Progress in Materials Science 119 (2021) 100795

Contents lists available at ScienceDirect

Progress in Materials Science

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

Emerging metallic systems for additive manufacturing: In-situ

alloying and multi-metal processing in laser powder bed fusion
S.L. Sing 1, S. Huang 1, G.D. Goh , G.L. Goh , C.F. Tey , J.H.K. Tan , W.Y. Yeong *
Singapore Centre for 3D Printing, School of Mechanical & Aerospace Engineering, Nanyang Technological University, Singapore

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

Keywords: While significant progress has been made in understanding laser powder bed fusion (L-PBF) as
Additive manufacturing well as the fabrication of various materials using this technology, there is still limited adoption in
3D printing the industry. One of the key obstacles identified is the lack of materials that can truly manufacture
Powder bed fusion
functional parts directly with L-PBF. This paper covers the emerging research on in-situ alloying
Selective laser melting
Processing parameters
and multi-metal processing. A comprehensive overview of the underlying scientific topics behind
In-situ alloying them is presented. The current state of research and progress from different perspectives (the
Multi-material materials and L-PBF processing parameters) are reviewed in order to provide a basis for follow-up
Machine learning research and development of these approaches. Defects, especially those associated with these
two material processing routes, are also elucidated by discussing the mechanisms of their for­
mation, including the main influencing factors, and the tendency for them to occur. Future
research trends and potential topics are illustrated. The final part of this paper summarizes
findings from this review and outlines the possibility of in-situ alloying and multi-metal processing
using L-PBF.

1. Introduction

Laser powder bed fusion (L-PBF), also commonly known as selective laser melting (SLM), is an additive manufacturing (AM)
technique [1,2] that has shown promising potential when applied to alloys [3] and ceramics [4,5]. It facilitates near net shape
manufacturing as the process starts from the preparation of computer aided design (CAD) data files which are subsequently aligned
with processing parameters and meshed into two-dimensional (2D) stacked layers by computer software. The production process
includes a loop of depositing layers of powder onto a substrate plate or previously processed layers, selectively melting the powder
with a high energy laser beam according to each layer profile, lowering the platform by one layer thickness, and then recoating a new
layer of powder. The process ends with the deposition and melting of the last sliced layer of the three-dimensional (3D) components
[6]. Fig. 1 illustrates the schematic of L-PBF process and principles. The white fonts indicate the machine parts and the black ones are
controllable parameters.
L-PBF allows quick 3D printing of parts with complex geometries directly from powders without the time consuming mold design
process [8–11]. Near net shape manufacturing makes them suitable for end applications or at least minimizes the extent of post-

* Corresponding author at: School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore. Singapore Centre for
3D Printing, Nanyang Technological University, Singapore. HP-NTU Digital Manufacturing Corporate Lab, Nanyang Technological University,
Singapore. NTU Institute for Health Technologies, Nanyang Technological University, Singapore.
E-mail address: [email protected] (W.Y. Yeong).
1
These authors contributed equally to this review article.

https://doi.org/10.1016/j.pmatsci.2021.100795
Received 7 August 2019; Received in revised form 17 November 2020; Accepted 5 March 2021
Available online 16 March 2021
0079-6425/© 2021 Elsevier Ltd. All rights reserved.
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 1. Schematic of a typical L-PBF build chamber and process. The white fonts illustrate the machine components while the black fonts illustrate
the key processing parameters [7].

processing. However, polishing and heat treatment are sometimes essential for L-PBF parts [12]. L-PBF results in superior properties in
parts compared to counterparts produced by conventional methods due to the ultrafine and graded microstructure attributed to the
rapid cooling and solidification cycles (with cooling rate of 103–106 ◦ C/s) during the process [13–15]. While L-PBF exhibits promising
potential in the production of parts with unique structure and properties even with metallic glass and metal matrix composite at a fair
cost [16–19], there are still limited applications of this process in the industry. One of the main challenges is the limitation of materials
available for L-PBF. As mentioned, there has been extensive work done on L-PBF processed materials, but most of these alloys were
designed for conventional methods. The widely known established materials for L-PBF are stainless steel, tool steel, Ti6Al4V, and
AlSi10Mg. Hence, there is a need for new alloys designed specifically for L-PBF. The capability of L-PBF to process powder mixtures has
opened new and exciting material research opportunities, especially for metal matrix composites which have been researched widely.
“In-situ alloying” is a term coined for using L-PBF to simultaneously fabricate functional parts and to create alloys via mixed powder
feedstocks. This strategy has the potential for rapid design and verification of new alloys. Moreover, the functionality of a part typically
requires more than one discrete material within the part. Hence, it is also important for L-PBF to be able to process multi-metal. While
there are great promises, defects unique to LPBF remain significant concerns. These include rough surface, porosity, and residual stress,
etc. Therefore, to achieve the L-PBF processing of multi-metal and in-situ alloying, a thorough understanding of both materials and L-
PBF processing parameters is necessary.
Several reviews that focused on the knowledge of the L-PBF process of metals or alloys are available [20–23]. The usage of powder
mixtures instead of pre-alloyed powder incurs changes in the physical properties of materials, thus affecting the melt pool’s thermal
and rheological behavior, resulting in the shift of the L-PBF processing window. A comprehensive understanding of the fabrication
process and its relationship with the part’s performance is necessary, and this work seeks to gain a thorough understanding of the
associated scientific and technological knowledge.
This review focuses on the state-of-the-art approaches within the perspectives of L-PBF, with focus on the in-situ alloying and multi-
metal processing strategies. Other additive manufacturing (AM) techniques are not within the scope of this review. Physical phe­
nomena in L-PBF will first be introduced to offer a basic understanding of the scientific challenges of these two approaches. An in-depth
review of the fabrication consideration related to in-situ alloying and multi-metal processing, including the materials and related
processing parameters, is then presented. Emphasis on parts’ physical properties and powder preparation techniques for in-situ alloying
are placed in this review. The review also addresses the defects unique to L-PBF associated with these emerging material systems. The
final part of this paper outlines the potentials and challenges for future research in these areas.

2. Laser powder bed fusion

L-PBF involves many physical phenomena such as absorption and reflection of laser radiation, heat transfers, phase transformation,
fluid flow, vaporization, emission of materials, and chemical reactions. Concurrently, powder to liquid and liquid to solid transitions
happens [24,25]. It is crucial to understand these physical phenomena and their relations to the successful manufacturing of defect-
free parts using L-PBF.
Heat absorption occurs during the interactions between laser and powder. Photons from a laser beam are reflected numerous times,
however, only a small fraction of its energy is absorbed by the surface of the powder particles [26]. As such, the laser and powder
characteristics are both critical in affecting the L-PBF melt pool formation and geometry [27]. The power density distribution of laser
source which often follows axisymmetric Gaussian profile is given by:

2
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

( )
fP r2
Pd = 2
exp − f 2 (1)
πrb rb

where f is the distribution factor, P is the total power of the heat source, rb is the radius of the heat source and r is the radial distance of
any point from the axis of the heat source. From Equation (1), the laser type and laser diameter, also known as spot size, are accounted
for by f and rb, respectively. These have been reviewed elsewhere [22]. As the laser characteristics are often not as easily varied as
powder characteristics, due to them being machine-specific, emphasis in this discussion would only be placed on the controllable laser
parameters such as power and scanning speed. It is interesting to note that the laser beam can penetrate deeper into a powder bed than
into a bulk material of the same composition due to the multiple photons reflections and absorptions [28]. Laser and powder in­
teractions for in-situ alloying and multi-metal processing during L-PBF will be further discussed in later sections.
Heat transfers including heat conduction, convection, and radiation, occur simultaneously with fluid flow. Fluid flow can be
explained using gravity, buoyancy, Marangoni flow, and vaporization. In L-PBF, Marangoni flow and vaporization are more significant
than gravity and buoyancy, hence, they are more discussed [29]. Marangoni flow is surface tension driven [30–32] and happens due to
the pulling of liquid with low surface tension towards liquid with higher surface tension. In the melt pool, temperature or composition
gradients may generate surface tension difference which results in Marangoni flow [33]. As lasers with Gaussian profiles are typically
used in L-PBF, the center of the melt pools has the highest temperature. Thus, the fluid flow within the melt pools tends to be radially
outward [21]. Marangoni number Ma, which is used to measure the strength of the convective flow, is given below:
dγ LΔT
Ma = − (2)
dT μα

where μ is the dynamic viscosity, α is the thermal conductivity of the material, L is the characteristics length of the melt pool (typically
taken as the width of the melt pool), ΔT is the difference between the maximum temperature inside the pool and the solidus tem­
perature of the material, and dT
dγ
is the sensitivity of the surface tension with respect to temperature.
The material vaporizes extensively when the temperature exceeds its boiling point, and the generated recoil pressures drive the
fluid motion in return. High recoil pressure results in the removal of molten material by expulsion while low recoil pressure facilitates
smoothing of the melt pool in L-PBF [34–36]. Keyhole mode is formed when high recoil pressure is exerted on the melt pool with the
laser beam penetrating the material up to a certain layer thickness in L-PBF. Additionally, selective evaporation of volatile elements
changes the local and global material composition [25,37]. This is especially important for in-situ alloying and will be discussed in later
sections.
It is important to understand the physical phenomena, hence, a thorough understanding of the variation in material properties due
to the influence of processing parameters is important. The key processing parameters involved in L-PBF include laser power (P),
scanning speed (v), hatch spacing (h), layer thickness (d). Extensive research has been carried out on various alloys to study their effect
on part microstructures and properties [38–40]. The relationship between densification and processing parameters for in-situ alloying
and multi-metal processing resembles the one in typical processing of pre-alloyed powders, except there are additional complications
for the emerging systems. Such complications that are dependent on the processing parameters would be detailed in later sections.
The individual parameters P, v, h, and d can be combined into one equation:
P
ε= (3)
v∙h∙d

where ε represents volumetric energy density [41]. Since each material has specific laser absorbance and thermal properties, extensive
experimental work must be carried out to obtain processing windows for new material systems that can lead to the production of parts

Fig. 2. Macro-morphology of (a) keyhole and (b) conduction melt pools [44].

3
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

with desired properties. Due to the specific nature of these processing windows, performance degradation can occur if they are applied
to other materials [42,43].
Energy input determines the heat transition modes during L-PBF. Keyhole mode and/or conduction mode (Fig. 2), distinguished by
whether the boiling point is exceeded in the melt pools can occur during L-PBF. Also shown in Fig. 2, a typical melt pool can be divided
into three zones including deposition zone (DZ) on the top, re-melting zone (RZ) on the nether part, and heat-affected zone (HAZ) in
the peripheral area of the melt pool. The length of RZ in keyhole mode is larger compared to conduction ones, penetrating deeper into
the previous layers.
These two heat transition modes are related to the densification behavior in a L-PBF process. Yang et al. plotted a parameters map
for high relative density parts and revealed a limited region for keyhole mode whilst a wide region for conduction mode [44]. Sig­
nificant vaporization of materials and resultant keyhole pores lead to relatively high porosity for parts formed from keyhole mode.
Stable melt tracks with fewer defects are more attainable in conduction mode. Yadroitsev et al. asserted that the maximum temperature
of melt pools significantly increases with laser power and only slightly decreases with scanning speed [45]. However, Yang et al.
investigated the threshold for keyhole mode and highlighted the importance of laser energy density associated with scanning speed,
laser power, and layer thickness, among which scanning speed is the most significant factor, followed by laser power and layer
thickness [44]. Lower scanning speed indicates longer interactions between laser and materials with higher energy input which may
facilitate keyhole mode. Higher laser power, which results in higher energy input per second, and thinner layer thickness, which results
in lesser materials, show a similar tendency to facilitate keyhole mode [46]. Extremely low scanning speed, high laser power, and thin
layer thickness result in instability of the scanning tracks. In such cases, L-PBF is accompanied by an increase in the melt volume and a
decrease in the melt viscosity. Irregularity appears as the Marangoni effect becomes more significant. The vapor recoil pressure also
results in distortion of the melt tracks. The irregularity of the melt tracks, which leads to poor powder depositions, and keyhole pores
lead to low relative density. Overlapping hatches are expected to eliminate defects upon repeated re-melting of the part during
subsequent laser scanning. Hatch spacing, which is the distance between the laser passes, is a significant variable in determining the
energy density. Hence, the porosity and pore morphology vary due to different overlapping conditions of laser scanning tracks in L-
PBF. Likewise, the layer thickness is essential to layer fusion and hence, relative density. When a thicker layer thickness is used, more
input energy is needed in melting the powder particles and less heat is left for re-melting of the previous layer. This results in poorer
interlayer bonding and more residual pores. Qiu et al. simulated melt pool behavior and asserted that the interaction between the laser
beam and melt pool is particularly violent when thick powder layers were processed. This is characterized by significant melt splashing
and the increased velocity of melt flow within the melt pools. The increase in porosity and surface roughness as a result of the increase
in powder layer thickness is also validated by experiments [47].
Mishra et al. showed that there exists a simple linear relation of hatch spacing or layer thickness to the overlapping cross-sections of
the melt pools and thus to the relative density [48]. Many research reported a direct increase in relative density with the hatch spacing
or layer thickness [49–51]. However, in some research, there exists a threshold for hatch spacing [52,53] and layer thickness [54,55]
respectively. Below this threshold, decreasing them would lower the relative density of the parts. Tang et al. established that sufficient
overlap of melt pools to avoid incomplete melting is obtained if the following geometric criterion is fulfilled [56–58]:
( )2 ( )2
h d
+ ≤1 (4)
W D

where h is the hatch spacing, d is the layer thickness, W is the melt pool width, and D is the melt pool depth.
As concluded in research done by Yang et al., high densification is more achievable in conduction mode as a wider region for
parameter selection is available [44]. Nevertheless, careful optimization is necessary to avoid defects and imperfection.
The melt pool dimensions become too small when the wetting ability is not enough, the contact area between the melt pool and the
substrate then becomes considerably limited. This leads to unfavorable wetting, flowing, and spreading characteristics. Melt pools with
greater circumference-to-length ratios are supposed to show a more stable behavior [24,59]. Therefore, balling occurs due to the
longer melt pools which tend to become discontinuous with necking. This transition is also associated with laser power and layer
thickness as they also affect the energy density. Lower laser power leads to lower energy input and resembles higher scanning speed.
Although larger layer thickness enables a bigger melt pool [47,54,60], the melt pool is far away from the substrate, leading to a
relatively smaller contact area between the melt pool and substrate that results in the balling effect. The balling effect would be
discussed in later sections.
Detailed discussions about the formation of different regions within the processing window of a material have been addressed
elsewhere [22]. The determination of processing parameters for different alloys, during in-situ alloying and multi-metal processing,
thus requires in-depth investigations and would be discussed in later sections.

3. In-situ alloying using laser powder bed fusion

3.1. Aluminium and its alloys

In-situ alloying of aluminium based alloys has been employed to accomplish several tasks. It has been mainly used for producing 3D
printable aluminium alloys with low thermal expansion, crack-free, and can be age-hardened.
The addition of silicon into aluminium is known to improve its 3D printability by reducing the coefficient of thermal expansion
(CTE). A study on eutectic Al12Si (wt%) has been done via in-situ alloying in which mixed elemental powder has been utilized [61].

4
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

The method of homogenizing the powder blend is by using a tumbling mixer. By adjusting the L-PBF parameters to manipulate the rate
of cooling, silicon solubility in aluminium can be manipulated. X-ray diffraction (XRD) data shows that employment of low power or
high scanning speed can lead to a faster cooling rate and improve silicon solubility (Fig. 3).
The microstructure of this alloy mainly consists of aluminium rich α-Al phase and silicon phase. Fine microstructures can be seen
but it becomes coarser with increasing energy density, ε (Fig. 6). The coarsening mechanism can be attributed to nano-sized silicon
agglomeration and rejection of silicon from aluminium due to longer melt pool retention. The tensile properties are highly dependent
on the microstructure - low ε leads to crescent shape porosity while high ε leads to silicon agglomeration, both of which deteriorate the
ultimate tensile strength (UTS) and ductility.
Meanwhile, in-situ alloyed Al18Si (wt%) was studied along with its wear properties. [62]. A maximum relative density of 96% was
obtained. Instead of using elemental powder blend, pre-alloyed Al12Si (wt%) powder was mixed with silicon to obtain the desired
composition. Part with the best wear property was found to be those containing the most amount of nano-sized silicon. The wear
resistance was found to be superior to that of hot extruded classical hypereutectic AlSiCu alloy [63]. Due to the in-situ alloying process,
either unmelted silicon or large silicon precipitates are inevitably formed when ε is too low or too high. The best wear resistance can be
achieved when the porosity is reduced to a minimum and the majority of silicon are retained in nano-size. In a study, AlSi10Mg powder
was blended with silicon powder to achieve the desired total silicon contents of 25 wt% (Al25Si) and 50 wt% (Al50Si) [64]. It is found
that parts with high relative density of more than 99% can be achieved, however, the optimal processing parameters are dependent on
the silicon content. Furthermore, considerable refinement of the microstructure and primary silicon particles have been achieved
compared to cast materials. The tailorability of CTE with adjustment of silicon content has been shown. The CTE is reduced by 0.2 ×
10− 6 1/K per wt% silicon, which gave a total CTE reduction of 43% for Al50Si.
Contrary to conventional belief, there was an inverse relationship between wear resistance and primary silicon phase size. The
larger silicon phase was speculated to form fine silicon particles layer due to cracking and dislodging when subjected to wear, acting
like a shield [65]. It was shown that the primary silicon phase size is related to the cooling rate and a slow cooling rate leads to a larger
primary silicon phase from L-PBF [66]. The same authors also subjected the in-situ alloyed Al50Si to heat treatment at a temperature
between 300 and 600 ◦ C, followed by water quench [67]. It was found that the porosity increases after a heat treatment at the
temperature of 550 ◦ C and 600 ◦ C. Moreover, as-built Al50Si did not show a lamellar eutectic structure, but the characteristic eutectic
structure appears after a heat treatment process at 600 ◦ C (which is above the eutectic temperature). Like the effect of the sustained
melt pool, heat treatment leads to silicon diffusion out from the supersaturated Al(Si) phase. With increasing heat treatment tem­
perature, silicon crystallite size increases, and the mean residual stress was reduced. The observed silicon crystallite size, as well as
residual stress, are also homogenized after heat treatment in all heat treated samples.
As compared to pre-alloyed powder as feedstock, in-situ alloying of Al12Si requires higher ε to achieve similar relative density [68],
possibly due to the positive enthalpy of formation. The highest relative density obtained via in-situ alloying (99%) is still lower than
that of pre-alloyed feedstock (99.5%) for Al12Si used in the study. For parts of similar relative density, those produced via pre-alloyed
feedstock possess finer cellular microstructure, as observed using scanning electron microscopy (SEM). This is shown in Fig. 4 [69],
while in-situ alloyed ones have nano-sized agglomeration that resembles annealed parts (due to relatively higher energy density
needed). The significant differences in melting point and positive enthalpy of formation between aluminium and silicon lead to the
need for higher ε when in-situ alloyed, which can limit the type of obtainable microstructure due to the narrower processing window.
Moreover, the amount of unmelted silicon particles in in-situ alloyed part is more prominent (Fig. 5). The effect of these unmelted
particles is discussed in later sections.
Despite successes in printing AlSi based alloys that were originally designed for casting, wrought aluminium alloys remain a
challenge to be processed with L-PBF. Wrought aluminium alloys (such as the 2000, 5000, 6000, and 7000 series) are highly in demand
in the AM industry due to their strength as a result of precipitation hardening. The issue is that alloying components that enable
precipitation hardening also leads to a larger solidification range in the wrought aluminium alloys. This subsequently leads to hot
tearing during the solidification of columnar L-PBF parts due to epitaxial grain growth from the previous layer. Small equiaxed grains

Fig. 3. XRD analysis of the powder mixture and L-PBF processed samples with variations of (a) laser power and (b) laser scanning speed [61].

5
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 4. SEM images of microstructure of L-PBF processed Al-12Si with several laser power (a) 180 W, (b) 240 W and (c) 300 W [61].

Fig. 5. (a) Representative microstructure of the as a solidified alloy. The phase contrast in these microstructures is Si - grey and Al - white. (b) High
magnification SEM image obtained on the BD-TD plane of the AS alloy, revealing cellular solidification within each of the laser melted pool. Hatch
overlap region is marked by 1, and region 2 shows coarsening of Si phase just outside these overlaps. The phase contrast is Si -white and Al -grey. (c)
Inverse pole figure map obtained on the BD-TD plane of the AS specimen. Black solid lines indicate some of the melt pool borders. (d) High
magnification SEM image for the HS alloy (captured from the BD-TD plane), revealing a uniform distribution of Si particles [69].

Fig. 6. Nanoparticle Enhanced Powder (nano-sized Zr on wrought Al alloy powder) [70].

6
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

are preferred over columnar grains for wrought aluminium parts produced by L-PBF to accommodate thermal contraction strain that
leads to hot cracking. Nevertheless, the conquest usually revolves around material-specific parameters optimization. Achieving small
equiaxed grains remains a challenge. Essentially, Martin et al. picked up the challenge of printing Al 7075 and 6061 by inducing
heterogeneous nucleation sites to achieve equiaxial fine grains via the addition of 1 vol% nano-sized hydrogen stabilized zirconium
[70]. Zirconium forms Al3Zr nucleation phase when melted with aluminium which can act as nucleation sites. However, pre-alloying
of wrought aluminium alloys with zirconium via gas atomization process cannot solve the challenge as rapid grain coarsening can
occur during the atomization process. As such, in-situ alloying provides a viable alternative. The study by Martin et al. used electrostatic
assembly technique of zirconium on pre-alloyed Al 7075 and 6061 powders to achieve homogenous mixture and prevent settling.
Fig. 6 shows the resultant particle after electrostatic assembly. The usage of pre-alloyed wrought aluminium alloy powder + 1 vol%
zirconium resulted in crack-free Al 6061 (Fig. 7) and Al 7075 (Fig. 8) parts. Al 7075 was also fabricated with the addition of silicon to
eliminate hot cracking [71]. The addition of 4 wt% silicon refines grains and lowers the CTE to achieve crack-free parts. Grains in the 〈2
0 0〉 orientation was reduced and superseded by the 〈1 1 1〉 orientation. However, hot cracking prevention via silicon addition is at the
expense of ductility as compared to wrought Al 7075 that is not modified. Also, subsequent T6 heat treatment on the silicon modified
Al 7075 dissolves the fine microstructure and deteriorates strength.
Zirconium was also added to AlCuMg powder to reduce cracks that are formed during L-PBF [72]. Formation of cracks is prevented
and pronounced structural modification is observed due to precipitation of Al3Zr. These precipitates act as effective heterogeneous
nuclei which led to grain refinement. The addition of zirconium also led to an increase in tensile strength and yield strength, but a
decrease in elongation.
Due to the flexibility of the in-situ alloying technique, the Al-xCu alloys system has been in-situ alloyed with L-PBF as an effort for
compositional optimization before opting for the gas atomization process [73]. The melting point difference in aluminium and copper
leads to incomplete diffusion of copper which is subsequently seen as macro-elemental segregation. The inherent short melt pool
duration during L-PBF also adds to the problem of incomplete copper diffusion. Compositional and microstructural gradients can be
seen around the partially diffused copper particles. In general, the solidified melt pools consist of three distinct zones – the high cooling
rate zone, the low cooling rate zone, and the heat-affected zone. Among all compositions tested, Al33Cu (wt%) exhibits the highest
compressive strength of about 1000 MPa. The compressive failure of these alloys is controlled by the shear mode buckling of lamellar
eutectic structure. Hence, the addition of copper leads to a nano-scale eutectic structure that is beneficial for improving the
compressive strength. The Al33Cu contains high Al2Cu phase content and a larger volume fraction of fine eutectic structure, resulting
in high compressive strength. Samples from L-PBF with pre-heated powder bed also contains more Al2Cu intermetallic, resulting in
50% ultimate tensile strength improvement and better elongation despite having a coarser microstructure as compared to the one
without preheating [74]. In another study by Bartkowiak et. al, the in-situ alloying of Al-xCu was approached with a powder blend of
bigger aluminium particles with smaller copper particles (3 μm) [75]. The powder blending process was done with a tumbling mixer
and resulted in a homogenous blend without powder agglomeration. However, only single tracks had been studied and it was shown
that the powder mixture achieved homogenous microstructure with no visible macro-segregation after laser melting. There was also no
porosity observed in the melt tracks. A similar approach was also done on Al-xZn within the same research and homogenous cross-
section single tracks were also obtained.
Aside from application specific materials development, interesting utilization of the in-situ alloying process on the development of
anchorless selective laser melting (ASLM) has also been done. Support structures are typically needed in an L-PBF process to prevent
the warpage of overhang structures resulting from residual stress, but it is not required in ASLM. In principle, ASLM can be achieved by
preheating the powder bed to a high temperature (about 50% of the melting point of material) to achieve stress relief [76,77].
However, solid-state sintering of unprocessed powder can occur at this high temperature and lead to powder recoating issues. To
overcome this problem, mixed and blended powders with a higher melting point than the alloyed product can be used. One choice of
raw powder to be used is the pre-cursor powder of eutectic/hyper/hypo eutectic alloy. In fact, in-situ alloying of Al12Si had been used
to demonstrate ASLM [76]. In the study, the L-PBF processing capability of pre-alloyed and in-situ alloy Al12Si has been compared. The
Al12Si elemental powder blend can withstand substrate pre-heating of 380 ◦ C but the pre-alloyed Al12Si agglomerates after preheating

Fig. 7. Micrographs of etched Al6061, processed as received. Large cracks are observed in the absence of Zr (left). With the addition of Zr
nanoparticles, no cracking is observed, but there is some residual porosity (right). Rows indicate increasing magnification [70].

7
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 8. Micrographs of etched Al7075, processed as received. Large networks of cracks are observed in the absence of Zr (left). With the addition of
Zr nanoparticles, no cracking is observed, but there is some residual porosity (right). Rows indicate increasing [70].

at a temperature of only 100 ◦ C. With the high pre-heating temperature, 10 mm horizontal unsupported overhang structures with only
1 mm upward warpage at the tip (Fig. 9) were successfully fabricated. Furthermore, ASLM was also achieved via in-situ alloying of pre-
alloyed Al1.2 Mg (wt%) and Si14.8Cu6.5Ni(wt%) mixed to form Al 339 [77]. In-situ alloying with a preheating temperature of 380 ◦ C
successfully produced a horizontal unsupported overhang of 10 mm with only 0.1 mm warpage. Meanwhile, an attempt to use pre-
alloyed Al339 powder with 380 ◦ C preheating temperature led to powder agglomeration and resulted in recoater collision during
L-PBF.
The previous works relevant to in-situ alloying of aluminium based alloys with L-PBF are compiled in Supplementary Materials
Table S1.

3.2. Titanium and its alloys

3.2.1. Ti6Al4V based

Using Ti6Al4V in original composition, the effect of precursor powder types (mixed powder for in-situ alloying or pre-alloyed
powder) on microstructures and properties have been studied for L-PBF [78]. In the study, pre-alloyed Ti6Al4V powder was used
as a benchmark. Three different elemental powder mixtures were then prepared for study, namely: (1) simply mixing powders of
similar size, (2) mixing large titanium powder with smaller vanadium and aluminium powder, and (3) satellite smaller vanadium and
aluminium particles on larger titanium powder. It was found that the usage of (2) and (3) powder mixtures for L-PBF gave better
density, homogeneity, and hardness as compared to the usage of (1). However, the powder mixture (3) did not fare as well as the
powder mixture (2), mainly because the technique was not optimized in which the elemental vanadium powder used in this study is
not small enough to adhere to the titanium powder. Nevertheless, the key finding is that proper mixing and dispersing of small alloying
particles with parental elements can produce parts comparable to that of pre-alloyed powder feedstock.
Using a similar technique as Simonelli et al. [78], the same group attempts to suppress the highly columnar prior-β formation of
Ti6Al4V through the addition of iron, which promotes columnar to equiaxed transition (CET) [79]. The choice of element addition was
based on the growth restriction factor (Q) in which high Q favours constitutional supercooling that can lead to CET. Iron is a viable
choice as it demonstrates relatively high Q value, exhibits unfavourable kinetics to form unwanted intermetallic compounds with
titanium, and exhibits similar thermo-physical properties as titanium. The iron addition of up to 3 wt% leads to significant grain
refinement with little increase in benefits at 4 wt% (Fig. 10). The iron rejection during solidification inhibits the side growth of
columnar grain which reduces the competitive growth mechanism. Consequently, more columnar grains persist, leading to grain
refinement and less prominent texture as compared to L-PBF built Ti6Al4V. The addition of 3 wt% iron leads to retention of metastable
β grains and forms desirable laminar α + β microstructure (while maintaining fine prior beta) upon suitable heat treatment.

Fig. 9. Unsupported overhang components produced using the ASLM process from elemental Al + Si12 powder [76].

8
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 10. Optical micrographs of as-built samples revealing the microstructure: (a) Ti6Al4V, (b) (Ti6Al4V)-2Fe, (c) (Ti6Al4V)-3Fe, and (d)
(Ti6Al4V)-4Fe. The build direction is indicated by the white arrows [79].

To extend the existing L-PBF materials library, an attempt to manufacture Ti6Al4V + 10Mo (wt%) had been done [80]. Without the
addition of molybdenum, epitaxial growth from partial re-melting accompanied by stable planar solidification leads to large columnar
β grains in Ti6Al4V. However, no columnar β grain was formed after molybdenum addition as it leads to instability of the planar
solidification front and induces cellular solidification. As a result, the growth of β grains is toward the melt pool centre. The as-built
Ti6Al4V + 10Mo (wt%) was found to be fully β phase. Incompletely diffused molybdenum particles smaller than 10 μm were found to
be randomly distributed throughout the matrix but are too minute to be picked up by XRD. The incomplete diffused molybdenum can
lead to molybdenum-rich zones in which the β phase is fully stabilized. Depending on the temperature of heat treatment followed by
quenching, the microstructure can consist of either α + β or fully β phase. With the heat treatment process, the homogeneity of Ti6Al4V
+ 10Mo (wt%) can be improved, and it gets better with higher temperatures. Similar work was also done by mixing molybdenum and
iron (which are both β titanium stabilizer) into Ti6Al4V to form Ti10V2Fe5.5Al (wt%) [81]. With different heat treatment processes, it

Fig. 11. Backscattered electron SEM images showing the Pd powder (bright region) “doped” on Ti6Al4V powder [84].

9
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

was shown that the degree of α and β phases can be tailored to achieve a wide variety of mechanical properties.
Functionalization of implants through the improvement of anti-microbial properties had been done via in-situ alloying of Ti6Al4V
+ 1Cu (at%) [82]. The UTS of Ti6Al4V improved from 1243 to 1550 MPa after the minor addition of copper. Copper addition also
resulted in decreased E. coli growth by an order of magnitude and decreased S. aureus growth by two orders of magnitude. A similar
study was also done by Maxpherson et al. using Ti6Al4V + 5Cu (wt%) [83]. It was found that copper addition leads to a microstructural
change from α’ to extremely fine α + β Widmanstätten structure. Traces of Ti2Cu were also detected in the matrix that leads to strength
improvement. In addition, Ti2Cu also serves to promote antibacterial properties. In-situ alloyed Ti6Al4V + 5Cu (wt%) shows moderate
anti-bacterial property which is inferior to that of casted counterpart, which can be caused by the high cooling rate of L-PBF that leads
to lesser Ti2Cu formed. Maxpherson et al. also did studies on in-situ alloyed Ti6Al4V + 0.5Ag (wt%). The resultant alloy shows minimal
changes in microstructure but leads to three times increased ductility while retaining the same strength as Ti6Al4V. The silver addition
did show anti-bacterial property but is inferior to the in-situ alloyed Ti6Al4V + 5Cu (wt%) parts.
To improve corrosion resistance, nano-sized platinum was added into pre-alloyed Ti6Al4V powder [84]. A novel dual centrifugal
mixer was used in this study to avoid the possibility of powder agglomeration (low energy blender) and particle morphology change
(when a planetary ball mill is used). The bending of one batch of powder can be done in just several minutes with nano-sized platinum
being “doped” evenly on the surface of Ti6Al4V particles (Fig. 11). In-situ alloying of Ti6Al4V + Pd powder mixture resulted in ho­
mogenous platinum distribution with occasional platinum-rich areas in the form of lines (Fig. 12). Platinum addition does not affect
the mechanical properties significantly but leads to improvement in corrosion properties. It can eliminate the active anodic region by
shifting the corrosion potential into the passive region for Ti6Al4V, achieving improved corrosion resistance.
In short, in-situ alloying of Ti6Al4V with elements like molybdenum, iron, silver, copper, and platinum have been done to develop
new materials that improve biomedical related properties and to reduce the texture strength. Moreover, the pre-alloyed Ti6Al4V
system had been chosen as a benchmark to study the effect of precursor powder on the quality of products made via in-situ alloying. A
table to summarize notable works relevant to in-situ alloying of Ti6Al4V derived alloys with L-PBF can be found in Supplementary
Materials Table S2.

3.2.2. Titanium with β stabilizers

In-situ alloying of titanium alloys with β stabilizer via L-PBF is mainly for the making of biomedical implants. Biomedical titanium
alloys with β stabilizer are the largest group of alloys to undergo in-situ alloying. Biomedical implants play an important role to replace
missing, damaged or weakened biological structures [85]. Their impact on life is paramount for their ability to improve people’s
quality of life and even extend a person’s life span. Unfortunately, structural implants like hip joint replacements cannot always outlast
a human’s life. For instance, a comprehensive review by Ridzwan et al. shows that almost 10% of hip joint replacements done need to
undergo revision surgery and this number can increase depending on the patient’s condition. More than half of these revision surgeries
are due to the loosening of implants and can largely be attributed to the effect of “stress shielding” [86]. The “stress shielding” effect is
in accordance with the Wolff’s law which states that the bone in a healthy person or animal will adapt to the loads in which it is placed
[87]. Most of the current commercial implant materials, e.g. 316L stainless steel, CoCrMo, and Ti6Al4V were intended for other in­
dustries like the aerospace and marine industries. They are not specifically designed as implant materials, thereby potentially causing
problems to the receiving patients. CoCrMo and 316L stainless steel have high elastic moduli of more than 200 GPa while that of
human cortical bone range from 17.6 to 28 GPa [88], potentially leading to a detrimental problem of osteolysis. Despite having a lower
elastic modulus for Ti6Al4V, around 130 GPa [89], it is still many times higher than that of human cortical bone. This will lessen the
load being borne by the bone, leading to bone resorption and ultimately implant loosening. Moreover, implant loosening can also be
attributed to other causes like wear, high joint space fluid pressure, bacterial infection, and genetics [90]. In addition to implant
loosening, the alloying elements in these commonly used implant materials have been associated with cytotoxicity, for example, nickel
in 316L stainless steel, cobalt, and chromium in CoCrMo, as well as aluminium and vanadium in Ti6Al4V. The potential harm brought
about by these elements is not further discussed here since a comprehensive review on the toxicity of these elements has been done by
Wapner [91].

Fig. 12. Back scattered electron SEM images showing (a) microstructure and (b) magnified region of Ti6Al4V-0.2Pd fabricated by L-PBF where Pd
rich region can be seen as bright region [84].

10
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

To improve the quality and longevity of implants, one of the directions is the development of new materials. Therefore, new
β-titanium alloys composed of non-toxic elements such as niobium, tantalum, zirconia, molybdenum, and tin (Fig. 13) were since
designed to achieve lower elastic modulus, greater strength and better corrosion resistance [92]. β-titanium stabilizers like niobium
and tantalum have high biocompatibility and are the primary choices of elements for consideration when it comes to biocompatible
β-titanium alloys. Elements like zirconia, molybdenum, and tin can be added to modify phases and microstructures of resultant alloys if
needed.
An alternative way of implant’s bone modulus matching can be realized via proper manufacturing route to introduce porosity [93],
lattice structure [39,94,95], or achieve topological optimization [96]. With the free-form fabrication capability of AM, systematically
induced lattice structures (like homogenous porous material) and near-net-shape topologically optimized implants can be realized.
When paired with suitable alloy systems (such as low modulus β-titanium), AM can potentially produce functionally graded implants to
meet stiffness and strength criteria while achieving excellent osseointegration property at the same time. L-PBF is one of the AM
techniques that is suitable.
Low modulus β-titanium is often metastable and requires a high cooling rate during the manufacturing process to retain the β phase
[97]. It is often difficult to achieve a uniform cooling rate in a bulk part manufactured with conventional manufacturing techniques
like casting and forging. L-PBF is inherently designed for this due to its quench-like rapid laser melting and cooling cycle, aside from
being able to manufacture a near-net-shape β-titanium part straight from a CAD file. L-PBF’s unique thermal characteristics and history
offer the opportunity to tailor the microstructure of an alloy, potentially achieving different material properties as compared to those
produce via conventional methods [98].
In-situ alloying using L-PBF has been popular to achieve β-titanium using powder mixtures. This is due to the following reasons: (1)
as opposed to using pre-alloyed powder, in-situ alloying of powder mixture allows for flexible and rapid adjustment of alloy compo­
sitions. Different alloy compositions can be designed and manufactured within a short time frame for a quick study on the link between
compositional variation and material properties. Optimal compositions for many alloys have not yet been found for L-PBF which
allows more research opportunities using in-situ alloying, (2) pre-alloyed β-titanium powder containing refractory β stabilizers such as
niobium, tantalum, and molybdenum is not easily available due to the huge differences in melting point and density between the
stabilizers and titanium. The differences in material properties lead to challenges during the typical gas atomization process that
manufacturing the pre-alloyed powder feedstock. Moreover, small batch production of pre-alloyed powder can be expensive. In-situ
alloying has the potential to enable compositionally graded components. However, it is good to take note that achieving components
with compositional gradient remains a challenge for L-PBF which will be discussed further for multi-metal processing in later sections.
Past studies showed that promising properties can be obtained for β-titanium alloys produced via in-situ alloying. In-situ alloying of
β-titanium typically focused on the reduction of elastic modulus to reduce the stress shielding effect. The reduction of elastic modulus
had been done with several approaches which include (1) optimization of β-stabilizer addition into titanium, (2) addition of other
element(s) into binary titanium alloys with existing β-stabilizer and (3) porosity induction in β-titanium alloys. Despite the lack of pre-
alloyed powder, the study on alloys of different compositions as a mission to search for optimal biomedical materials is possible due to
the flexibility of in-situ alloying. The approaches for elastic modulus reduction were done via in-situ alloying with L-PBF and the key
findings of each research are as follow:

(1) Optimization of β-stabilizer addition into titanium

Fig. 13. Biocompatibility of metals (a): cytotoxicity of pure metals, and (b): relationship between polarization resistance and biocompatibility of
pure metals, CoCr alloy, and stainless steels [92].

11
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Low Young’s modulus with enough strength can be, in principle, obtained when full β-phase is retained with minimal alloying
components [38]. Past studies had shown that this minimal point can be obtained at 25.5 at% or 40 wt% niobium addition for binary
TiNb alloys upon quenching. Below the stated minimum niobium addition, undesired α’’ and ω phases will be formed, decreasing yield
strength and increasing the Young’s modulus. Above the stated minimum niobium addition, full β-phase can be retained, and the
Young’s modulus will increase slowly with additional niobium. As such, most of the past studies on L-PBF uses minimal niobium
addition of 40 wt% to retain the single β-phase. Experimentally, the minimum Young’s modulus was found to be at the composition of
Ti42Nb (wt%) [99]. Parameters optimization issues aside, most research done on L-PBF of TiNb are on completely metastable β instead
of near β compositions. In a typical binary low modulus titanium alloy with non-intermetallic forming β stabilizing element, it is known
that two instances of Young’s modulus minima can be seen with varying amount of β stabilizer (Fig. 14). The second minimum of near
42 wt% niobium is of greater interest in the community for its lower modulus as compared to that of the first minimum. Depending on
the thermal history, titanium alloy can exhibit phases like α, α’, α’’, metastable β, stable β, and ω through the addition of beta stabilizer
[100–102]. The α and α’ phases can be associated with as-quenched alloys with the minor addition of niobium while α’’, β, and ω
phases can be associated with that when niobium addition is near that of fully β stabilized composition [97]. The α’ and α’’ martensites,
as well as the metastable ω phase, are formed when the alloy is rapidly quenched in which their formations are highly dependent on the
amount of β-stabilizer added [103].
In-situ alloyed Ti-xNb using L-PBF where x = 0, 15, 25, 45 at% were studied by Wang et al. [104]. It was found that Ti25Nb has the
lowest Young’s modulus among other compositions tested due to its correspondence to full β retention with minimal niobium addition.
This composition also demonstrated the best in-vitro appetite forming capability, possibly due to it having the highest volume fraction
of β-titanium phase, which is associated with appetite nucleation. The addition of more niobium leads to more un-melted niobium
volume and decreases the volume fraction of the β-titanium phase. Nevertheless, the in-situ alloyed Ti40.5Nb (wt%) does achieve a
comparatively low average Young’s modulus of 77 GPa despite the existence of unmelted Nb particles [105]. Meanwhile, a study by
Huang et. al utilizes in-situ alloying to study Ti-xTa alloy of different compositions (x = 0, 10, 30, 50 wt%) [106]. The study found that
Ti30Ta possesses the lowest modulus (71 GPa) and the highest yield strength (920 MPa) among all compositions. The composition is a
near β composition where the phase constituent is α’’ dominated, indicating that it is not necessary to retain the full β phase for elastic
modulus minimization. By keeping the volumetric ε at a constant, it was found that the type of porosity changes from keyhole induced
porosity to lack of fusion porosity. The additional tantalum particles at higher compositions require additional energy to melt, hence
leading to insufficient energy for melting and fusing between layers when the volumetric ε were kept constant. Tantalum addition also
leads to more unmelted tantalum particles which can act as inoculants, causing the prior β-grains to undergo a columnar to equiaxed
transition. A similar study by Zhao et. al. investigated a narrower range of in-situ alloyed Ti-xTa compositions (x = 6, 12, 18, 25 wt%)
[107]. The composition for elastic modulus minimization was found to be at Ti25Ta, a composition similar to the study by Huang et. al.
[106]. The Ti25Ta alloy is α’ phase dominated, which is also reported by Brodie et. al. [108]. Despite being far from β-stabilized
compositions, in-situ alloyed Ti25Ta shows a low elastic modulus of 89 GPa with a high yield strength of 1029 MPa. It was also found
that Ti25Ta has the best corrosion resistance in ringer’s solution among all other compositions tested due to the formation of stable
Ta2O3 passive film. The Cl− anion in ringer’s solution can intensify the anodic dissolution of TiO2 passivating film. The addition of
tantalum causes a rise in Ta2O5 passive layer, leading to a more promising candidate for biomedical implants application. Despite
having some unmelted tantalum, in the study done by Sing et al., the tantalum dissolution has been proven to be sufficient for the
retention of low modulus β-phase. The in-situ alloyed Ti50Ta successfully achieved a low elastic modulus of 75.77 GPa with a high yield
strength of 882.77 MPa [109].
With laser re-melting, the texture strength of Ti42Nb is reduced where prior β-grains close to that of equiaxed grains can be
observed. Re-melting-induced texture weakening leads to an improvement of yield strength from 426 to 545 MPa while maintaining an
elastic modulus of 65 GPa. The strength improvement comes with a reduction of elongation from 25% to 11%. These mechanical
properties were achieved despite the alloy consisting of mostly α’ phase [110]. Another notable finding in this study is on the better

Fig. 14. Young’s modulus variation of as-quenched titanium-based alloy with the addition of beta stabilizer [99].

12
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

choice of ε in the range of 362 J/mm3 for the in-situ alloyed titanium-tantalum system, a much higher value as compared to a previous
study by Sing et. al (144 J/mm3) [109]. Laser re-melting strategy certainly shows promising potential to advance the in-situ alloying
using L-PBF. The main drawback of the strategy is the significantly increased build time, primarily due to the additional time needed
for laser scanning.
The study on near β titanium-niobium system is limited and only two studies were done via in-situ alloying with L-PBF. One research
team had printed parts with the composition of Ti35Nb (wt%) as an effort to study the effect of inhomogeneity on mechanical and
corrosion properties [111]. The as-fabricated and solution-treated (1000 ◦ C at 24 hrs followed by air cooling) samples were compared
in the study. XRD analysis shows that the as-printed part is of majority β phase with unmelted niobium particles, achieving a Young’s
modulus of 84.7 GPa. It is speculated that the concentrated laser energy source is the cause of the incomplete melting/diffusion of
niobium particles. Meanwhile, the solution-treated sample shows a similar XRD analysis results but without the XRD peaks from
niobium, indicating that better homogeneity was achieved. Large unmelted niobium particles act as arrest zones for shear bands and
eventually act as stress concentrators for grain bound sliding, leading to poor ductility of as-printed parts as compared to heat-treated
parts (38.5 – 47.3%). A slight decrease in yield strength was detected after the heat treatment. The heat-treated part also shows better
corrosion resistance due to its better chemical homogeneity. Unfortunately, no information on the change in elastic modulus was
shown after heat treatment. In another study, near β Ti20Nb (at%) was manufactured and traces of α’’ martensite were seen in XRD
[112]. It was found that the martensite traces decrease with increased ε, which is caused by better dissolution of niobium into the
matrix. No further studies were done on the alloy system in that research, possibly due to the presence of unmelted niobium. There is a
general lack of study on as-fabricated near β binary titanium alloy systems and their mechanical behaviors are not well understood.
Although not explicitly stated as an alloy designed for low modulus biocompatible implant, the work on the in-situ alloyed titanium-
rhenium system has laid the groundworks for the future L-PBF research on the system. Rhenium has almost twice the melting point of
titanium (3185 ◦ C and 1668 ◦ C respectively), which leads to incompletely melting of rhenium particles that can be reduced with
increasing ε [113]. The addition of 5.66 wt% rhenium in titanium greatly increases the yield strength (461 – 1038 MPa) while also
reduces the elastic modulus (119 – 103 GPa). However, the ductility of the alloy is largely sacrificed down to a value of only 2%. The
phase constituent after rhenium addition is mainly of α’ phase as the rhenium addition in the study is insufficient to retain the β phase.
Further studies on titanium-rhenium alloys were made on 2 and 4 wt% rhenium addition into titanium [114]. It was found that
rhenium addition can lead to the refinement of lath phase α/α’ microstructure which can be one of the factors behind the strength

Fig. 15. SEM and EDS images of polished Ti42Nb tensile test specimen (a) 10x SE mode (b) 50x SE mode (c) 1000x BSE mode and (d) EDS mapping
of region shown in (c) [119].

13
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

improvement. The nature of rhenium as β-Ti stabilizer leads to β phase retention around the incompletely diffused rhenium particles
with the formation of the brittle ω phase. The ω phase, coupled with refined microstructure and enhanced oxygen level from rhenium
addition eventually leads to brittle fracture characteristics. The brittle characteristic deteriorates fatigue crack growth resistance of
TiRe when compared to titanium. For the alloy system to be useful in structural application, further studies are required to ensure its
safe use with predictable failure from fatigue loading.
In some cases, in-situ alloying technique serves as a quick tool to screen through the effect of element(s) addition on a part’s
biocompatibility. Using in-situ alloying, Wang et. al. found that the addition of niobium into titanium leads to better in-vitro apatite
forming capability as compared to its pure titanium counterpart [104]. A similar study shows that seeding of MC3T3-E1 osteoblast
precursor cell line on in-situ alloyed Ti25Nb (wt%) leads to better cell spread and proliferation when benchmarked with titanium parts,
in addition to a superior in-vitro appetite forming capability [115]. Furthermore, in-situ alloyed Ti25Nb (wt%) scaffold demonstrated
impressive immunomodulatory properties [116]. The scaffold was also implanted into a rabbit’s tibia and shown to better promote
bone regeneration and osseointegration as compared to the pure titanium counterpart. As for Ti50Ta (wt%), Huang et. al. found that
using the same lattice structures, Ti50Ta possesses similar biological responses to titanium and Ti6Al4V based on the cell proliferation
tendency of osteosarcoma cell line SAOS-2 [106]. Meanwhile, Yan et. al. demonstrated that Ti15Ta10.5Zr does not induce infection
and osteomyelitis in the surrounding of rat’s tibia after five weeks of implantation in rat tibia’s fracture [117]. All of these observations
point toward the promising potential of in-situ alloyed titanium-based alloys as a biocompatible material for implant applications.
Pre-alloyed titanium with β-stabilizer is still not widely available and has very limited attempts to study on them. Despite that, a
study on Ti45Nb (wt%) where pre-alloyed powder feedstock was used has been made [118]. XRD patterns on the L-PBF part show
β-titanium peaks with broadening characteristics possibly due to residual stress in the part. The part showed a compressive strength of
723 MPa. Unfortunately, no Young’s modulus was reported in the study. Another study was done on Ti42Nb (wt%) processed by L-PBF
using pre-alloyed powder [119]. Parts built were more homogenous (contains no un-melted niobium) than those manufactured via in-
situ alloying and they have a relative density of more than 99.5%. Using SEM and energy-dispersive X-ray spectroscopy (EDS), Fig. 15
shows the polished surface of Ti42Nb and occasional microstructural inhomogeneity can be seen in Fig. 15(c). This feat can be
achieved with a relatively low Ed of 40 J/mm3 as compared to those done via in-situ alloying (214 J/mm3 when optimized [105]). Full-
β matrix was obtained and the alloy also shows better compressive yield strength as compared to tensile yield strength (674.08 MPa –
831.58 MPa).

(2) Addition of other element(s) into binary titanium alloy with existing β-stabilizer

The addition of tin to near β titanium-niobium composition (forming Ti37Nb6Sn (wt%)) has been experimented with to achieve
parts with low elastic modulus [120]. The tin addition suppresses hard and brittle ω precipitation in TiNb of near β composition, in
which the precipitation can lead to an increase in Young’s modulus. The study also found that increased tin leads to more preferential
[100] grain growth towards the build direction. The optimal ε value was found to be at 125 J/mm3 where the lowest Young’s modulus
(66 GPa) and high yield strength (~775 MPa) were obtained. It is interesting to take note that, unmelted niobium particles with an
average size of more than 30 μm were found to initiate secondary crack at the boundary, hence compromising the ductility. It was
further suggested that the low modulus of this alloy can be attributed to the uniform α’’-colony within the β-matrix. Another study
utilizes in-situ allying to develop new material for biomedical application by the addition of zirconium into titanium-tantalum alloy
[117]. Ti5Ta-xZr (x = 1.5, 5.5, 10.5, 15.5 wt%) were studied and it was found that zirconium can stabilize the β phase in these alloys.
The phase constituent of Ti15Ta10.5Zr is almost pure β, exhibiting an impressively low elastic modulus value of 42.93 GPa. A high
yield strength of 768.61 MPa is also obtained, demonstrating the vast potential of TiTaZr alloys system as a candidate for biomedical
implant material. Despite the outstanding properties, there are limited studies on the in-situ alloying process involving zirconium. The
speculated reason is due to the tendency of zirconium powder to undergo autoignition in the atmosphere, hence needing proper fa­
cilities and intensive care to handle.

(3) Induced porosity in β-titanium alloy

Macro-porous Ti40Nb (40 wt%) bulk L-PBF samples were produced via in-situ alloying of mechanically alloyed (MA) powder [93].
A relatively slow scan speed of 35 μm/s was used in the study. Traces of α and α’’ phases are found to be embedded in a β matrix based
on transmission electron microscopy (TEM) observation. No macro-segregation of niobium particles was reported due to the powder
preparation technique. Owing to the porous structure (~17%), elastic modulus as low as 33 GPa was achieved. Nevertheless, there is
no information reported on tensile mechanical strength. Sing et al. also used in-situ alloying to fabricate titanium-tantalum porous
structures and concluded that the L-PBF parameters can affect the physical and mechanical properties of these structures [39]. In
another study of in-situ alloyed porous titanium-molybdenum, it was discussed that the porosity level of the manufactured part has a
significant influence on corrosive properties [121]. The part’s corrosion resistance was found to be higher in parts with lower porosity.
Nevertheless, the parts with lower porosity values were also found to be more homogenous with lesser un-melted molybdenum
particles. It is hence difficult to conclude that porosity value alone had caused the change in corrosion resistance, as compositional
segregation can also play a part in corrosion resistance.
To summarize, most low modulus biocompatible titanium alloys manufactured by L-PBF were done using in-situ alloying. However,
all β-titanium stabilizers used are refractory and the porosity-segregation dilemma remains a challenge to be solved. More research in
this area is needed in order to address the porosity-segregation dilemma and better understand the suitability of in-situ alloyed
biocompatible titanium alloys as implant candidates. A table to summarize notable previous works relevant to L-PBF titanium alloys

14
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

with biocompatible β-stabilizer (mostly in-situ alloyed) can be seen in Supplementary Materials Table S3.

3.2.3. Other titanium systems

Aside from the widely used Ti6Al4V alloy and trending biomedical titanium alloys with β-stabilizer, other titanium alloys were also
explored for in-situ alloying using L-PBF. The Ti-185 alloy is a high strength and economical titanium alloy but the formation of β-flecks
(macro-segregation of iron) via ingot casting restricted its commercial use. As such, an attempt to produce Ti-185 with L-PBF has been
made due to its high cooling rate that can potentially suppress the macro-segregation of elements [122]. As the composition is
relatively new for L-PBF, a ball milled powder mixture was used in the study. It was found that the L-PBF Ti-185 resulted in good
mechanical properties (1890 MPa compressive strength and 11% elongation) due to the absence of β-flecks, fine-grained structure,
high dislocation density, distribution of nano-sized α-phase within β-matrix and high oxygen content. Heat treatment on the samples
was also conducted and it leads to homogenous spreading of α-phase that further improves the mechanical properties (2380 MPa
compressive strength and 26% elongation).
Although scarce, research on in-situ alloying of nickel-titanium has also been explored as a cost-effective alternative to pre-alloyed
nickel-titanium powder. Work has been done by Wang et. al to study the feasibility of making nickel-titanium parts via in-situ alloying
with different AM techniques [123]. The rapid solidification of melt pools inherent to L-PBF promotes the formation of TiNi, which is
responsible for the shape memory effect in nickel-titanium shape memory alloys. However, this also led to the formation of unwanted
brittle phases as well as metastable phases. On the other hand, a study by Zhang et al. shows that in-situ alloying of nickel-titanium is
feasible when the scanning speed is slow enough for densification (more than 99.5% relative density) and homogenization to form TiNi
phase [124]. With increasing scanning speed, an unwanted Ti2Ni phase can form. No further study on recovery strain has been made,
prompting the need for more research on the topic.
Furthermore, compositionally modified nickel-titanium alloy was done through the addition of niobium, mainly as an attempt to
improve the alloy’s toughness and biocompatibility [125]. Alloy with the composition of Ni44.0Ti12.2Nb (at%) was achieved via the
mixture of prealloyed nickel-titanium powder and niobium powder. Samples were then built using L-PBF and annealed at a tem­
perature of 850 ◦ C for 0.5 h. The retained niobium particles caused diffusion that accelerates the formation of eutectic phase and β-Nb

Fig. 16. The microstructure of the Ti22Al25Nb specimens after undergoing different heat treatment conditions: (a) 1250 ◦ C, 2.5 h; (b) 1250 ◦ C, 4 h;
(c) 1350 ◦ C, 2.5 h; (d) 1350 ◦ C, 3.5 h [127].

15
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

phase. The ultra-fine eutectic phase and nanoscaled β-Nb phase formation eventually lead to excellent compression properties of 1640
MPa (yield strength) and 39% (compressive strain), a significant improvement as compared to the cast counterpart (yield strength of
960 MPa). A recovered strain of 8.2% was recorded from compressive unloading, up from a compressive strain of 21%. The elastic
strain recovery is 5.6% and the hyperelastic strain recovery is 2.6%. Preliminary studies by Grigoriev et. al. [126] and Polozov et. al.
[127] manufactured Ti2AlNb-based (O phase) intermetallic part via in-situ alloying with L-PBF. To achieve optimization of such an
intermetallic part, in-situ alloying was first done with Ti5Al (wt%), followed by Ti6Al7Nb (wt%) and finally Ti22Al25Nb (at%). Two
types of laser were utilized for these studies: the Gaussian profile laser (400 W, 80 μm spot diameter) and the top-hat profile laser
(1000 W, 700 μm spot diameter). Grigoriev et. al found that at constant volumetric ε, high powered top-hat laser (950 W) can achieve
in-situ alloyed Ti5Al with better homogeneity but it is also more prone to cracking. However, Polozov et. al. found that the usage of
Gaussian profile laser at 275 W with constant ε can achieve alloy composition nearer to the designed composition due to milder
aluminium evaporation. They then studied the in-situ alloying of Ti6Al7Nb and found the presence of partially molten niobium which
can be reduced with heat treatment at 1350 ◦ C (or more than 1200 ◦ C [128]). The resultant Ti6Al7Nb exhibited an ultimate tensile
strength of 850 MPa which is similar to that made by metal injection moulding with sintering. Unfortunately, the L-PBF Ti6Al7Nb
exhibits poor elongation of only 2% which might be caused by the large pores formed. A large amount of partially melted niobium is
found in in-situ alloyed Ti22Al25Nb by Grigoriev et al. Further heat treatments were done to manipulate the volume fraction of O phase
where the hardness value increases with O phase fraction from 338.6 HV (as built) to 353.3 HV (1250 ◦ C homogenized) towards 358.
3HV (1250 ◦ C homogenized, 950 ◦ C aged). Further study by Polozov et. al. shows that the heat treatment time can influence the extent
of niobium dissolution and O phase volume fraction (Fig. 16). Nevertheless, the Ti22Al25Nb built here is not usable for engineering
applications due to the significant number of cracks present in the bulk samples.
To reduce the anisotropy of L-PBF parts, Barriobero-Vila et. al. approached the problem through the addition of a rare earth
element, lanthanum [129]. In-situ alloying of Ti2La (wt%) exhibits less pronounced texture with a combination of equiaxed α grains
and elongated tortuous α grains. Instead of the typical columnar β to α transformation, the reduced texture strength of Ti2La due to the
formation of α nucleation at the solidification front. Further heat treatment of T2La leads to new α phase formation and extensive
globularization, along with microstructural refinement that can be controlled by adjusting the cooling rate as seen in Fig. 17. The main
microstructure forming mechanism of Ti2La begins with the poor solubility of lanthanum in titanium α/β, which leads to lanthanum
rejection to L1/β interface that causes constitutional supercooling where α nucleation happens via a peritectic path. A similar effect of
lanthanum addition on the microstructure was demonstrated with in-situ alloying of Ti1.4Fe1La (wt%). This shows the potential in
avoiding epitaxial growth during L-PBF.

Fig. 17. Electron backscatter diffraction (EBSD) mapping showing the effect of heat treatment on L-PBF as-built parts from 950 ◦ C (L1 + β field)
down to room temperature, forming finer α grains with increasing cooling rate: (a) cooling with a 5 ◦ C min− 1, (b) cooling with a 100 ◦ C min− 1. The
scale bars indicate a length of 50 μm. Black lines in EBSD mapping indicate high angle grain boundaries (>10◦ ) [129].

16
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

The compilation of other titanium alloys not classified as Ti6Al4V based or titanium with β stabilizer manufactured via in-situ
alloying can be seen in Supplementary Materials Table S4.

3.3. Other material systems

While aluminium and titanium-based alloys make up most of the research done for in-situ alloying for L-PBF, there are some works
done on noteworthy materials using this technique as well. The research trend mirrors those using pre-alloyed powders, which aims to
expand the materials library for L-PBF.
FeNiSi soft magnetic alloy was manufactured using in-situ alloying by L-PBF. Coated powder (Fig. 18) is used instead of powder
mixture to combine the advantages of chemical homogeneity and lower cost [130]. With low scanning speed, racks appeared in the
fabricated samples which led to poor ductility. On the other hand, large irregular pores appeared with higher scanning speed. It is also
found that the soft magnetic properties of the samples became worse with increasing laser scanning speed due to the macro-structural
(porosity) and crystallography (grain size) effect.
In-situ alloying by L-PBF is used to study the effect of different contents of aluminium in Mg3Zn (ZK30) alloy [131]. While
aluminium has low solubility in magnesium, it is shown that L-PBF can improve the solid solubility due to the rapid solidification and
cooling rate. During solidification, the solute produces constitutional supercooling at the liquid/solid interface due to the slow
diffusion which results in grain refinement. In a similar study done using dysprosium instead of aluminium, it was also found that grain
size decreases significantly with increased dysprosium contents [132]. Grain refinement is influenced by grain nucleation and grain
growth. Due to dysprosium additions, the MgZnDy phase has a higher melting point compared to MgZn2, which is also formed during
L-PBF. This leads to lower temperature differences between the secondary MgZnDy phase which precipitated first and the α-mag­
nesium matrix. In another study done using the MgAl alloys system, it is found that equiaxed grain is transformed from dendrites under
a high temperature gradient [133]. High scanning speed leads to a high liquid cooling rate that results in refinement of the solid phase.
Titanium has also been added to magnesium alloy AZ61 to improve corrosion resistance by promoting the formation of aluminium
enriched α eutectic phase and suppressing the formation of Mg7Al12 phase [134].
Even with near full density, L-PBF produced zinc has relatively poor strength and needs to be improved. With a minor addition of
alloying elements, the mechanical properties of Zn can potentially be improved. In-situ alloying of zinc with magnesium was done by
Yang et. al. [135]. As compared to pure zinc, the Zn3Mg (wt%) has significantly enhanced mechanical properties due to grain
refinement and precipitation strengthening. The elastic modulus of the alloys also gradually increases with magnesium addition,
whereby a value of 48.2 ± 4.2 GPa can be obtained for Zn3Mg. Zinc alloys with 1–4 wt% addition of magnesium also have a suitable
degradation rate and good cytocompatibility. The degradation rates of these alloys are between 0.10 ± 0.04 mm year− 1 and 0.18 ±
0.03 mm year− 1, whereby the recommended value is 0.5 mm year− 1 or lesser [136]. Better corrosion resistance can also be obtained
with magnesium addition via grain refinement and the release of Mg2+ that would result in the formation of inert magnesium hydroxyl
carbonate. The improvement of mechanical properties via magnesium addition was clearly observed. The mechanical properties of
Zn3Mg are better as compared to pure zinc, even when the pure zinc has near 100% relative density (Zn3Mg only has 98.2% relative
density).
Another choice of alloying element for mechanical properties enhancement is silver [137]. Like the addition of magnesium, zinc
can be strengthened with silver due to the formation of precipitate and grain refinement via the effect of constitutional supercooling.
Depending upon the amount of silver, the corrosion rate of the Zn-xAg alloy can be decreased or increased. Minor silver addition leads
to grain refinement which increases corrosion resistance. However, too much silver can lead to excessive AgZn3 precipitation and leads
to galvanic corrosion.
A summary of in-situ alloying of other material systems is tabulated in Supplementary Materials Table S5.

4. Multi-metal processing using laser powder bed fusion

In L-PBF, the powder is deposited layer by layer onto a powder bed and a laser beam is utilized to selectively melt regions within

Fig. 18. Surface morphologies, (b) and (c), cross-sections of nickel-coated high silicon steel powder [130].

17
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

each layer to form 3D parts. Conventional L-PBF does not have a similar level of flexibility as directed energy deposition (DED) for
multi-metal processing due to the lack of proper deposition mechanisms. Hence, there has not been much research on multi-metal
printing using L-PBF. A brief overview of other multi-metal processing methods is given in Supplementary Materials Section 2.
To date, fabrication of different multi-metal parts using L-PBF have been explored such as Inconel 625/steel [138], tantalum-
tungsten/steel [138], tool steel/copper-chromium-zirconium [139], tool steel H13/copper [140], aluminium alloy/copper alloy
[14], 316L stainless steel/copper alloy [141], TiB2/Ti6Al4V [142], 316L stainless steel/MS1 maraging steel [143], Fe/Al12Si [144].
Besides L-PBF, electron beam powder bed fusion (EB-PBF) has also been used for fabricating Inconel 718/316L stainless steel [145] and
Ti6Al4V/copper [146] multi-metal parts. The presence of welding features of these parts is substantially smaller than those fabricated
by conventional welding methods. For ease of discussion, the sections are divided into the material used as structural material i.e.

Fig. 19. FE-SEM images showing the interface microstructure of L-PBF steel/bronze: (a) entire fusion zone (100×), (b) area A of entire fusion zone
(500×), (c) area B of entire fusion zone (600×), (d) area C of entire fusion zone (3000×), (e) area D of (c) (1500×), (f) schematic diagram of
dendritic cracks [148].

18
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

material with the higher strength, with the other materials in the multi-metal systems considered as functional material, i.e. material
used to enhance the functionality of the parts and is added onto the structural material.

4.1. Steels

In a study conducted using 316L stainless steel/C18400 copper alloy, a substantial amount of iron and copper diffusion was
observed at the interface, which suggests good metallurgical bonding [14]. The diffusion of elements was assisted by the convective
forces within the melt pools. Due to the rapid cooling in L-PBF, the melt pools underwent supercooling, which means that the heat
removal rate exceeded the heat of fusion releasing rate. However, no further study was done on L-PBF process parameters.
Similar results were obtained in studies using 316L stainless steel/Cu10Sn [147,148]. Respective optimised parameters for 316L
stainless steel and Cu10Sn were used and microscopic cracks were found near the boundary of the interface and 316L stainless steel but
not at the boundary between the interface and Cu10Sn. This is attributed to the difference in physical properties between steel and
bronze, such as the coefficient of thermal expansion which caused the tearing of the steel. Furthermore, dendritic crack that is sub­
stantially perpendicular to the boundary of the interface and 316L stainless steel also observed due to the higher thermal conductivity
of Cu10Sn which concentrated a large amount of heat in the interface which led to gradual increase in thermal stress, causing the
cracks to propagate [148]. Field emission scanny electron microscopy (FE-SEM) images of the interface are shown in Fig. 19.
Form the XRD results, it could be concluded that no intermetallic compounds are formed at the interface (Fig. 20 (a)). The
microhardness decreases from the 316L stainless steel region through the interface to the Cu10Sn region (Fig. 20 (b)). Similar ob­
servations were made using the 316L stainless steel/C18400 material combination [14].
A similar study was conducted on the fabrication of 316L stainless steel/Cu10Sn multi-metal components using L-PBF. The opti­
mised parameters from other studies were used for 316L stainless steel while they optimise the parameters for Cu10Sn fabrication on
316L stainless steel [149]. Multi-metal samples formed using the optimised parameters are shown in Fig. 21. Using optimised pa­
rameters for interface formation can results in good interfacial bonding with minimal cracks formation. It is also concluded that the
multi-metal specimens could withstand a considerable degree of torsional stress in bending and shear directions.
Morphology of CuSn/18Ni300 multi-metal system was also studied (Fig. 22). Two kind of melt pools were observed in the 18Ni300
region which include columnar and cellular cells. These cells grew from the melt track boundaries and stopped growing when they
reached the interface. At the interface, some small pores can be observed.
Using EDS, the mixing of α-Fe and α-Cu phase is observed at the interface, with some iron, copper, nickel, chromium, and titanium
found in the CuSn region, indicating that diffusion has occurred during L-PBF (Fig. 23).
Bai et al. investigated the interface of the 316L stainless steel/C52400 copper system at both sides, i.e. at both the regions nearer to
316L stainless steel and C52400 [151]. Cracks are observed at the interface near the 316L stainless steel side due to stress mismatch,
regardless of which material is deposited first (Fig. 24).
It is observed that C52400 is brought to the 316L side and vice versa at each of the interfaces which results in the mixing of both
materials. However, the same material tends to segregate due to the difference in their surface tension and the rapid cooling (Fig. 25).
Similarly, both interfaces are studied for the Inconel 718/316L stainless steel system using the sandwich design where Inconel 718
is built between 316L stainless steel [152]. It is shown that at both interfaces, uneven morphology is observed which can increase the
total interface area and joint strength. Likewise, cracks are observed at the 316L stainless steel sides close to the interfaces.
Shakerin et al. studied the deposition of maraging steel (MS1) onto H13 tool steel using L-PBF to form MS1-H13 bimetals. It was
concluded that a very narrow interface was formed between the two types of steel without any cracks or discontinuities [153]. The
EBSD analysis of the interface revealed a continuous joining between the two materials. The maraging steel exhibits a very fine
microstructure in the first layer solidified on top of the H13 due to the rapid heat transfer form the melt pools to the substrate. After a
few layers, the microstructure become slightly coarser due to the higher temperature of the previously deposited layers (Fig. 26 (a)).
Further analysis using SEM clearly shows an interface region between the two materials and the precipitates in the H13 appeared to be

Fig. 20. (a) XRD pattern of the 316L stainless steel, Cu10Sn and interface (b) Vickers microhardness along the interface [148].

19
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 21. 316L stainless steel/Cu10Sn multi-metal parts formed by L-PBF [149].

Fig. 22. Morphology of CuSn/18Ni300 multi-metal system (a) CuSn/18Ni300 interface (b) high magnification of area A (c) high magnification of
area B (d) high magnification of area C [150].

coarsened due to partial melting and subsequent mixing with molten powder, hence forming a partially melted zone (Fig. 26 (c)) which
also happens in fusion welding. Beyond the interface, the MS1 microstructure consists of similar morphology with melt pools ac­
commodating coarse and fine equiaxed cells as well as columnar cells of submicron sizes.
L-PBF techniques are favourable for the printing of intricate overhanging features such as those found in lattice structures
[2,154,155]. This is because the overhanging regions can be supported by the unconsolidated powders bed from previous layers.
Besides, L-PBF has better in-plane print resolution due to its smaller laser spot diameter of ~0.08 to 0.1 mm. Fabrication of lattice
structures using multi-metal by L-PBF has been shown using 316L stainless/Cu10Sn (Fig. 27) [149].

20
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 23. EDS results of CuSn/18Ni300 interface [150].

Fig. 24. Cracks distribution at (a) the C52400/316L interface and (b) 316L/C52400 interface [151].

4.2. Aluminium and its alloys

Demir & Previtali studied the multi-metal processing feasibility using pure iron and Al12Si. In the samples, the main defect within
the deposited material appears as large cracks rather than porosity due to the lack of fusion or excessive melting. The proportion of
cracks is influenced only by laser power and is attributed to the formation of FeAl intermetallic [144]. A study was conducted using
AlSi10Mg/C18400 copper alloy system to investigate the interface characteristics [141]. Focused ion beam (FIB) imaging was used to
observe the interface which identified distinct copper-rich and aluminium-rich regions with an intermixed region at the interface due
to the movement of elements assisted by diffusion (Fig. 28). This is indicative of the dilution effect that are also observed in other multi-
metal processing studies.
Like the 316L stainless steel/Cu10Sn system, cracks were also found in segments of the interface for the AlSi10Mg/C18400 system
which is also attributed to the difference in thermal coefficients. However, using XRD, it is concluded that Al2Cu intermetallic is formed
which aggravated the cracks formation. The intermetallic compounds formed in the interface also led to anomalous readings for
microhardness due to the precipitation of such compounds that are harder but more brittle (Fig. 29).
Böhm et al. derived a theoretical model for the composition profile for multi-metal processing by L-PBF using AlSi10Mg and pure
aluminium [156]. The model takes into account the dilution ratio as a function of the penetration depth of the melt pool, which is in
turn dependent on the process parameters. A schematic of the resulting staircase composition profile across the interface due to the
remelting of prior layers is shown in Fig. 30.
Multi-metal parts are fabricated by L-PBF of AlSi10Mg on AlCuNiFeMg substrate. SEM-EDS elemental analysis was conducted along

21
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 25. (a) Melt behavior at the 316L/C52400 interface (b) at the C52400/316L interface [151].

Fig. 26. (a) EBSD unique color grain map and (b) scanning electron micrographs of the MS1-H13 interface with (c), (d) and (e) as magnified images
of the different zones [153].

Fig. 27. 316 L/Cu10Sn lattice structure specimens with different unit sizes [149].

22
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 28. FIB image of Cu/Al interface; (b) SEM image of Cu/Al interface; (c) FIB image of copper; (d) SEM image of AlSi10Mg [141].

Fig. 29. Microhardness variation along the Al/Cu interface [141].

the interface and it showed that successful partial melting of the substrate and the mixing of the AlSi10Mg with the AlCuNiFeMg
occurred, resulting in sound metallurgical bonding at the interface [157].

4.3. Titanium and its alloys

In order to understand the effect of processing parameters for multi-metal processing, Wang et al. studied the formation mecha­
nisms of TiB2 tracks on Ti6Al4V by L-PBF [158]. The melt pool geometries and surface morphology of the melt tracks were analyzed. It
is found that the shape of the melt pools depends on the laser-powder interaction. When energy transfer efficiency is larger than the
material absorptivity, keyhole is formed, leading to the formation of deep and narrow melt pools.
Wei et al. studied the formation of multi-metal system using Ti5Al2.5Sn/Ti6Al4V. Due to the optimized parameters used, good
metallurgical bonding was observed at the interface without any pores, cracks and lack of fusions [159]. At the interface, diffusion of
elements is observed as indicated by the gradual decrease in tin from Ti5Al2.5Sn to Ti6Al4V region, and vice versa for the vanadium
(Fig. 31).

23
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 30. Staircase composition profile is depicted in the layers within the interface [156].

Fig. 31. Microstructure characterizations of the as-built Ti5Al2.5Sn/Ti6Al4V multi-metal sample (a) OM image (b) EDS line scanning result
showing the element distribution around the Ti5Al2.5Sn/Ti6Al4V interface (c) Schematic diagram showing the formation mechanism of the element
diffusion (d) SEM image of the Ti5Al2.5Sn layers (e) SEM image of the Ti6Al4V layers and (f) SEM image showing the microstructure around the
Ti5Al2.5Sn/Ti6Al4V interface [159].

Due to the presence of the interface, there is no abrupt change in the microhardness between the two titanium alloys (Fig. 32),
which agrees with the results observed in other material systems.
The multi-functional concept for L-PBF produced parts is shown possible by fabrication of cellular structures using NiTi/Ti6Al4V
multi-metal system [160]. This allowed the allocation of specific functions that are localized and the design concept for a multi-metal
hip implant is shown in Fig. 33. SEM images show the successful transitions from NiTi to Ti6Al4V with nano-indentation results
showing an interface region between the two materials.

5. Challenges in in-situ alloying and multi-metal processing by laser powder bed fusion

Functional applications of L-PBF parts largely depends on whether they are defect-free. However, many metallurgical defects in L-
PBF, such as balling, porosity, cracking, and oxide inclusions remain as challenges. Furthermore, with in-situ alloying and multi-metal
processing, there are additional complications such as loss of alloying elements, intermetallic phases, and unmelted particles. Though
efforts have been made to develop the solutions, many unknowns are still involved. In the remaining parts of this section, each of these
defects is elucidated by highlighting the mechanism of their formation, main influencing factors, and the potential remedial measures.

24
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 32. Microhardness distributions around the Ti5Al2.5Sn/Ti6Al4V interface of the dissimilar titanium alloy samples under different processing
states: (a) as-built (b) after heat treatment [159].

Fig. 33. Design concept of a NiTi/Ti6Al4V multi-material hip implant [160].

5.1. Balling

Balling is a phenomenon which decreases the surface quality of L-PBF parts. It is a typical L-PBF defect and can be divided into two
types: (1) ellipsoidal balls of about 500 μm that are formed as a result of broken melt tracks. These are resulted from the insufficient
wetting ability to the substrate previous layer, as discussed previously. Both severe oxidations of the melt pool or low energy input
could also result in worsened wetting characteristics [60]. Hence, the atmosphere during L-PBF is typically controlled by flooding of
inert gases, such as argon or nitrogen. Stacking of these large balls results in scratching of the recoating blade or roller and hinders the
movement of the recoater, thus, stop the production process. (2) The spherical balls of about 10 μm are formed due to a reduction in
surface energy of liquid at short length scales by splashing of melt pools [161]. Both types of balls impede the uniform deposition of
new powder on the previous layer and tend to cause defects, such as porosity and crack, which would be addressed in later sections.
Balling effect is mainly affected by the wetting characteristics of the melt pools, which are in turn affected by the processing
parameters. Hence, the selection of these parameters must take into consideration the material physical properties. Balling effect also
depends on the powder preparation methods which determine the powder feedstock quality. The feedstock affects the thermal

25
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 34. The effect of melt pool depth-to-width ratio on pore formation [164].

26
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

absorptivity and transfer behavior which can cause significant evaporative recoil pressure in the melt pools. This causes melt insta­
bility, which results in balling [162].
Balling can deteriorate the surface quality of L-PBF components, hence, in addition to parameters optimization, re-melting, and
post-process treatments, including polishing, may be needed.

5.2. Porosity

Porosity is a common feature in many metallurgical processes, which include L-PBF. The pore-forming mechanisms for L-PBF
include lack of fusion porosity caused by insufficient melting and gas entrapment. Lack of fusion is due to insufficient overlapping of
melt tracks or penetration of the laser into the powder layers due to low laser power, high scanning speed, large hatch spacing, and
large layer thickness. The sources of entrapped gases can be the gas used to flood the chamber (shielding gas), vapor due to vapor­
ization of the powder bed, and gases trapped within the powder particles during the powder production [163]. Gas entrapment can
also be due to turbulent flow in the melt pools.
Porosity is also commonly encountered in in-situ alloying with L-PBF and is shown in a study by Fischer et. al [105]. Ti40.5Nb (wt%)
was studied and a laser with discontinuous scan mode was used. As the volumetric ε increases, the cause of porosity transits from lack
of fusion to having a minimum amount of porosity and finally to keyhole formation. The increment in ε changes the melt pool for­
mation mechanism from conduction mode to keyhole melting, subsequently causing an increased aspect ratio of the melt pool and lead
to a higher probability of keyhole porosity formation. An illustration of the effect of melt pool aspect ratio on keyhole porosity is shown
in Fig. 34.
It is found that for mixed powders, each component powders must be melted to enable chemical mixing in a liquid state as
compared to using pre-alloyed powders [64]. For the case of aluminium–silicon alloys, silicon has a higher melting point which makes
it more difficult to create sufficient melt pools when the silicon fraction increases. Higher laser power is necessary to melt the silicon
powder which can cause melt pool instability and results in the keyhole induced porosity (Fig. 35). Melt pool instability is due to fluid
forces that leads to effect such as Reyleigh instability, causing discontinued melt tracks.
The two main pore forming mechanisms can be identified by the location of the pores, with the former type outside the melt pool
while the latter inside the melt pool. Bubbles or pores are formed in a way to balance the vapour pressure with surface forces, resulting
in surface movements [165]. Both types can be resulted from the change in the melt pool flow pattern which is controlled by forces
such as gravitational force, vapor pressure and thermal capillary forces exerted on the metallic/gaseous interface. The aforementioned
balling effect deteriorates the recoating of the new layer and may result in unmelted powder particles entrapped in defective positions
such as the neighbouring melt tracks.
Cave-like pores and open pores formed on the top surface of the track is illustrated in Fig. 36. The cave-like pores appear to be
formed by the unstable melt flow moving away from laser scanning direction where the melt was solidified in the air before fusion with
the previous layer thus creating a hole beneath. Open pores or discontinuities on the L-PBF produced surfaces could be due to the
splashing of the melt pools or lack of materials filling in at some localised sites. The cave-like pores mainly remain as pores because of
the coverage, while the open pores could be either filled with melt materials during the next scanning or remain as pores. Simulations
done by Qiu et al. [165] and Khairallah et al. [166] proved the existence of these pores due to the unstable melt.
Materials vaporization during L-PBF results in keyhole pores formation. These pores are typically reported to be spherical and
located at the tail end of the melt pools if they failed to escape [167]. Fig. 37 illustrates the entrapping process of keyhole pores and
their positions for laser welding. Large bubbles were intermittently formed at the bottom of the keyhole as the shape and depth of the
keyhole fluctuated violently. They end up floating and remain as pores in the rear of the melt pools if they fail to escape before the melt
pool solidifies.
Surface moisture on the powder particles can release hydrogen during L-PBF. Hydrogen gas is mainly released near the

Fig. 35. Micrograph of AlSi25 sample produced at 400 W and 2000 mm/s, some melt pools are marked by dotted lines, and keyhole porosity is
indicated by red arrows, BD is building direction [64].

27
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 36. SEM micrographs showing the variation of the top surface structure of L-PBF Ti6Al4V samples with powder layer thickness 60 μm [47].

Fig. 37. Schematic representation of melt flows in the melt pool and weld bead geometry, showing keyhole pore formation and distribution [168].

Fig. 38. Schematic overview of the interaction zone between laser radiation and powder [169].

28
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

solidification front of the melt pool as the dissolution of hydrogen decreases with decreasing temperature (Fig. 38). For a high so­
lidification rate, i.e. high scanning speed, the solubility is increased [169]. Hydrogen porosity can be reduced when internal drying
process is used [65,169]. Internal drying process can be achieved by re-melting with a lower laser power used during the first scanning.
In such a case, a higher drying temperature is attained which lowers the moisture level, thus achieving internal drying process and
reducing the pore density. Gases entrapped inside the powder particles during the powder atomization process can lead to pores as the
gas is easily entrapped in the melts and can be eliminated by refining the powders [163].
The above porosity formation discussions are all based on the stability of the melt pools and/or the fluid flow forces. In addition, Gu
et al. also used different shielding gases (argon, helium, and nitrogen) to demonstrate the importance of melt flow on gas formation
[170]. The unstable fluid flow in the melt pool can be identified by the range of the parameters. Kasperovich et al. showed that
scanning speed has the most dominant influence on the porosity fractions (Fig. 39) [167]. The laser power has a significant effect on
the resulting defect densities and also shows a distribution with an optimal parameter range allowing minimal defects (Fig. 39c). High
scanning speed and low laser power are associated with the lack of fusion pores while keyhole pores are attributed to low scanning
speed and high laser power. While the hatch distance was found to be least sensitive (Fig. 39b), it should be noted that only the porosity
in the melt pools is analysed.

5.3. Crack

The crack types in L-PBF can be categorized into hot crack and cold crack. Hot crack, also known as solidification crack, can be
formed at the late stage of solidification when grains form continuous skeletons. Two phenomena account for hot crack initiation, (1)
temperature differences during solidification with large temperature range resulting in insufficient convection in the liquid region and
(2) solid deformation. Crack initiation and growth are prevented and the effect of entrapped liquid is reduced by fine equiaxed semi-
solid structures which allow easier grain rotation and deformation than dendrites.
Cracks caused by residual stresses are more common for L-PBF. Residual stresses in L-PBF are caused by two possible mechanisms,

Fig. 39. Influence of the L-PBF processing parameters on the total porosity (2D measurements) as a function of (a) scan speed v, (b) hatch spacing h,
and c) laser power P. The volume energy density Ev, is included in the upper abscissae [167].

29
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 40. Temperature gradient mechanism in L-PBF [174].

temperature gradient mechanism (TGM) and cool-down phase of the molten top layers. The first mechanism is illustrated in Fig. 40. A
steep temperature gradient is developed just beneath the laser spot when a laser beam is stroked onto the surface layer due to rapid
heating and slow heat conduction. Compressive stresses in the upper layers are formed as the expansion of the upper layer is restricted
by the cooler underlying solidified layer. The compressive stresses in the material cause plastic deformation when the yield strength is
reached, which remain in the absence of mechanical constrain. During cooling, this compressive state is converted into residual tensile
stresses and a bending angle towards the layer beam develops. Those residual stresses may induce cracking [171,172]. The melted top
layers tend to shrink due to the thermal contraction. This deformation is again prohibited by the underlying layers, thus introducing

Fig. 41. 316L stainless steel/Cu10Sn specimens by L-PBF. (a) parameters used in L-PBF (b) Group I with slightly cracked interface (c) Group II with
severe cracks at the interface and over melting at top surfaces (d) Group III with well-fused surfaces with varying degree of cracks [149].

30
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

tensile stresses in the top layer and compressive stresses below [173,174].
Cracking and interlayer de-bonding can occur as residual stresses become larger in the direction perpendicular to the scanning
direction than along scanning direction [69]. The level of residual stress depends on material properties especially the elastic modulus
and CTE. Thermal stresses in multiphase materials were discussed because of internal constraints caused by different CTE between the
phases. If the thermal expansion is hindered by external constraints or uneven temperature distributions, bending stresses (σb ) pro­
portional to the average CTE are generated [175]:
∫
6 L/2
σb = − 2 Eαt ΔTX ’ dx (5)
L − L/2

where L is the layer thickness, E is the elastic modulus, αt is the average CTE of the composite, ΔT is the temperature difference and X’ is
equal to x − L/2, x is the distance measured from the layer surface. The average CTE of a multi-phase material (αt ) is derived as follows
based on the Turner Equation:
α1 V1 K1 + α2 V2 K2 + α3 V3 K3 + ⋯
αt = (6)
V1 K1 + V2 K2 + V3 K3 + ⋯

where V is the volume fraction, K is the bulk modulus and αi is the CTE of each phase. According to Equation (6), lower elastic modulus
and average CTE which depends on the particular CTE and volume fraction of the material constituents, favor less stress. Taking into
consideration of the compatibility between the dissimilar metals for multi-metal processing, close CTE between them is preferred. This
is also true for in-situ alloying by L-PBF as the constituent element particles that are unmelted can have restricted expansion and/or
contraction, causing residual stresses. Crack formation was found to be reduced with increasing laser power and/or decreasing
scanning speed, which correlates with increasing energy density. However, increasing energy input leads to the formation of keyhole
porosity. Thus, there must be a balance between crack formation and porosity [64].
Crack formation is also attributed to high thermal stresses during L-PBF as a result of steep thermal gradients, fast cooling rates as
well as CTE mismatch between the substrate and in-situ alloying materials [64]. Mercelis et al. established a relationship between
residual stresses and the height of the sample and substrate, scanning strategy, and heating conditions in 316L stainless steel [176].
Although the quantitative results cannot be simply extrapolated to other materials, most qualitative results can be generalized. For
instance, the length and direction of scanning vectors can redistribute the stress profiles. Residual stresses can also be controlled by the
preheating of the build substrate due to the decrease in the temperature gradient. This can be useful during multi-metal and in-situ
alloying. Cracks at the interface and surface warpage are observed in 316L stainless/Cu10Sn samples, as shown in Fig. 41. The crack
initiations at the interface are said to be caused by rapid solidification of the liquid phase copper which sharply tore the solid phase
steel due to a large difference in CTE [149]. Liu et al. also attributed the cracks formation due to the infiltration of copper into the
austenitic grain boundaries of steel [14].
Densification is facilitated by high temperature as the effect of low viscosity on wettability is compromised. However, cracks caused
by residual stress is thus induced due to the high energy input and temperature gradient. Zhu et al. demonstrated that the thermal
shrinkage during laser melting was increased with increased laser energy input under a cool-down phase mechanism [177].
L-PBF processing parameters should be manipulated with care to reduce the residual stresses. It is noted that the powder prepa­
ration also has an effect on residual stresses. Powder distribution will affect the thermal behavior, thus influencing the stability of the
melt pools. However, more detailed research is needed as for now only AlMangour et al. found that ball-milled powders induce larger
and more severe cracks than directly mixed powders, although the latter revealed deep groove and pores [178].
Cracks and porosity can be formed due to the presence of uncompressed powders. It can be minimized by increasing the packing
density of the powder with a roller compressor during powder spreading [179]. For in-plane multi-metal parts, the formation of cracks
could also be due to the solidification-order of the dissimilar materials. Anstaett et al. investigated the in-plane joining of tool steel
1.2709 (X3NiCoMoTi18-9-5) and 2.1293 CuCr1Zr (CCZ) and found that cracks form only when CCZ is scanned first but no crack is

Fig. 42. Equilibrium vapor pressure of metals at different temperatures [181].

31
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

observed when the tool steel is scanned first [139]. The crack formation within the CCZ section could be attributed to the inclusion of
iron particles from the tool steel. Cracks can form when the multi-metal parts are not processed at an appropriate cooling rate. Beal
et al. have shown the copper favours the formation of cracks within the linearly mixed compositional gradient of H13 tool steel and
copper as it increases the cooling rate of the multi-metal parts [140]. Chen et al. showed that microcracks can form at high laser energy
input due to the huge thermal gradient at the interface; whereas pores can form at low laser energy input due to the inability of the high
viscosity metal liquid to flow into the voids between the particles at a lower operating temperature [142]. The formation of large cracks
between the two dissimilar materials can also be due to the low compatibility and miscibility of the two materials [144].

5.4. Loss of alloying element

Significant vaporization for volatile elements with higher equilibrium vapor pressure such as magnesium, zinc, lithium, and
aluminium can occur during L-PBF due to the high temperature of melt pools. Fig. 42 plots the equilibrium vapor pressure of some
metals at different temperatures. Selective vaporization of elements will result in a change in the composition of the components and
degrade the mechanical properties, such as reducing strength, elongation, hardness [180], corrosion resistance, and increasing crack
susceptibility [181], which would affect the 3D printability. According to Collur et al., the vaporization mechanism can be divided into
three stages, transport of vaporization elements from the bulk to the surface of the melt pool, followed by the vaporization at the
liquid/vapor interface and transport of the vaporized species into the surrounding gaseous phases [182].
Like welding, the change in composition during L-PBF depends on the vaporization rate and the volume of the melt pools. Although
the rate of vaporization increases with laser power, the change in composition is most pronounced at low laser powers because of the
small size and, consequently, the high surface-to-volume ratio of the melt pools. The high scanning speed ensures a shorter time span of
the melt pools and reduces vaporization [183].
In a study to in-situ alloy Al50Si (wt%), the resultant part shows higher silicon content due to aluminium evaporation [65]. It was
observed that the extent of aluminium evaporation increases with the input ε. The hyper-eutectic composition resulted in micro­
structure with primary silicon and eutectic silicon. Like in-situ alloyed Al12Si, increasing ε leads to sustained melt pools which causes
silicon precipitation out from the saturated Al(Si) solid solution. By lowering the cooling rate, the size of the primary silicon phase
increases from a mean diameter of 3.12 – 7.31 μm with laser power from 260 W to 350 W. During L-PBF for Al 7075, evaporation of
zinc and magnesium can occur and reduces the overall strength, however, the problem of volatile elements evaporation can be
compensated via adjustment for the raw material used [71]. If ε was to be increased further to promote melting and mixing, the recoil
pressure from vaporized metal can lead to the formation of keyhole porosities. During the in-situ alloying of nickel-titanium, an
exothermic reaction between titanium and nickel can lead to turbulent melt pools and it was hypothesized that selective ejection of
particles might have occurred during L-PBF, leading to depletion of titanium in the manufactured part [123]. In another study done
using a magnesium–aluminium alloys system, it is found that with high energy input, the powder was evaporated due to the low
boiling point of magnesium. These vaporized magnesium then oxidized to form MgO. Lack of fusion was obtained at low energy input
[133].

5.5. Oxide inclusions

The densification and performance of L-PBF parts can deteriorate due to oxide inclusions. During L-PBF, oxidation can occur on the
surface of the melt pools or as a result of entrapped gas due to the turbulent flow in the keyhole position through the reaction with the
oxygen impurity in the shielding gas. Simonelli et al. compared the composition of starting particles of 316L stainless steel, AlSi10Mg,
and Ti6Al4V directly with the laser spatters generated during L-PBF. Oxide films on the surfaces of 316L stainless steel and AlSi10Mg
spatters are observed, however, there was no oxidation on that of Ti6Al4V [42]. The oxides are underpinned by surface enrichment of

Fig. 43. Typical defects of in-situ alloyed TiNb binary alloy with increasing energy density a) porosity b) unmelted Nb [105].

32
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

the most volatile element present in the alloys because no oxides were observed in the bulk microstructure of the spattered material. If
these elements (manganese and silicon in 316L stainless steel, magnesium in AlSi10Mg) have high affinity with oxygen, oxides can be
formed. Magnesium oxide films in aluminium alloys contain spinel (MgO.Al2O3) and MgO depending on the magnesium contents.
However, the protective oxide layers are thinner in 316L stainless steel and titanium alloys than in aluminium alloys and have a
negligible effect on L-PBF, because they can be disrupted and stirred into the melt pool or vaporized by the laser beam [184]. Since
aluminium has a greater affinity to oxygen and can form thicker oxide films, with dissociation pressure of alumina being rather low
(10− 52 pO2) at the melting point of aluminium, oxide films cannot be avoided at low oxygen concentration. Therefore, most re­
searchers focus on aluminium components formed by L-PBF. There are two sides for oxide films, wetted side and dry side. The wetted
sides may nucleate intermetallic compounds while the unwetted dry sides are potential nucleation sites for porosity. Hence, most oxide
films are at the pore boundaries [57]. As aluminium dendrites may not grow through the dry side of oxide films, oxides may also be
distributed at the grain boundaries. Oxide films are always folded dry side to dry side as observed in convectional castings and laser-
welded aluminium alloys [185,186].
Measures must be taken to eliminate the oxide films because they not only degrade the densification behavior but can also initiate
cracks [187]. High energy density is required to break up the oxide films. However, only the oxide films at the bottom of the melt pools
can be interrupted due to stronger Marangoni flow. Those at the sides of the melt pools remain intact, creating the oxides “walls”.
Therefore, new methods of controlling the oxidation process and disrupting oxide films formed within the components, for example,
scanning strategy optimization, should be further developed. The higher energy density employed seems to be more effective in
breaking up the oxide films though it induces a higher affinity to oxygen. However, no comparable data has been offered for in-situ
alloying and multi-metal processing by L-PBF.

5.6. Interfacial phases and unmelted particles

Interfacial phases and unmelted particles are common challenges for in-situ alloying and multi-metal processing by L-PBF. During
the in-situ alloying of titanium-niobium, the amount of unmelted niobium decreases as the ε increases which leads to the formation of
parts with better compositional homogeneity [105]. The trend of porosity and unmelted niobium with increasing ε are illustrated in
Fig. 43. When a certain minimum ε threshold is crossed, ε increment generally leads to a better dissolution of niobium particles but
induces more keyhole porosity. It is of interest to note that the issue with in-situ alloying of titanium-niobium via L-PBF is that, the
processing parameters that produced minimum porosity do not necessarily lead to a homogenous part (Fig. 44). However, the for­
mation of Marangoni convection induces capillary forces which are said to aid the dispersion of magnesium–aluminium alloys,
improving homogeneity [133]. This leads to a seemingly limited process window in which the part’s quality can only be partially
optimized. It is also hard to decide on the optimal combination of keyhole porosity percentage and the amount of unmelted niobium in
the L-PBF parts. Such a dilemma between the need for unmelted particle reduction and keyhole porosity reduction remains a challenge,
leading to a porosity-segregation dilemma that needs to be addressed.
Similarly, the in-situ alloyed Ti7.5Mo (wt%) system studied by Kang et. al. shows a similar influence of ε on the number of unmelted
molybdenum particles and porosity [188]. The ε increment improved the part’s densification to a maximum of 99.7% while minimized
the amount of unmelted molybdenum particles. However, the porosity type shown in the optimized part contains keyhole porosity,
indicating that further reduction of unmelted molybdenum particles with increased ε will only lead to more keyhole porosity for­
mation. This observation reinforces the need to address the porosity-segregation dilemma. A study on in-situ alloyed Ti50Ta (wt%) by
Sing et. al. also points towards a similar trend of density and unmelted tantalum particles with increased ε [94]. The optimized parts in
the study contain unmelted particles which signify the possibility for improvement. It is difficult to eliminate the compositional
segregation caused by unmelted tantalum particles while obtaining parts with no porosity. The issue of unmelted particles and
compositional segregation is commonly seen in the in-situ alloying technique. This issue, however, is especially prominent when it
comes to the in-situ alloying of biocompatible β-titanium. The main reason behind such occurrence is that biocompatible β-titanium
stabilizers are mostly composed of refractory elements, in which the β-titanium stabilizers’ melting point is very different from that of
titanium. The properties of titanium and different β-titanium stabilizers are listed in Table 1.

Fig. 44. SEM of Ti40.5Nb at different energy density a) 39 J/mm3 b) 214 J/mm3 [105].

33
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Table 1
Melting point, boiling point, density, and the lower limit of Moeq β -Stabilization wt% associated with biocompatible elements commonly utilized in
the in-situ alloying process via L-PBF.
Element Melting Point (◦ C) Boiling Point (◦ C) Density (g/cm3) Moeq Based Critical β -Stabilization Lower Limit (wt%)

Ti 1668 3287 4.51 N/A

Mo 2623 4639 10.3 11.8
Nb 2468 4927 8.57 42.1
Ta 3017 5458 16.4 53.6

For reference, the molybdenum equivalent, Moeq (wt%) within the alloys can be calculated using [189]:
Moeq = 1.0Mo + 0.67V + 0.44W + 0.28Nb + 0.22Ta + 2.9Fe + 1.6Cr + 0.77Cu + 1.11Ni + 1.43Co + 1.54Mn + 0Sn + 0Zr − 1.0Al (7)

The critical lower limit of Moeq for β stabilization is 6.25 at % or 11.8 wt%.
The mechanical properties of the multi-metal interfaces mostly depend on three factors, namely (1) the type of constituent phase(s),
(2) the thickness of the detrimental phases, and (3) the morphology of the interfacial phases. Brittle intermetallic compounds normally
act as crack initiation sites and impair the ductility of materials [190].
The diffusion of elements within the interface can result in intermetallic formation, as shown in a study using AlSi10Mg/C18400
copper alloy [141]. Al2Cu intermetallic compounds are formed which are found to be detrimental to the mechanical properties of
multi-metal structures. Interfaces with thick detrimental intermetallic layers normally fail in a brittle manner. Generally, iron­
–titanium intermetallic compounds have been proven to be the most detrimental to bond properties regardless of the type of joining
techniques used. This is due to the formation of cracks at the interfaces as a result of the intermetallic compounds [191–193]. To
improve the mechanical properties of the iron–titanium interface, copper [194,195], nickel [196,197], and silver-based [198,199]
interlayers have been incorporated to suppress the formation of iron–titanium intermetallic compounds between the titanium and steel
alloys. However, copper-titanium and nickel-titanium intermetallic compounds are still found to be detrimental at the interface as they
normally result in brittle failure across the interface. In contrast, AgTi is the only intermetallic that did not have a detrimental effect on
the bond strength as the fracture of Ti/AgTi/Ag/SS interface happened within the ductile silver interlayer [198,199]. In addition, the
hardness of the silver-titanium intermetallic (164 HV) [198] is also much lower compared to that of the copper-titanium (530–670 HV)
[191], nickel-titanium (~662 HV) [197,200], and iron- intermetallic compounds (>1000 HV) [201]. Other frequently found detri­
mental phases includes σ-FeCrV phase (forms when stainless steel is excessively enriched with vanadium and/or chromium content)
[202,203], TiC (forms when titanium is bonded directly to steels with more than ~0.16w% carbon), and λ(Cr,Fe)2Ti [204,205].
For the joining of Inconel 718 and 316L stainless steel using electron beam powder bed fusion (EB-PBF), Hinojos et al. have found
that the amount of precipitate formed during the process is much lesser compared to conventional joining methods [145]. However, a
high amount of carbide and precipitate formation in regions like the fusion zone can reduce the corrosion resistance and cause
embrittlement. In a separate study, Demir et al. explained that the formation of cracks at the multi-metal interface between Fe/Al12Si
is due to the high fragility of the iron-aluminium intermetallic [144].
Apart from the inherent properties of the interfacial phases, the quantity of those phases also has a severe effect on the bond
properties. The structural properties of multi-metal parts weaken when an excessive amount of detrimental phase is formed due to the
disproportionate mixing of interlayers and base alloys. Therefore, the tensile strength of the diffusion bonded interfaces decreases with
an increase in intermetallic thickness [194,197]. In other words, the titanium/steel interface which contains iron–titanium inter­
metallic compounds can exhibit a high bond strength if the quantity of intermetallic compounds is greatly reduced. As such, the friction

Fig. 45. Overview of the scanning strategy used for the different samples. The building (BD), scanning (SD), and transverse direction (TD) are
indicated. Sample A is scanned with long unidirectional vectors; sample B with long bidirectional vectors; sample C is first scanned with long
bidirectional vectors in TD and secondly scanned with long bidirectional vectors in SD; sample D is scanned with island strategy with 90◦ rotation
but without shift and sample E is scanned with island strategy with 90◦ rotation and 1 mm shift between the layers [216].

34
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

welded, and explosion welded titanium/steel interfaces are typically stronger compared to other joining techniques and failure
typically happens at the base alloy instead of the interface [203,206–209]. The formation of intermetallic compounds can be controlled
by restricting the dilution between the materials and decreasing their interaction time. The interaction time can be reduced by
lowering laser power, increasing scanning speed, or hatch spacing which controls the melt pool dimensions during L-PBF.

6. Potential research for in-situ alloying and multi-metal processing by laser powder bed fusion

6.1. Novel fabrication strategies

Microstructure developed during L-PBF tends to be strongly textured due to the directional heat flux and large thermal gradients
[210]. While processing parameters that affect the energy input are proven to affect microstructure textures [211], manipulations by
scanning strategies are supposed to be easier. To achieve a full density with controllable microstructures, multiple scanning strategies
have been developed [212]. Three main basic strategies, namely unidirectional, bidirectional, chessboard, or island strategies, are well
developed for L-PBF (Fig. 45a, b, and d). There are some variants derived from these three basic strategies with rotation for a certain
angle (e.g. 60◦ [213], 67◦ [214], or 90◦ [215]) or shift for a certain distance between the layers [216]. However, the research done by
Robinson et al. concluded that hatch rotation has little effect on the density, residual stresses, top surface roughness, and strength of L-
PBF parts [174]. Unidirectional scanning is the simplest strategy and generally leads to the least densification and strongest texture
while other variants show superior densification behavior [217]. The improvement on densification of other strategies depends more
on materials and other processing parameters. For instance, an insignificant change was reported in work done by Thijs et al. [216]. In
their study, five strategies (unidirectional scanning, bidirectional scanning, re-scanning based on bidirectional scanning with 90◦
rotation of the scanning direction, island scanning, shifting the position of the scan vectors between the layers based on island) are used
in AlSi10Mg, with a relative density of 99.0%, 98.9%, 99.4%, 98.2% and 98.7% obtained, respectively. As full density has been
achieved in unidirectional scanning strategy, an insignificant improvement was witnessed. However, the effect on microstructure
texture is obvious. The texture index (TI), defined by
∫
TI = (f (g) )2 dg, (8)
eulerspace

where f is the orientation distribution function as a function of the Euler space coordinates g. For isotropic materials, TI is equal to
unity) for these strategies was revealed to be 1.974, 1.982, 1.266, 1.127 and 1.079, respectively, which is consistent with the fact that
rotation of scanning direction always eases texture by changing the heat flux.
Layer thickness is one of the key parameters in multi-metal fabrication as it affects the penetration depth of the melt pools. A larger
layer thickness will lead to thicker penetration depth and hence the formation of a greater amount of intermetallic compounds which
results in lower mechanical strength of the multi-metal parts. As such, layer thickness should be kept to a minimum to avoid the
formation of detrimental phases.
The layer and powder thicknesses can differ substantially for the first few build layers when the process has not reached the steady-
state thicknesses. Fig. 46 and Fig. 47 depict how the powder thickness varies across the initial build layers. In the beginning, the
powder will have a powder thickness of tp1 which is the physical gap width between the substrate and the lower tip of the recoater
blade. After the loose powder in the first layer is melted, it consolidates into a dense layer with a thickness of tm1. Thereafter, the build
platform is dropped by a fixed displacement, z, and the powder is then deposited over the newly formed layer. If the thickness of the
consolidated layer (tm1) is smaller than the layer-wise displacement (z), then the thickness of the unconsolidated powder (tp2, tp2 … tpm)
and layer thickness (tm2, tm3 … tmn) in the subsequent layers will increase towards steady-state values. The powder thickness and layer
thickness for each build layers can be obtained using the following formula [218]:
∑n− 2
tpn = tp1 + (z − c × tp1 ) i=0 (1 − c)i (9)

Fig. 46. Illustration of layer thickness variation [218].

35
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 47. Powder and layer thickness variation across the initial build layers [218].

tmn = c × tpn (10)

where c is the consolidation ratio defined in Equation (10). Given that the initial powder thickness (tp1) and platform displacement (z)
are known, experimental measurements of the layer thickness (tmn) is required to obtain the consolidation ratio, c, through Equations
(9) and (10). Similarly, the consolidation ratio can also be written as the product of the powder packing density (p) and a material loss
coefficient (l).
c = tmn /tpn = p.(1 − l) (11)

The consolidation ratio, c, is equal to the packing density, p under ideal conditions where there is no material loss. The powder
particles can be ejected from the powder bed during the laser scanning process due to the spattering phenomenon. The spattering event
can be attributed to the 1) induced argon gas flow towards the melt pool and 2) the intense metal vapor jet above the melt pool
[219–222]. The induced argon flow carried the powder particles towards the melt pool while the intense metal vapor jet imparts an
upward momentum to eject the particles away from the melt pool. Considering the typical ranges for the powder density (0.4–0.6) and
the consolidation ratio (~0.3) [218], the powder lost due to spattering can be calculated to be 0.25–0.5.
Fig. 47 illustrates the variations of layer thickness and powder thickness using the equation above for an initial powder thickness
(50 μm), constant platform displacement of 50 μm, and a consolidation ratio of 0.34. It can be observed that both powder and layer
thicknesses rise gradually and reach approximately 90% of their steady-state value within 6 build layers. For the first print layer, a low
layer thickness and low laser energy input may be used to minimize the melt pool penetration and limit the dilution between dissimilar
materials. With a few layers of dissimilar materials across the build, the composition of the top layer shifts towards that of the
deposited material, and dilution control is no longer essential. Thereafter, the deposition rates can be increased by increasing the layer
thickness and penetration depth.
Re-melting strategies, which refers to layer scanning twice or more before depositing a new powder layer, had shown their po­
tentials in reducing porosity and surface roughness since they eliminate surface contaminants, removes oxide films, and produces a
clean solid–liquid interface at the atomic level, especially when other parameters cannot achieve full density. In order to reduce re­
sidual stress and cracks, Li et al. proposed that re-melting should be carried out at a lower laser energy density [223]. Smaller melt
pools are formed because lesser laser energy is absorbed by the solidified bulk compared to powder materials even with the same
scanning parameters, producing finer dendrites due to higher cooling rates. Nevertheless, investigation of re-melting on the un-melted
previous layer and layers below is rare due to the complex thermal history. Yasa et al. studied multiple re-melting parameters and
concluded that their effects on densification depend on careful parameter optimization. They reported that re-melting could remedy
the porosity defects during the L-PBF process [224]. Similar conclusions are drawn in a study conducted by Yu et al. [225]. In addition,
they also found re-melting decreases the surface roughness of L-PBF parts. Re-melting can also be applied only to the last layer or the
outer skin of the part if the aim is to reduce the roughness or to enhance the surface properties. Re-melting was done at every layer to
improve the overall part’s homogeneity for in-situ alloying of Ti6Al4V + 1Cu (at%). As the ε required for Ti6Al4V and copper are
similar, there is no unmelted copper found. Only copper enriched areas of up to 35 wt% can be observed due to incomplete diffusion of
molten copper. Due to the density difference, the copper enriched area is often found at the bottom of the melt pools [82]. While
research on scanning strategies has been carried out using single material systems, it is of interest to see the effect of different scanning
strategies on the bonding and interface for multi-material processing due to the difference in melt pools and other physical phenomena
resulted from these. It has been demonstrated that remelting strategies could not improve the bonding strength between CuSn and
18Ni300 as the interface became brittle, however, this requires further research for other material systems [150].
There is a need to come up with creative unconventional strategies to deal with the issue of compositional segregation without
compromising the part’s density to propel the in-situ alloying technique via L-PBF. To resolve this issue, a preliminary study by Huang
et. al. demonstrated the effectiveness of laser re-melting strategy to address the porosity-segregation dilemma of in-situ alloyed binary
alloys with elements of drastically different melting points [110]. The essence behind the laser re-melting strategy is to have multiple
laser passes on the same layer, effectively giving more opportunity for the dissolution of refractory alloying elements without resorting
to excessive ε that leads to keyhole porosity. Ti42Nb (wt%) parts were built in the study and it was found that laser re-melting strategy

36
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

could enhance the part’s homogeneity while maintaining the amount of keyhole porosity. The laser re-melting strategy was also
studied by Brodie et. al. on in-situ alloyed Ti25Ta (wt%) [108]. It was shown that the effectiveness of the re-melting strategy to
approach the porosity-segregation dilemma is highly dependent on the laser parameters. When the single scan parameters are opti­
mized, pore-free parts can be achieved with a large amount of area percentage of unmelted tantalum (~0.75%). Utilization of laser re-
melting further reduces the area percentage unmelted tantalum to ~0.25% without inducing additional porosity. Besides having a
higher melting point, tantalum possesses higher thermal conductivity as compared to titanium. Tantalum particles essentially cool
faster and have limited time for diffusion, in which the re-melting strategy provides a second chance for dissolution. Further inves­
tigation was also done by Brodie et al. on the Ti25Ta (wt%) manufactured via LPBF using conventional strategy and laser re-melting
strategy to understand their low cycle fatigue behavior [226]. It was found that the keyhole-induced porosity is the dominant fatigue
crack initiation site. No evidence has been found on unmelted tantalum particles to initiate fatigue crack, hence prompting the need for
further investigation on the effect of localized inhomogenity. In some cases, the laser re-melting strategy can lead to more keyhole
porosity at a lower scanning speed (or when the ε is relatively high). For the fabrication of Ti37Nb6Sn (wt%) alloy via in-situ alloying,
the increased ε leads to improved homogeneity but induces higher porosity due to more tin evaporation [120]. Moreover, it was found
that the narrow melt pool produced by L-PBF only encircles a few powders and can contribute to macro-segregation due to localized
inhomogeneity of the powder mixture [123]. Meanwhile, in-situ alloying of Al-xCu leaves unmelted copper particles which can
deteriorate tensile properties by acting as crack nucleation sites. To mitigate this problem, Martinez et al. proceeded with the pre­
heating of the powder bed [74]. A density of more than 99.1% can be achieved for in-situ alloyed Al-xCu manufactured from both
preheated and non-preheated powder beds. However, the preheated powder bed resulted in samples that are more chemically ho­
mogenous. A study by Huang et al. demonstrated the possibility to approach the porosity-segregation dilemma of in-situ alloying via
melt pool manipulation [293]. Through the unique usage of high power top-hat profile laser coupled with high-speed stripe scanning
strategy, the authors managed to create big and slow-moving melt pools that simaltaneously homogenizes and densifies the resultant
parts.
Compositional segregation is a common issue when the in-situ alloying technique is used. One innovative approach by Zafari & Xia
is to leverage such an inherent issue of in-situ alloying to achieve hybrid titanium alloy (HYTA) with enhanced mechanical properties
[227]. As opposed to having a homogenous material, Ti6Al4V (wt%) and Ti5Al5V5M3Cr (wt%) powder of equal mass were mixed and
in-situ alloyed to intentionally achieve full dense parts with blocks of material portion retaining the composition of its raw materials.
The approach leads to the retention of α’ (Ti6Al4V) and β (Ti5Al5V5Mo3Cr) phases where each phase undergoes a series of work
hardening mechanisms progressively that eventually improves the ultimate tensile strength to more than 1200 MPa and elongation to
failure to more than 15%. The strain mismatch or stress concentration at α’/β interface is relieved by the activation of twinning at the
interface as well as the smooth transfer of deformation from β to α’ phase, hence achieving simultaneous strengthening and
improvement in ductility. Additional study on HYTA has been done using equal mass of titanium (wt%), Ti6Al4V (wt%), and
Ti5Al5V5Mo3Cr (wt%) powder [228]. The ternary HYTA has a microstructure that is much more complex but with each phase having
subtle differences in strength, eventually eliminating double yielding phenomena in binary HYTA to achieve a high yield strength of
~1000 MPa (ultimate tensile strength ~ 1100 MPa). The formation of highly heterogeneous microstructure leads to a large variety of
deformation mechanisms such as slip, twinning, strain-induced martensitic transformation and grain rotation activated. Accompanied
by smooth transfer of deformation between phases, enhanced ductility of ~20% elongation to failure was achieved. Further heat
treatment on HYTA can achieve tailored strength and ductility as needed, making the HYTA system a promising material to be used.
Despite the tendency for incomplete melting of constituents during in-situ alloying, the parts produced using this approach have been
shown to have comparable performance to parts formed using conventional pre-alloyed powders. Zhao et al. fabricated Cu15Ni8Sn
samples using pre-alloyed powder and powder mixtures of copper-tin and nickel. Both types of powder were able to produce samples
with high density of more than 99% with similar microstructures and phases [229]. Chen et al. also found that the samples produced by
in-situ alloying possess higher hardness and ultimate tensile strength compared to similar materials fabricated using pre-alloyed
powder [230].

Fig. 48. Schematic representation of three different orientations used.

37
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Vertical, inclined, and horizontal building orientations (Fig. 48) are popular for parts with high length-to-height ratios for L-PBF.
Rashid et al. coined a novel term, energy per layer (El), to predict the relative density while trying to quantify the relationship between
relative density and build orientation [231]. El is given by:
El = E × d × Ap (12)

where El is laser energy density supplied, d is powder layer thickness and Ap is printing area per layer. It was found that energy per layer
in the range of 504–895 J yielded ≥99.8% relatively dense Al12Si L-PBF samples. The anisotropy caused by building orientation is also
investigated. For example, the fatigue resistance is higher in the horizontal direction than in the inclined and vertical direction [232].
Tolosa et al. compared mechanical tests of 316L stainless steel in all the possible build orientations in L-PBF and not only in three main
directions [233]. For different materials, the detailed comparison between different build orientations needs further investigation.
However, due to the nature of L-PBF, most of the multi-metal parts produced have variation along the vertical direction, which restricts
the build orientation choices.

6.2. Preventing intermetallic compounds formation

A series of compatible interlayers can be introduced to the multi-metal interface to eliminate all the brittle phases. The binary phase
diagrams of all the connected pairs of elements in Fig. 49 are free of brittle phases. In other words, a ductile fracture within the
interface or weaker base alloy can be achieved using suitable choices of interlayers [205,234]. Theoretically, there is no restriction on
the number of interlayers and the composition of each interlayer. In other words, alloys may also be used as interlayers. However, it
would be impractical to use too many interlayers for joining multi-metal structures as many alloys’ compositions will have to be
prepared. As such, the DED process is among the most efficient way to achieve a wide range of compositions across the multi-metal
interfaces as it allows mixing and deposition of distinct powders at arbitrary compositions [192,202,235].
Nonetheless, Tey et al. have attempted to prevent the intermetallic formation during multi-metal processing using L-PBF by
introducing a copper alloy interlayer between Ti6Al4V and 316L stainless steel [237]. However, it was found that the titanium/copper
interface contained three detrimental phases which limited the strength of the multi-metal part. Despite this, this method of intro­
ducing an interlayer is envisioned to improve multi-metal bonding.

6.3. New material systems

As mentioned previously, the materials available for L-PBF are limited due to restrictions by the feedstock production methods as
well as the 3D printability of pre-existing materials. As such, huge research effort has been devoted to expanding the materials library.
One of the emerging research areas is high entropy alloys (HEAs). HEAs have high mixing configurational entropies that stabilise solid
solutions based on simple underlying face-centered cubic (FCC), body-centered cubic (BCC), or hexagonal close-packed (HCP)
structures. Using the composition-based definition, an HEA is composed of five or more elements with a concentration between 50 and
35 atomic percentage (at%). Some of the commonly used constituent elements of HEAs include copper, chromium, cobalt, iron,
manganese, nickel, tantalum, titanium, molybdenum, niobium, vanadium, zirconium, tungsten, zinc, aluminium, silicon and boron.
HEAs exhibit superior mechanical properties at high temperatures and exceptional strength, ductility, and fracture toughness at
cryogenic temperatures; hence their popularity in multiple industries that require extreme temperature applications. For L-PBF, many
studies have been conducted on HEAs, including CoCrFeMnNi [238–241], FeCoCrNi [242–244], FeCoNiCrN [243], AlCoCrFeNi [245],
FeMnCoCrSi [246]. Systematic and comprehensive reviews have been done on HEAs processed by L-PBF [247,248].
As an emerging materials group for L-PBF, most of the works conducted have used pre-alloyed powders as feedstocks. However, in-
situ alloying provides the opportunities for rapid modification of test compositions for the HEAs which lower the cost for production.
To test the feasibility of compositions using more a more cost-effective method, Luo et al. used powder mixtures of AlCrCuFeNi pre-
alloyed powder with nickel powder [249]. Chen et al. fabricated CoCrFeMnNi HEA samples using a mixture of CoCrFeNi pre-alloyed

Fig. 49. Binary combinations of elements with no intermetallic compounds [236].

38
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 50. Schematic illustration of the rapid alloy development methodology using powder mixtures for the L-PBF, where “…n” is a placeholder for
additional elements [250].

powder and manganese elemental powder and showed the feasibility of in-situ alloying to form HEAs [230]. Ewald et al. used in-situ
alloying to carry out a study on the AlCCoFeMnNi alloy system [250]. The experimental approach is shown in Fig. 50. The study
showed the possibility of processing complex powder blends with different morphologies, sizes by L-PBF. The manufacturing of parts
with complex geometry, such as lattice structures is also shown in the same study.
Unlike conventional joining techniques, AM features some unique capabilities which make it attractive for multi-metal processing.
Firstly, with the appropriate dispensing mechanism, material compositions can be varied across the build layers by mixing different
powders in various proportions. This allows the user to introduce any arbitrary number of interlayers between the base alloys. As such,
it is possible to directly manipulate the composition, microstructure, and properties across the interface during the build process.
Secondly, by varying the material distribution within and across each build layer, multi-metal interfaces with a complex
geometrical design and material variation in 3D can potentially be created (Fig. 51). This could allow interlocking mechanisms to be
incorporated at the multi-metal interface for improved interfacial fracture toughness [251–253]. In comparison, friction welding and
explosion welding are restricted to the bonding of near-planar and/or planar surfaces only. As a result, the distribution of material can
only be varied in one-dimension (1D). Other joining techniques like laser welding, electron beam welding, diffusion bonding, brazing,
transient liquid phase bonding, and friction stir welding have the capability to process non-planar surfaces featuring material vari­
ations in 2D.
Thirdly, the repeated stacking of melt pools within and across the build layers naturally induces microscale waviness (Fig. 51). This
waviness could potentially increase the resistance to crack propagation and enhance fracture toughness [251–253]. In contrast, joining
processes such as diffusion bonding, brazing, and transient liquid phase bonding produce joints with continuous and planar layers of
intermetallic compounds which makes crack propagation relatively easy.
Lastly, the layer thickness and penetration depth of the melt pool in AM techniques can be varied to control the interfacial
composition gradient (Fig. 52). Low layer thickness and shallow penetration depth can be used to produce a narrow interface with
limited dilution while a large layer thickness and penetration depth can be used to produce a thicker interface with a gradual transition
in properties. By dynamically changing the penetration depth at selected layers, it is also possible to homogenize or create an arbitrary
functional gradient across the interface (Fig. 52). Such flexibility in altering the interface is not available in other joining techniques.
Even though the above mentioned capabilities are available to both the DED and L-PBF processes, each process excels differently in
the respective areas. Specifically, DED excels in varying the material across and within each build layer while the L-PBF process re­
quires an additional dispensing mechanism to realise such benefit [179,254,255]. However, the better resolution of the L-PBF process
makes it superior in limiting interfacial dilution. Additionally, the L-PBF process also features higher cooling rates (106–107 K/s [37])
than DED due to its smaller laser spot diameter (~80–100 μm), lower layer thickness (20–100 μm), and higher laser scanning velocities
(102–103 mm/s). Hence, the microstructure of the L-PBF interface is likely to be finer and may contain metastable phases which differs
from a DED interface. Furthermore, the L-PBF may also produce a unique composite-like interface due to imperfect mixing within the

Fig. 51. Illustration of interfaces with 1D, 2D, and 3D material variations.

39
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 52. Effect of melt pool dimension on composition gradient.

rapidly solidified melt pools [256]. The composite-like structure may contain regions belonging to the undiluted base alloy which
could potentially act as a ductile reinforcement for the brittle interface. This, however, needs to be verified through further research.
Due to the lack of tensile test data for AM multi-metal parts, little is known about its capabilities and bond strength. In particular,
the multi-metal processing potential of L-PBF has been largely overlooked thus far. Nevertheless, AM techniques possess a unique
combination of characteristics that could make them compatible with multi-metal processing.

6.4. New powder deposition methods

New powder deposition mechanisms for L-PBF are needed to achieve multi-metal processing and in-situ alloying. It is important for
the active development of new deposition methods to achieve functionally graded materials (FGM) as well. For instance, Liu et al.
introduced a partition to the powder recoating mechanism so that two different powder materials can be deposited at any build layer to
produce multi-metal parts [14]. Wang et al. simply incorporated two hoppers into the build chamber (Fig. 53). However, these
methods only allow materials to be varied across the build layers but not within each layer. In order to achieve intra-layer material
variation, a nozzle-based system has been incorporated into L-PBF so that different powders can be selectively deposited within each
build layer [14,141,255].
Unlike typical L-PBF, a multiple powder delivery system was designed to enable multi-metal processing (Fig. 54). This system
includes two sets of powder feeders and two powder hoppers that can load two metals at the sample time. The release of powder onto
the powder bed can be controlled selectively by a switch.
Recognizing the needs from the industries to have both multi-metal processing capability in single layer and across the layers, a six
channel ultrasonic powder delivery system is developed as shown in Fig. 55 [257].
Another prototype for multi-metal deposition is demonstrated for L-PBF [144], as shown in Fig. 56. The method is used to fabricate
specimens using pure iron and Al12Si, with components that consist of pure iron and Al12Si regions with a transition region of in-situ
alloyed materials from these two constituents.
An innovative multi-metal powder deposition system that utilizes a vibration plate for powder metering was developed for EB-PBF

Fig. 53. Schematic of multi-metal L-PBF with two hoppers in the build chamber [158].

40
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 54. Schematic of the multi-material SLM machine [147].

Fig. 55. Schematic of the multi-material set up with six channel ultrasonic dispenser with (b) and (c) showing the 3D model and photograph of the
ultrasonic powder dispenser array [257].

[258], as shown in Fig. 57. The utilization of a vibrating plate reduces the need for revolute pair, making it highly reliable in a dusty
environment. The system was developed for the fabrication of Ti6Al4V/Ti47Al2Cr2Nb graded materials. Due to the similar process
nature, the innovation can be adapted for the L-PBF process with proper modification.
Zhang et al. developed an in-situ powder mixing system for L-PBF that allowed two different powders to be mixed and dispensed in a
selected ratio (Fig. 58). The core part of the system is the powder mixing and feeding system that has three sub-vibration systems to
provide stable powder flows. Both types of powders can be mixed evenly by rotating the mixing box and then the mixture is released
into the lower hopper for dispensing to the build platform [259]. This system also enables multi-metal processing as the mixing system

41
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 56. Concept of using multi-material transition zone for assembling Al-alloys with steel. (b) In-house built prototype L-PBF system with multi-
material processing capability. (c) Design of the powder feeder system. (d) Working principle of the powder feeder for mixing powders. (e) Cali­
bration curves of the delivered powder mass (m) of pure Fe and (f) Al12Si as a function of applied voltage (A) and vibration time (tv). Error bars
depict 95% confidence interval for the mean [144].

can be deactivated while keeping the dispensing system operational which allows for the two materials to be deposited independently.
Recycling of materials from multi-metal processing is challenging as the unconsolidated powder cannot be easily recovered as they
are interspersed within the common powder bed. To address this issue, Wei et al. incorporated a vacuum system to remove loose
powder from selective regions in the powder bed which can help avoid cross-contamination of different powders when removing the
materials [179,257]. A similar concept is also proposed by Chivel [260]. Demir & Previtali suggested that AM industry can refer to
electronics and pharmaceutical industries for powder recycling and separation methods. For example, the magnetic powder separation
method can be used and separation based on shape and density can also be employed [144]. Gravity separation techniques can also be
derived from the mining industry to achieve separation by density difference [261]. Horn et al. investigated the possibility of using

42
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 57. Multi-metal deposition system for EB-PBF (a) Powder supply with vibration plate into the mixing box. (b) Application of mixed powder
onto the powder bed [258].

Fig. 58. In-situ powder mixing system for L-PBF (a) detailed schematic of the system (b) delivering powders into a mixing box (c) mixing powders by
motor controlled rotation of the box with (d) and (e) showing the mechanism of the mixing systems where powders were both vertically and
horizontally merged and horizontally separated and vertically merged again respectively [259].

43
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

binary powder mixtures of different particle size distributions which can be sorted and recycled by screening after L-PBF. Firstly, they
conditioned the powders in accordance with DIN 6615 to secure sortable powder size distributions that are sufficiently different from
one another. After processing by L-PBF, the powders were sorted using two screening stages. It is found that the sorting process yields
an economic solution for powder recycling as the efficiency is only 10% lower than typical recycling methods used for single material
processing [262].

6.5. Simulations

Due to the complexity of the physical phenomena during L-PBF, simulations have been popular as ways to understand the process.
Multiple research have been conducted using the discrete element method [263], thermo-fluid dynamics [264,265], finite element
method [266,267], and phase-field model [268,269]. Comprehensive reviews on simulations for L-PBF have been published
[270–272]. Despite extensive work on simulations for L-PBF, limited information is available for multi-metal processing and in-situ
alloying.
A multi-layer finite element model was proposed to investigate the thermal behavior for TiB2/Ti6Al4V multi-metal processing by L-
PBF [142]. A schematic of the simulations is shown in Fig. 59.
In the model, the element birth and death method was applied to simulate the multi-layers during L-PBF. The simulation results
showed that the maximum temperature gradient was located at the interface between TiB2 and Ti6Al4V, and the interface temperature
and melt pool lifetime are important for the wettability at the interface. In another study, molecular dynamics were used to simulate
the L-PBF melting process between iron and aluminium for multi-metal processing [273]. It is observed that aluminium experienced a
higher melting rate compared to iron and the two metals can fuse with each other after melting. A discrete element method was used to
simulate the melt pools’ evolvement and melt track morphology for multi-metal deposited in the same and across different layers
(Fig. 60) [274].
The simulations can visualise the interface between 316L stainless steel and Cu10Sn. Due to their different thermal properties, melt
pool development appear differently when the same energy density was applied and 316L stainless steel melted first which results in
the melting of Cu10Sn by conduction and convection flow from the liquid 316L stainless steel. Phase migration at the interface was
found to be related to the convection within the melt pools which contributed to the mixing and elemental diffusion (Fig. 61).
Meanwhile, multi-material simulation has been done based on the EB-PBF technology with a multi-scale simulation, in which, the

Fig. 59. Schematic overview of the L-PBF physical model (a) the established finite element model (b) and laser scanning strategy (c) during the L-
PBF process (point 1 at the center of the Ti6Al4V layer; point 2 at the center of the TiB2 layer) [142].

44
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Fig. 60. Framework for multi-track, multi-layer, and multi-metal L-PBF modelling [274].

Fig. 61. Simulations for multi-metal L-PBF (a) thermal boundary conditions of the calculation domain (b) 316L and Cu10Sn powders with clear
boundary (c) laser beam applied on the boundary with 175 W and 800 mm/s (d) track morphology after solidification [274].

Fig. 62. Initial powder beds for different sceneroir with Case 1 and 5 for single material, Case 2, 3 and 4 for in-situ alloying and Case 6 and 7 for
multi-metal processing [276].

melting of particles with different materials was considered [275]. Nevertheless, the homogenization process of in-situ alloying
through particle mixing process based on Marangoni convection as well as particle dissolution process based on diffusion are not well
studied.

45
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Sun et al. presented a mesoscopic model based on a volume of fluid method. In their work, they simulated L-PBF with mixed and
unmixed dissimilar powders which make their model applicable for both multi-metal processing and in-situ alloying. The initial
powder beds for different scenarios are shown in Fig. 62. For mixed powders, non-homogenous temperature distribution is observed
due to the different thermal-physical properties of the two materials. The model also showed partially melted or unmelted particles
embedded in a fully melted matrix which results in significant fluctuation of the track width. For the melting of unmixed powder beds,
an asymmetrical melt pool is observed at the interface [276].
There has been limited available literature on simulations for in-situ alloying by L-PBF. However, numerous researches have been
done on simulations for metal matrix composites using L-PBF [277–282]. Like in-situ alloying, powder mixtures are typically used in
composites, and thus, several aspects of these simulations can be adapted in future research.

6.6. Machine learning and artificial intelligence

While simulations allow a deeper understanding of L-PBF, there are still unresolved challenges such as process variability and lack
of standards. Such variations result in difficulty in process planning which restricted the widespread adoption of L-PBF in the industry.
Machine learning and artificial intelligence can be used to enhance and complement simulations in achieving higher part quality
[283]. Machine learning is often utilized to capture process behavior using measurement data.
Typically for processing parameters optimisation, experiments are done to identify the process windows to avoid certain defects.
This is time-consuming and cost-intensive, in addition to material and even equipment specific. Hence, optimised parameters usually
cannot be generalised for L-PBF. A data-driven framework is used to detect layer-wise anomalies observed by thermal imaging for L-
PBF [284]. A spatial statistics model is then used with supervised and unsupervised machine learning techniques to detect defects
within the layer.
Machine learning methods were applied to measured data to establish input and output relationships between the process pa­
rameters and surface texture. Predictive models using genetic programming and neural network modelling for process planning
purposes are achieved for L-PBF of Inconel 625 [285]. Deep neural networks are used to interpret the image classification problem
using melt pools in order to predict the defects due to L-PBF [286,287]. Comprehensive reviews on machine learning for additive
manufacturing are available [288,289].
By combining simulations and machine learning, it is possible to predict or optimise process parameters to achieve desired part
properties. Self-learning was also made possible [290,291]. A conceptual framework on the combination of statistical analysis,
mathematical modelling, and machine learning techniques has been proposed [292]. These will be useful for multi-metal processing
and in-situ alloying for L-PBF as more variables such as materials selection, powder size distribution, and differences in material
properties are involved which make the process more complex.

7. Summary

An analysis of the available literature on in-situ alloying and multi-metal processing by L-PBF has revealed their importance and
explores the possibilities they offer in expanding L-PBF applications. L-PBF will be increasingly employed for fabricating high quality,
low cost, repeatable, and reliable functional parts. With these new approaches, L-PBF will enhance manufacturing and engagement in
the industry. In-situ alloying allows the expansion of the materials library that is available for processing by L-PBF as it also allows rapid
feasibility studies of different compositions of alloys. It has also been highlighted that in-situ alloyed parts have comparable, if not
superior, part properties compared to the same materials formed using pre-alloyed powder by L-PBF. With new materials with
comparable performance, it should encourage a higher adaptation rate in the industry. For multi-metal processing, it allows for
functional parts that often require a combination of different materials properties to be directly fabricated using L-PBF. As such, with
multi-metal processing capability, the industry can then enhance their manufacturing capabilities with the direct fabrication of a wider
range of products. This review brings deep insights into the fabrication, challenges, and potentials in using in-situ alloying and multi-
metal processing by L-PBF.
Fabrication using these routes is considered from the perspectives of materials and processing parameters. Although much progress
has been achieved in modifying the 3D printability of metals, limited alloys are available nowadays. While various methods have been
adopted to prepare the feedstock, more efficient and economic processes are in demand, especially for industrial production. This
review also addressed the influence of heat transfer and fluid flow in in-situ alloying and multi-metal processing. Defects associated
with L-PBF need to be carefully controlled as complexity is added to the process for these approaches. Balling effect, porosity, crack,
loss of alloying elements, oxide inclusions, intermetallic phases, and unmelted particles are all discussed. More investigations are
necessary to overcome these challenges.
Ultimately, L-PBF offers to advance materials science and manufacturing technology in the future. By using in-situ alloying and
multi-metal processing, it becomes more promising for high-performance products. However, research in this field is relatively new
and in the infancy stage, there is a need for accelerated trials and investigations for real-time applications to come to fruition.

CRediT authorship contribution statement

S.L. Sing: Conceptualization, Methodology, Writing - original draft, Writing - review & editing, Visualization, Supervision, Project
administration. S. Huang: Conceptualization, Methodology, Writing - original draft, Writing - review & editing. G.D. Goh: Writing -
review & editing. G.L. Goh: Writing - review & editing. C.F. Tey: Writing - original draft. J.H.K. Tan: Writing - original draft. W.Y.

46
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

Yeong: Conceptualization, Supervision, Project administration, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Acknowledgement

The authors acknowledge the support from the National Research Foundation, Prime Minister’s Office, Singapore under its
Medium-Sized Centre funding scheme. S. Huang also acknowledge the Nanyang President’s Graduate Scholarship from Nanyang
Technological University, Singapore for funding the Ph.D. studies.

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pmatsci.2021.100795.

References

[1] Liu ZH, Zhang DQ, Chua CK, Leong KF. Crystal structure analysis of M2 high speed steel parts produced by selective laser melting. Mater Charact 2013;84:
72–80.
[2] Sing SL, Yeong WY, Wiria FE, Tay BY. Characterization of titanium lattice structures fabricated by selective laser melting using an adapted compressive test
method. Exp Mech 2016;56:735–48.
[3] Herzog D, Seyda V, Wycisk E, Emmelmann C. Additive manufacturing of metals. Acta Mater 2016;117:371–92.
[4] Sing SL, Yeong WY, Wiria FE, Tay BY, Zhao Z, Zhao L, et al. Direct selective laser sintering and melting of ceramics: a review. Rapid Prototyp J 2017;23:
611–23.
[5] Yap CY, Chua CK, Dong ZL, Liu ZH, Zhang DQ, Loh LE, et al. Review of selective laser melting: Materials and applications. Appl Phys Rev 2015;2:041101.
[6] Bogue R. Nanocomposites: a review of technology and applications. Assembly Automat 2011;31:106–12.
[7] Yu WH, Sing SL, Chua CK, Kuo CN, Tian XL. Particle-reinforced metal matrix nanocomposites fabricated by selective laser melting: a state of the art review.
Prog Mater Sci 2019;104:330–79.
[8] Wu W, Tor SB, Chua CK, Leong KF, Merchant A. Investigation on processing of ASTM A131 Eh36 high tensile strength steel using selective laser melting. Virt
Phys Prototyp 2015;10:187–93.
[9] Meng L, Zhao J, Lan X, Yang H, Wang Z. Multi-objective optimisation of bio-inspired lightweight sandwich structures based on selective laser melting. Virt
Phys Prototyp 2020;15:106–19.
[10] Concli F, Gilioli A. Numerical and experimental assessment of the mechanical properties of 3D printed 18-Ni300 steel trabecular structures produced by
Selective Laser Melting – a lean design approach. Virt Phys Prototyp 2019;14:267–76.
[11] Anandan S, Hussein RM, Spratt M, Newkirk J, Chandrashekhara K, Misak H, et al. Failure In metal honeycombs manufactured by selective laser melting of 304
L stainless steel under compression. Virt Phys Prototyp 2019;14:114–22.
[12] Sing SL, Huang S, Yeong WY. Effect of solution heat treatment on microstructure and mechanical properties of laser powder bed fusion produced cobalt-
28chromium-6molybdenum. Mater Sci Eng, A 2020;769:138511.
[13] Yang J, Yu H, Yin J, Gao M, Wang Z, Zeng X. Formation and control of martensite in Ti-6Al-4V alloy produced by selective laser melting. Mater Des 2016;108:
308–18.
[14] Liu ZH, Zhang DQ, Sing SL, Chua CK, Loh LE. Interfacial characterization of SLM parts in multi-material processing: metallurgical diffusion between 316L
stainless steel and C18400 copper alloy. Mater Charact 2014;94:116–25.
[15] Yang C, Zhao YJ, Kang LM, Li DD, Zhang WW, Zhang LC. High-strength silicon brass manufactured by selective laser melting. Mater Lett 2018;210:169–72.
[16] Attar H, Bönisch M, Calin M, Zhang L-C, Zhuravleva K, Funk A, et al. Comparative study of microstructures and mechanical properties of in situ Ti–TiB
composites produced by selective laser melting, powder metallurgy, and casting technologies. J Mater Res 2014;29:1941–50.
[17] Zhang LC, Xu J, Eckert J. Thermal stability and crystallization kinetics of mechanically alloyed TiC/Ti-based metallic glass matrix composite. J Appl Phys
2006;100:033514.
[18] Liu LH, Yang C, Wang F, Qu SG, Li XQ, Zhang WW, et al. Ultrafine grained Ti-based composites with ultrahigh strength and ductility achieved by equiaxing
microstructure. Mater Des 2015;79:1–5.
[19] Kühn U, Mattern N, Gebert A, Kusy M, Boström M, Siegel U, et al. Nanostructured Zr-and Ti-based composite materials with high strength and enhanced
plasticity. J Appl Phys 2005;98:054307.
[20] DebRoy T, Wei HL, Zuback JS, Mukherjee T, Elmer JW, Milewski JO, et al. Additive manufacturing of metallic components – process, structure and properties.
Prog Mater Sci 2018;92:112–224.
[21] Gu DD, Meiners W, Wissenbach K, Poprawe R. Laser additive manufacturing of metallic components: materials, processes and mechanisms. Int Mater Rev
2013;57:133–64.
[22] Olakanmi EO, Cochrane RF, Dalgarno KW. A review on selective laser sintering/melting (SLS/SLM) of aluminium alloy powders: processing, microstructure,
and properties. Prog Mater Sci 2015;74:401–77.
[23] Sercombe TB, Li X. Selective laser melting of aluminium and aluminium metal matrix composites: a review. Mater Technol 2016;31:77–85.
[24] Yadroitsev I, Gusarov AV, Yadroitsava I, Smurov I. Single track formation in selective laser melting of metal powders. J Mater Process Technol 2010;210:
1624–31.
[25] Markl M, Körner C. Multiscale modeling of powder bed-based additive manufacturing. Annu Rev Mater Res 2016;46:93–123.
[26] AlMangour B, Grzesiak D, Borkar T, Yang J-M. Densification behavior, microstructural evolution, and mechanical properties of TiC/316L stainless steel
nanocomposites fabricated by selective laser melting. Mater Des 2018;138:119–28.
[27] Ahn IH, Moon SK, Hwang J, Bi G. Characteristic length of the solidified melt pool in selective laser melting process. Rapid Prototyp J 2017;23:370–81.
[28] Wang XC, Laoui T, Bonse J, Kruth J-P, Lauwers B, Froyen L. Direct selective laser sintering of hard metal powders: experimental study and simulation. Int J Adv
Manuf Technol 2002;19:351–7.
[29] Scharowsky T, Osmanlic F, Singer RF, Körner C. Melt pool dynamics during selective electron beam melting. Appl Phys A 2013;114:1303–7.
[30] Arafune K, Hirata A. Thermal and solutal Marangoni convection in In–Ga–Sb system. J Cryst Growth 1999;197:811–7.

47
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

[31] Kovalev OB, Gurin AM. Multivortex convection of metal in molten pool with dispersed impurity induced by laser radiation. Int J Heat Mass Transf 2014;68:
269–77.
[32] Chan CL, Mazumder J, Chen MM. Effect of surface tension gradient driven convection in a laser melt pool: three-dimensional perturbation model. J Appl Phys
1988;64:6166–74.
[33] Rongy L, De Wit A. Steady Marangoni flow traveling with chemical fronts. J Chem Phys 2006;124:164705.
[34] Fabbro R, Hamadou M, Coste F. Metallic vapor ejection effect on melt pool dynamics in deep penetration laser welding. J Laser Appl 2004;16:16–9.
[35] Fabbro R, Slimani S, Doudet I, Coste F, Briand F. Experimental study of the dynamical coupling between the induced vapour plume and the melt pool for
Nd–Yag CW laser welding. J Phys D Appl Phys 2006;39:394.
[36] Liu Y, Yang Y, Mai S, Wang D, Song C. Investigation into spatter behavior during selective laser melting of AISI 316L stainless steel powder. Mater Des 2015;87:
797–806.
[37] Loh LE, Chua CK, Yeong WY, Song J, Mapar M, Sing SL, et al. Numerical investigation and an effective modelling on the Selective Laser Melting (SLM) process
with aluminium alloy 6061. Int J Heat Mass Transf 2015;80:288–300.
[38] Mahamood RM, Akinlabi ET. Effect of the scanning speed of treatment on the microstructure, microhardness, wear, and corrosion behavior of laser metal-
deposited Ti–6AL–4V/TiC composite. Mater Sci 2017;53:76–85.
[39] Sing SL, Wiria FE, Yeong WY. Selective laser melting of lattice structures: a statistical approach to manufacturability and mechanical behavior. Rob Comput
Integr Manuf 2018;49:170–80.
[40] Thijs L, Verhaeghe F, Craeghs T, Humbeeck JV, Kruth J-P. A study of the microstructural evolution during selective laser melting of Ti–6Al–4V. Acta Mater
2010;58:3303–12.
[41] Saedi S, Shayesteh Moghaddam N, Amerinatanzi A, Elahinia M, Karaca HE. On the effects of selective laser melting process parameters on microstructure and
thermomechanical response of Ni-rich NiTi. Acta Mater 2018;144:552–60.
[42] Simonelli M, Tuck C, Aboulkhair NT, Maskery I, Ashcroft I, Wildman RD, et al. A study on the laser spatter and the oxidation reactions during selective laser
melting of 316L stainless steel, Al-Si10-Mg, and Ti-6Al-4V. Metall Mater Trans A 2015;46:3842–51.
[43] Yap CY, Chua CK, Dong ZL. An effective analytical model of selective laser melting. Virt Phys Prototyp 2016;11:21–6.
[44] Yang J, Han J, Yu H, Yin J, Gao M, Wang Z, et al. Role of molten pool mode on formability, microstructure and mechanical properties of selective laser melted
Ti-6Al-4V alloy. Mater Des 2016;110:558–70.
[45] Yadroitsev I, Krakhmalev P, Yadroitsava I. Selective laser melting of Ti6Al4V alloy for biomedical applications: temperature monitoring and microstructural
evolution. J Alloy Compd 2014;583:404–9.
[46] Kuo CN, Chua CK, Peng PC, Chen YW, Sing SL, Huang S, et al. Microstructure evolution and mechanical property response via 3D printing parameter
development of Al-Sc alloy. Virt Phys Prototyp 2020;15:120–9.
[47] Qiu C, Panwisawas C, Ward M, Basoalto HC, Brooks JW, Attallah MM. On the role of melt flow into the surface structure and porosity development during
selective laser melting. Acta Mater 2015;96:72–9.
[48] Mishra P, Ilar T, Brueckner F, Kaplan A. Energy efficiency contributions and losses during selective laser melting. J Laser Appl 2018;30:032304.
[49] Yang Y, Wen S, Wei Q, Li W, Liu J, Shi Y. Effect of scan line spacing on texture, phase and nanohardness of TiAl/TiB 2 metal matrix composites fabricated by
selective laser melting. J Alloy Compd 2017;728:803–14.
[50] Li R, Shi Y, Wang Z, Wang L, Liu J, Jiang W. Densification behavior of gas and water atomized 316L stainless steel powder during selective laser melting. Appl
Surf Sci 2010;256:4350–6.
[51] Zhang S, Wei Q, Cheng L, Li S, Shi Y. Effects of scan line spacing on pore characteristics and mechanical properties of porous Ti6Al4V implants fabricated by
selective laser melting. Mater Des 2014;63:185–93.
[52] Yadroitsev I, Thivillon L, Bertrand P, Smurov I. Strategy of manufacturing components with designed internal structure by selective laser melting of metallic
powder. Appl Surf Sci 2007;254:980–3.
[53] Di W, Yongqiang Y, Xubin S, Yonghua C. Study on energy input and its influences on single-track, multi-track, and multi-layer in SLM. Int J Adv Manuf Technol
2012;58:1189–99.
[54] Ma M, Wang Z, Gao M, Zeng X. Layer thickness dependence of performance in high-power selective laser melting of 1Cr18Ni9Ti stainless steel. J Mater Process
Technol 2015;215:142–50.
[55] Gu D, Shen Y. Processing and microstructure of submicron WC–Co particulate reinforced Cu matrix composites prepared by direct laser sintering. Mater Sci
Eng, A 2006;435:54–61.
[56] Tang M, Pistorius PC, Beuth J. Geometric model to predict porosity of part produced in powder bed system. Mater Sci Technol Proc 2015,:129–35.
[57] Tang M, Pistorius PC. Oxides, porosity and fatigue performance of AlSi10Mg parts produced by selective laser melting. Int J Fatigue 2017;94:192–201.
[58] Tang M. Inclusions, Porosity, and Fatigue of AlSi10Mg Parts Produced by Selective Laser Melting. Carnegie Mellon University; 2017.
[59] Gusarov AV, Yadroitsev I, Bertrand P, Smurov I. Heat transfer modelling and stability analysis of selective laser melting. Appl Surf Sci 2007;254:975–9.
[60] Li R, Liu J, Shi Y, Wang L, Jiang W. Balling behavior of stainless steel and nickel powder during selective laser melting process. Int J Adv Manuf Technol 2011;
59:1025–35.
[61] Kang N, Coddet P, Dembinski L, Liao H, Coddet C. Microstructure and strength analysis of eutectic Al-Si alloy in-situ manufactured using selective laser melting
from elemental powder mixture. J Alloy Compd 2017;691:316–22.
[62] Kang N, Coddet P, Liao H, Baur T, Coddet C. Wear behavior and microstructure of hypereutectic Al-Si alloys prepared by selective laser melting. Appl Surf Sci
2016;378:142–9.
[63] Gode C, Yilmazer H, Ozdemir I, Todaka Y. Microstructural refinement and wear property of Al–Si–Cu composite subjected to extrusion and high-pressure
torsion. Mater Sci Eng, A 2014;618:377–84.
[64] Hanemann T, Carter LN, Habschied M, Adkins NJE, Attallah MM, Heilmaier M. In-situ alloying of AlSi10Mg+Si using Selective Laser Melting to control the
coefficient of thermal expansion. J Alloy Compd 2019;795:8–18.
[65] Kang N, Coddet P, Chen C, Wang Y, Liao H, Coddet C. Microstructure and wear behavior of in-situ hypereutectic Al–high Si alloys produced by selective laser
melting. Mater Des 2016;99:120–6.
[66] Kang N, Coddet P, Liao H, Coddet C. Macrosegregation mechanism of primary silicon phase in selective laser melting hypereutectic Al – High Si alloy. J Alloy
Compd 2016;662:259–62.
[67] Kang N, Coddet P, Ammar M-R, Liao H, Coddet C. Characterization of the microstructure of a selective laser melting processed Al-50Si alloy: effect of heat
treatments. Mater Charact 2017;130:243–9.
[68] Prashanth KG, Scudino S, Klauss HJ, Surreddi KB, Löber L, Wang Z, et al. Microstructure and mechanical properties of Al–12Si produced by selective laser
melting: effect of heat treatment. Mater Sci Eng, A 2014;590:153–60.
[69] Suryawanshi J, Prashanth KG, Scudino S, Eckert J, Prakash O, Ramamurty U. Simultaneous enhancements of strength and toughness in an Al-12Si alloy
synthesized using selective laser melting. Acta Mater 2016;115:285–94.
[70] Martin JH, Yahata BD, Hundley JM, Mayer JA, Schaedler TA, Pollock TM. 3D printing of high-strength aluminium alloys. Nature 2017;549:365.
[71] Montero-Sistiaga ML, Mertens R, Vrancken B, Wang X, Van Hooreweder B, Kruth J-P, et al. Changing the alloy composition of Al7075 for better processability
by selective laser melting. J Mater Process Technol 2016;238:437–45.
[72] Zhang H, Zhu H, Nie X, Yin J, Hu Z, Zeng X. Effect of Zirconium addition on crack, microstructure and mechanical behavior of selective laser melted Al-Cu-Mg
alloy. Scr Mater 2017;134:6–10.
[73] Wang P, Deng L, Prashanth KG, Pauly S, Eckert J, Scudino S. Microstructure and mechanical properties of Al-Cu alloys fabricated by selective laser melting of
powder mixtures. J Alloy Compd 2018;735:2263–6.
[74] Martinez R, Todd I, Mumtaz K. In situ alloying of elemental Al-Cu12 feedstock using selective laser melting. Virt Phys Prototyp 2019;14:242–52.

48
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

[75] Bartkowiak K, Ullrich S, Frick T, Schmidt M. New developments of laser processing aluminium alloys via additive manufacturing technique. Phys Procedia
2011;12:393–401.
[76] Vora P, Mumtaz K, Todd I, Hopkinson N. AlSi12 in-situ alloy formation and residual stress reduction using anchorless selective laser melting. Addit Manuf
2015;7:12–9.
[77] Vora P, Martinez R, Hopkinson N, Todd I, Mumtaz K. Customised alloy blends for in-situ Al339 alloy formation using anchorless selective laser melting.
Technologies 2017;5:24.
[78] Simonelli M, Aboulkhair NT, Cohen P, Murray JW, Clare AT, Tuck C, et al. A comparison of Ti-6Al-4V in-situ alloying in Selective Laser Melting using simply-
mixed and satellited powder blend feedstocks. Mater Charact 2018;143:118–26.
[79] Simonelli M, McCartney DG, Barriobero-Vila P, Aboulkhair NT, Tse YY, Clare A, et al. The influence of iron in minimizing the microstructural anisotropy of Ti-
6Al-4V produced by laser powder-bed fusion. Metall Mater Trans A 2020;51:2444–59.
[80] Vrancken B, Thijs L, Kruth JP, Van Humbeeck J. Microstructure and mechanical properties of a novel β titanium metallic composite by selective laser melting.
Acta Mater 2014;68:150–8.
[81] Huber F, Papke T, Scheitler C, Hanrieder L, Merklein M, Schmidt M. In situ formation of a metastable β-Ti alloy by laser powder bed fusion (L-PBF) of
vanadium and iron modified Ti-6Al-4V. Metals 2018;8:1067.
[82] Krakhmalev P, Yadroitsev I, Yadroitsava I, de Smidt O. Functionalization of biomedical Ti6Al4V via in situ alloying by Cu during laser powder bed fusion
manufacturing. Materials 2017;10:1154.
[83] Macpherson A, Li X, McCormick P, Ren L, Yang K, Sercombe TB. Antibacterial titanium produced using selective laser melting. JOM 2017;69:2719–24.
[84] Qiu C, Fones A, Hamilton HGC, Adkins NJE, Attallah MM. A new approach to develop palladium-modified Ti-based alloys for biomedical applications. Mater
Des 2016;109:98–111.
[85] Sing SL, An J, Yeong WY, Wiria FE. Laser and electron-beam powder-bed additive manufacturing of metallic implants: a review on processes, materials and
designs. J Orthop Res 2016;34:369–85.
[86] Ridzwan MIZ, Shuib S, Hassan AY, Shokri AA, Ibrahim MNM. Problem of stress shielding and improvement to the hip implant designs: a review. J Med Sci
2007;7:460–7.
[87] Wolff J. Das gesetz der transformation der knochen. A Hirshwald 1892;1:1–152.
[88] Li Y, Yang C, Zhao H, Qu S, Li X, Li Y. New developments of Ti-based alloys for biomedical applications. Materials 2014;7:1709–800.
[89] Niinomi M. Recent metallic materials for biomedical applications. Metall Mater Trans A 2002;33:477.
[90] Sundfeldt M, Carlsson VL, Johansson BC, Thomsen P, Gretzer C. Aseptic loosening, not only a question of wear: a review of different theories. Acta
Orthopaedica 2006;77:177–97.
[91] Wapner KL. Implications of metallic corrosion in total knee arthroplasty. Clin Orthop Relat Res 1991;271:12–20.
[92] Kuroda D, Niinomi M, Morinaga M, Kato Y, Yashiro T. Design and mechanical properties of new β type titanium alloys for implant materials. Mater Sci Eng, A
1998;243:244–9.
[93] Zhuravleva K, Bönisch M, Prashanth K, Hempel U, Helth A, Gemming T, et al. Production of porous β-type Ti–40Nb alloy for biomedical applications:
comparison of selective laser melting and hot pressing. Materials 2013;6:5700–12.
[94] Sing SL, Wiria FE, Yeong WY. Selective laser melting of titanium alloy with 50 wt% tantalum: effect of laser process parameters on part quality. Int J Refract
Metal Hard Mater 2018;77:120–7.
[95] Yang Y, Wang G, Liang H, Gao C, Peng S, Shen L, et al. Additive manufacturing of bone scaffolds. Int J Bioprint 2019;5:148.
[96] Al-Tamimi AA, Peach C, Bartolo P. Topology optimization of metallic locking compression plates produced using electron beam melting. Int J Adv Manuf
Technol 2018;104:195–210.
[97] Aleixo GT, Afonso C, Coelho A, Caram R. Effects of omega phase on elastic modulus of Ti-Nb alloys as a function of composition and cooling rate. Solid State
Phenomena: Trans Tech Publ; 2008. p. 393–8.
[98] Sun Z, Tan X, Tor SB, Chua CK. Simultaneously enhanced strength and ductility for 3D-printed stainless steel 316L by selective laser melting. NPG Asia Mater
2018;10:127–36.
[99] Ozaki T, Matsumoto H, Watanabe S, Hanada S. Beta Ti alloys with low Young’s modulus. Mater Trans 2004;45:2776–9.
[100] Hon Y-H, Wang J-Y, Pan Y-N. Composition/phase structure and properties of titanium-niobium alloys. Mater Trans 2003;44:2384–90.
[101] Collings EW. The physical metallurgy of titanium alloys. Metals Park Ohio. 1984;3.
[102] Long M, Rack HJ. Titanium alloys in total joint replacement—a materials science perspective. Biomaterials 1998;19:1621–39.
[103] Kim HY, Miyazaki S. Martensitic transformation and superelastic properties of Ti-Nb base alloys. Mater Trans 2015;56:625–34.
[104] Wang Q, Han C, Choma T, Wei Q, Yan C, Song B, et al. Effect of Nb content on microstructure, property and in vitro apatite-forming capability of Ti-Nb alloys
fabricated via selective laser melting. Mater Des 2017;126:268–77.
[105] Fischer M, Joguet D, Robin G, Peltier L, Laheurte P. In situ elaboration of a binary Ti–26Nb alloy by selective laser melting of elemental titanium and niobium
mixed powders. Mater Sci Eng, C 2016;62:852–9.
[106] Huang S, Sing SL, de Looze G, Wilson R, Yeong WY. Laser powder bed fusion of titanium-tantalum alloys: Compositions and designs for biomedical
applications. J Mech Behav Biomed Mater 2020;108:103775.
[107] Zhao D, Han C, Li Y, Li J, Zhou K, Wei Q, et al. Improvement on mechanical properties and corrosion resistance of titanium-tantalum alloys in-situ fabricated
via selective laser melting. J Alloy Compd 2019;804:288–98.
[108] Brodie EG, Medvedev AE, Frith JE, Dargusch MS, Fraser HL, Molotnikov A. Remelt processing and microstructure of selective laser melted Ti25Ta. J Alloy
Compd 2020;820:153082.
[109] Sing SL, Yeong WY, Wiria FE. Selective laser melting of titanium alloy with 50 wt% tantalum: microstructure and mechanical properties. J Alloy Compd 2016;
660:461–70.
[110] Huang S, Sing SL, Yeong WY. Selective laser melting of Ti42Nb composite powder and the effect of laser re-melting. Key Eng Mater 2019;801:270–5.
[111] Wang JC, Liu YJ, Qin P, Liang SX, Sercombe TB, Zhang LC. Selective laser melting of Ti–35Nb composite from elemental powder mixture: microstructure,
mechanical behavior and corrosion behavior. Mater Sci Eng, A 2019;760:214–24.
[112] Nagase T, Hori T, Todai M, Sun S-H, Nakano T. Additive manufacturing of dense components in beta-titanium alloys with crystallographic texture from a
mixture of pure metallic element powders. Mater Des 2019;173:107771.
[113] Chlebus E, Kuźnicka B, Dziedzic R, Kurzynowski T. Titanium alloyed with rhenium by selective laser melting. Mater Sci Eng, A 2015;620:155–63.
[114] Majchrowicz K, Pakieła Z, Brynk T, Romelczyk-Baishya B, Płocińska M, Kurzynowski T, et al. Microstructure and mechanical properties of Ti–Re alloys
manufactured by selective laser melting. Mater Sci Eng, A 2019;765:138290.
[115] Zhao D, Han C, Li J, Liu J, Wei Q. In situ fabrication of a titanium-niobium alloy with tailored microstructures, enhanced mechanical properties and
biocompatibility by using selective laser melting. Mater Sci Eng 2020;C(110784).
[116] Liang H, Zhao D, Feng X, Ma L, Deng X, Han C, et al. 3D-printed porous titanium scaffolds incorporating niobium for high bone regeneration capacity. Mater
Des 2020;194:108890.
[117] Yan L, Yuan Y, Ouyang L, Li H, Mirzasadeghi A, Li L. Improved mechanical properties of the new Ti-15Ta-xZr alloys fabricated by selective laser melting for
biomedical application. J Alloy Compd 2016;688:156–62.
[118] Schwab H, Prashanth K, Löber L, Kühn U, Eckert J. Selective laser melting of Ti-45Nb alloy. Metals 2015;5:686–94.
[119] Schulze C, Weinmann M, Schweigel C, Keßler O, Bader R. Mechanical properties of a newly additive manufactured implant material based on Ti-42Nb.
Materials 2018;11:124.
[120] Chen W, Chen C, Zi X, Cheng X, Zhang X, Lin YC, et al. Controlling the microstructure and mechanical properties of a metastable β titanium alloy by selective
laser melting. Mater Sci Eng, A 2018;726:240–50.

49
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

[121] Xie F, He X, Cao S, Mei M, Qu X. Influence of pore characteristics on microstructure, mechanical properties and corrosion resistance of selective laser sintered
porous Ti–Mo alloys for biomedical applications. Electrochim Acta 2013;105:121–9.
[122] Azizi H, Zurob H, Bose B, Reza Ghiaasiaan S, Wang X, Coulson S, et al. Additive manufacturing of a novel Ti-Al-V-Fe alloy using selective laser melting. Addit
Manuf 2018;21:529–35.
[123] Wang C, Tan XP, Du Z, Chandra S, Sun Z, Lim CWJ, et al. Additive manufacturing of NiTi shape memory alloys using pre-mixed powders. J Mater Process
Technol 2019;271:152–61.
[124] Zhang B, Chen J, Coddet C. Microstructure and transformation behavior of in-situ shape memory alloys by selective laser melting Ti–Ni mixed powder. J Mater
Sci Technol 2013;29:863–7.
[125] Liu S, Han S, Zhang L, Chen L-Y, Wang L, Zhang L, et al. Strengthening mechanism and micropillar analysis of high-strength NiTi–Nb eutectic-type alloy
prepared by laser powder bed fusion. Compos B Eng 2020;200:108358.
[126] Grigoriev A, Polozov I, Sufiiarov V, Popovich A. In-situ synthesis of Ti2AlNb-based intermetallic alloy by selective laser melting. J Alloy Compd 2017;704:
434–42.
[127] Polozov I, Sufiiarov V, Popovich A, Masaylo D, Grigoriev A. Synthesis of Ti-5Al, Ti-6Al-7Nb, and Ti-22Al-25Nb alloys from elemental powders using powder-
bed fusion additive manufacturing. J Alloy Compd 2018;763:436–45.
[128] Bolzoni L, Weissgaerber T, Kieback B, Ruiz-Navas EM, Gordo E. Mechanical behaviour of pressed and sintered CP Ti and Ti–6Al–7Nb alloy obtained from
master alloy addition powder. J Mech Behav Biomed Mater 2013;20:149–61.
[129] Barriobero-Vila P, Gussone J, Stark A, Schell N, Haubrich J, Requena G. Peritectic titanium alloys for 3D printing. Nat Commun 2018;9:3426.
[130] Kang N, El Mansori M, Guittonneau F, Liao H, Fu Y, Aubry E. Controllable mesostructure, magnetic properties of soft magnetic Fe-Ni-Si by using selective laser
melting from nickel coated high silicon steel powder. Appl Surf Sci 2018;455:736–41.
[131] Shuai C, He C, Feng P, Guo W, Gao C, Wu P, et al. Biodegradation mechanisms of selective laser-melted Mg-xAl-Zn alloy: grain size and intermetallic phase. Virt
Phys Prototyp 2018;13:59–69.
[132] Long T, Zhang X, Huang Q, Liu L, Liu Y, Ren J, et al. Novel Mg-based alloys by selective laser melting for biomedical applications: microstructure evolution,
microhardness and in vitro degradation behaviour. Virt Phys Prototyp 2017;13:71–81.
[133] Zhang B, Liao H, Coddet C. Effects of processing parameters on properties of selective laser melting Mg–9%Al powder mixture. Mater Des 2012;34:753–8.
[134] Shuai C, Yang W, Yang Y, Gao C, He C, Pan H. A continuous net-like eutectic structure enhances the corrosion resistance of Mg alloys. Int J Bioprint 2019;5:
207.
[135] Yang Y, Yuan F, Gao C, Feng P, Xue L, He S, et al. A combined strategy to enhance the properties of Zn by laser rapid solidification and laser alloying. J Mech
Behav Biomed Mater 2018;82:51–60.
[136] Erinc M, Sillekens WH, Mannens RGTM, Werkhoven RJ. Applicability of existing magnesium alloys as biomedical implant materials; 2009.
[137] Shuai C, Xue L, Gao C, Yang Y, Peng S, Zhang Y. Selective laser melting of Zn–Ag alloys for bone repair: microstructure, mechanical properties and degradation
behaviour. Virt Phys Prototyp 2018;13:146–54.
[138] Jones JB, Cooper DE, Wimpenny DI, Gibbons GJ. Gateways toward dissimilar multi-material parts. In: RAPID 2012 and 3D Imaging Conferences & Exposition;
2012.
[139] Anstaett C, Seidel C, Reinhart G. Fabrication of 3D multi-material parts using laser-based powder bed fusion. In: Proceedings of the 28th Annual International
Solid Freeform Fabrication Symposium; 2017.
[140] Beal VE, Erasenthiran P, Hopkinson N, Dickens P, Ahrens CH. Fabrication of x-graded H13 and Cu powder mix using high power pulsed Nd: YAG laser.
Proceedings of Solid Freeform Fabrication Symposium; 2004.
[141] Sing SL, Lam LP, Zhang DQ, Liu ZH, Chua CK. Interfacial characterization of SLM parts in multi-material processing: intermetallic phase formation between
AlSi10Mg and C18400 copper alloy. Mater Charact 2015;107:220–7.
[142] Chen C, Gu D, Dai D, Du L, Wang R, Ma C, et al. Laser additive manufacturing of layered TiB2/Ti6Al4V multi-material parts: understanding thermal behavior
evolution. Opt Laser Technol 2019;119:105666.
[143] Nadimpalli VK, Dahmen T, Valente EH, Mohanty S, Pedersen DB. Multi-material additive manufacturing of steels using laser powder bed fusion. In: Euspen’s
19th International Conference & Exhibition: The European Society for Precision Engineering and Nanotechnology; 2019. p. 240–3.
[144] Demir AG, Previtali B. Multi-material selective laser melting of Fe/Al-12Si components. Manuf Lett 2017;11:8–11.
[145] Hinojos A, Mireles J, Reichardt A, Frigola P, Hosemann P, Murr LE, et al. Joining of Inconel 718 and 316 Stainless Steel using electron beam melting additive
manufacturing technology. Mater Des 2016;94:17–27.
[146] Terrazas CA, Gaytan SM, Rodriguez E, Espalin D, Murr LE, Medina F, et al. Multi-material metallic structure fabrication using electron beam melting. Int J Adv
Manuf Technol 2014;71:33–45.
[147] Chen J, Yang Y, Song C, Wang D, Wu S, Zhang M. Influence mechanism of process parameters on the interfacial characterization of selective laser melting
316L/CuSn10. Mater Sci Eng, A 2020.
[148] Chen J, Yang Y, Song C, Zhang M, Wu S, Wang D. Interfacial microstructure and mechanical properties of 316L/CuSn10 multi-material bimetallic structure
fabricated by selective laser melting. Mater Sci Eng, A 2019;752:75–85.
[149] Chen K, Wang C, Hong Q, Wen S, Zhou Y, Yan C, et al. Selective laser melting 316L/CuSn10 multi-materials: processing optimization, interfacial
characterization and mechanical property. J Mater Process Technol 2020;283:116701.
[150] Zhang M, Yang Y, Wang D, Song C, Chen J. Microstructure and mechanical properties of CuSn/18Ni300 bimetallic porous structures manufactured by selective
laser melting. Mater Des 2019;165:107583.
[151] Bai Y, Zhang J, Zhao C, Li C, Wang H. Dual interfacial characterization and property in multi-material selective laser melting of 316L stainless steel and C52400
copper alloy. Mater Charact 2020;167:110489.
[152] Mei X, Wang X, Peng Y, Gu H, Zhong G, Yang S. Interfacial characterization and mechanical properties of 316L stainless steel/inconel 718 manufactured by
selective laser melting. Mater Sci Eng, A 2019;758:185–91.
[153] Shakerin S, Hadadzadeh A, Amirkhiz BS, Shamsdini S, Li J, Mohammadi M. Additive manufacturing of maraging steel-H13 bimetals using laser powder bed
fusion technique. Addit Manuf 2019;29:100797.
[154] Gorny B, Niendorf T, Lackmann J, Thoene M, Troester T, Maier HJ. In situ characterization of the deformation and failure behavior of non-stochastic porous
structures processed by selective laser melting. Mater Sci Eng, A 2011;528:7962–7.
[155] Yan C, Hao L, Hussein A, Young P, Raymont D. Advanced lightweight 316L stainless steel cellular lattice structures fabricated via selective laser melting. Mater
Des 2014;55:533–41.
[156] Böhm C, Werz M, Weihe S. Dilution ratio and the resulting composition profile in dissimilar laser powder bed fusion of AlSi10Mg and Al99.8. Metals 2020;10:
1222.
[157] Hadadzadeh A, Amirkhiz BS, Shakerin S, Kelly J, Li J, Mohammadi M. Microstructural investigation and mechanical behavior of a two-material component
fabricated through selective laser melting of AlSi10Mg on an Al-Cu-Ni-Fe-Mg cast alloy substrate. Addit Manuf 2020;31:10937.
[158] Wang R, Gu D, Chen C, Dai D, Ma C, Zhang H. Formation mechanisms of TiB2 tracks on Ti6Al4V alloy during selective laser melting of ceramic-metal multi-
material. Powder Technol 2020;367:597–607.
[159] Wei K, Zeng X, Li F, Liu M, Deng J. Microstructure and mechanical property of Ti-5Al-2.5Sn/Ti-6Al-4V dissimilar titanium alloys integrally fabricated by
selective laser melting. JOM 2020;72:1031–8.
[160] Bartolomeu F, Costa MM, Alves N, Miranda G, Silva FS. Additive manufacturing of NiTi-Ti6Al4V multi-material cellular structures targeting orthopedic
implants. Opt Laser Technol 2020;134:106208.
[161] Gu D, Shen Y. Balling phenomena in direct laser sintering of stainless steel powder: metallurgical mechanisms and control methods. Mater Des 2009;30:
2903–10.

50
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

[162] Loh LE, Liu ZH, Zhang DQ, Mapar M, Sing SL, Chua CK, et al. Selective laser melting of aluminium alloy using a uniform beam profile. Virt Phys Prototyp 2014;
9:11–6.
[163] Qi H, Azer M, Ritter A. Studies of standard heat treatment effects on microstructure and mechanical properties of laser net shape manufactured Inconel 718.
Metall Mater Trans A 2009;40:2410–22.
[164] Zhou J, Tsai H-L. Porosity formation and prevention in pulsed laser welding. J Heat Transfer 2007;129:1014–24.
[165] Panwisawas C, Qiu CL, Sovani Y, Brooks JW, Attallah MM, Basoalto HC. On the role of thermal fluid dynamics into the evolution of porosity during selective
laser melting. Scr Mater 2015;105:14–7.
[166] Khairallah SA, Anderson AT, Rubenchik A, King WE. Laser powder-bed fusion additive manufacturing: physics of complex melt flow and formation
mechanisms of pores, spatter, and denudation zones. Acta Mater 2016;108:36–45.
[167] Kasperovich G, Haubrich J, Gussone J, Requena G. Correlation between porosity and processing parameters in TiAl6V4 produced by selective laser melting.
Mater Des 2016;105:160–70.
[168] Katayama S. Introduction: fundamentals of laser welding. Handbook of Laser Welding Technologies. Elsevier; 2013. p. 3–16.
[169] Weingarten C, Buchbinder D, Pirch N, Meiners W, Wissenbach K, Poprawe R. Formation and reduction of hydrogen porosity during selective laser melting of
AlSi10Mg. J Mater Process Technol 2015;221:112–20.
[170] Dai D, Gu D. Effect of metal vaporization behavior on keyhole-mode surface morphology of selective laser melted composites using different protective
atmospheres. Appl Surf Sci 2015;355:310–9.
[171] Chahal V, Taylor RM. A review of geometric sensitivities in laser metal 3D printing. Virt Phys Prototyp 2020;15:227–41.
[172] Salmi A, Atzeni E. Residual stress analysis of thin AlSi10Mg parts produced by Laser Powder Bed Fusion. Virt Phys Prototyp 2020;15:49–61.
[173] Kempen K, Thijs L, Vrancken B, Buls S, Van Humbeeck J, Kruth JP. Producing crack-free, high density M2 Hss parts by selective laser melting: pre-heating the
baseplate. In: Proceedings of the 24th International Solid Freeform Fabrication Symposium; 2013. p. 131–9.
[174] Robinson JH, Ashton IRT, Jones E, Fox P, Sutcliffe C. The effect of hatch angle rotation on parts manufactured using selective laser melting. Rapid Prototyp J
2019;25:289–98.
[175] Gåård A, Krakhmalev P, Bergström J. Microstructural characterization and wear behavior of (Fe, Ni)–TiC MMC prepared by DMLS. J Alloy Compd 2006;421:
166–71.
[176] Mercelis P, Kruth J-P. Residual stresses in selective laser sintering and selective laser melting. Rapid Prototyp J 2006;12:254–65.
[177] Zhu HH, Lu L, Fuh JYH. Study on shrinkage behaviour of direct laser sintering metallic powder. Proc Inst Mech Eng, Part B: J Eng Manuf 2006;220:183–90.
[178] AlMangour B, Grzesiak D. Selective laser melting of TiC reinforced 316L stainless steel matrix nanocomposites: Influence of starting TiC particle size and
volume content. Mater Des 2016;104:141–51.
[179] Wei C, Li L, Zhang X, Chueh Y-H. 3D printing of multiple metallic materials via modified selective laser melting. CIRP Ann 2018;67:245–8.
[180] Ramasamy S, Albright CE. CO2 and Nd: YAG laser beam welding of 6111–T4 aluminum alloy for automotive applications. J Laser Appl 2000;12:101–15.
[181] Zhao H, DebRoy T. Weld metal composition change during conduction mode laser welding of aluminum alloy 5182. Metall Mater Trans B 2001;32:163–72.
[182] Collur MM, Paul A, DebRoy T. Mechanism of alloying element vaporization during laser welding. Metall Trans B 1987;18:733–40.
[183] Pastor M, Zhao H, Martukanitz R, DebRoy T. Porosity, underfill and magnesium lose during continuous wave Nd: YAG laser welding of thin plates of aluminum
alloys 5182 and 5754. Welding J 1999;78. 207-S-16-S.
[184] Louvis E, Fox P, Sutcliffe CJ. Selective laser melting of aluminium components. J Mater Process Technol 2011;211:275–84.
[185] Cao X, Wallace W, Immarigeon J-P, Poon C. Research and progress in laser welding of wrought aluminum alloys. II. Metallurgical microstructures, defects, and
mechanical properties. Mater Manuf Processes 2003;18:23–49.
[186] Katayama S, Seto N, Mizutani M, Matsunawa A. Formation mechanism of porosity in high power YAG laser welding. In: Proceedings of the International
Congress on Applications of Lasers and Electro-optics: ICALEO: Springer-Verlag; 2000. p. 200016.
[187] Read N, Wang W, Essa K, Attallah MM. Selective laser melting of AlSi10Mg alloy: process optimisation and mechanical properties development. Mater Des
2015;65:417–24.
[188] Kang N, Li Y, Lin X, Feng E, Huang W. Microstructure and tensile properties of Ti-Mo alloys manufactured via using laser powder bed fusion. J Alloy Compd
2019;771:877–84.
[189] Lu Y, Dong Y, Jiang L, Wang T, Li T, Zhang Y. A criterion for topological close-packed phase formation in high entropy alloys. Entropy 2015;17:2355–66.
[190] Tan C, Zhang X, Dong D, Attard B, Wang D, Kuang M, et al. In-situ synthesised interlayer enhances bonding strength in additively manufactured multi-material
hybrid tooling. Int J Mach Tools Manuf 2020;155:103592.
[191] Ting W, Zhang B-G, Feng J-C. Influences of different filler metals on electron beam welding of titanium alloy to stainless steel. Trans Nonferr Met Soc China
2014;24:108–14.
[192] Bobbio LD, Otis RA, Borgonia JP, Dillon RP, Shapiro AA, Liu Z-K, et al. Additive manufacturing of a functionally graded material from Ti-6Al-4V to Invar:
experimental characterization and thermodynamic calculations. Acta Mater 2017;127:133–42.
[193] Sahasrabudhe H, Harrison R, Carpenter C, Bandyopadhyay A. Stainless steel to titanium bimetallic structure using LENS™. Addit Manuf 2015;5:1–8.
[194] Kundu S, Ghosh M, Laik A, Bhanumurthy K, Kale GB, Chatterjee S. Diffusion bonding of commercially pure titanium to 304 stainless steel using copper
interlayer. Mater Sci Eng, A 2005;407:154–60.
[195] Tomashchuk I, Sallamand P, Andrzejewski H, Grevey D. The formation of intermetallics in dissimilar Ti6Al4V/copper/AISI 316 L electron beam and Nd: YAG
laser joints. Intermetallics 2011;19:1466–73.
[196] Muralimohan CH, Ashfaq M, Ashiri R, Muthupandi V, Sivaprasad K. Analysis and characterization of the role of Ni interlayer in the friction welding of titanium
and 304 austenitic stainless steel. Metall Mater Trans A 2016;47:347–59.
[197] Sam S, Kundu S, Chatterjee S. Diffusion bonding of titanium alloy to micro-duplex stainless steel using a nickel alloy interlayer: interface microstructure and
strength properties. Mater Des 2012;40:237–44.
[198] Lee JG, Lee M-K. Microstructure and mechanical behavior of a titanium-to-stainless steel dissimilar joint brazed with Ag-Cu alloy filler and an Ag interlayer.
Mater Charact 2017;129:98–103.
[199] Deng Y, Sheng G, Xu C. Evaluation of the microstructure and mechanical properties of diffusion bonded joints of titanium to stainless steel with a pure silver
interlayer. Mater Des 2013;46:84–7.
[200] Cheepu M, Ashfaq M, Muthupandi V. A new approach for using interlayer and analysis of the friction welding of titanium to stainless steel. Trans Indian Inst
Met 2017;70:2591–600.
[201] Kundu S, Sam S, Chatterjee S. Interfacial reactions and strength properties in dissimilar titanium alloy/Ni alloy/microduplex stainless steel diffusion bonded
joints. Mater Sci Eng, A 2013;560:288–95.
[202] Reichardt A, Dillon RP, Borgonia JP, Shapiro AA, McEnerney BW, Momose T, et al. Development and characterization of Ti-6Al-4V to 304L stainless steel
gradient components fabricated with laser deposition additive manufacturing. Mater Des 2016;104:404–13.
[203] Ishida K, Gao Y, Nagatsuka K, Takahashi M, Nakata K. Microstructures and mechanical properties of friction stir welded lap joints of commercially pure
titanium and 304 stainless steel. J Alloy Compd 2015;630:172–7.
[204] Aleman B, Gutiérrez L, Urcola JJ. Interface microstructures in diffusion bonding of titanium alloys to stainless and low alloy steels. Mater Sci Technol 1993;9:
633–41.
[205] Song TF, Jiang XS, Shao ZY, Fang YJ, Mo DF, Zhu DG, et al. Microstructure and mechanical properties of vacuum diffusion bonded joints between Ti-6Al-4V
titanium alloy and AISI316L stainless steel using Cu/Nb multi-interlayer. Vacuum 2017;145:68–76.
[206] Mudali UK, Rao BMA, Shanmugam K, Natarajan R, Raj B. Corrosion and microstructural aspects of dissimilar joints of titanium and type 304L stainless steel.
J Nucl Mater 2003;321:40–8.
[207] Chiba A, Nishida M, Morizono Y, Imamura K. Bonding characteristics and diffusion barrier effect of the TiC phase formed at the bonding interface in an
explosively welded titanium/high-carbon steel clad. J Phase Equilibria 1995;16:411–5.

51
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

[208] Nishida M, Chiba A, Honda Y, Hirazumi J-i, Horikiri K. Electron microscopy studies of bonding interface in explosively welded Ti/steel clads. ISIJ Int 1995;35:
217–9.
[209] Liao J, Yamamoto N, Liu H, Nakata K. Microstructure at friction stir lap joint interface of pure titanium and steel. Mater Lett 2010;64:2317–20.
[210] Wei HL, Elmer JW, DebRoy T. Origin of grain orientation during solidification of an aluminum alloy. Acta Mater 2016;115:123–31.
[211] Li XP, Van Humbeeck J, Kruth JP. Selective laser melting of weak-textured commercially pure titanium with high strength and ductility: a study from laser
power perspective. Mater Des 2017;116:352–8.
[212] Sing SL, Yeong WY. Laser powder bed fusion for metal additive manufacturing: perspectives on recent developments. Virt Phys Prototyp 2020;15:359–70.
[213] Thijs L, Montero Sistiaga ML, Wauthle R, Xie Q, Kruth J-P, Van Humbeeck J. Strong morphological and crystallographic texture and resulting yield strength
anisotropy in selective laser melted tantalum. Acta Mater 2013;61:4657–68.
[214] Gu D, Guo M, Zhang H, Sun Y, Wang R, Zhang L. Effects of laser scanning strategies on selective laser melting of pure tungsten. Int J Extreme Manuf 2020;2:
025001.
[215] AlMangour B, Grzesiak D, Yang J-M. Scanning strategies for texture and anisotropy tailoring during selective laser melting of TiC/316L stainless steel
nanocomposites. J Alloy Compd 2017;728:424–35.
[216] Thijs L, Kempen K, Kruth J-P, Van Humbeeck J. Fine-structured aluminium products with controllable texture by selective laser melting of pre-alloyed
AlSi10Mg powder. Acta Mater 2013;61:1809–19.
[217] Aboulkhair NT, Everitt NM, Ashcroft I, Tuck C. Reducing porosity in AlSi10Mg parts processed by selective laser melting. Addit Manuf 2014;1:77–86.
[218] Bidare P, Maier RRJ, Beck RJ, Shephard JD, Moore AJ. An open-architecture metal powder bed fusion system for in-situ process measurements. Addit Manuf
2017;16:177–85.
[219] Bidare P, Bitharas I, Ward RM, Attallah MM, Moore AJ. Fluid and particle dynamics in laser powder bed fusion. Acta Mater 2018;142:107–20.
[220] Matthews MJ, Guss G, Khairallah SA, Rubenchik AM, Depond PJ, King WE. Denudation of metal powder layers in laser powder bed fusion processes. Acta
Mater 2016;114:33–42.
[221] Guo Q, Zhao C, Escano LI, Young Z, Xiong L, Fezzaa K, et al. Transient dynamics of powder spattering in laser powder bed fusion additive manufacturing
process revealed by in-situ high-speed high-energy x-ray imaging. Acta Mater 2018;151:169–80.
[222] Gunenthiram V, Peyre P, Schneider M, Dal M, Coste F, Koutiri I, et al. Experimental analysis of spatter generation and melt-pool behavior during the powder
bed laser beam melting process. J Mater Process Technol 2018;251:376–86.
[223] Li XP, Kang CW, Huang H, Sercombe TB. The role of a low-energy–density re-scan in fabricating crack-free Al85Ni5Y6Co2Fe2 bulk metallic glass composites
via selective laser melting. Mater Des 2014;63:407–11.
[224] Yasa E, Deckers J, Kruth JP. The investigation of the influence of laser re-melting on density, surface quality and microstructure of selective laser melting parts.
Rapid Prototyp J 2011;17:312–27.
[225] Yu WH, Sing SL, Chua CK, Tian XL. Influence of re-melting on surface roughness and porosity of AlSi10Mg parts fabricated by selective laser melting. J Alloy
Compd 2019;792:574–81.
[226] Brodie EG, Richter J, Wegener T, Niendorf T, Molotnikov A. Low-cycle fatigue performance of remelted laser powder bed fusion (L-PBF) biomedical Ti25Ta.
Mater Sci Eng, A 2020;798:140228.
[227] Zafari A, Xia K. High Ductility in a fully martensitic microstructure: a paradox in a Ti alloy produced by selective laser melting. Mater Res Lett 2018;6:627–33.
[228] Zafari A, Lui EW, Jin S, Li M, Molla TT, Sha G, et al. Hybridisation of microstructures from three classes of titanium alloys. Mater Sci Eng, A 2020;788:139572.
[229] Zhao C, Wang Z, Li D, Xie M, Kollo L, Luo Z, et al. Comparison of additively manufacturing samples fabricated from pre-alloyed and mechanically mixed
powders. J Alloy Compd 2020;830:154603.
[230] Chen P, Li S, Zhou Y, Yan M, Attallah MM. Fabricating CoCrFeMnNi high entropy alloy via selective laser melting in-situ alloying. J Mater Sci Technol 2020;43:
40–3.
[231] Rashid R, Masood SH, Ruan D, Palanisamy S, Rahman Rashid RA, Elambasseril J, et al. Effect of energy per layer on the anisotropy of selective laser melted
AlSi12 aluminium alloy. Addit Manuf 2018;22:426–39.
[232] Brandl E, Heckenberger U, Holzinger V, Buchbinder D. Additive manufactured AlSi10Mg samples using Selective Laser Melting (SLM): microstructure, high
cycle fatigue, and fracture behavior. Mater Des 2012;34:159–69.
[233] Tolosa I, Garciandía F, Zubiri F, Zapirain F, Esnaola A. Study of mechanical properties of AISI 316 stainless steel processed by “selective laser melting”,
following different manufacturing strategies. Int J Adv Manuf Technol 2010;51:639–47.
[234] Li P, Li J, Xiong J, Zhang F, Raza SH. Diffusion bonding titanium to stainless steel using Nb/Cu/Ni multi-interlayer. Mater Charact 2012;68:82–7.
[235] Hofmann DC, Roberts S, Otis R, Kolodziejska J, Dillon RP, Suh J-o, et al. Developing gradient metal alloys through radial deposition additive manufacturing.
Sci Rep 2014;4:5357.
[236] Wang T, Zhang B, Chen G, Feng J. High strength electron beam welded titanium–stainless steel joint with V/Cu based composite filler metals. Vacuum 2013;
94:41–7.
[237] Tey CF, Tan X, Sing SL, Yeong WY. Additive manufacturing of multiple materials by selective laser melting: Ti-alloy to stainless steel via a Cu-alloy interlayer.
Addit Manuf 2020;31:100970.
[238] Li R, Niu P, Yuan T, Cao P, Chen C, Zhou K. Selective laser melting of an equiatomic CoCrFeMnNi high-entropy alloy: Processability, non-equilibrium
microstructure and mechanical property. J Alloy Compd 2018;746:125–34.
[239] Ren J, Mahajan C, Liu L, Follette D, Chen W, Mukherjee S. Corrosion behavior of selectively laser melted CoCrFeMnNi high entropy alloy. Metals 2019;9:1029.
[240] Kim Y-K, Choe J, Lee K-A. Selective laser melted equiatomic CoCrFeMnNi high-entropy alloy: Microstructure, anisotropic mechanical response, and multiple
strengthening mechanism. J Alloy Compd 2019;805:680–91.
[241] Xu Z, Zhang H, Li W, Mao A, Wang L, Song G, et al. Microstructure and nanoindentation creep behavior of CoCrFeMnNi high-entropy alloy fabricated by
selective laser melting. Addit Manuf 2019;28:766–71.
[242] Lin D, Xu L, Jing H, Han Y, Zhao L, Minami F. Effects of annealing on the structure and mechanical properties of FeCoCrNi high-entropy alloy fabricated via
selective laser melting. Addit Manuf 2020;32:101058.
[243] Song M, Zhou R, Gu J, Wang Z, Ni S, Liu Y. Nitrogen induced heterogeneous structures overcome strength-ductility trade-off in an additively manufactured
high-entropy alloy. Appl Mater Today 2020;18:100498.
[244] Sun Z, Tan XP, Descoins M, Mangelinck D, Tor SB, Lim CS. Revealing hot tearing mechanism for an additively manufactured high-entropy alloy via selective
laser melting. Scr Mater 2019;168:129–33.
[245] Niu PD, Li RD, Yuan TC, Zhu SY, Chen C, Wang MB, et al. Microstructures and properties of an equimolar AlCoCrFeNi high entropy alloy printed by selective
laser melting. Intermetallics 2019;104:24–32.
[246] Agrawal P, Thapliyal S, Nene SS, Mishra RS, McWilliams BA, Cho KC. Excellent strength-ductility synergy in metastable high entropy alloy by laser powder bed
additive manufacturing. Addit Manuf 2020;32:101098.
[247] Han C, Fang Q, Shi Y, Tor SB, Chua CK, Zhou K. Recent advances on high-entropy alloys for 3D printing. Adv Mater 2020;1903855.
[248] Chen S, Tong Y, Liaw PK. Additive manufacturing of high-entropy alloys: a review. Entropy 2018;20:937.
[249] Luo S, Zhao C, Su Y, Liu Q, Wang Z. Selective laser melting of dual phase AlCrCuFeNix high entropy alloys: Formability, heterogeneous microstructures and
deformation mechanisms. Addit Manuf 2020;31:100925.
[250] Ewald S, Kies F, Hermsen S, Voshage M, Hasse C, Schleifenbaum JH. Rapid alloy development of extremely high-alloyed metals using powder blends in laser
powder bed fusion. Materials 2019;12:1706.
[251] Cordisco FA, Zavattieri PD, Hector Jr LG, Bower AF. Toughness of a patterned interface between two elastically dissimilar solids. Eng Fract Mech 2012;96:
192–208.
[252] Cordisco FA, Zavattieri PD, Hector Jr LG, Carlson BE. Mode I fracture along adhesively bonded sinusoidal interfaces. Int J Solids Struct 2016;83:45–64.

52
S.L. Sing et al. Progress in Materials Science 119 (2021) 100795

[253] Li B-W, Zhao H-P, Qin Q-H, Feng X-Q, Yu S-W. Numerical study on the effects of hierarchical wavy interface morphology on fracture toughness. Comput Mater
Sci 2012;57:14–22.
[254] Al-Jamal OM, Hinduja S, Li L. Characteristics of the bond in Cu-H13 tool steel parts fabricated using SLM. CIRP Ann 2008;57:239–42.
[255] Yang S, Evans JRG. A multi-component powder dispensing system for three dimensional functional gradients. Mater Sci Eng, A 2004;379:351–9.
[256] Hegge HJ, De Hosson JTM. The influence of convection on the homogeneity of laser-applied coatings. J Mater Sci 1991;26:711–4.
[257] Wei C, Sun Z, Chen Q, Liu Z, Li L. Additive manufacturing of horizontal and 3D functionally graded 316L/Cu10Sn components via multiple material selective
laser melting. J Manuf Sci Eng 2019;141:081014.
[258] Guo C, Ge W, Lin F. Dual-material electron beam selective melting: hardware development and validation studies. Engineering 2015;1:124–30.
[259] Zhang X, Chuen Y-h, Wei C, Sun Z, Yan J, Li L. Additive manufacturing of three-dimensional metal-glass functionally gradient material components by laser
powder bed fusion with in situ powder mixing. Addit Manuf 2020;33:101113.
[260] Chivel Y. New approach to multi-material processing in selective laser melting. Phys Procedia 2016;83:891–8.
[261] Chapter 16 - Gravity Separation. In: Gupta A, Yan D, editors. Mineral Processing Design and Operations (Second Edition). Amsterdam: Elsevier; 2016. p.
563–628.
[262] Horn M, Prestel L, Schmitt M, Binder M, Schlick G, Seidel C, et al. Multi-material additive manufacturing – recycling of binary metal powder mixtures by
screening. Procedia CIRP 2020;93:50–5.
[263] Wu Y-C, San C-H, Chang C-H, Lin H-J, Marwan R, Baba S, et al. Numerical modeling of melt-pool behavior in selective laser melting with random powder
distribution and experimental validation. J Mater Process Technol 2018;254:72–8.
[264] Leitz K-H, Grohs C, Singer P, Tabernig B, Plankensteiner A, Kestler H, et al. Fundamental analysis of the influence of powder characteristics in Selective Laser
Melting of molybdenum based on a multi-physical simulation model. Int J Refract Metal Hard Mater 2018;72:1–8.
[265] Ding X, Wang L. Heat transfer and fluid flow of molten pool during selective laser melting of AlSi10Mg powder: Simulation and experiment. J Manuf Processes
2017;26:280–9.
[266] Bruna-Rosso C, Demir AG, Previtali B. Selective laser melting finite element modeling: validation with high-speed imaging and lack of fusion defects
prediction. Mater Des 2018;156:143–53.
[267] Arısoy YM, Criales LE, Özel T. Modeling and simulation of thermal field and solidification in laser powder bed fusion of nickel alloy IN625. Opt Laser Technol
2019;109:278–92.
[268] Wang X, Chou K. Microstructure simulations of Inconel 718 during selective laser melting using a phase field model. Int J Adv Manuf Technol 2019;100:
2147–62.
[269] Radhakrishnan B, Gorti SB, Turner JA, Acharya R, Sharon JA, Staroselsky A, et al. Phase field simulations of microstructure evolution in IN718 using a
surrogate Ni–Fe–Nb alloy during laser powder bed fusion†. Metals 2019;9:14.
[270] Kyogoku H, Ikeshoji T-T. A review of metal additive manufacturing technologies: mechanism of defects formation and simulation of melting and solidification
phenomena in laser powder bed fusion process. Mech Eng Rev 2020;7:19–00182.
[271] Schoinochoritis B, Chantzis D, Salonitis K. Simulation of metallic powder bed additive manufacturing processes with the finite element method: a critical
review. Proc Inst Mech Eng, Part B: J Eng Manuf 2015;231:96–117.
[272] Tan JHK, Sing SL, Yeong WY. Microstructure modelling for metallic additive manufacturing: a review. Virt Phys Prototyp 2020;15:87–105.
[273] Sorkin A, Tan JL, Wong CH. Multi-material modelling for selective laser melting. Procedia Eng 2017;216:51–7.
[274] Gu H, Wei C, Li L, Han Q, Setchi R, Ryan M, et al. Multi-physics modelling of molten pool development and track formation in multi-track, multi-layer and
multi-material selective laser melting. Int J Heat Mass Transf 2020;151:119458.
[275] Yan W, Ge W, Smith J, Lin S, Kafka OL, Lin F, et al. Multi-scale modeling of electron beam melting of functionally graded materials. Acta Mater 2016;115:
403–12.
[276] Sun Z, Chuen Y-h, Li L. Multiphase mesoscopic simulation of multiple and functionally gradient materials laser powder bed fusion additive manufacturing
processes. Addit Manuf 2020;35:101448.
[277] Li X, Willy HJ, Chang S, Lu W, Herng TS, Ding J. Selective laser melting of stainless steel and alumina composite: experimental and simulation studies on
processing parameters, microstructure and mechanical properties. Mater Des 2018;145:1–10.
[278] Gan J, Gao H, Wen S, Zhou Y, Tan S, Duan L. Simulation, forming process and mechanical property of Cu-Sn-Ti/diamond composites fabricated by selective
laser melting. Int J Refract Metal Hard Mater 2020;87:105144.
[279] Xia M, Gu D, Ma C, Zhang H, Dai D, Chen H, et al. Fragmentation and refinement behavior and underlying thermodynamic mechanism of WC reinforcement
during selective laser melting of Ni-based composites. J Alloy Compd 2019;777:693–702.
[280] Gu D, Yang Y, Xi L, Yang J, Xia M. Laser absorption behavior of randomly packed powder-bed during selective laser melting of SiC and TiB2 reinforced Al
matrix composites. Opt Laser Technol 2019;119:105600.
[281] Dai D, Gu D, Ge Q, Li Y, Shi X, Sun Y, et al. Mesoscopic study of thermal behavior, fluid dynamics and surface morphology during selective laser melting of Ti-
based composites. Comput Mater Sci 2020;177:109598.
[282] Gu D, Xia M, Dai D. On the role of powder flow behavior in fluid thermodynamics and laser processability of Ni-based composites by selective laser melting. Int
J Mach Tools Manuf 2019;137:67–78.
[283] Goh GD, Sing SL, Yeong WY. A review on machine learning in 3D printing: applications, potential, and challenges. Artif Intell Rev 2020.
[284] Mahmoudi M, Ezzat AA, Elwany A. Layerwise anomaly detection in laser powder-bed fusion metal additive manufacturing. J Manuf Sci Eng 2019;141:031002.
[285] Özel T, Altay A, Kaftanoğlu B, Leach R, Senin N, Donmez A. Focus variation measurement and prediction of surface texture parameters using machine learning
in laser powder bed fusion. J Manuf Sci Eng 2020;12:011008.
[286] Kwon O, Kim HG, Ham MJ, Kim W, Kim G-H, Cho J-H, et al. A deep neural network for classification of melt-pool images in metal additive manufacturing.
J Intell Manuf 2020;31:375–86.
[287] Kunkel MH, Gebhardt A, Mpofu K, Kallweit S. Quality assurance in metal powder bed fusion via deep-learning-based image classification. Rapid Prototyp J
2019;26:259–66.
[288] Meng L, McWilliams B, Jarosinski W, Park H-Y, Jung Y-G, Lee J, et al. Machine learning in additive manufacturing: a review. JOM 2020;72:2363–77.
[289] Qi X, Chen G, Li Y, Cheng X, Li C. Applying neural-network-based machine learning to additive manufacturing: current applications, challenges, and future
perspectives. Engineering 2019;5:721–9.
[290] Rudolph J-P, Emmelmann C. Self-learning calculation for selective laser melting. Procedia CIRP 2018;67:185–90.
[291] Renken V, Albinger S, Goch G, Neef A, Emmelmann C. Development of an adaptive, self-learning control concept for an additive manufacturing process. CIRP J
Manuf Sci Technol 2017;19:57–61.
[292] Baturynska I, Semeniuta O, Martinsen K. Optimization of process parameters for powder bed fusion additive manufacturing by combination of machine
learning and finite element method: a conceptual framework. Procedia CIRP 2018;67:227–32.
[293] Huang S, Narayan RL, Tan JHK, Sing SL, Yeong WY. Resolving the porosity-unmelted inclusion dilemma during in-situ alloying of Ti34Nb via laser powder bed
fusion. Acta Mater 2021;204:116522.

53