Acta Materialia 103 (2016) 128–140

Contents lists available at ScienceDirect

Acta Materialia
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a c t a m a t

Full length article

Toughening of aluminum matrix nanocomposites via spatial arrays of

boron carbide spherical nanoparticles
Lin Jiang a, Hanry Yang a, Joshua K. Yee a, Xuan Mo a, Troy Topping a,b, Enrique J. Lavernia a,
Julie M. Schoenung a,*
a Department of Chemical Engineering and Materials Science, University of California, Davis, CA 95616, USA
b Department of Mechanical Engineering, California State University, Sacramento, Sacramento, CA 95819, USA

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

Article history: To enhance the toughness of metal matrix nanocomposites, we demonstrate a strategy that involves the
Received 3 July 2015 introduction of spatial arrays of nanoparticles. Specifically, we describe an approach to synthesize a mi-
Received in revised form 28 September crostructure characterized by arrays of fiber-like nanoparticle-rich (NPR) zones that contain spherical
2015
nanoparticles of boron carbide (sn-B4C) embedded in an ultrafine grained (UFG) aluminum alloy matrix.
Accepted 29 September 2015
Available online 6 November 2015
A combination of cryomilling and hot-extrusion was used to obtain this particular microstructure, and
the mechanical behavior and operative strengthening and deformation mechanisms were investigated
in detail. When compared to an equivalent unreinforced material, the presence of the array of NPR zones
Keywords:
Metal matrix nanocomposites contributed to a 26% increase in tensile strength. Moreover, when compared to a nanocomposite con-
Microstructural toughening taining a homogeneous distribution of nanoparticles, a 30% increase in toughness was observed. High
Size effects nanohardness values obtained for the NPR zones and the observation of “pull-out” phenomena on frac-
Aluminum ture surfaces, suggest that the NPR zones behave as “hard” fiber-like units that can effectively sustain
Nanoparticles tensile loading and thereby enhance the strengthening efficiency of sn-B4C. Also, the presence of the array
of NPR zones surrounded by nanoparticle-free (NPF) zones led to an enhancement in strength with limited
loss in ductility. This behavior was rationalized on the basis of a low value of the Schmid factor in regions
adjacent to NPR zones, coupled with the ease of dislocation movement in NPF zones. Finally, the ratio
of the plastic zone size to the size of the “hard” NPR zones is proposed as an important factor that governs
the overall toughness of the nanocomposite.
© 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction surprisingly, there is an extensive body of literature describing efforts

aimed at developing strategies to attain homogenous distribu-
Nanocomposites represent a recent approach to engineer ma- tions of nanoreinforcements (e.g., carbon nanotubes (CNTs), B4C
terials with heretofore unattainable combinations of physical and nanoparticles) in metallic matrices [6–8]. Interestingly, the idea of
mechanical properties to meet the ever increasing performance controlling the distribution of reinforcing particles at the me-
demands of advanced technologies [1]. The unique properties of soscale to engineer unique microstructural architectures has emerged
novel reinforcing phases, such as carbon nanofibers and spherical as an innovative approach to surmount current property trade-
nanoparticles, were originally thought to provide an opportunity offs [9,10,11]. Some notable examples of these architectures include:
to surmount the well documented tradeoff between strength and tri-modal [12], ring-structure [13], multi-filament [14], and layered
ductility that has plagued traditional metal matrix composites [1,2]. structures [10,11,15], which represent a drastic departure from the
However, despite encouraging early results, progress has been dis- original idea that a homogenous distribution of reinforcing par-
appointing, partly because decreasing the size of the reinforcement ticles is a prerequisite to optimal mechanical behavior [6,16,17]. An
into the nanoscale introduces other challenges, such as particle clus- example involving the development of a fiber-like geometry by con-
tering, which can degrade fatigue and fracture response [2–5]. Not trolling the stacking of CNTs during consolidation is provided by the
work of Jiang et al. [10,11]. In these studies, two-dimensional flakes
with adsorbed CNTs were used as building blocks instead of spher-
* Corresponding author. ical powders to attain a two-dimensional (2-D) distribution of CNTs
E-mail address: [email protected] (J.M. Schoenung). [10,11,18–20]. The documented increase in tensile strength and

http://dx.doi.org/10.1016/j.actamat.2015.09.057
1359-6454/© 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
 L. Jiang et al./Acta Materialia 103 (2016) 128–140 129

plasticity was ascribed to toughening by the layered arrangement A and 50 vol.% of Powder C were mixed in a V-blender to obtain
[10]. An additional example is provided by the work of Peng et al. Powder D. Powder D has the same sn-B4C content (2.5 vol.%) as
[9], in which Al2O3 fiber agglomerates, arranged as spheres, were Powder B.
randomly distributed in a fiber-free Al matrix. Their results showed Two types of the nanocomposites were prepared using the same
enhanced damage tolerance when compared to the same material consolidation processes but different starting cryomilled powders,
but with randomly distributed fibers [9]. Moreover, the hybrid com- as shown schematically in Fig. 1. Powder A was consolidated as a
posite materials composed of granules of cemented carbide [21–24] cryomilled Al for comparison as a baseline with the nanocomposites.
have also been highlighted as models of how the distribution of par- Powder B was consolidated into bulk nanocomposites with homo-
ticulate reinforcements can affect mechanical behavior. This hybrid geneously distributed sn-B4C (Nanocomposite I). Powder D was
particulate material can be described as a “composite within a com- consolidated into bulk nanocomposites with fiber-like sn-B4C rich
posite” since the granules, containing ceramic particles within a zones (Nanocomposite II). Powders A, B and D were each consoli-
metal binder, are bonded within a metal matrix [21,22]. Note that dated using the following experimental approach. First, the powders
the hybrid composites exhibited a superior combination of frac- were encapsulated in AA 6061 cans, and then hot vacuum
ture toughness and high-stress wear resistance than conventional degassed at 500 °C for 20 h to remove the stearic acid and any phys-
cemented carbides [21,22,24]. In addition to experimental studies, ically adsorbed gases, such as moisture, nitrogen and hydrogen [28].
there are also numerical simulations that suggest that it is possi- The degassed powder was then consolidated by hot isostatic press-
ble to obtain strengthening with an inhomogeneous particle ing (HIP) at 400 °C and 124 MPa, followed by extrusion at 400 °C
distribution [25]. with an extrusion ratio of 10:1.
Interestingly, review of the published literature shows that most
studies involving the influence of non-homogenous dispersions of 2.2. Characterization
reinforcing particles on mechanical behavior involve micron sized
particles, and hence information on the influence of non- The distribution of the sn-B4C in the Al matrix, as well as the
homogenous nanoparticle distributions is almost non-existent sn-B4C powder morphology, was studied using a FEI 430 nano-
[26–28]. Accordingly, the goal of the present work is to provide scanning electron microscope (NanoSEM) equipped with a back-
insight into the following two questions. First, is it feasible to develop scatter detector. Sample cross-sections were polished using a JEOL
a strategy to obtain spatially engineered arrays of nanoparticles in SM-09010 cross-section ion polisher. Ion polishing was used in lieu
a metallic matrix? Second, what is the influence of such nanoparticle of traditional mechanical polishing to preserve any subsurface mi-
arrays on deformation and fracture mechanisms? To address these crostructure features. Statistics on frequency and cumulative
questions, we used a cryomilling and consolidation approach to fab- percentage of sn-B4C interparticle distance in the nanocomposites
ricate a microstructure that contains fiber-like arrays of spherical were measured by ImageJ software (version 1.49) based on SEM
B4C nanoparticles (sn-B4C) embedded in an ultrafine grained (UFG) images.
5083 Al matrix, such that the nanocomposite contains alternating Specimens for transmission electron microscopy (TEM) and
arrays of “hard” fiber-like 5083 Al zones containing a high density scanning TEM (STEM) were mechanically ground and dimpled
of sn-B4C nanoparticles and a “soft” UFG 5083Al matrix. For com- to a thickness of ∼20 μm, and then further thinned to electron
parison, a 5083 Al matrix nanocomposite with a homogeneous transparency using a Gatan PIPS 691 ion-milling system cooled by
distribution of the sn-B4C was also fabricated. 5083 Al was se- liquid nitrogen. High resolution TEM (HRTEM) specimens were
lected as a matrix based on the availability of extensive published prepared using a FEI Scios 3-D dual beam (an electron beam and a
data on this system, and sn-B4C was selected as reinforcement due focused ion beam (FIB)) instrument operated at 2.0 kV and 23 pA
to its extremely high hardness of ∼30 GPa at room temperature for the ion beam at the final thinning stage. Grain size statistics
coupled with low density (2.5 g cm−3) and remarkable chemical in- and interface structure were determined from STEM/TEM/HRTEM
ertness [12,26]. micrographs acquired on a JEOL JEM-2500SE microscope oper-
ated at 200 kV. The Schmid factor and strain contouring of the
specimen sectioned from the center of the extruded sample along
2. Experimental the extrusion direction were recorded by orientation imaging
microscopy (OIM) using the electron backscatter diffraction (EBSD)
2.1. Preparation of nanocomposites technique integrated into the FEI Scios 3-D dual beam microscope
(controlled using TSL software and analyzed by Channel
The synthesis approach used to fabricate the nanocomposites in 5 software).
this study is schematically shown in Fig. 1. The starting powder was Tensile tests were carried out along the extrusion direction using
gas atomized Al–4.5Mg–0.57Mn–0.24Fe (5083 Al), produced by an Instron 8801 universal testing machine using dog-bone-shaped
Valimet Inc. (Stockton, CA). The spherical nanoparticles of B4C (sn- specimens with a gauge length of 12 mm and a diameter of 3 mm.
B4C) were fabricated and supplied by the Hefei Kaier Nanometer Compression tests were conducted using cylindrical specimens 8 mm
Energy & Technology Co., LTD (Hefei, China), and exhibit an average in height and 5 mm in diameter, conforming to ASTM E8M stan-
cross-sectional nanoparticle size of 50 nm. As shown in Fig. 1, the dards for subsize specimens. All of the mechanical testing was
as-received sn-B4C powder exhibited a narrow size distribution and performed at a strain rate of 10−3 s−1, and the strain was measured
perfect crystal structure. Cryomilling of 5083 Al powder, sn-B4C using a video extensometer with a resolution of 5 μm. Each mate-
powder and 0.2 wt.% stearic acid (CH3(CH2)16CO2H) was carried out rial was tested five times to obtain average strength, average strain
for 12 h in liquid nitrogen using a modified 1S Szegvari attritor. A and standard deviation values for both tensile and compressive
rotational speed of 180 RPM and a ball-to-powder ratio of 30:1 were behavior.
used. A complete description of the cryomilling process and the ra-
tionale for selecting these parameters can be found in a prior study 3. Results
[27]. As shown in Fig. 1, three distinct powders were prepared by
cryomilling for 12 h: (a) 5083 Al powder (designated hereafter as 3.1. Microstructure of the as-designed nanocomposites
Powder A); (b) 2.5 vol.% sn-B4C/5083 Al composite powder (desig-
nated hereafter as Powder B); (c) 5 vol.% sn-B4C/5083 Al composite SEM and TEM microstructures of the nanocomposites with
powder (designated hereafter as Powder C). Fifty (50) vol.% of Powder various reinforcement geometrical arrays are provided in Fig. 2. In
130 L. Jiang et al./Acta Materialia 103 (2016) 128–140

Fig. 1. Schematic of the preparation of the nanocomposites with different structures. Optical microscopy (OM) was used to characterize the morphology of the initial 5083
Al powder. SEM was used to characterize Powders A, B, and C. TEM was use to characterize the spherical B4C nanoparticle (sn-B4C) powder, and HRTEM was used to char-
acterize the lattice of the sn-B4C powder. NPR: nanoparticle-rich; NPF; nanoparticle-free.

Nanocomposite I, the sn-B4C are homogeneously distributed through- NPR zones. Also, some of the sn-B4C nanoparticles in the NPR zones
out the matrix material. This random distribution originates from were located within UFG interiors (Fig. 3b) as a result of the
cryomilled Powder B and indicates the efficiency of cryomilling cryomilling and hot-consolidation, which is consistent with our pre-
for dispersing nanoparticles in a metal matrix. In Nanocomposite vious study [29]. The HRTEM of the sn-B4C/Al interface (Fig. 3c,d),
II (Fig. 2g–i), the sn-B4C concentration in the entire cross section suggests that the interfaces of sn-B4C/Al were free of obvious defects.
is the same as in the homogeneous structure (2.5 vol.% sn-B4C), For the interface between the intragranular sn-B4C and Al, the (1-
but the fiber-like sn-B4C rich zones are locally enriched in concen- 11)Al planes are shown to be parallel with (024) B4C planes (Fig. 3c).
tration (5 vol.% sn-B4C). The sn-B4C rich zones (dark grey regions In contrast, there was no crystallographic orientation relationship
in Fig. 2d and g) originate from the cryomilled Powder C, and the between the intergranular sn-B4C and Al matrix. More details on
sn-B4C free zones (light grey regions in Fig. 2d and g) originate these interfacial orientations can be found in our previous re-
from the cryomilled Powder A. In contrast, a spherical-like array search [29].
of sn-B4C rich zones results from the consolidation by HIPing of
Powder D (50 vol.% Powder A and 50 vol.% Powder C), as shown 3.2. Mechanical behavior
schematically in Fig. 1. As we can see in Fig. 2d–f, the average
diameter of the nanoparticle rich (NPR) zones in this material is Representative tensile behavior for Nanocomposites I and II is
approximately 30 μm. After extrusion (i.e., in Nanocomposite II), shown in Fig. 4a. The corresponding numerical results are pre-
these NPR zones become elongated resulting in fiber-like struc- sented in Table 1; the small standard deviation values highlight the
tures with alternating regions of nanoparticle free (NPF) and NPR reproducibility of the results. Nanocomposite II exhibited a tensile
zones, as seen in Fig. 2g–i. The spacing between NPR zones varies yield strength (YS) of 761 ± 8.4 MPa, and elongation of 2%. Com-
within the range of 2–10 μm. Note that the sn-B4C are still well pared to the unreinforced UFG 5083 Al consolidated using an
dispersed within the NPR zones and that most of the sn-B4C in identical process (see Fig. 4a), the fiber-like structure contributes
the zones are individual particles, as seen in Fig. 2f and i. This to a 36.3% increase in tensile YS. More importantly, compared to
represents a key difference between a tailored distribution and the properties of Nanocomposite I, tensile YS: 699 ± 9.0 MPa and
clustering/agglomeration commonly seen in nanoparticle rein- elongation: 1.5%, the fiber-like structure still resulted in enhanced
forced composites. tensile properties.
Fig. 3 shows the STEM/TEM microstructure of Nanocomposite The compressive behavior of Nanocomposites I and II is shown
II. As seen in Fig. 3a, the sn-B4C were uniformly distributed in the in Fig. 4b, with numerical values provided in Table 1. The
 L. Jiang et al./Acta Materialia 103 (2016) 128–140 131

nanocomposite with spherical-like arrays of the sn-B4C (Fig. 2d–f) arrays were alternately performed on the NPR and NPF zones, large
was not tested for mechanical behavior because it was not fully dense fluctuations in hardness and modulus were found in Nanocomposite
after HIP. The compressive properties reveal a trend that is consis- II. In Nanocomposite I, however, the fluctuations in hardness and
tent with that of the tensile properties. The compressive yield modulus were much smaller, as shown in Fig. 5c,d. When aver-
strength of Nanocomposite II is approximately 60 MPa higher than aged, it is observed that these two structures exhibit almost the same
that corresponding to Nanocomposite I. Moreover, both the tensile average elastic modulus (Fig. 5c), but the average hardness of
and compressive strain of Nanocomposite II is higher than that of Nanocomposite II is approximately 23.7% higher than that of
Nanocomposite I. Additional insight into the mechanical behavior Nanocomposite I, as shown by the dashed lines in Fig. 5d. This ob-
can be gained by considering the toughness (energy absorbed per servation suggests that the geometrical arrangement of the
volume of material), which can be estimated from numerical in- reinforcing particles influences the overall mechanical response of
tegration of the data in a measured stress–strain experiment [30]. the nanocomposites.
Accordingly, on the basis of the integrated area under the entire
tensile stress–strain curve to the fracture point, the as-calculated 3.3. Deformation characteristics
toughness of Nanocomposite I is 7.54 ± 0.18 J m−3, whereas the as-
calculated toughness of Nanocomposite II corresponds to The tensile fracture surfaces of Nanocomposite II, shown in Fig. 6,
9.75 ± 0.23 J m−3; the toughness of the fiber-like structure is thus reveal several interesting features. First, the fracture surface is highly
about 1.3 times that of the homogeneous structure. Note that both irregular, showing evidence of pull-out (Fig. 6a and c). This is in con-
Nanocomposite I and Nanocomposite II were consolidated by the trast to when the reinforcing particles are homogeneously
same process and have the same content of the sn-B4C and equiv- distributed, in which case the fracture surface is relatively smooth
alent grain sizes (see Table 1). (Fig. 6b). Second, examination of cross sections from pull-out zones
Fig. 5 shows the nanoindentation properties for the suggests that these regions contain a high concentration of sn-B4C
nanocomposite samples, and the results are summarized as follows. (black dots in Fig. 6e,f). Third, cracks were observed at interfaces
The values of the hardness and modulus of Nanocomposite II, and between NPR and NPF zones, as indicated by Fig. 6e,f.
the hardness and modulus of the NPR zones (dark-grey regions in TEM studies (Fig. 7) of NPR and NPF zones revealed that the dis-
Fig. 5b) are higher than those corresponding to the NPF zones (light- location density increases with increasing concentration of
grey regions in Fig. 5b). Thus, when the periodic nanoindentation nanoparticles. As seen in the higher magnification image of an NPR

Fig. 2. Representative SEM images of: (a–c) nanocomposite with randomly distributed sn-B4C (Nanocomposite I), (d–f) nanocomposite with spherical-like arrays of the
sn-B4C, and (g–i) nanocomposite with fiber-like arrays of the sn-B4C (Nanocomposite II). Note that the sn-B4C are uniformly distributed in the NPR zones.
132 L. Jiang et al./Acta Materialia 103 (2016) 128–140

Fig. 3. Representative STEM/TEM images of Nanocomposite II: (a) STEM and (b) TEM to show intragranular or intergranular sn-B4C as indicated by arrows; (c) HRTEM shows the
interface orientation between an intragranular sn-B4C and the Al matrix; (d) HRTEM shows there is no interface orientation between an intergranular sn-B4C and the Al matrix.

Fig. 4. Representative (a) tensile and (b) compressive behavior of Nanocomposites I and II.

Table 1
Summary of experimental compressive/tensile properties and calculated yield strength values for the materials.

Sample/Structure B4C (Vol.%) Grain size (nm) Calculated values Experimental values

ΔσH−P Δσdis ΔσL-T σc Compression Tension

Yield strength Strain Yield strength Strain Toughness

Nanocomposite I 2.5 189 ± 29 345 MPa 211 MPa 8.2 MPa 695 MPa 705 ± 9.8 MPa 17% 699 ± 9.0 MPa 1.5% 7.54 ± 0.18 J m−3
Nanocomposite II 2.5 175 ± 38 359 MPa 211 MPa 8.2 MPa 709 MPa 765 ± 9.2 MPa 20% 761 ± 8.4 MPa 2% 9.75 ± 0.23 J m−3
NPR zone 5 129 ± 24 417 MPa 266 MPa 19.5 MPa 833 MPa – – – – –
NPF zone 0 221 ± 43 319 MPa 123 MPa 0 MPa 572 MPa 547 ± 9.3 MPa 30% 558 ± 8.9 MPa 2.6% 14.28 ± 0.31 J m−3

ΔσH−P: Yield strength contributed from the grain refinement; Δσdis: Yield strength contributed from the dislocation strengthening;
ΔσL-T: Yield strength contributed from the load bearing of the sn-B4C; σc: Yield strength of the Nanocomposite;
NPR: nanoparticle-rich; NPF: nanoparticle-free (5083 Al matrix). Standard deviation values were calculated on the basis of five sets of data.
 L. Jiang et al./Acta Materialia 103 (2016) 128–140 133

Fig. 5. Nanoindentation behavior of 2.5 vol.% sn-B4C/Al nanocomposites: corresponding optical microscopy of: (a) Nanocomposite I and (b) Nanocomposite II after nanoindentation;
(c) elastic modulus variation in the nanocomposites; (d) hardness variation in the nanocomposites.

Fig. 6. Representative fracture surface of Nanocomposite II: (a, c) red circles indicate the pull-out fiber-like NPR zones, (d–f) slice views of a pull-out fiber-like NPR zone, in-
dicating cracks were generated at the interface of the NPR and NPF zones. (b) Fracture surface of Nanocomposite I. (g) Schematic of load transfer from the NPF to the NPR
zones and forming pull-out zones that are sn-B4C rich. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

zone (Fig. 7b), the grains near a sn-B4C contain a high density of figure, grain color (in the web version) corresponds to different
dislocations. As seen in a typical HRTEM image of the grains in the Schmid factor values. As seen in Fig. 8c, the NPR and NPF zones are
NPR zones (Fig. 7c), the local edge dislocation density is on the order characterized by different Schmid factor distributions. The grains
of 3.3 × 10−16 m−2, whereas in the NPF zones, the grains typically exhibit in the NPF zones are mostly blue (in the web version), which cor-
a much lower dislocation density (about 7.3 × 10−15 m−2), as shown relates to a higher average Schmid factor; whereas some grains in
in Fig. 7d. Additionally, a high concentration of dislocations is also the NPR zones show different colors, as shown in Fig. 8c. It follows
observed in grains located at NPF/NPR interfaces after the sample that grains in the NPF zones have higher Schmid factor values com-
was deformed to approximately 2% strain, as seen in Fig. 7e,f. pared to grains in the NPR zones, when the applied loading (P) is
To provide information on the deformation characteristics of the along the extrusion direction. Fig. 9 shows the strain contour map
grains in the NPR/NPF zones, the distribution of Schmid factors for Nanocomposite II reconstructed using the Channel 5 software.
(Fig. 8) and strain contours (Fig. 9) in the fiber-like structure were In this figure, red (in the web version) grains correspond to the
examined. Fig. 8c shows the Schmid factor distribution map for highest strain values, whereas blue (in the web version) is associ-
Nanocomposite II reconstructed with the Channel 5 software. In this ated with the lowest strain values. These results suggest that when
134 L. Jiang et al./Acta Materialia 103 (2016) 128–140

Fig. 7. Representative TEM images of Nanocomposite II: (a) NPR and NPF zones; (b) dislocations near a sn-B4C in the NPR zones; representative high resolution TEM images
to show: (c) a high density of dislocations in the NPR zone, and (d) a low density of dislocations in the NPF zone; (e) bright field and (f) weak beam dark field TEM to show
high density of dislocations generated in a grain in the interface region between a NPF and a NPR zone after tensile test, inset is the corresponding diffraction pattern.

Fig. 8. (a) Schematic of the factors related to Schmid factor value. (b) EBSD reconstruction map to show grain size. (c) EBSD reconstruction map to show Schmid factor
value distribution in the NPR/NPF zones.
 L. Jiang et al./Acta Materialia 103 (2016) 128–140 135

Fig. 9. (a) EBSD reconstruction map and (b) diagram to show strain contouring in the NPR/NPZ zones after 2% compression test. The black region is un-indexed area due to
the noise from sn-B4C or secondary phases.

Nanocomposite II is deformed, strain reaches a maximum value in strain, whereas, faceted microparticles are prone to act as strain/
the NPF zones rather than in the NPR zones, as seen in Fig. 9. stress concentrators (the regions with rainbow color (in the web
version) in Fig. 10a).
4. Discussion The second principle involves achieving a significant strength/
deformability difference between “hard” and “soft” regions by
4.1. Design of spatial arrangement of reinforcing particles tailoring the distribution of sn-B4C. On one hand, without the pres-
ence of the sn-B4C, grains in NPF zones can deform plastically due
The results described in previous sections suggest that the spatial to the ease of dislocation movement. On the other hand, the sn-
arrangement of reinforcing nanoparticles (e.g., homogeneous or fiber- B4C can enhance the strength of the NPR zones since the TDI zones
like) determines the mechanical behavior of the nanocomposite. around the sn-B4C are harder than the zones far away from the sn-
Additional examples of possible reinforcement arrays are provid- B4C. It is apparent that higher composite strength can be obtained
ed in Fig. 10. In this figure, the background represents the matrix, using smaller particles, for a given volume fraction, provided that
whereas the blue (in the web version) regions represent the tran- particle–matrix interfaces are sufficiently strong. As the particle size
sitional dislocation interaction (TDI) region around the reinforcement decreases to the nanoscale, dislocation strengthening caused by the
phase. The TDI regions contain not only geometrically necessary presence of B4C particles dominates the overall response. The density
dislocations (GNDs), which are “punched” from the surface of re- of GNDs (ρGND) can be estimated by Refs. [32,33]:
inforcement particles and reach some equilibrium distance [31,32],
as shown in the inset of Fig. 11c, and form from the mismatch in 8 f ε y 12 f ( ΔC ΔT )
ρGND = ρEM + ρCTE = + (1)
elastic modulus and coefficient of thermal expansion (CTE), but bd (1 − f ) bd
also any dislocation loops that may originate from Orowan
strengthening. where ρEM is the density of GNDs due to elastic modulus mis-
It is important to note that these and other possible spatial match; ρCTE is the density of GNDs due to CTE mismatch; f is the
arrays of reinforcing particles do not necessarily result in en- volume fraction of reinforcement particles; b is the magnitude of
hanced mechanical properties. Accordingly, there are two principles the Burgers vector; d is the mean reinforcement particle diame-
that need to be considered in order to optimize toughness, for ter; εy is the yielding strain (0.2%); ΔC is the CTE mismatch between
example. First, we should take stress concentration issues into the B4C particle and the Al matrix [34]; and ΔT is the maximum tem-
consideration. Particle size, morphology and concentration in NPR perature change during thermomechanical processing. Results
zones can significantly affect the stress distribution. For example, obtained using Eq. (1) are summarized in Fig. 11c as a function of
a spherical nanoparticle geometry is preferred over a faceted ge- B4C particle diameter. For example, in the case of 5 vol.% sn-B4C par-
ometry on the basis of stress concentration considerations [41]. ticles, a dramatic increase in dislocation density is observed for
Hence, at a given overall level of plastic strain, nanocomposites particles less than 60 nm. This results in much higher dislocation
with spherical nanoparticles tend to experience a more uniform density in NPR zones compared to NPF zones as indicated by
136 L. Jiang et al./Acta Materialia 103 (2016) 128–140

Fig. 10. Schematic of the distributions of reinforcing particles and dislocation interaction zones in composites: (a) conventional structure with homogeneously distributed
faceted particles; (b) nanostructured nanocomposites with homogeneously distributed spherical nanoparticles. Nanocomposites with tailored distributions of spherical
nanoparticles: (c) spherical NPR zones; (d) fiber-like NPR zones. The shaded blue areas represent the matrix regions influenced by dislocations and the rainbow areas rep-
resent stress/strain concentrators. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 11. (a) Schematic of a crack propagating and fracturing through a faceted microparticle or nucleating from the dislocation punched zone-matrix boundary, or nucle-
ating from the interface between NPR and NPF zones; (b) SEM image showing a crack propagating along the interface between NPR and NPF zones; the arrows indicate
that the cracks went around the matrix near the sn-B4C. (c) Density of geometrically necessary dislocations in the dislocation punched zones as it varies with the diameter
of the sn-B4C (assuming volume fraction equals 0.05). The dislocation punched zone (dp) around a sn-B4C with a diameter (d) is illustrated in the inset of (c). (d) Statistics
on frequency and cumulative percentage of sn-B4C interparticle distance in the NPR zones of Nanocomposite II, and (e) Nanocomposite I.
 L. Jiang et al./Acta Materialia 103 (2016) 128–140 137

Fig. 7a–d. As a consequence, together with Orowan strengthening (compressive YS: 765 ± 9.2 MPa) of the yield strength of
from sn-B4C, one should expect strengthening of the metal matrix Nanocomposite II.
originating from NPR zones. Thus, we consider the evaluation of Nanocomposite II by treat-
ing the fiber-like NPR zones as fiber-like reinforcing units for which
the aspect ratio of the NPR zones, s, no longer has a value of 1. We
4.2. Influence of spatial arrangement of sn-B4C on strengthening can then use the modified shear lag model to approximate a value
of s for Nanocomposite II, given the yield strength:
In the case of nanocomposites with homogeneously distrib-
uted sn-B4C (i.e., Nanocomposite I), the increase in yield strength σ c = σ NPF + 1 2σ NPF ⋅ fNPR ⋅ s (7)
of the metal matrix that is attributable to Hall-Petch (ΔσH-P) and dis-
location strengthening (Δσdis) mechanisms can be expressed as where fNPR is volume fraction of the NPR zones (50 vol. % used in
follows [35,36]: this study); σNPF(572 MPa) is the yield strength of the NPF zones.
The value of s was estimated to be about 1.35 by using the exper-
Δσ matrix = Δσ H−P + Δσ dis imental value of 765 MPa as the yield strength. These results provide
(2)
= k y ⋅ D−1 2 + (Δσ ORBC )2 + (Δσ ORS )2 + (Δσ GND )2
4 support to the suggestion that the fiber-like NPR zones behaves as
units of reinforcement with an aspect ratio greater than 1 when sub-
where ky is a material constant (0.15 MPa × m1/2 for the Al–Mg alloy jected to deformation.
[37]) and D is the mean grain size; Δσ ORS
is the contribution from Note that the yield strength of the NPR zones (σNPR) was esti-
Orowan strengthening from the presence of cryomilling-induced dis- mated to be 833 MPa (Table 1). There is a yield strength difference
persoids such as oxides and nitrides, and it is evaluated to be of 261 MPa between the NPR and NPF zones. This is consistent
123 MPa for the cryomilled 5083 Al alloy matrix [38]; Δσ OR B4C
is the with the nanoindentation results (Fig. 5), which show that the
contribution of the Orowan strengthening from the presence of sn- NPR zones have much higher hardness values compared with
B4C; Δσ OR
B4C
can be calculated using the Orowan–Ashby equation in those for the NPF zones. Such strength/hardness difference results
Ref. [38]: from dislocation strengthening by sn-B4C as discussed above (about
143 MPa contribution) and the slight grain size difference between

Δσ B4C OR = M
0.4Gb In d b ( ) (3)
NPR and NPF (about 98 MPa contribution), as shown in Table 1.
The slight grain size difference can be attributed to the presence
π (1 − v ) λ
12
of sn-B4C, which suppresses grain growth during the cryomilling
where M = 3.06 is the mean orientation factor for fcc Al, G is the and thermal consolidation processes, as observed in our previous
shear modulus of the pure Al matrix (26.4 GPa); b is the Burgers study [40].
vector (2.86 × 10−10 m); v is the Poisson’s ratio; d = 2 3 ⋅ d and
( )
λ = d π 4 f − 1 . ΔσGND is the stress contribution due to GNDs and
4.3. Influence of spatial arrangement of sn-B4C on toughening
can be expressed as follows [39]:
4.3.1. Intrinsic toughening: grain deformation
Δσ GND = αGb ρGND (4) The toughness of the nanocomposites is related to the energy
released during matrix grain deformation, crack nucleation and crack
where α is the constant that is equal to 1.25 [39]; Δσ OR
B4C
and ΔσGND propagation. Intrinsic toughening is primarily associated with matrix
were calculated to be 121.5 MPa and 121.8 MPa, respectively, for grain deformation; as such, it is effective against the initiation of
Nanocomposite I, which contains 2.5 vol.% sn-B4C with a size of cracks [30].
50 nm. Thus, for Nanocomposite I, Δσdis and ΔσH-P were calculated The toughness data obtained herein can be understood by taking
to be 211 MPa and 345 MPa, respectively, by using the parameters into account two factors. First, the spatial arrangement of sn-B4C
listed in Table 1. Moreover, the sn-B4C can also contribute to the affects grain deformation as revealed by an analysis of the Schmid
yield strength from load bearing considerations, and the incre- factor value distribution in the nanocomposites. Under a static
ment in yield strength can be evaluated using the modified shear loading condition, the critical shear stress required to activate slip
lag model [35]: depends on the orientation of a particular grain, and hence the
Schmid factor represents a measure of how difficult it is to acti-
Δσ L −T = 1 2(σ 0 + Δσ matrix ) f ⋅ s (5) vate slip. The value of the Schmid factor can be calculated based
on the following equation [41]:
where σ0 = 130 MPa is the yield strength of coarse grained 5083 Al
and s is the aspect ratio of the reinforcement. Thus, for a spherical τc
m= = cos λ ⋅ cos α (8)
reinforcement the value of s is 1. Δσ L−T was calculated to be σs
8.2 MPa, indicating the low load bearing capability of sn-B4C. The
yield strength of the nanocomposites can then be estimated as where m is Schmid factor, τc is critical shear stress for activating dis-
follows [35]: location slip, σs is yield stress, λ is the angle between the applied
load direction and the slip direction, and α is the angle between
σ c = σ 0 + Δσ matrix + Δσ L −T (6) the applied load direction and the axis normal to the slip plane. The
m value is determined from the cosine values of the α and λ angles
In the case of Nanocomposites I and II, they both have the same as schematically shown in Fig. 8a; a higher value of the Schmid factor
content of sn-B4C and similar average grain sizes, thus they have means that grains are easily deformed along the applied loading
almost the same calculated yield strength. The yield strength of direction. These results indicate that grains located in NPF zones
Nanocomposite I is estimated to be 695 MPa. The slight underes- are more easily deformed relative to grains in NPR zones (Fig. 8),
timation relative to the experimental value (compressive YS: suggesting that the presence of the sn-B4C influences grain texture
705 ± 9.8 MPa) is likely due to the many simplifying assumptions evolution during hot-extrusion. Generally, extrusion leads to Al grains
made in the formulation of the above equations. However, there is aligned with <111> along the extrusion direction, resulting in a fiber
a large gap between the estimated value and experimental value texture and the grains with such texture have high Schmid factor
138 L. Jiang et al./Acta Materialia 103 (2016) 128–140

values [42]. Presumably, the sn-B4C interfere with grain rotation schematically shown in Fig. 12a, whereas outside of R0, the mag-
during extrusion [29]. Thus, the grains in NPF zones have higher nitude of the stress concentration goes to zero. To estimate the
deformability and behave as strain bearing “soft” units due to the characteristic dimension R0, the KIC of the NPR zones was calcu-
ease of dislocation movement and absence of the sn-B4C. lated by applying the following equation [46].
Second, the sn-B4C distribution influences matrix grain defor-
mation via a mechanism involving dislocation punched zones, the ⎧ σ yαβdE ⎫ − 61
K IC = 0.77 ⎨ 2 ⎬
V (10)
size of which is designated as dp (see Fig. 11c). The presence of dis- ⎩ (1 − υ ) ⎭⎪
location punched zones around the sn-B4C significantly affects failure
mechanisms. According to the simulation result from Suh et al. where β=1.67 and α=0.5 for full constraint (plane strain) condi-
[32,33], for a fixed volume fraction, as particle size decreases, a tran- tion, Vf is volume fraction of sn-B4C, respectively, E is elastic modulus,
sition occurs from a particle/matrix interface-dominated failure and υ is Poisson’s ratio. By combining Eqs. (9) and (10) the char-
initiation to a dislocation punched zone-matrix boundary domi- acteristic dimension R0 of the crack tip’s “plastic zone” is estimated
nated failure initiation, in the case of a well-bonded particle/ as follows:
matrix interface. This suggests that the matrix grains in the
dislocation punched zones deform differently from matrix grains 0.26d 2E 2
R0 = (11)
and can give rise to crack initiation. In our case, crack initiation is π V f1 3
likely to occur at dislocation punched zone-matrix boundaries, as
schematically shown in Fig. 11a. Evidence of this behavior can be From Eq. (11), it can be seen that the sn-B4C has significant effects
found in Fig. 11b, in which micro-cracks followed an interface on the characteristic dimension R0 of the nanocomposites. As seen
between NPR and NPF zones and most of these cracks nucleated in Fig. 12b,c, the R0 sharply increases with decreasing volume frac-
near the sn-B4C as indicated by arrows. Note that in NPR zones, sn- tion (Vf) and increasing size (d) of sn-B4C. Thus, the R0 value for NPR
B4C have small interparticle distances and thus could result in zones with 5 vol.% is much smaller than that for NPF zones without
overlapping dislocation punched zones. As seen in Fig. 11d, the sn-B4C. In Nanocomposite II, these two building blocks are alter-
average interparticle distance (ID) in the NPR zones is 79.8 nm and nately distributed, which results in periodic variations in the values
approximately 55% of the IDs are less than 75 nm, which is equal of R0 at a microscale. Such variations, which correspond to the length
to the punched zone size of the sn-B4C, as schematically shown in scale of the NPF/NPR zones, thus activate “microstructural” tough-
the inset of Fig. 11c (dp ≈ 1.4d [32,33]). In contrast, the average ID ening mechanisms. On the one hand, the “soft” NPF zones can
in the Nanocomposite I is more than 149.2 nm and only 11% of the increase the stability against crack propagation into the “hard” NPR
IDs are less than 75 nm (1.4d), as seen in Fig. 11e. Thus, the over- zones [44]. One the other hand, suppression of tensile instability
lapping of dislocation punched zones is more likely to occur in the requires a mechanism to limit shear band extension in the NPF zones.
NPR zones. Such overlap of dislocation punched zones leads to NPR Shear bands initiated in plastically “soft” zones can be arrested in
zones acting as a single fiber-like unit. Consequently, the disloca- surrounding “hard” zones of higher yield strength or stiffness [45].
tions tended to concentrate at the interfaces of the NPR and NPF In principle, it should be possible to design the microstructural length
zones (Fig. 7e,f) when the nanocomposite was subjected to defor- scales (the thickness of NPR or NPF zones and parameters for the
mation, such that cracks initiated at those interfaces (Fig. 6e,f), which sn-B4C, Vf, d etc.) to match the characteristic length scale R0 for achiev-
promoted the pull-out of the NPR zones (Fig. 6a and c). Thus, the ing an optimized toughening effect.
ID within the NPR zones influences toughness. On one hand, a small One principle is to match the R0 with L, the thickness of “hard”/
value of ID may lead to a high stress concentration; on the other “soft” zones (L ≈ R0) to increase the complexity of crack propagation
hand, a large value will prevent NPR zones from acting as fiber- routes in the composites [44]. This design principle is rationalized
like units. in Nanocomposite II based on the following facts. If the thickness
of the NPR/NPF zones is too large when compared to the value of
R0, the cracks generated in the NPR zones will not be affected by
4.3.2. Extrinsic toughening: crack deflection the NPF zones and moreover, any shear bands generated in the “soft”
Extrinsic toughening mechanisms, such as crack deflection and NPF zones will not be arrested by the “hard” NPR zones. In con-
bridging, act principally on the wake of a crack to reduce (shield) trast, a geometrical arrangement with small values of thickness for
the local stresses/strains experienced at a crack tip [39]. A com- the NPR/NPF zones will experience a uniform stress or strain re-
posite made from fiber constituents can have a surprisingly high sponse at the microscale, similar to a structure with homogeneously
toughness if a crack is repeatedly deflected at fiber/matrix inter- distributed nanoparticles.
faces [39]. In our fiber-like structure in Nanocomposite II, crack By adjusting the volume fraction and size of the sn-B4C in the
deflection at NPR and NPF zones allows for fiber-like NPR pull-out NPR zone, we can vary the value of R0 so as to match it with the
behavior (Fig. 6), which contributed significantly to the toughness. thickness of the NPF/NPR zones. For the current study, the parti-
As schematically shown in Fig. 12a, a crack generated at NPR/ cle diameter of sn-B4C (d) is 50 nm and Vf in NPR zones is 0.05.
NPF interfaces will either propagate along the interface or traverse Thus, the corresponding R0 of NPR zones was estimated to be 6.9 μm,
the NPR/NPF zones. The length scale of the NPR/NPF zones as shown in Fig. 12b. In Nanocomposite II, the volume fraction of
affects the crack deflection response. As schematically shown in the NPR zones (originating from Powder C) and NPF zones (origi-
Fig. 12a, the crack propagation will be affected by the “plastic zone” nating from Powder A) are both 50%, and thus the thicknesses of
around the crack tip. The characteristic dimension R0 of a crack tip’s the NPR and NPF zones are almost the same. By analyzing an array
“plastic zone” is a material length scale related to fracture tough- of micrographs (such as those in Fig. 3a or Fig. 2g), it was found
ness (KIC) and for a mode I opening crack, it can be estimated as [42]: that the thicknesses of NPR/NPF zones, L, varied over the range
2–10 μm. These values match with the calculated value of R0, which
2
1 ⎛ K IC ⎞ is on the order of 6.9 μm. Under these conditions, cracks were pref-
R0 = (9)
2π ⎜⎝ σ y ⎟⎠ erentially deflected at the interfaces of NPR and NPF zones and
resulted in the pull-out of NPR zones during tensile failure, as
R0 is associated with the maximum spatial extension of shear shown in Fig. 6. This lengthens the crack propagation route, as sche-
bands originating at an opening crack tip [43,44]. Within R0, stress matically shown in Fig. 6g, thus enhancing the toughness of
concentration decreases with distance away from the crack tip, as Nanocomposite II.
 L. Jiang et al./Acta Materialia 103 (2016) 128–140 139

Fig. 12. (a) Schematic of “plastic zone” R0 and related stress field around a crack tip at the interface of NPR and NPF zones. Variation in the length scale of the “plastic zone”
R0 with: (b) the volume fraction (assuming d equals 50 nm), and (c) the diameter of B4C (assuming volume fraction equals 0.05).

Additionally, for a given volume fraction, the R0 also varies with showed both enhanced strength and toughness compared to
the diameter of the B4C particles, even if they are no longer in the nanocomposites with the same composition but in which the
nm regime, until a plateau is reached, as shown in Fig. 12c. The nanoparticles were homogeneously distributed throughout the
plateau occurs when R0 in the particle-rich zone reaches the value matrix. Additionally, these two structures showed obvious dif-
for the particle-free zone, corresponding to the situation where the ferences in nanoindentation response. Larger fluctuations in
influence of reinforcement particles on the length of the plasticity hardness/elastic modulus were found in the sample with fiber-
zone (and consequently crack deflection capability) can be ignored. like NPR zones.
Assuming 5 vol.% B4C in the particle rich zones, KIC = 30 MPa m1/2 2. Although sn-B4C showed poor direct load-carrying capacity due
andσc = 560 MPa in the particle free zones, this plateau occurs at a to the low aspect ratio and small size (only about 8 MPa en-
value of approximately 400 μm, and a corresponding particle size hancement with a concentration of 2.5 vol.% B4C nanoparticles),
of 600 nm, This behavior is in contrast to that for the finer sn-B4C, they contributed to remarkable dislocation strengthening ability
which generate local plastic deformation within a length scale less in the Al matrix (over 200 MPa enhancement from 2.5 vol.% sn-
than 10 μm. This calculation highlights the advantage of using na- B4C), due to the generation of geometrically necessary dislocations
noscale particles for the spatial design of reinforcements in metal and the activation of Orowan mechanisms. With the fiber-like
matrix composites. distribution of sn-B4C, additional microstructural strengthen-
ing was obtained. More dislocations were pinned by spherical
5. Conclusions nanoparticles located inside the NPR zones and contributed to
the large variation in hardness values between the NPR and
Spherical B4C nanoparticles (sn-B4C) were assembled into fiber- nanoparticle-free (NPF) zones. Due to this variation, the stress
like arrays within an ultrafine grained 5083 Al matrix. The influence was transferred from “soft” NPF zones to “hard” NPR zones, and
of this unique sn-B4C distribution on the strengthening and defor- some of the fiber-like NPR zones were pulled out at the frac-
mation characteristics of the nanocomposite was investigated in ture surface during the tensile loading. The above facts support
detail, with the following findings. that NPR zones act as load-bearing fiber-like “hard” units and
result in a higher strength value.
1. At a given volume fraction (2.5 vol.% sn-B4C), nanocomposites 3. The presence of the fiber-like NPR zones also enhanced ductil-
reinforced with fiber-like nanoparticle-rich (NPR) zones ity by affecting the behavior of nearby matrix grains. The NPR
140 L. Jiang et al./Acta Materialia 103 (2016) 128–140

zones showed smaller Schmid factor values compared to the NPF [18] L.L. Cao, Z.Q. Li, G.L. Fan, L. Jiang, D. Zhang, W.J. Moon, Y.S. Kim, The growth of
carbon nanotubes in aluminum powders by the catalytic pyrolysis of
zones because the formation of preferred orientations within the
polyethylene glycol, Carbon 50 (2012) 1057–1062.
grains due to the extrusion was suppressed by the presence of [19] L. Jiang, G.L. Fan, Z.Q. Li, X.Z. Kai, D. Zhang, Z.X. Chen, S. Humphries, G. Heness,
sn-B4C. In contrast, the NPF zones had relatively more move- W.Y. Yeung, An approach to the uniform dispersion of a high volume fraction
able dislocations and enhanced grain texture with preferential of carbon nanotubes in aluminum powder, Carbon 49 (2011) 1965–1971.
[20] Z.Q. Li, L. Jiang, G.L. Fan, Y. Xu, D. Zhang, Z.X. Chen, S. Humphries, High volume
deformation in the extrusion direction. Thus, strain could be ac- fraction and uniform dispersion of carbon nanotubes in aluminium powders,
commodated by the “soft” NPF zones, resulting in higher ductility Micro Nano Lett. 5 (2010) 379–381.
in the nanocomposite. [21] X. Deng, B.R. Patterson, K.K. Chawla, M.C. Koopman, Z. Fang, G. Lockwood, A.
Griffo, Mechanical properties of a hybrid cemented carbide composite, Int. J.
4. Improved toughness was also observed in the nanocomposite Refract. Met. Hard Mater. 19 (2001) 547–552.
with fiber-like NPR zones, and was explained as follows. First, [22] Z. Fang, G. Lockwood, A. Griffo, A dual composite of WC-Co, Metall. Mat. Trans.
the nanoscale size (50 nm) of the sn-B4C and well-bonded in- A 30 (1999) 3231–3238.
[23] R. Saha, E. Morris, N. Chawla, S.M. Pickard, Hybrid and conventional particle
terface of sn-B4C/matrix suppressed crack nucleation inside the reinforced metal matrix composites by squeeze infiltration casting, J. Mater.
NPR zones. Together with the overlapping of dislocation punched Sci. Lett. 21 (2002) 337–339.
zones caused by the small interparticle distance between the sn- [24] V.C. Nardone, J.R. Strife, K.M. Prewo, Microstructurally toughened particulate-
reinforced aluminum matrix composites, Metall. Mat. Trans. A 22A (1991)
B4C, crack nucleation occurred at the interfaces of the NPR and 171–182.
NPF during the deformation. Second, the crack propagation was [25] J. Segurado, C. González, J. Llorca, A numerical investigation of the effect of
governed by the ratio of the plastic zone size, R0, to the size of particle clustering on the mechanical properties of composites, Acta Mater. 51
(2003) 2355–2369.
the fiber-like NPR zones. When the thickness of the NPR/NPF
[26] M.M. Balakrishnarajan, P.D. Pancharatna, R. Hoffmann, Structure and bonding
zones (L) is on the order of R0, the crack propagation distance in boron carbide: the invincibility of imperfections, New J. Chem. 31 (2007)
increases, thus contributing to the overall increase in the tough- 473–485.
ness of the nanocomposite. [27] D.B. Witkin, E.J. Lavernia, Synthesis and mechanical behavior of nanostructured
materials via cryomilling, Prog. Mater. Sci. 51 (2006) 1–60.
[28] Z. Zhang, S. Dallek, R. Vogt, Y. Li, T. Topping, Y. Zhou, J. Schoenung, E. Lavernia,
Degassing behavior of nanostructured Al and its composites, Metall. Mat. Trans.
Acknowledgments A 41 (2010) 532–541.
[29] L. Jiang, H. Wen, H. Yang, T. Hu, T. Topping, D. Zhang, E.J. Lavernia, J.M.
The authors acknowledge financial support from the Office of Schoenung, Influence of length-scales on spatial distribution and interfacial
characteristics of B4C in a nanostructured Al matrix, Acta Mater. (2015) 327–343.
Naval Research (ONR) under the guidance of Dr. Lawrence Kabacoff [30] U.G.K. Wegst, H. Bai, E. Saiz, A.P. Tomsia, R.O. Ritchie, Bioinspired structural
(ONR N00014-12-1-0237), and the Defense University Research In- materials, Nat. Mater. 14 (2015) 23–36.
strumentation Program (DURIP) under grant N00014-13-1-0668. [31] X.Z. Kai, Z.Q. Li, G.L. Fan, Q. Guo, Z.Q. Tan, W.L. Zhang, Y.S. Su, W.J. Lu, W.J. Moon,
D. Zhang, Strong and ductile particulate reinforced ultrafine-grained metallic
composites fabricated by flake powder metallurgy, Scr. Mater. 68 (2013)
References 555–558.
[32] Y.S. Suh, S.P. Joshi, K.T. Ramesh, An enhanced continuum model for size-
dependent strengthening and failure of particle-reinforced composites, Acta
[1] Y.T. Zhu, X.Z. Liao, Nanostructured metals – retaining ductility, Nat. Mater. 3 Mater. 57 (2009) 5848–5861.
(2004) 351–352. [33] J.C. Shao, B.L. Xiao, Q.Z. Wang, Z.Y. Ma, K. Yang, An enhanced FEM model for
[2] N. Chawla, Y.L. Shen, Mechanical behavior of particle reinforced metal matrix particle size dependent flow strengthening and interface damage in particle
composites, Adv. Eng. Mater. 3 (2001) 357–370. reinforced metal matrix composites, Compos. Sci. Tech. 71 (2011) 39–45.
[3] E. Ma, Instabilities and ductility of nanocrystalline and ultrafine-grained metals, [34] D.C. Dunand, A. Mortensen, On plastic relaxation of thermal stresses in
Scr. Mater. 49 (2003) 663–668. reinforced metals, Acta Metall. Mater 39 (1991) 127–139.
[4] K. Lu, The future of metals, Science 328 (2010) 319–320. [35] Z.H. Zhang, T. Topping, Y. Li, R. Vogt, Y.Z. Zhou, C. Haines, J. Paras, D. Kapoor,
[5] R.O. Ritchie, The conflicts between strength and toughness, Nat. Mater. 10 (2011) J.M. Schoenung, E.J. Lavernia, Mechanical behavior of ultrafine-grained Al
817–822. composites reinforced with B4C nanoparticles, Scr. Mater. 65 (2011) 652–655.
[6] D.J. Lloyd, Particle-reinforced aluminum and magnesium matrix composites, [36] S. Scudino, G. Liu, M. Sakaliyska, K.B. Surreddi, J. Eckert, Powder metallurgy of
Int. Mater. Rev. 39 (1994) 1–23. Al-based metal matrix composites reinforced with β-Al3Mg2 intermetallic
[7] Z. Wang, T.K. Chen, D.J. Lloyd, Stress-distribution in particulate-reinforced particles: analysis and modeling of mechanical properties, Acta Mater. 57 (2009)
metal-matrix composites subjected to external load, Metall. Mater Trans. A 24 4529–4538.
(1993) 197–207. [37] D.J. Lloyd, S.A. Court, Influence of grain size on tensile properties of Al-Mg alloys,
[8] L. Jiang, K.K. Ma, H. Yang, M.J. Li, E.J. Lavernia, J.M. Schoenung, The Mater. Sci. Tech. Ser. 19 (2003) 1349–1354.
microstructural design of trimodal aluminum composites, Jom-Us 66 (2014) [38] B.Q. Han, Z. Lee, S.R. Nutt, E.J. Lavernia, F.A. Mohamed, Mechanical properties
898–908. of an ultrafine-grained Al-7.5 Pct Mg alloy, Metall. Mater. Trans. A 34 (2003)
[9] H.X. Peng, Z. Fan, J.R.G. Evans, Novel MMC microstructure with tailored 603–613.
distribution of the reinforcing phase, J. Microsc.-Oxford 201 (2001) 333–338. [39] F. Tang, I.E. Anderson, T. Gnaupel-Herold, H. Prask, Pure Al matrix composites
[10] L. Jiang, Z. Li, G. Fan, L. Cao, D. Zhang, Strong and ductile carbon nanotube/ produced by vacuum hot pressing: tensile properties and strengthening
aluminum bulk nanolaminated composites with two-dimensional alignment mechanisms, Mater. Sci. Eng. A 383 (2004) 362–373.
of carbon nanotubes, Scr. Mater. 66 (2012) 331–334. [40] K.H. Chung, J. He, D.H. Shin, J.M. Schoenung, Mechanisms of microstructure
[11] L. Jiang, Z. Li, G. Fan, L. Cao, D. Zhang, The use of flake powder metallurgy to evolution during cryomilling in the presence of hard particles, Mater. Sci. Eng.
produce carbon nanotube (CNT)/aluminum composites with a homogenous CNT A 356 (2003) 23–31.
distribution, Carbon 50 (2012) 1993–1998. [41] X.L. Nan, H.Y. Wang, L. Zhang, J.B. Li, Q.C. Jiang, Calculation of Schmid factors
[12] J. Ye, B.Q. Han, Z. Lee, B. Ahn, S.R. Nutt, J.M. Schoenung, A tri-modal aluminum in magnesium: analysis of deformation behaviors, Scr. Mater. 67 (2012)
based composite with super-high strength, Scr. Mater. 53 (2005) 481–486. 443–446.
[13] J.C. Wong, M. Paramsothy, M. Gupta, Using Mg and Mg–nanoAl2O3 concentric [42] C. Jiao, Z.K. Yao, Y.F. Han, Extrusion deformation texture of SiCW/Al composites,
alternating macro-ring material design to enhance the properties of magnesium, J. Mater. Sci. Tech. 8 (1992) 25–29.
Compos. Sci. Tech. 69 (2009) 438–444. [43] G. Ravichandran, A. Molinari, Analysis of shear banding in metallic glasses under
[14] S.Y. Qin, G.D. Zhang, Preparation of high fracture performance SiCp-6061Al/ bending, Acta Mater. 53 (2005) 4087–4095.
6061Al composite, Mater. Sci. Eng. A 279 (2000) 231–236. [44] R.D. Conner, Y. Li, W.D. Nix, W.L. Johnson, Shear band spacing under bending
[15] L. Jiang, Z. Li, G. Fan, D. Zhang, A flake powder metallurgy approach to Al2O3/Al of Zr-based metallic glass plates, Acta Mater. 52 (2004) 2429–2434.
biomimetic nanolaminated composites with enhanced ductility, Scr. Mater. 65 [45] D.C. Hofmann, J.-Y. Suh, A. Wiest, G. Duan, M.-L. Lind, M.D. Demetriou, W.L.
(2011) 412–415. Johnson, Designing metallic glass matrix composites with high toughness and
[16] Z.Y. Fan, A microstructural approach to the effective transport properties of tensile ductility, Nature 451 (2008) 1085–1089.
multiphase composites, Philos. Mag. A 73 (1996) 1663–1684. [46] A.B. Pandey, B.S. Majumdar, D.B. Miracle, Effect of aluminum particles on the
[17] Z. Zhou, H.X. Peng, Z. Fan, D.X. Li, MMCs with controlled non-uniform fracture toughness of a 7093/SiC/15p composite, Mater. Sci. Eng. A 259 (1999)
distribution of submicrometre Al2O3 particles in 6061 aluminium alloy matrix, 296–307.
Mater. Sci. Tech. Ser. 16 (2000) 908–912.