadium ELI (Extra Low Interstitial) with Powder Bed Fusion, A. F3001–14,
[15] A.R. Balachandramurthi, J. Moverare, S. Mahade, R. Pederson, Additive
West Conshohocken, PA, 2014.
manufacturing of alloy 718 via electron beam melting: effect of post-treatment on
[51] L.C. Ardila, F. Garciandia, J.B. González-Díaz, P. Álvarez, A. Echeverria, M.
the microstructure and the mechanical properties, Materials 12 (1) (2019) 68.
M. Petite, R. Deffley, J. Ochoa, Effect of IN718 recycled powder reuse on
[16] N. Read, W. Wang, K. Essa, M.M. Attallah, Selective laser melting of AlSi10Mg
properties of parts manufactured by means of selective laser melting, Phys.
alloy: process optimisation and mechanical properties development, Mater. Des.
Procedia 56 (2014) 99–107.
65 (2015) 417–424.
[52] R. Cunningham, A. Nicolas, J. Madsen, E. Fodran, E. Anagnostou, M.D. Sangid, A.
[17] B. Li, H. Wang, J. Jie, Z. Wei, Effects of yttrium and heat treatment on the
D. Rollett, Analyzing the effects of powder and post-processing on porosity and
microstructure and tensile properties of Al–7.5Si–0.5Mg alloy, Mater. Des. 32 (3)
properties of electron beam melted Ti-6Al-4V, Mater. Res. Lett. 5 (7) (2017)
(2011) 1617–1622.
516–525.
[18] R. Casati, M. Hamidi Nasab, M. Coduri, V. Tirelli, M. Vedani, Effects of platform
[53] P. Bidare, I. Bitharas, R.M. Ward, M.M. Attallah, A.J. Moore, Laser powder bed
pre-heating and thermal-treatment strategies on properties of AlSi10Mg alloy
fusion in high-pressure atmospheres, Int. J. Adv. Manuf. Technol. 99 (1) (2018)
processed by selective laser melting, Metals 8 (11) (2018) 954.
543–555.
[19] X. Liu, C. Zhao, X. Zhou, Z. Shen, W. Liu, Microstructure of selective laser melted
AlSi10Mg alloy, Mater. Des. 168 (2019), 107677.

13
P. Moghimian et al. Additive Manufacturing 43 (2021) 102017

[54] H.P. Tang, M. Qian, N. Liu, X.Z. Zhang, G.Y. Yang, J. Wang, Effect of powder [83] H. Gruber, P. Karimi, E. Hryha, L. Nyborg, Effect of Powder Recycling on the
reuse times on additive manufacturing of Ti-6Al-4V by selective electron beam Fracture Behavior of Electron Beam Melted Alloy 718, Powder Metall. Prog. 18
melting, JOM Orig. 67 (3) (2015) 555–563. (1) (2018) 40–48.
[55] I. Gibson, D. Rosen, B. Stucker, Additive Manufacturing Technologies: 3D [84] Y.N. Zhang, X. Cao, P. Wanjara, M. Medraj, Oxide films in laser additive
Printing, Rapid Prototyping, and Direct Digital Manufacturing, Springer, New manufactured Inconel 718, Acta Mater. 61 (17) (2013) 6562–6576.
York, 2014. [85] H.J.T. Ellingham, Reducibility of oxides and sulfides in metallurgical processes,
[56] D.B. Williams, C.B. Carter, Transmission Electron Microscopy: A Textbook for J. Soc. Chem. Ind. 63 (1944) 125–133.
Materials Science (no. v. 1), Springer, 2009. [86] A.N.D. Gasper, B. Szost, X. Wang, D. Johns, S. Sharma, A.T. Clare, I.A. Ashcroft,
[57] W.D. Callister, Materials Science and Engineering: An Introduction, John Wiley & Spatter and oxide formation in laser powder bed fusion of Inconel 718, Addit.
Sons Incorporated,, 2007. Manuf. 24 (2018) 446–456.
[58] I. Utke, S. Moshkalev, P. Russell, Nanofabrication Using Focused Ion and Electron [87] H. Yu, S. Hayashi, K. Kakehi, Y.-L. Kuo, Study of Formed Oxides in IN718 Alloy
Beams: Principles and Applications, Oxford University Press, 2012. during the Fabrication by Selective Laser Melting and Electron Beam Melting,
[59] P. Nandwana, W.H. Peter, R.R. Dehoff, L.E. Lowe, M.M. Kirka, F. Medina, S. Metals 9 (1) (2019) 19.
S. Babu, RecyclaBility Study On Inconel 718 and Ti-6Al-4V Powders For Use In [88] T. Sanviemvongsak, D. Monceau, B. Macquaire, High temperature oxidation of IN
Electron Beam Melting, Metall. Mater. Trans. B Orig. 47 (1) (2015) 754–762. 718 manufactured by laser beam melting and electron beam melting: effect of
[60] V.P.R. Niñerola, Powder recyclability in electron beam melting for aeronautical surface topography, Corros. Sci. 141 (2018) 127–145.
use, Aircr. Eng. Aerosp. Technol.: Int. J. 87 (2) (2015) 147–155. [89] Powder recycling in laser powder bed fusion: improving processing, maintaining
[61] G. Shanbhag, M. Vlasea, The effect of reuse cycles on Ti-6Al-4V powder quality, Elementum 3D, 2020.
properties processed by electron beam powder bed fusion, Manuf. Lett. 25 (2020) [90] D. Gu, Laser Additive Manufacturing of High-Performance Materials, Springer,
60–63. Berlin Heidelberg, 2015.
[62] S. Ghods, E. Schultz, C. Wisdom, R. Schur, R. Pahuja, A. Montelione, D. Arola, [91] N.T. Aboulkhair, I. Maskery, C. Tuck, I. Ashcroft, N.M. Everitt, The microstructure
M. Ramulu, Electron beam additive manufacturing of Ti6Al4V: Evolution of and mechanical properties of selectively laser melted AlSi10Mg: The effect of a
powder morphology and part microstructure with powder reuse, Materialia 9 conventional T6-like heat treatment, Mater. Sci. Eng.: A 667 (2016) 139–146.
(2020), 100631. [92] M. Naeem, Laser Processing of Reflective Materials, Laser Tech. J. 10 (1) (2013)
[63] S. Ghods, R. Schur, E. Schultz, R. Pahuja, A. Montelione, C. Wisdom, D. Arola, 18–20.
M. Ramulu, Powder reuse and its contribution to porosity in additive [93] S. Ochiai, Mechanical Properties of Metallic Composites, CRC Press,, 1993.
manufacturing of Ti6Al4V, Materialia 15 (2021), 100992. [94] N.T. Aboulkhair, N.M. Everitt, I. Ashcroft, C. Tuck, Reducing porosity in
[64] Y. Cao, M. Delin, F. Kullenberg, L. Nyborg, Surface modification of Ti-6Al-4V AlSi10Mg parts processed by selective laser melting, Addit. Manuf. 1–4 (2014)
powder during recycling in EBM process, Surf. Interface Anal. n/a (n/a) (2020) 77–86.
1–5. [95] S.H. Huo, M. Qian, G.B. Schaffer, E. Crossin, 21 - aluminium powder metallurgy,
[65] A. Montelione, S. Ghods, R. Schur, C. Wisdom, D. Arola, M. Ramulu, PowdEr in: R. Lumley (Ed.), Fundamentals of Aluminium Metallurgy, Woodhead
Reuse In Electron Beam Melting Additive Manufacturing of Ti6Al4V: particle Publishing, 2011, pp. 655–701.
microstructure, oxygen content and mechanical properties, Addit. Manuf. 35 [96] K. Riener, S. Oswald, M. Winkler, G.J. Leichtfried, "Influence of storage
(2020), 101216. conditions and reconditioning of AlSi10Mg powder on the quality of parts
[66] Y. Sun, M. Aindow, R.J. Hebert, The effect of recycling on the oxygen distribution produced by laser powder bed fusion (LPBF), Addit. Manuf. 39 (2021), 101896.
in Ti-6Al-4V powder for additive manufacturing, Mater. High. Temp. 35 (1–3) [97] A. Raza, T. Fiegl, I. Hanif, A. MarkstrÖm, M. Franke, C. Körner, E. Hryha,
(2018) 217–224. Degradation of AlSi10Mg powder during laser based powder bed fusion
[67] M. Yan, M.S. Dargusch, T. Ebel, M. Qian, A transmission electron microscopy and processing, Mater. Des. 198 (2021), 109358.
three-dimensional atom probe study of the oxygen-induced fine microstructural [98] A.H. Maamoun, M. Elbestawi, G.K. Dosbaeva, S.C. Veldhuis, Thermal post-
features in as-sintered Ti–6Al–4V and their impacts on ductility, Acta Mater. 68 processing of AlSi10Mg parts produced by Selective Laser Melting using recycled
(2014) 196–206. powder, Addit. Manuf. 21 (2018) 234–247.
[68] M. Velasco-Castro, E. Hernández-Nava, I.A. Figueroa, I. Todd, R. Goodall, The [99] P.E. Carrion, A. Soltani-Tehrani, S.M. Thompson, N. Shamsaei, Effect of powder
effect of oxygen pickup during selective laser melting on the microstructure and degradation on the fatigue behavior of additively manufactured As-built Ti-6Al-
mechanical properties of Ti–6Al–4V lattices, Heliyon 5 (12) (2019) 02813. 4V, Solid Free. Fabr. Proc. (2018) 1366–1372.
[69] N. Kazantseva, P. Krakhmalev, I. Yadroitsev, A. Fefelov, A. Merkushev, [100] M. Lutter-Günther, C. Gebbe, T. Kamps, C. Seidel, G. Reinhart, Powder recycling
M. Ilyinikh, N. Vinogradova, I. Ezhov, T. Kurennykh, Oxygen and nitrogen in laser beam melting: strategies, consumption modeling and influence on
concentrations in the Ti-6Al-4V alloy manufactured by direct metal laser sintering resource efficiency, Prod. Eng. Orig. 12 (3) (2018) 377–389.
(DMLS) process, Mater. Lett. 209 (2017) 311–314. [101] Metal powder reuse regimes and impact on part reproducibility – Part I, 2018.
[70] R. Montanari, G. Costanza, M.E. Tata, C. Testani, Lattice expansion of Ti–6Al–4V [102] M. Boulos, Plasma power can make better powders, Met. Powder Rep. 59 (5)
by nitrogen and oxygen absorption, Mater. Charact. 59 (3) (2008) 334–337. (2004) 16–21.
[71] W. Wei, W. Wu, S. Fan, X. Duan, In-situ laser additive manufacturing of Ti6Al4V [103] A. Okello, V. Samper, Effective Powder Reuse Strategies, 2020.
matrix composites by gas–liquid reaction in dilute nitrogen gas atmospheres, [104] G. Gaustad, E. Olivetti, R. Kirchain, Improving aluminum recycling: a survey of
Mater. Des. 202 (2021), 109578. sorting and impurity removal technologies, Resour., Conserv. Recycl. 58 (2012)
[72] C. Zeng, H. Wen, H. Bellamy, P.T. Sprunger, P.J. Schilling, S.M. Guo, Titanium 79–87.
and nitrogen interactions under laser additive manufacturing conditions, Surf. [105] H.K. Rafi, N.V. Karthik, H. Gong, T.L. Starr, B.E. Stucker, Microstructures and
Coat. Technol. 378 (2019), 124955. mechanical properties of Ti6Al4V parts fabricated by selective laser melting and
[73] O.A. Quintana, J. Alvarez, R. Mcmillan, W. Tong, C. Tomonto, Effects of reusing electron beam melting, J. Mater. Eng. Perform. Orig. 22 (12) (2013) 3872–3883.
Ti-6Al-4V powder in a selective laser melting additive system operated in an [106] M.J. Donachie, Titanium: A Technical Guide, second ed., ASM International,
industrial setting, JOM Orig. 70 (9) (2018) 1863–1869. 2000.
[74] V. Seyda, N. Kaufmann, C. Emmelmann, Investigation of aging processes of Ti- [107] S.S. Al-Bermani, M.L. Blackmore, W. Zhang, I. Todd, The Origin of
6Al-4 V powder material in laser melting, Phys. Procedia 39 (2012) 425–431. Microstructural Diversity, Texture, and Mechanical Properties in Electron Beam
[75] M. Balmayer, F. Larouche, F. Marion, Metal powder atomization manufacturing Melted Ti-6Al-4V, Metall. Mater. Trans. A, Orig. 41 (13) (2010) 3422–3434.
processes, Patent Appl. PCT/CA2016/051240, 2017. [108] R. Monckton, Low-energy Electron Interactions with Water, University of
[76] Investigating the effects of multiple re-use of Ti6Al4V powder in additive Manchester, Manchester, UK, 2014.
manufacturing (AM), 2016. [109] W.K. Camp, et al., Electron Microscopy of Shale Hydrocarbon Reservoirs: AAPG
[77] A.A. Saha, S.K. Mitra, M. Tweedie, S. Roy, J. McLaughlin, Experimental and Memoir 102, American Association of Petroleum Geologists, 2013.
numerical investigation of capillary flow in SU8 and PDMS microchannels with [110] R. Xu, O.A. Di Guida, Comparison of sizing small particles using different
integrated pillars, Microfluid. Nanofluid. 7 (4) (2009) 451–465. technologies, Powder Technol. 132 (2) (2003) 145–153.
[78] A. Strondl, O. Lyckfeldt, H. Brodin, U. Ackelid, Characterization and control of [111] R.M. Mahamood, Laser Metal Deposition Process of Metals, Alloys, and Composite
powder properties for additive manufacturing, JOM Orig. 67 (3) (2015) 549–554. Materials, Springer International Publishing, 2017.
[79] H. Masuo, Y. Tanaka, S. Morokoshi, H. Yagura, T. Uchida, Y. Yamamoto, [112] E.J. Garboczi, N. Hrabe, Particle shape and size analysis for metal powders used
Y. Murakami, Influence of defects, surface roughness and HIP on the fatigue for additive manufacturing: technique description and application to two gas-
strength of Ti-6Al-4V manufactured by additive manufacturing, Int. J. Fatigue atomized and plasma-atomized Ti64 powders, Addit. Manuf. 31 (2020), 100965.
117 (2018) 163–179. [113] K. Riener, N. Albrecht, S. Ziegelmeier, R. Ramakrishnan, L. Haferkamp, A.
[80] L. Denti, A. Sola, S. Defanti, C. Sciancalepore, F. Bondioli, Effect of Powder B. Spierings, G.J. Leichtfried, Influence of particle size distribution and
Recycling in Laser-based Powder Bed Fusion of Ti-6Al-4V, 19, 2, pp. 1213–2489. morphology on the properties of the powder feedstock as well as of AlSi10Mg
[81] K. Kakko, J. Porkka, K. Kupi, R. Ruohomaa, N. Raitanen, V. Luotonen, Evaluation parts produced by laser powder bed fusion (LPBF), Addit. Manuf. 34 (2020),
of Powder Aging Effects on Process and Properties of Ti-6Al-4V ELI, 2019. 101286.
[82] D. Deng, Additively Manufactured Inconel 718: Microstructures and Mechanical
Properties. Department of Management and Engineering, Linköping University,
2018.

14