Stacking faults in formation of silver nanodisks

Article abstract of DOI:10.1021/jp0303826

Structural investigations by using high-resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) of silver nanodisks with different sizes are presented. The disks have a face centered cubic (fcc) crystal structure and their flat surfaces are (111). Stacking faults parallel to the (111) planes are frequently observed for the nanodisks. A unique (111) stacking fault model which is parallel to the flat (111) disk surface has been proposed to explain the observed 1/3{422} forbidden reflections in [111] SAED pattern and the corresponding 37×{422} supperlattice fringes in the [111] HRTEM image. It is suggested that the presence of the stacking faults may be the key in the formation and growth of the disk morphology. This study may provide an insight to synthetically controlling particle shape and size through defect engineering.

Full text of DOI:10.1021/jp0303826

©
Copyright 2003 by the American Chemical Society
VOLUME 107, NUMBER 34, AUGUST 28, 2003
LETTERS
Stacking Faults in Formation of Silver Nanodisks
†
‡
†
,‡
,†
V. Germain, Jing Li, D. Ingert, Z. L. Wang,* and M. P. Pileni*
Laboratoire LM2N., UniVersit e´ Pierre et Marie Curie (Paris VI), B.P. 52, 4 Place Jussieu,
F-752 31 Paris Cedex 05, France, and School of Materials Science and Engineering,
Georgia Institute of Technology, Atlanta, Georgia 30332-0245
ReceiVed: March 28, 2003; In Final Form: June 28, 2003
Structural investigations by using high-resolution transmission electron microscopy (HRTEM) and selected
area electron diffraction (SAED) of silver nanodisks with different sizes are presented. The disks have a face
centered cubic (fcc) crystal structure and their flat surfaces are (111). Stacking faults parallel to the (111)
planes are frequently observed for the nanodisks. A unique (111) stacking fault model which is parallel to the
flat (111) disk surface has been proposed to explain the observed 1/3{422} forbidden reflections in [111]
SAED pattern and the corresponding 3×{422} supperlattice fringes in the [111] HRTEM image. It is suggested
that the presence of the stacking faults may be the key in the formation and growth of the disk morphology.
This study may provide an insight to synthetically controlling particle shape and size through defect engineering.
5
In the past few years, research on inorganic nanocrystals has
had a tremendous development because of their great potential
for applications. This involves biology, electronics, transport,
and information technology and implies producing nanocrystals
with well-defined sizes, shapes, and crystallinity. Several
challenges have to be faced. We must produce large amounts
and well-controlled materials and be able to manipulate these
triangular and hexagonal shapes, have been synthesized using
different methods. These silver nanoprisms and nanodisks have
the optical properties different from the silver nanospheres and
4
,5
have a potential application in optics. We have been able to
tune the silver nanodisk size from 30 to 120 nm, resulting in
drastic changes in the optical properties.
8
Both silver nanoprisms and nanodisks mentioned above
synthesized by different groups using different methods have a
flat platelike shape with the flat top surface parallel to the (111)
plane of their fcc crystal structure. Also, both show the presence
of the 1/3{422} reflections in the [111] electron diffraction
pattern, which should be forbidden for perfect fcc structure.
This phenomenon has been observed by many authors in Au
or Ag platelike crystals, and a number of explanations for the
occurrence of such forbidden reflections have been previously
1
nanocrystals. Several approaches to produce these nanomate-
rials with various shapes have been undertaken in the past
decade. Transmission electron microscopy is a powerful and
unique technique for structure characterization.2
4
,5
It is well-known that the physical and chemical properties of
the nanoscale crystals are usually closely related to the nano-
particle shape and size. For this reason, the shape- and size-
controlled procedure during the nanoparticles synthesis has
become a new and interesting research focus. Very recently,
3
,9-12
suggested,
which will be elaborated in below.
3
-7
several groups have made nanodisks morphology particles.
In this paper, based on detailed structure characterization for
the silver nanodisks with different sizes by using HRTEM and
SAED analyses, we give an alternative explanation about the
formation of these 1/3{422} forbidden reflections: It is the (111)
stacking faults lying parallel to the (111) surface and extending
across the entire disk that are responsible for creating the
The silver nanoprisms with the perfectly triangular and truncated
4
triangular shapes, as well as the silver nanodisks with truncated
*
To whom correspondence should be addressed.
Universit e´ Pierre et Marie Curie (Paris VI).
Georgia Institute of Technology.
†
‡
1
0.1021/jp0303826 CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/05/2003

8718 J. Phys. Chem. B, Vol. 107, No. 34, 2003
Letters
Figure 2. (A) HRTEM image at 400 kV of a silver nanodisk in the
[
111] orientation (sample E). The measured distance is three times the
distance of {422}. In the insert, Fourier transform of the same area.
B) [111] HRTEM image of silver nanodisk at the edge region of the
(
disk. The 3 × {422} fringes extend over the edge area. (C) Sketch of
the fcc stacking of 3 layers of atoms along the [111] direction. The
projected positions of the A atoms build a bigger unit cell (3 × {422}
distance).
Figure 1. TEM images of silver nanodisks. From A to F the hydrazine
content increases from r ) 4.7 to 11.8. (A) r ) 4.7, (B) r ) 6.3, (C)
)
7.1, (D) r ) 7.9, (E) r ) 9.8, (F) ) 11.8, (G) Schematic drawing of
a typical nanodisk with the crystallographic orientation. (H) A typical
111] SAED pattern of silver nanodisk in the sample E. The bright
[
spots correspond to the {220} Bragg reflections, and the circled spot
is 2h 20. The weaker spots closer to the center are attributed to 1/3{422}
forbidden reflections; the boxed spot is 1/3 22 4h .
1/3{422} reflections in the [111] pattern. We also propose that
the stacking faults play an important role for the formation and
growth of the platelike silver nanodisks. This will provide an
insight to synthetically controlling particle shape and size
through defect design.
The nanodisk sizes are tuned by the amount of reducing agent
added during the synthesis (see Appendix I). Figure 1 shows
the TEM13 images for 6 different samples differing by their
sizes. Sphericals and highly faceted large particles with a rather
wide size distribution are produced (Figure 1). In the following,
we will concentrate our effort on the large and faceted
nanocrystals.
Figure 3. SAED patterns for all samples. All patterns show the
1
1
/3{422} forbidden reflections. In Figure 4, parts A, B, and D, the
/3{422} forbidden reflections occur in the circles in which the image
contrast was adjusted. In Figure 4, parts B and F, the hexagonal box
marked the 1/3{422} forbidden reflections. (Note: In Figure 4B the
pattern does not come from one nanodisk; therefore, the pattern is not
as perfect as others; but the same periodicity as marked by the hexagonal
box caused by the 1/3{422} forbidden reflections is still visible.) In
Figure 4, parts C and E one of the 1/3{422} forbidden reflections is
marked by the circle.
given in Figure 1H, showing a relationship between reciprocal
space and real space. This HRTEM image is not expected for
a regular [111] HRTEM image of the fcc crystal. Usually, the
[111] HRTEM image of the fcc crystal is built by 3 sets of the
{220} spacing with a 6-fold symmetry. For silver, the {220} d
spacing is 1.44 Å. Because this interfringe distance is beyond
the resolution of our microscope, the [111] HRTEM image of
the silver crystal will be not observable. However, for these
samples, HRTEM images built by 3 sets of 3 × {422} spacing
were frequently observed, this indicated that there is a super-
lattice caused by a bigger periodicity than a regular [111] unite
cell. We have observed that these 3 × {422} fringes have the
following features: (a) They build a perfect lattice and extend
across the entire disk, include the region with a thickness
gradient (for example, at the edge of the disks, see the Figure
3B); (b) There is not any isolated “island” in the entire image
which do not show these 3 × {422} fringes.
Figure 1G shows the selected area electron diffraction
SAED) pattern from the sample shown in Figure 1E. The
(
electron beam was perpendicular to the flat surface of the disk
lying on the TEM grid. The diffraction pattern exhibits 6 bright
and sharp spots in a 6-fold symmetry corresponding to the {220}
reflections of the fcc silver single-crystal orientated in the [111]
direction, indicating that the flat surface of the disks is parallel
5
to the (111) habit plane. In previous work, the crystallography
of the disk is described as a flat (111) face on the top with
truncated faces as shown in Figure 1H. Figure 2A shows a
HRTEM image of the [111] orientated disk (same sample as in
Figure 2) with his flat (111) plane lying on TEM grid. The image
shows a perfect lattice and exhibits 6-fold symmetry. This is
well demonstrated by the Fourier transform shown in the insert
of Figure 2A. The lattice spacing is 2.50 Å, corresponding to
the 3 × {422} lattice spacing of the fcc silver crystal which
can be built by 3 sets of 3 × {422} spacing. In the following,
we name these fringes as 3 × {422} supperlattice fringes. This
HRTEM observation is consistent with the diffraction pattern
In Figure 1H, the diffraction pattern shows 1/3{422} reflec-
tions (as marked by the box) which are formally forbidden for
the perfect fcc structure. This observation is the same as those

Letters
J. Phys. Chem. B, Vol. 107, No. 34, 2003 8719
intersect with the Ewald sphere, and the 1/3{422} reflections
will be very easy to observe in the [111] diffraction pattern. It
is impossible to distinguish between these two models because
in both cases the projected crystal potential along [111] is
identical. The model described above is for a single stacking
fault. If two or more (111) stacking faults exist within the
nanodisk, because a different stacking fault will produce the
same supperlattice periodicity, the same effect will appear in
the HRTEM image and diffraction pattern, unless there are three
stacking faults produced by intrinsic/extrinsic of A, B, and C
layers, respectively.
The change in the intensity of the 1/3{422} forbidden
reflections observed with various samples (Figure 3) is due to
the fact that the intensity of the 3 × {422} supperlattice fringes
depend on the different stacking sequence and thickness of the
disks. By adjusting the [111] stacking sequence, the projected
potential of the 3 × {422} superlattice changes and this results
in different intensities of 1/3{422} forbidden reflections.
At this step, there is still one question open: is there a link
between the stacking fault and the disk growth? We propose
that the (111) stacking fault(s) play an important role to the
formation and growth of the (111) disk. It is likely that the
growth in parallel to the stacking fault plane is the fastest, and
the existence of the stacking fault may be the key for the
formation and growth of the diskette morphology. This has been
confirmed by Figure 4A, in which the (111) stacking fault planes
are apparent. The silver nanodisks observed in six of our samples
are directly related to the presence of (111) stacking faults. A
Figure 4. (A) TEM image of the silver nanodisk taken in side view,
showing the contrast from (111) stacking faults and a preferential growth
along the stacking faults. The orientation relationship of the SAED
pattern and the nanodisk is shown in Figure 3, parts A and B. (B) A
typical SAED pattern of silver nanodisk at 100 kV in the [01 1h ]
orientation (side view).
that have been observed in previous work for these silver
5
4
nanodisks as well as in the silver nanoprism. In Figure 3, the
SAED of the 6 samples showing the same spots is presented.
However, it can be note that the intensities differ from one
sample to other. The 1/3{422} forbidden reflections on the [111]
14
similar example has been reported by other authors: The (111)
microtwins play an important role in the evolution of the
microstructure of the nano-Ge precipitates in Al-Ge and
Ag-Ge alloys; An accelerated growth along the (111) twin
plane leads to triangular or hexagonal nanoplate precipitates.
We also suggest that the (111) stacking fault model proposed
SAED pattern have been observed previously in platelike Au
and Ag nanocrystals.3
,4,9-12
Figure 4A shows the TEM images
of disks in side view. The typical [0 1h 1] SAED pattern obtained
for a beam direction parallel to the (111) disk surface is shown
in Figure 4B.
4
here could be applied to the work reported by others. That work
reported the photoinduced conversion of silver nanospheres to
nanoprisms. The silver nanoprisms were formed from the small
spherical silver particles in the colloidal suspension and grow
with increased fluorescent irradiation. The similar [111] dif-
fraction pattern with 1/3{422} forbidden reflections was reported
for their silver nanoprisms. We suggest that photo irradiation
induced (111) stacking fault in small silver particles and the
growth was accelerated along (111) stacking fault, resulting in
the formation of the triangle silver nanoprisms.
In this letter, the structural study of silver nanodisks differing
by their sizes is presented. The nanodisks have a fcc crystal
structure with (111) stacking fault(s) lying parallel to the (111)
surface and extending across entire disk. We highlight that the
A number of models for the occurrence of such forbidden
3,9-12
reflections have been previously suggested
(see Appendix
II). Model 1 can be excluded because all over the nanodisk the
same lattice parameter is observed, whereas it is not the case
3
,10,17
for a fractional unit cell along [111] direction. Models
2,
3
, and 4 involve a twin parallel to the flat surface and can be
ruled out because the SAED in the [01 1h ] direction does not
exhibit a mirror symmetry (Figure 4B) as expected for twins
parallel to the 111 face. Model 5 could explain the above
observation. However, it requires a spacing between twins in
the order of ∼1 nm. Figure 4A shows a spacing of 5 nm, a
value far away from the 1 nm required. From these data presence
of twins can be rejected as a large density of stacking fault.
Let consider two types of stacking fault. (i) Stacking fault I
(111) stacking fault is the reason for the occurrence of the
1
/3{422} forbidden reflections and the 3 × {422} superlatice
1
8
(
intrinsic): An A layer is removed and the stacking sequence
fringes. We conclude that the stacking faults may be the key
for the formation and growth of the nanodisks. Therefore, by
controlling the nucleation, growth, and elimination of the defect
is then ABCBCABCABC... In this case the projected potential
in A positions will be lower than that in B and C positions by
one atom. (ii) Stacking fault II (extrinsic): An A layer is added
that means the insertion of an extra (111) layer (for instance,
the A layer) into the regular stacking sequence of ABCAB-
CABC ..., a sequence of AB(A)CABCABC...will be caused.
In this case, the projected potential in A positions will be higher
than that in B and C positions by one atom.
(stacking fault), the shape and size controlling of the nanocrystal
will be possible. This will provide an insight to the synthetically
controlling particle shape and size of the nanocrystals.
Appendix I
For a single stacking fault, a strong supperlattice (or forbid-
den) diffraction effect could be induced, similar to the effect
caused by a single surface mono-atomic step as mentioned in
the model 1. Because of the subnanometer thickness of the
stacking fault, their reciprocal points are strongly elongated
along the [111] direction. Therefore, they are very easy to
The nanodisks are obtained using the method described by
8
Maillard and al. Briefly, a micellar solution of 60% 0.1 M
1
5
15
Ag(AOT) and 40% 0.1 M Na(AOT) (AOT) 2 ethyl-hexyl
suffossucinate) with a water content w ) [H2O]/[AOT] ) 2 is
mixed with 0.1M Na(AOT) solution. The nanocrystals are
coated with dodecanethiol (40 µl per mL) and extracted from
9

8720 J. Phys. Chem. B, Vol. 107, No. 34, 2003
Letters
the solution. The hydrazine16 content, r, defined as the ratio
by the very thin twin-like layer, and the 1/3{422} reflections
may appear in a [111] pattern.
1
[N2H4]/[AOT], varying from 4.7 to 11.8, controls the nanodisk
size.
Model 6. Surface reconstruction may results in extra reflec-
1
2
tions. However, surface reconstruction can only be easily
observed for a clean surface under ultrahigh vacuum condition.
For our sample synthesized in solution and passivated with
polymer, we can rule out the possibility.
Appendix II
A brief summary is given as follows. For easy presentation,
the fcc structure is taken as a stacking of hexagonal close-packed
(
111) layers along the [111] direction following a sequence of
References and Notes
ABCABCABC..., where A, B, and C represent the three possible
stacking positions for the (111) planes.
Model 1. For a perfect crystal with clean surfaces, mono-
atomic surface steps on the (111) surface, which caused a
fractional unit cell along the [111] direction and resulted in an
incomplete ABC stacking sequence, break the regular fcc
(1) Pileni, M. P. Nature Mater. 2003, 2, 145-150.
(2) Wang, Z. L. J. Phys. Chem. 2000, 104, 1153.
(3) Kirkland, A. I.; Jefferson, D. A.; Duff, D. G.; Edwards, P. P.;
Gameson, I.; Johnson, B. F. G.; Smith, D. J. Proc. R. Soc. London A 1993,
440, 589-609.
(4) Jin, R.; Cao, Y.W.; Mirkin, C. A.; Kelly, K. L.; Schatz, G. C.;
9
Zheng, J. G. Science 2001, 294, 1901.
5) Maillard, M.; Giorgio, S.; Pileni, M. P. AdV. Mater. 2002, 14, 15,
084.
interference pattern and give rise to the 1/3{422} reflections.
(
Model 2. For the platelets containing a single twin on the
1
1
(111) plane parallel to the upper and lower (111) surface of the
(
6) Chen, S.; Fan, Z.; Carroll, D. L. J. Phys. Chem. 2002, 106, 10777.
plate, if the number of (111) layers is not equal to 3n (n is an
integer), the 1/3{422} forbidden reflections will appear; other-
wise not. (In this model the (111) twin boundary is not in the
center of the crystal.)3
Model 3. The effect on HRTEM image and diffraction pattern
caused by a twin in the fcc bicrystalline silicon nanowires has
been reported. When the twin boundary lies parallel to the
surface and perpendicular to electron beam and the twin
boundary is in the center of the crystal (central twin), the
(7) Pastoriza-Santos, I.; Liz-Marzan, M. L. Nano Lett. 2002, 2, 903.
(
8) Maillard, M.; Giorgio, S.; Pileni, M. P. J. Phys. Chem. 2003, 107,
1, 1466-2470.
9) Cherns, D. Philos. Mag. 1974, 30, 549.
10) Kirkland, A. I.; Jefferson, D. A.; Duff, D. G.; Edwards, P. P. Inst.
Phys. Conf. Ser. 1989, 98, 375.
(
,10
(
(
55
(
11) Tanaka, N.; Cowley, J. M. Mater. Res. Soc. Symp. Proc. 1985, 41,
1
1
7
12) Heyraud, J. C.; M e´ tois, J. J. Surf. Sci. 1980, 100, 519.
(13) TEM images at moderate resolution and SAED patterns were
recorded using a Hitachi HF-2000 field emission gun (FEG) transmission
electron microscope (TEM) operating at 200 kV; and HRTEM was carried
out using a JEOL 4000EX microscope operating at 400 kV.
1
/3{422} forbidden reflections and 3 × {422} fringes appear.
Model 4. Two or more parallel twins on (111) planes which
(14) Dahmen, U.; Hethering, C. J. D.; Radmilovic, V.; Johnson, E. E.
are parallel to the upper and lower surface will always cause
J.; Xiao, S. Q.; Luo, C. P. Proc. 15th Int. Congr. Electron. Microsc. 2002,
131.
(15) Petit, C.; Lixon, P.; Pileni, M. P. J. Phys. Chem. 1993, 97, 12974.
the 1/3{422} forbidden reflections, regardless the number of
the (111) layers.3,10
(16) Sodium di(2-ethyl-hexyl)sulfoccinate Na(AOT), isooctane, hydra-
Model 5. High-density stacking faults on (111) plane could
form microtwins. For example, the insertion of one extra (111)
layer (such as the B layer) causes a stacking sequence of ABC-
zine and dodecanethiol were from Sigma, Fluka, Prolabo (France) and
Janssen Chemicals, respectively. The products were used without further
purification.
(
17) Carim, A. H.; Lew, K. K.; Redwing, J. M. AdV. Mater. 2001, 1489.
(B)ABCABC.... Here, the ABC(B)A is a twin-like thin slide.
(18) Hirsch, P.; Howie, A.; Nicholson, R. B.; Pashley, D. W.; Whelan,
In this case, the reciprocal lattice points in the first-order Laue
zone will be elongated because of the shape factor introduced
M. J. Electron Microscopy of Thin Crystals; Robert E. Krieger Publishing
Company: New York, 1977.