of 5
1
Appendix
A
for
Substantial
tensile ductility in sputtered Zr
-
N
i
-
Al nano
-
sized metallic glass
Rachel Liontas, Mehdi Jafary
Zadeh, Qiaoshi Zeng, Yong
-
Wei Zhang, Wendy L. Mao, Julia R. Greer
Uniaxial compression on Zr
-
Ni
-
Al specimens
In addition to the tensile experiments discussed in detail, we also examined the mechanical properties of
the as
-
sputtered Zr
-
Ni
-
Al metallic glass via uniaxial compression experiments on cylindrical nano
-
compress
ion specimens. Specimen were fabricated by FIB milling of the same ~4.5
μ
m thick sputtered film
analyzed by TEM in Figure 2. The specimens were fabricated with diameters ranging from 225 nm to 1560
nm, and corresponding heights such that the aspect ratio w
as ~1:3 so as to avoid buckling during
compression. The specimens were compressed at a strain rate of 1
×
10
-
3
s
-
1
using the Hysitron PI
-
85 nano
-
mechanical testing instrument in
-
situ in the Versa 3D scanning electron microscope. The resultant
compression dat
a is displayed in Figure
A
.1
.
These specimens exhibited two very distinct regimes of
mechanical behavior depending on their size: ductile behavior and homogeneous flow was observed for
“small” specimens with initial diameters
555 nm, while shear banding and localized failure was obse
rved
for “large” specimens with initial diameters
890 nm. These differences in mechanical response are apparent
from the stress
-
strain curves, which show uniform plastic loading for the “small” specimens (Figure
A.1
(a)),
and many strain bursts from the
catastrophic shear banding events for the “large” specimens (Figure
A.1
(b)).
The post
-
compression images of the specimens also illustrate the differences in mechanical response with the
“small” specimens exhibiting homogeneous deformation near the top of
the specimens (Figure
A.1
(c), (d))
and the “large” specimens exhibiting clear shear bands (Figure
A.1
(e), (f))
.
Overall, the compression data
indicates a transition from homogeneous
-
like flow to shear banding at a specimen diameter between 555 and
890 nm
.
2
Additional Molecular Dynamics Results
In order to determine the appropriate temperature to conduct the annealing for the MD studies of Figure
8, it was
necessary to first determine the glass transition temperature (T
g
) of the MD
-
formed metallic glass.
Towards that aim, during the quenching process of forming the MD samples, the potential energy per atom
was monitored as a function of temperature for cooli
ng rates of
10
10
K/s, 10
12
K/s, and 10
13
K/s, as shown
in Figure
A
.2
. As in Caprion et al.
[1]
, we determined T
g
as the crossover between extrapolations of the high
and low temperature curves of potential energy. The resultant T
g
values
are denoted by dashed lines in F
igure
A.2
, with values of ~
950 K for the cooling rate of 10
10
K
/s, ~1000 K for the cooling rate of 10
12
K/s, and
~
1050 K for the cooling rate of 10
13
K/s. Since the MD annealing was performed on the sample originally
formed with a cooling rate of 10
13
K/s, we conducted the annealing at ~80% of the ~1050 K T
g
, which w
as
~850 K.
Figure A.1.
Uniaxial compression experiments on FIB
-
fabricated Zr
-
Ni
-
Al specimens, including
engineering stress
-
strain results on
(a)
“small” specimens with diameters
555 nm and
(b)
“large”
specimens with diameters
890 nm; post
-
compression imag
es on specimens with initial diameters of
(c)
492 nm,
(d)
540 nm,
(e)
1300 nm, and
(f)
1560 nm.
3
As shown in Figure 7 (f) and Figure 8 (d), we primarily utilized the potential energy per atom and the
percentage of Al atoms centered in full icosahedron (
Voronoi
index Al <0 0 12
0>) as measures of atomic
-
level structure and ordering in the MD samples, however many other
Voronoi
cells can also be mea
s
ured.
Some additional
Voronoi
cells, or short
-
range order (SRO) clusters, were monitored as a function of metallic
glass cooling rate, and these results are shown in Figure
A
.
3
. Only stable SROs with a dominant
percentage
were included in Figure
A
.3
. It can be observed that by increasing the cooling rate, the percentage of
dominant and stable SROs decrease, which means there is a corresponding increase in the percentage of
other unstable SROs, or liquid
-
like clusters.
The Al <0 0 12
0>, or the
fracti
on of Al atoms centered in full
icosahedron, is the most sensitive to changes in cooling rate and was thus selected for the analysis of ordering
included in Figures 7 (f) and Figure 8 (d).
F
igure A
.2
.
Variation of potential energy as a function of temperature during the formation of
Zr
55
Ni
30
Al
15
metallic glass quenched from 2300 K with cooling rates of 10
10
K/s, 10
12
K/s, and 10
13
K/s. The vertical dashed
-
lines denote the glass transition temperature (T
g
) of the alloy.
4
SRIM estimates of Ga
+
ion penetration from FIB
To quantify the amount of Ga potentially implanted into the specimens as a result of the FIB
fabrication,
The Stopping and Range of Ions in Matter
(SRIM) software was used. For this calculation, the energy of the
incoming Ga ions was set at 30 keV
at an incoming angle of 89.9999 degrees (glancing angle)
,
1
,000,000
Ga
ions were used for the calcula
tion, the target was
Zr
55
Ni
25
Al
20
with an assumed bulk density of 6.33 g cm
-
3
.
The
results of these calculations are shown in Figure
A.4
.
As discussed in section 4.2 and can be observed in
Figure
A.4
~78% of the implanted Ga is present within the first 10
nm of the free surface of the specimen,
and ~98% of the implanted Ga is present within the first 20 nm of the free surface of the specimen.
Due to
the range of FIB currents and exposure times used to fabricate the specimen, we did not have a quantitative
number for the dose, which would have allowed calculation of Ga concentration in the specimens
.
However,
the energy dispersive x
-
ray spectroscopy (EDS) results done on a tensile specimen shown in Figure 2 (g)
indicate a Ga concentration of 1.9 ± 0.9 at. %,
and this value should be considered an overestimate of the
actual Ga concentration as EDS is more sensitive to surface atoms
and as demonstrated in Figure
A.4
there is
a much higher concentration
of Ga close to the sample surface.
F
igure A.3
.
Percentage of Zr, Ni, and Al atoms present as the c
entral atom in various
Voronoi
cells, or
SRO clusters, as a function of metallic glass cooling rate.
5
References
for Appendix A
[1] D. Caprion, and H.R. Schober,
Influence of the quench rate and the pressure on the glass transition
temperature in selenium, J. Chem. Phys.
117 (2002) 2814
-
2818.
F
igure A
.4
.
Fract
ion of implanted Ga ions as a function of distance from the free surface of Zr
-
Ni
-
Al
metallic glass exposed to
30 keV glancing angle Ga
-
ion irradiation. Each bar covers a 1 nm range of
depth. These results were obtained from SRIM calculations.