close

Вход

Забыли?

вход по аккаунту

?

j.matchar.2018.08.025

код для вставкиСкачать
Materials Characterization 145 (2018) 29–38
Contents lists available at ScienceDirect
Materials Characterization
journal homepage: www.elsevier.com/locate/matchar
Solidification pattern, microstructure and texture development in Laser
Powder Bed Fusion (LPBF) of Al10SiMg alloy
T
⁎
Hong Qina, Vahid Fallaha, , Qingshan Donga, Mathieu Brochub, Mark R. Daymonda,
Mark Gallerneaulta
a
b
Department of Mechanical and Materials Engineering, Queen's University, Kingston, ON K7L 3N6, Canada
Department of Materials, McGill University, Montreal, QC H3A 0C5, Canada
A R T I C LE I N FO
A B S T R A C T
Keywords:
Additive manufacture
Laser Powder Bed Fusion
AlSi alloys
Solidification microstructure
Crystallographic texture
Cellular structure
A comprehensive analysis of solidification patterns and microstructural development is presented for an
Al10SiMg sample produced by Laser Powder Bed Fusion (LPBF). Utilizing a novel scanning strategy that involves
counter-clockwise rotation of the scan vector by 67° upon completion of each layer, a relatively randomized
cusp-like pattern of protruding/overlapping scan tracks has been produced along the build direction. We show
that such a distribution of scan tracks, as well as enhancing densification during LPBF, reduces the overall
crystallographic texture in the sample, as opposed to those normally achieved by commonly-used bidirectional
or island-based scanning regimes with 90° rotation. It is shown that, under directional solidification conditions
present in LPBF, the grain structure is strictly columnar throughout the sample and that the grains' orientation
aligns well with that of the α-Al cells. The size evolution of cells and grains within the melt pools, however, is
shown to follow opposite patterns. The cells'/grains' size distribution and texture in the sample are explained via
use of analytical models of cellular solidification as well as the overall heat flow direction and local solidification
conditions in relation to the LPBF processing conditions. Such a knowledge of the mechanisms upon which
microstructural features evolve throughout a complex solidification process is critical for process optimization
and control of mechanical properties in LPBF.
1. Introduction
Laser Powder Bed Fusion (LPBF) is one of the most popular metal
Additive Manufacturing (AM) processes that is based on a layer-bylayer manufacturing consisting of re-melting/solidification of the
layered powder by a high energy input [1–4]. This technique has attracted increased interest in the production of high value, low production volume parts where it has strong advantages in creating complex geometries [1,2], particularly in applications where thermal and
cooling fluid management would be impossible to create using conventional casting and machining operations [3]. Al10SiMg is commonly
used in aerospace and transportation applications on account of its low
cost, excellent (specific density) mechanical properties and good
thermal conductivity [4]. Applying LPBF to these materials increases
the cooling rate during processing yielding a finer microstructure [4–7]
and thus an improvement in the as-produced (as-cast) mechanical
properties compared to the same alloy cast conventionally [7]. Cooling
rates as high as 106 Ks−1 have been found to exist during LPBF [5,8],
while the cooling rates attainable via conventional casting processes are
⁎
normally about 3–4 orders of magnitude lower, e.g. ~300 Ks−1 in strip
casting of Al alloys [9].
A complex thermal cycle develops within the LPBF part being built,
which is caused by a highly concentrated heat source (generated by a
focused laser beam) that rapidly moves along a finely resolved scanning
pattern in each layer [4,10]. Such a heat cycle may give rise to development of an unsteady heat flow, a condition that is further exacerbated in LPBF of metals with higher thermal conductivity and reflectivity, such as Al [11]. As a result, melt pools generated in LPBF of
Al alloys are known to fluctuate substantially in size throughout the
process [5,11]. Due to such uncertainty in melt-pool sizes as well as the
lower absorption of laser light due to high reflectivity, LPBF of Al alloys
is normally performed at relatively high laser powers, under which
condition the melt pools penetrate occasionally as deep as four layers
underneath the current track [11]. Also, the selected scanning regime
can further enhance such melt pool instabilities. Due to the repetitive
nature of commonly-used scanning regimes within successive layers
along the build direction, local heat concentrations normally occur as a
result of coinciding/overlapping scanning vectors and their start/end
Corresponding author.
E-mail address: vahid.fallah@queensu.ca (V. Fallah).
https://doi.org/10.1016/j.matchar.2018.08.025
Received 17 May 2018; Received in revised form 24 July 2018; Accepted 15 August 2018
Available online 18 August 2018
1044-5803/ © 2018 Elsevier Inc. All rights reserved.
Materials Characterization 145 (2018) 29–38
H. Qin et al.
in this study based on maximizing the interval number of printed layers,
N (as defined and explained in ref. [17]), with the same deposition
direction. It is shown by Guan et al. [17] that the larger N value the
lower the residual stresses and higher the densities within the build;
this results in improved mechanical properties. Recently, the hatch
angle of 67° has been adopted in the standard operating parameters of
some commercial SLM systems including the Renishaw AM400. Due to
such successive rotation in subsequent layers, as opposed to a simple
bidirectional regime, the local heat concentration will be reduced
across the part being built. Such local heat concentrations are rather
unsteady [13] and normally occur due to coinciding/overlapping
scanning vectors and their start/end points within successive layers
along the build direction. Such conditions give rise to the formation of
occasional key-hole melt pools during the LPBF processing thus promoting formation of microscopic solidification/shrinkage porosity [5].
The relative density was measured using Archimedes principle by
weighing the sample in air as well as in methanol. The measured
density, ~99%, competes with the highest Archimedes densities reported for LPBF of Al10SiMg alloy, such as those reported by Thijs et al.
[5].
Referring to the specimen coordinates shown in Fig. 1(b), X-Y (top
view), Y-Z (side view) and X-Z (front view) sections were cut from the
as-fabricated part and used for the microstructural examinations. The
samples were prepared using common metallographic techniques for
aluminum alloys. As a reference, one specimen was etched in Keller's
reagent while the other samples were ion-milled using a PECS II system.
Ion-milling was used rather than electropolishing in order to avoid the
possibility of silica contamination. Microstructural analysis of the asfabricated Al10SiMg sample cross-sections was performed on a FEGSEM (FEI Nova NanoSEM 400), using backscatter and secondary electron imaging. Grain sizes and crystallographic orientation information
were obtained using ECC imaging (under 10 keV and 7 mm working
distance) and EBSD analysis was carried out with a step size of 0.5 μm.
EBSD was carried out using a Bruker e-FlashHR detector and the acquired data was analyzed in an Oxford Instruments HKL Channel 5
software.
points [5,12]. Such heat concentrations, that are rather unsteady [13],
give rise to the formation of occasional key-hole melt pools during LPBF
processing thus enhancing porosity formation [5,12,14]. This is commonly observed in LPBF of Al10SiMg under highly repetitive scanning
strategies such as the so-called uni- or bi-directional regimes
[5,12,14–16]. To boost densification in LPBF, researchers have rather
utilized island-based scanning regimes where each layer is divided into
several islands. The scan vectors are then alternated by 90° in orientation amongst neighbouring islands in each layer, while also being
slightly shifted in location with respect to the corresponding islands
within the previous layer. Particularly, Thijs et al. [5] have shown that
utilization of such an island-based scanning strategy significantly enhances densification, as well as inducing an overall reduction in the
modes and intensities of crystallographic texture within the LPBF part.
To further explore the effect of scanning regimes on enhancing heat
flow stability and uniformity during LPBF, in this study, we utilize a
new strategy where the bidirectional scan vectors are rotated counterclockwise by 67° with respect to the previous layer.
Furthermore, thanks to the unique combination of analytical,
sample preparation and electron microscopy techniques utilized in this
study, we have been able to conduct a comprehensive and highly detailed analysis of cells/grains structure and texture development as well
as the evolution of various types of Si phase during LPBF. It is of a great
importance to understand the solidification patterns and the resultant
microstructural features in relation to the processing conditions, as such
features are the major contributors to the final strength in the LPBF
parts [5,15]. Obtaining such a knowledge can be a crucial factor in
tailoring the LPBF process design in order to effectively control the
microstructure, residual stresses and defects formation, and thus to
engineer the final physical and mechanical properties.
This study presents a fundamental approach to accurately resolve
the solidification and phase transformation patterns that lead to the
microstructural development during one of the more complex LPBF
processing conditions; i.e. an LPBF process that utilizes a more isotropic
scanning pattern throughout the part, presented here for the first time.
The focus of the present study is to obtain a solid knowledge of microstructure development in LPBF of Al10SiMg alloy, while the mechanical properties of printed parts will be investigated in a future
study. The test specimens, prepared using a variety of techniques from
chemical etching to ion-milling, are characterized using high-resolution
Scanning Electron Microscopy (SEM) for a detailed observation of cells/
grain structure development. Moreover, Electron Channelling Contrast
(ECC) imaging and Electron Backscattered Diffraction (EBSD) techniques are effectively used in conjunction with analytical investigation of
solidification process in order to link the microstructural development
to the LPBF processing conditions employed.
3. Results
Here, we present our microstructural analysis and observations of
the evolution of cellular/dendritic network and grain structure/texture
development across the LPBF part being built under the above described processing conditions.
3.1. Cellular/dendritic structure
SEM examination of a Keller's solution etched cross-sections along
the side and front view, Fig. 2(c–d), revealed a repetitive cusp-like
pattern that has been observed before in LPBF processed Al alloys
[15,18–20]. Unlike those developed in LPBF using a simple unidirectional scan strategy [5], the cusp-like pattern seems to be distributed
rather randomly across the side and front views. Such a randomized
pattern is the result of the 67° rotation in the scanning direction of each
subsequent layer and the deep melt-pools that penetrate into the previous layers with a variable depth. The melt-pools during LPBF of
aluminum alloys are known to be normally deeper than the layer
thickness [5], while also being unsteady and sometimes varying up to
four times the layer thickness [11]. This is in accordance with the top
view microstructure, shown in Fig. 2(a), where a variety of melt pool
orientations (angled at ~67°, as annotated) appear in the image, indicating the interception of melt pools from multiple scanned layers in a
single cross-section through the sample.
Fig. 3(a) shows the cellular microstructure within a re-solidified
melt-pool. Three different zones are distinguishable across the meltpool: fine, coarse and the heat affected zone, HAZ, as shown in a higher
magnification in Fig. 3(a1-a2-a3), respectively. It is noted that Keller's
2. Experimental Procedure
A rectilinear sample (100 mm × 20 mm × 10 mm) was printed
using Al10SiMg powder starting stock with a size range of 20–63 μm
and an Mg content of approximately 0.3 wt%. Fig. 1(a) shows a secondary SEM surface shot of the powder stock with a cross-sectional
view of the internal microstructure captured in the optical micrograph
shown in Fig. 1(b). The high magnification secondary SEM image,
shown in Fig. 1(c), and the corresponding inset (Fig. 1(ci)) reveal the
cellular microstructure and the intercellular Si network including a
lamellar eutectic structure as well as Si particles (as labeled in
Fig. 1(ci)).
The sample, shown in Fig. 1(d), was printed in an Argon atmosphere
using a Renishaw AM400 operated with standard processing parameters provided by the manufacturer. The build plate temperature was
35 °C and a rod-style support structure of 3 mm was utilized underneath
the sample. As demonstrated in Fig. 1(b), to enhance densification in
the part, the bidirectional scan vectors were rotated counter-clockwise
by 67° upon completion of each layer. The hatch angle of 67° is adopted
30
Materials Characterization 145 (2018) 29–38
H. Qin et al.
Fig. 1. (a) Secondary SEM surface shot of Al10SiMg powder starting stock, (b) optical micrograph and (c) high-magnification secondary SEM image of the crosssectional view of the internal microstructure with the corresponding inset shown in (ci); (d) the printed sample and schematic representation of scanning strategy;
The bi-directional scan vectors in Layer n + 1 are rotated by 67° counter clockwise with respect to those at Layer n.
solution aggressively etches the α-Al matrix and leaves behind a continuous network of Si-rich inter-cellular phase, similar to those revealed
by Thijs et al. [5]. This, however, limits the ability to resolve more
detail about the nature of such a Si-rich network and its building blocks,
including a eutectic α-Al/Si structure as well as Si particles. Therefore,
other means of sample preparation and examination were subsequently
employed.
Fig. 3(b) is from a similar region as in Fig. 3(a) but was prepared by
ion milling, which appears to be less preferential in its removal of the αAl matrix and overall less reactive to the surface, when compared to the
effect of the Keller's etch. At low magnification (Fig. 3(b)), the revealed
microstructure depicts a cellular/dendritic structure similar to that revealed by the Keller's etch (Fig. 3(a)). However, at higher magnification, the fine, coarse and HAZ regions (shown in Fig. 3(b1-b2-b3), respectively) exhibit more microstructural details than their Keller's
etched counterparts, thus allowing more precise description of the
inter-cellular microstructure within the melt pools (as will be described
below).
The morphology of the cellular/dendritic structure revealed, as can
be seen in Fig. 3(a–b), varies from directional to non-directional inside
and across the melt pools. These two regions are denoted as “D” (Directional) and “ND” (Non-directional) in Fig. 4(a), in which the same
side-view microstructure as in Fig. 3(a) is shown at a higher
magnification. Here, we explain that these two morphologies are, in
reality, different views of the same cellular structure. Solidification
during laser material treatment normally occurs under constraint conditions where the laser beam is scanned at high velocities with respect
to the heat diffusion rate, α/rL, where α is the thermal diffusivity (K/
ρcp), rL the laser beam radius, K the thermal conductivity, ρ the density
and cp the specific heat capacity [21,22]. Under such conditions, it is
shown that the solidification direction is controlled primarily by the
heat flux constraint and is usually along the normal to the solid/liquid
interface, i.e. the melt-pool boundary. In this case, the local growth
velocity, Vs, can be determined knowing the angle θ between the
normal to the melt-pool boundary and the laser scan direction, as demonstrated in Fig. 4(b) [23,24]. For a cellular/dendritic structure with
developed side branches, however, not only the heat flux constraint but
also the preferred crystallographic directions may affect the resultant
growth orientations [22,24]. The local growth velocity, Vs, is then related to the laser scan velocity, V, by:
Vs =
V cos θ
cos φ
(1)
where φ is the angle between the normal to the melt pool boundary and
the cell/dendrite axis. Due to the normally large imposed temperature
gradients during laser processing of metals and particularly in LPBF
Fig. 2. Secondary Electron SEM micrographs showing the microstructures revealed by Keller's etch for cross-sections at (a) top, (b) side and (c) front views.
31
Materials Characterization 145 (2018) 29–38
H. Qin et al.
Fig. 3. Secondary Electron SEM micrographs showing the side-view microstructures revealed by (a) Keller's etch and (b) ion milling. High magnifications of a1, a2, a3
and b1, b2, b3 areas referring to fine, coarse and HAZ zones in (a) and (b), respectively; the term “MP” in (b) denotes Melt-Pool.
(~106 Km−1 [8]), the cells/dendrites are often free of side branches
thus growing primarily along the normal to the melt-pool boundary, i.e.
φ = 0 [21,24]. Therefore, as Wu et al. [25] has also pointed out, the
cells appearing non-directional in Fig. 4(a) are simply a cross-section of
the directional cells within the microstructure which have grown relatively perpendicular to a melt pool boundary which is orientated on
an angle to the side view cross-section. The rotation of scanning direction by 67° in each subsequent layer explains the relatively random
location of melt pool boundaries with respect to the viewing crosssection. To further explain the variety of cells orientations in Fig. 4(a),
the front view microstructure (shown in Fig. 4(b)) is annotated to depict
three potential cell orientations with respect to a given side-view crosssection (the dotted line). Fig. 5(a & b) show more examples of regions
containing a mixture of cross-sectional and longitudinal views of cells/
dendrites, with magnified images shown in Fig. 5(c & d), respectively.
The cellular morphology revealed by deep Keller's etching, as shown in
Fig. 6, further reaffirms its viewing-angle dependency, i.e. the cells'
directional appearance gradually fades as the specimen is tilted from
−15° to +20°. Also, a 3D assembly of optical micrographs of microstructures revealed by Keller's etch is shown in Fig. 10(a) to help visualize the distribution of melt-pool boundaries as it relates to the
utilized scanning strategy.
At higher magnification, as shown in Fig. 5(c–d), the high clarity
ion-milled surfaces reveal a great deal of microstructural detail. It can
be seen that the intercellular regions are decorated with a discontinuous network of nano-sized Si-rich particles, as well as what
appears to be AleSi eutectic lamella which form predominantly at occasional “triple points” (denoted by an arrow on Fig. 5(d)). The regions
within the primary α-Al cells also contain a very fine network of Si-rich
particles, i.e. appearing as black dots across all cells in Fig. 5(c–d). It is
then suggested that, upon the last stage of solidification, the Si-rich
liquid trapped in-between the cells solidifies into an inter-cellular
mixture of AleSi eutectic and a Si-rich α-Al phase. The latter phase,
upon cooling from solidification temperatures, further undergoes a
precipitation of Si-rich particles. Similarly, the primary α-Al phase cells
that are solidified under high cooling rates are supersaturated with Si
and undergo a precipitation of fine, intra-cellular, Si-rich particles upon
post-solidification cooling. It is noteworthy that the Si-rich network
within the HAZ region (as shown in Fig. 3(b3)) has rather coarsened/
ripened due to the heat from an adjacent passing molten pool. It has
been shown that short periods of time at high temperature can lead to
Si-rich particle coarsening and a reduction in their number [26]. Under
these conditions, agglomeration and diffusion of Si is enhanced [27]. To
further confirm the variety of Si-rich phases identified above, their
distribution, crystallographic and morphological nature will be studied
in detail in an upcoming Transmission Electron Microscopy (TEM) investigation.
3.2. Grain Structure
In order to reveal and analyze the grain structure that contains the
primary α-Al cells, ECC imaging and EBSD analysis were performed.
This allows us to carry out an orientation comparison of the different
cellular regions in relation to the larger grain domains.
Smooth polished surfaces created by ion-milling allowed ECC imaging at high resolution, contrast and clarity, as can be seen in Fig. 7.
Particularly, the contrast from ECC was found to be sufficient to delineate both grain and cell boundaries, as demonstrated in the high
magnification images of the grains shown in Fig. 7(b & d). Fig. 7(a & c)
shows the grain structures revealed by ECC imaging of two different
32
Materials Characterization 145 (2018) 29–38
H. Qin et al.
3D printed component. Highly smooth ion-milled surfaces allowed
EBSD analysis to be performed with high indexing efficiency (> 95%)
thus resulting in the production of exceptionally high resolution maps.
Fig. 8(a–c) show the orientation maps obtained for top, side and frontview microstructures, respectively. The corresponding grain boundary
maps, shown in Fig. 8(d–f), were obtained for a minimum of 15° misorientation. The high-resolution grain boundary maps reveal both the
morphology of the Al grains and the melt-pool boundaries. Similarly to
the grain structures revealed in the ECC images, one can observe both
the cross-sectional and longitudinal views of the same columnar grains
within the melt-pools. The examples of these two different views of
columnar grains are labeled with dotted lines within the front-view
grain boundary map in Fig. 8(d). It can be seen that the grain structure
coarsens towards the center of the melt-pools. Also, the epitaxial
growth from the previous layers can be observed mainly in the regions
closer to the melt-pools centerline.
To analyze the texture development within the melt pools and
across the LPBF part, as demonstrated in Fig. 9, we obtained (100),
(110) and (111) pole figures from top-, side- and front-view microstructures as well as the inverse pole figures for various sample coordinates. As can be seen in the pole figures shown in Fig. 9, ⟨100⟩ is
the predominant texture in the sample while there are also weak ⟨110⟩
and ⟨111⟩ texture components present in either top, side or front view
microstructures. The inverse pole figures for all views demonstrate
strong intensities for ⟨100⟩ texture along the build direction and to a
lesser extent along the side- and front-view directions. Thus, the ⟨100⟩
texture has developed predominantly along the build direction. Weak
⟨110⟩ and ⟨111⟩ textures have developed primarily on a ~20–70°
angle with respect to the build direction, as can be seen from their
radial distribution in the top-view pole figures demonstrated in
Fig. 9(a). These textures are distributed rather uniformly with respect to
the build direction, which is due to the successive rotation of scanning
direction in each subsequent layer, as opposed to the four-fold anisotropy that normally develops in bidirectional or island-based scanning
strategies with 90° rotation (as demonstrated by Thijs et al. [5]). Thus,
the scanning strategy employed in this study further reduces the overall
texture in the LPBF parts by cutting back the anisotropy. A 3D assembly
of the orientation and grain boundary maps from various orthogonal
views is shown in Fig. 10(b–c) to assist with the visual perception of the
overall texture within the LPBF part.
Fig. 4. Secondary Electron SEM micrographs of (a) the side view and (b) the
front view microstructures; On the front view image in (b), the potential variations in the growth direction of cells are demonstrated along an imaginary
side-view cross-section (the dotted line), assuming a local scan vector, V. “D”
and “ND” refer to Directional and Non-directional, respectively.
regions within the side view cross-section. While the image in Fig. 7(c)
is predominantly along a longitudinal view of columnar grains, the
image in Fig. 7(a) shows a cross-sectional view of the same-type columnar grains. Under normal solidification conditions (the range of Vs
and G values) in LPBF and using the Hunt's criterion of columnar to
equiaxed transition (CET) [28], as depicted for Al10SiMg alloy by Chou
et al. [15], a fully columnar grain structure must develop throughout
the melt pools.
At higher magnification, looking at the cellular structure within
these grains, one can observe that the grain and cell trajectories are
largely parallel to each other; i.e., within the cross-sectional view of the
grains (Fig. 7(b)) one can see a cross-sectional view of the cells and,
similarly, for the longitudinal view of the grains (Fig. 7(d)), the cells
appear with the same columnar morphology and orientation. Very fine
grains can be observed only along the melt-pool boundaries where a
competitive growth mechanism has been active (as denoted by an
arrow in Fig. 7(a)). As can be seen in Fig. 7(c), the columnar grains have
grown predominantly perpendicular to the melt-pool boundaries and
towards the center of the melt-pool. Also, as indicated by arrows in
Fig. 7(c), a number of grains have continually grown through the meltpool boundary suggesting some degree of growth epitaxy.
Although the ECC imaging of ion-milled surfaces was able to resolve
(at varying resolution) the grain structure, EBSD analysis is required for
delineation of the grains' orientation and the overall crystallographic
and morphological texture within the melt-pools as well as across the
4. Discussion
Below, we investigate the mechanisms upon which the size and
morphology of the cellular network and grain structure develop during
LPBF processing, as well as the evolution of the crystallographic texture
in the part. Such mechanisms are discussed below as they relate to the
solidification conditions normally present during LPBF.
4.1. Cell Size Evolution
It was pointed out in our observations in the previous section that
the cells are coarser near the melt-pool boundaries and are increasingly
refined towards the center of the melt-pools, as can be seen in Fig. 3(b1)
vs. Fig. 3(b2). The cell size evolution within the melt-pools can be explained by means of analytical formula that incorporate the effect of
local solidification conditions, i.e. by accounting for variations of
growth velocity, Vs, and temperature gradient, G, along the solid-liquid
(S/L) interface (or the melt-pool boundary). A well-known model of
such kind is the geometrical model of cellular growth presented by
Hunt [29] which correlates the evolution of primary dendrite arm (or
cell) spacing, λ1, with the local solidification conditions along the advancing S/L interface as well as the alloy thermophysical properties:
λ1 = [2.83(k∆T0 DΓ )0.25] Vs−0.25G−0.5
(2)
The segment in the brackets represents the thermophysical
33
Materials Characterization 145 (2018) 29–38
H. Qin et al.
Fig. 5. Secondary Electron SEM images of microstructures (a–b) at two different areas within the side view cross-section revealed by ion-milling. Higher magnification images showing (c) the cross-sectional and (d) longitudinal view of the cellular/dendritic structure.
Fig. 6. Cellular morphology revealed by deep Keller's etching of the front-view microstructure; The same area is imaged by varying the specimen tilt from (a) −15° to
(b) 0° and then (c) +20°.
the S/L interface favors an increase in cell spacing towards the top/
center of the melt-pool, while the increasing Vs tends to refine it.
However, since normal melt-pool sizes in LPBF are about an order of
magnitude smaller than those in LMD, the variations in G within the
melt-pools are rather negligible and often reported in the literature as a
single value [5,8]. The cell spacing is thus primarily controlled by Vs
variations along the S/L interface, meaning a gradual refinement from
the melt-pool boundary towards the top/center. This is consistent with
the observations in this study, e.g. Fig. 3.
properties of the alloy, where k is the partition coefficient in the Al-rich
portion of AleSi phase diagram, ΔT0 the freezing range, D the liquid
diffusion coefficient of the solute atom (Si in this study) and Γ the
Gibbs–Thomson parameter. Furthermore, the above model has been
quantitatively verified by Fallah et al. [24] against the experimental
observations and phase-field simulations of solidification microstructure in Laser Metal Deposition (LMD) of TieNb alloys.
According to Eq. (2), the cell spacing is inversely related to both
growth velocity and temperature gradient along the S/L interface. As
explained earlier in Section 3.1 and according to Eq. (1), the local
growth velocity along the S/L interface for a cellular structure can be
approximated as a function of the laser scan velocity and the morphology (curvature) of the melt-pool boundary, i.e. Vs = V cos θ.
Therefore, considering the general morphology of the melt pools in
LPBF (as demonstrated in Fig. 11), the growth velocity varies from
nearly zero at the bottom of the melt-pools to a maximum of V (the laser
scan velocity) at the top of the melt-pools. Using finite element simulations of heat transfer during LMD, Fallah et al. [23] have shown that
the temperature gradient, G, is maximum at the bottom of the melt-pool
and decreases along the S/L interface towards the top (see the schematic representation shown in Fig. 11). Thus, the reduction of G along
4.2. Crystallographic Texture
The evolution of crystallographic texture is controlled by the following main factors: overall heat flow direction within the part being
built, growth direction and preferred growth orientation of the cells.
While the first two factors are primarily governed by the processing
conditions, the third important contribution is the alloy characteristics.
As discussed in Section 3.1, the growth direction along the S/L interface
is imposed by the local orientation of the melt-pool boundary which
itself is a function of the laser processing conditions [23]. The overall
heat flow direction is mainly dictated by the build direction, which is
34
Materials Characterization 145 (2018) 29–38
H. Qin et al.
Fig. 7. (a & c) ECC images of the side-view microstructure within the melt pools at two different regions; Higher magnification images of (b) the cross-sectional and
(d) the longitudinal view of the grains; “MPB” refers to Melt-Pool Boundary.
Fig. 8. (a–c) Orientation maps and (d–f) the corresponding grain boundary maps for top, side and front view microstructures, respectively; The grain boundaries
were identified with a minimum of 15° misorientation.
occurs specifically when the heat flow direction is aligned with the
growth direction of the cells. Such a condition is more likely met in
regions near the melt-pool centerline where the cells' growth direction
is more closely aligned with the overall heat flow direction. Under such
due to the relatively small melt-pool size compared to the part size as
well as the successive remelting of the previous layers during LPBF.
In cellular solidification of cubic materials, it is well-known that the
cells preferentially grow along the ⟨100⟩ crystal direction [30]. This
35
Materials Characterization 145 (2018) 29–38
H. Qin et al.
Fig. 9. Pole figures and inverse pole figures obtained from (a) top-, (b) side- and (c) front-view microstructures. The (100), (110) and (111) pole figures (left) and
inverse pole figures (right) along the build direction (Z,), side-view direction (Y) and front-view direction (X) are given. The relative intensity of the diffraction peaks
is indicated by the color scale.
Average Misorientation (KAM) maps shown in Fig. 12, the higher
average misorientation values around the melt-pool centerlines suggest
the presence of such a crystallographic texture. The exact microstructural nature of such high KAM values around the melt-pool centerline and their relationship with the crystallographic and morphological texture as well as with the grain size evolution within the part will
be the subject of a future study.
Weak ⟨110⟩ and ⟨111⟩ textures appear to be more randomized in
the rotating scanning strategy (employed in this study) than what has
been reported for the well-known bidirectional and island scan strategies with 90° rotation [5]; this refers to the circular distribution of such
orientation intensities in the top-view pole figures (Fig. 9(a)) as opposed to a four-fold anisotropy of intensity distribution reported by
Thijs et al. [5]. Thus, the 67° rotating scanning strategy that is chosen to
conditions, the growth near the melt-pool centerline is more likely to
occur epitaxially, and the absence of competitive growth gives rise to
the formation of larger columnar grains along the melt-pool centerline
and closer to its center (as can be seen in Fig. 7(c)). In addition, the
successive remelting of previous layers from various angles (due to the
successive rotation of scanning direction) promotes directionality of
heat flow and thus the alignment of columnar grains of ⟨100⟩ orientation along the build direction.
The conditions described above, together, give rise to the development of a moderately strong crystallographic texture along the build
direction; i.e., columnar grains of predominantly ⟨100⟩ orientation.
This is evident from the high intensities of (100) diffraction peaks in the
inverse pole figures obtained along the build direction (for all three
views), as can be seen in Fig. 9. Also, as demonstrated in the Kernel
Fig. 10. 3D representation of microstructure in the LPBF part compiled from three orthogonal views manifesting (a) the melt-pool boundaries distribution revealed
by Keller's Etch and optical microscopy, (b–c) the orientation and grain boundary maps, respectively, revealed by EBSD analysis of ion-milled surfaces.
36
Materials Characterization 145 (2018) 29–38
H. Qin et al.
protrude/overlap one another leave behind a relatively randomized
cusp-like pattern across the part being built. Under such conditions, we
show that, in addition to an enhanced densification, the overall crystallographic texture is reduced, as compared with the commonly used
bidirectional or island-based scanning regimes with 90° rotation.
Incorporating a unique combination of sample preparation and
electron microscopy techniques along with an in-depth data analysis,
the microstructural development is analyzed in detail as follows:
- A fully columnar grain structure has developed across the melt pools
possessing an orientation similar to that of the cells.
- It appears that nano-sized Si-rich phases have solidified in three
separate stages leading to formation of: (1) inter-cellular eutectic Si
lamella, (2) an inter-cellular Si-rich phase precipitated within the
inter-cellular α-Al phase upon cooling from solidification and (3) an
intra-cellular Si-rich phase precipitated within the α-Al primaries
upon cooling to room temperature.
- The cellular structure, developed under strictly directional solidification conditions, is increasingly refined as moving from the melt
pool boundary towards the melt pool centerline. Such refinement is
attributed to the dominant effect of growth velocity, Vs, in small
melt pools such as in LPBF, that changes from nearly zero to a value
close to the scanning speed near the melt pool center.
- The grain size evolution follows an opposite pattern to the cells and
increasingly coarsens moving away from the melt pool boundary
towards its centerline, where the overall heat flow direction (which
is along the build direction) is more closely aligned with the growth
direction (i.e. perpendicular to the melt pool boundary). Moreover,
under such conditions, the grains near the melt pool centerline tend
to grow epitaxially into large columnar grains creating a morphological texture along the build direction.
- A moderately strong ⟨100⟩ crystallographic texture has developed
along the build direction. This is due to the fact that, in regions near
the melt pool centerline, the overall heat flow direction is closely
aligned with the growth direction as well as with the preferred
growth orientation in Al, i.e. ⟨100⟩.
- Weak ⟨110⟩ and ⟨111⟩ textures are present on a ~20–70° angle to
the build direction while being distributed radially and fairly uniformly, as opposed to a four-fold anisotropic distribution that normally develops under commonly used bidirectional and island-based
scanning regimes with 90° rotation. The scanning strategy utilized in
this study further reduces the overall crystallographic texture in the
Fig. 11. Schematic representation of variation of Vs and G along the melt-pool
boundary. Tm denotes the alloy melting temperature; The isotherm Tm represents the estimated locus of the melt-pool boundary; The temperature gradient along the melt-pool boundary is then defined as G = ΔT/dx, estimated by
the locus of the two isotherms shown above, i.e. Tm and Tm + ΔT.
enhance part densification, also reduces the strength of the overall
texture and promotes a more isotropic LPBF part beyond what could be
achieved by 90° island-based scanning strategies.
4.3. Morphological Texture and Grain Size Evolution
Larger grains form towards the melt-pool center (Fig. 8(e)) which is
attributed to the successive remelting of previous layers as well as the
alignment of heat removal direction with the growth direction [31,32].
Particularly, the latter enhances the alignment of cells growing near the
melt-pool center thus minimizing the competitive growth and grain
boundary formation. In addition to the above described ⟨100⟩ crystallographic texture, this also creates a morphological texture featuring
the large columnar grains which extend mainly along the build direction.
5. Conclusions
The detailed microstructural characterization and analysis presented in this study allows for an in-depth understanding of the unique
cellular and grain structure development under a novel scanning
strategy that aims at enhancing densification during LPBF. Under a
scanning regime that incorporates a counter-clockwise rotation of the
scan vector by 67° upon each successive layer, the scan tracks that
Fig. 12. Kernel Average Misorientation (KAM) maps obtained from (a) top-view, (b) side-view and (c) front-view microstructures. The high-angle grain boundaries
are imposed on the KAM maps.
37
Materials Characterization 145 (2018) 29–38
H. Qin et al.
LPBF part.
[15]
Acknowledgements
This research was supported by the NSERC Discovery Grant program and the Queen's University Faculty of Engineering and Applied
Science Dean's Research Fund.
[16]
References
[18]
[17]
[19]
[1] I. Yadroitsev, I. Smurov, Selective laser melting technology: from the single laser
melted track stability to 3D parts of complex shape, Phys. Procedia 5 ( (2010)
551–560.
[2] H. Schleifenbaum, W. Meiners, K. Wissenbach, C. Hinke, Individualized production
by means of high power Selective Laser Melting, CIRP J. Manuf. Sci. Technol. 2 (3)
(2010) 161–169.
[3] M. Gebler, A.J.M.S. Uiterkamp, C. Visser, A global sustainability perspective on 3D
printing technologies, Energy Policy 74 (2014) 158–167.
[4] D. Herzog, V. Seyda, E. Wycisk, C. Emmelmann, Additive manufacturing of metals,
Acta Mater. 117 (2016) 371–392.
[5] L. Thijs, K. Kempen, J.P. Kruth, J.V. Humbeeck, Fine-structured aluminium products with controllable texture by selective laser melting of pre-alloyed AlSi10Mg
powder, Acta Mater. 61 (5) (2013) 1809–1819.
[6] E. Brandl, U. Heckenberger, V. Holzinger, D. Buchbinder, Additive manufactured
AlSi10Mg samples using Selective Laser Melting (SLM): microstructure, high cycle
fatigue, and fracture behavior, Mater. Des. 34 (2012) 159–169.
[7] N. Read, W. Wang, K. Essa, M.M. Attallah, Selective laser melting of AlSi10Mg
alloy: process optimisation and mechanical properties development, Mater. Des.
(1980–2015) 65 (2015) 417–424.
[8] M. Tang, P.C. Pistorius, S. Narra, J.L. Beuth, Rapid solidification: selective laser
melting of AlSi10Mg, JOM 68 (3) (2016) 960–966.
[9] V. Fallah, D.J. Lloyd, M. Gallerneault, Processing and characterization of continuous-cast AlMgSc (Zr) sheets for improved strength, Mater. Sci. Eng. A 698
(2017) 88–97.
[10] A.B. Spierings, K. Dawson, T. Heeling, P.J. Uggowitzer, R. Schäublin, F. Palm,
K. Wegener, Microstructural features of Sc-and Zr-modified Al-Mg alloys processed
by selective laser melting, Mater. Des. 115 (2017) 52–63.
[11] A.B. Spierings, K. Dawson, M. Voegtlin, F. Palm, P.J. Uggowitzer, Microstructure
and mechanical properties of as-processed scandium-modified aluminium using
selective laser melting, CIRP Ann. 65 (1) (2016) 213–216.
[12] K. Kempen, L. Thijs, J.V. Humbeeck, J.P. Kruth, Processing AlSi10Mg by selective
laser melting: parameter optimisation and material characterisation, Mater. Sci.
Technol. 31 (8) (2015) 917–923.
[13] J.C. Ion, Laser Processing of Engineering Materials: Principles, Procedure and
Industrial Application, Butterworth-Heinemann, Oxford, 2005.
[14] M. Cloots, P.J. Uggowitzer, K. Wegener, Investigations on the microstructure and
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
38
crack formation of IN738LC samples processed by selective laser melting using
Gaussian and doughnut profiles, Mater. Des. 89 (2016) 770–784.
R. Chou, A. Ghosh, S.C. Chou, M. Paliwal, M. Brochu, Microstructure and mechanical properties of Al10SiMg fabricated by pulsed laser powder bed fusion,
Mater. Sci. Eng. A 689 (2017) 53–62.
N.T. Aboulkhair, N.M. Everitt, I. Ashcroft, C. Tuck, Reducing porosity in AlSi10Mg
parts processed by selective laser melting, Addit. Manuf. 1 (2014) 77–86.
K. Guan, W. Zemin, G. Ming, L. Xiangyou, Z. Xiaoyan, Effects of processing parameters on tensile properties of selective laser melted 304 stainless steel, Mater. Des.
50 (2013) 581–586.
T.B. Sercombe, X. Li, Selective laser melting of aluminium and aluminium metal
matrix composites, Mater. Technol. 31 (2) (2016) 77–85.
K.G. Prashanth, S. Scudino, H.J. Klauss, Kumar Babu Surreddi, L. Löber, Z. Wang,
A.K. Chaubey, U. Kühn, J. Eckert, Microstructure and mechanical properties of
Al–12Si produced by selective laser melting: effect of heat treatment, Mater. Sci.
Eng. A 590 (2014) 153–160.
C. Weingarten, D. Buchbinder, N. Pirch, W. Meiners, K. Wissenbach, R. Poprawe,
Formation and reduction of hydrogen porosity during selective laser melting of
AlSi10Mg, J. Mater. Process. Technol. 221 (2015) 112–120.
W. Kurz, R. Trivedi, Microstructure and phase selection in laser treatment of materials, J. Eng. Mater. Technol. 114 (4) (1992) 450–458.
M. Zimmermann, M. Carrard, W. Kurz, Rapid solidification of Al-Cu eutectic alloy
by laser remelting, Acta Metall. 37 (12) (1989) 3305–3313.
V. Fallah, M. Alimardani, S.F. Corbin, A. Khajepour, Temporal development of meltpool morphology and clad geometry in laser powder deposition, Comput. Mater.
Sci. 50 (7) (2011) 2124–2134.
V. Fallah, M. Amoorezaei, N. Provatas, S.F. Corbin, A. Khajepour, Phase-field simulation of solidification morphology in laser powder deposition of Ti–Nb alloys,
Acta Mater. 60 (4) (2012) 1633–1646.
J. Wu, X.Q. Wang, W. Wang, M.M. Attallah, M.H. Loretto, Microstructure and
strength of selectively laser melted AlSi10Mg, Acta Mater. 117 (2016) 311–320.
C. Yan, L. Hao, A. Hussein, P. Young, J. Huang, W. Zhu, Microstructure and mechanical properties of aluminium alloy cellular lattice structures manufactured by
direct metal laser sintering, Mater. Sci. Eng. A 628 (2015) 238–246.
Y. Birol, Microstructural evolution during annealing of a rapidly solidified Al–12Si
alloy, J. Alloys Compd. 439 (1–2) (2007) 81–86.
J.D. Hunt, Steady state columnar and equiaxed growth of dendrites and eutectic,
Mater. Sci. Eng. 65 (1) (1984) 75–83.
J.D. Hunt, Cellular and primary dendrite spacings, Solidification and Casting of
Metals, the Metals Society, London, 1979, pp. 3–9.
M.E. Glicksman, Principles of Solidification, Springer, New York, 2011.
X.P. Li, X.J. Wang, M. Saunders, A. Suvorova, L.C. Zhang, Y.J. Liu, M.H. Fang,
Z.H. Huang, T.B. Sercombe, A selective laser melting and solution heat treatment
refined Al–12Si alloy with a controllable ultrafine eutectic microstructure and 25%
tensile ductility, Acta Mater. 95 (2015) 74–82.
M.M. Kirka, P. Nandwana, Y. Lee, R.R. Dehoff, Solidification and solid-state
transformation sciences in metals additive manufacturing, Scr. Mater. 135 (2017)
130–134.
Документ
Категория
Без категории
Просмотров
0
Размер файла
9 467 Кб
Теги
matchar, 2018, 025
1/--страниц
Пожаловаться на содержимое документа