Size effects in the oriented-attachment growth process: The case of cu nanoseeds

Article abstract of DOI:10.1021/ic900394z

Size effects in the oriented-attachment (OA) growth process of Cu nanoseeds were found. Monodispersed Cu nanoseeds with average diameters of 2.2, 3.4, and 5.2 nm were controllably synthesized by the reduction of copper acetate in a boiling solvent and usi

Full text of DOI:10.1021/ic900394z

Inorg. Chem. 2009, 48, 5117–5128 5117
DOI: 10.1021/ic900394z
Size Effects in the Oriented-Attachment Growth Process: The Case
of Cu Nanoseeds
Shuling Shen, Jing Zhuang, Xiangxing Xu, Amjad Nisar, Shi Hu, and Xun Wang*
Department of Chemistry, Tsinghua University, Beijing, 100084, P. R. China
Received December 18, 2008
Size effects in the oriented-attachment (OA) growth process of Cu nanoseeds were found. Monodispersed Cu
nanoseeds with average diameters of 2.2, 3.4, and 5.2 nm were controllably synthesized by the reduction of copper
acetate in a boiling solvent and using dodecanethiol (DT) as a stabilizer and sulfur source of sulfide. These Cu
nanoseeds were then treated under solvothermal conditions. When the diameters of Cu nanoseeds were smaller than
5 nm, Cu2S nanorods with lengths of ∼30-100 nm and diameters of ∼2-4 nm were obtained at lower temperatures,
and Cu2S nanodisks with diameters of ∼6-13 nm and thicknesses of ∼2-4 nm were obtained at higher
temperatures. Once the diameter of Cu nanoseeds was larger than 5 nm, only irregular particles were obtained,
regardless of other conditions. The uniformity, which related to the density of DT on the surface of Cu nanoseeds, was
the key for success of self-assembly of the final nanocrystals. High-resolution transmission electron microscopy
images demonstrated that these nanorods, nanodisks, and particles were formed by an OA process of Cu nanoseeds
into 1D, 2D, and 3D aggregates, which recrystallized into single crystals.
Introduction
even macrocrystals, in that it is based on the separation of
nucleationand the subsequentgrowth process tosome extent,
which are usually quite difficult to discern in the normal
crystal growth process. In this paper, we have synthesized Cu
nanoseeds with diameters of 2, 3, and 5 nm first; then, with
these nanoseeds as building blocks, OA processes were
carried out under solvothermal conditions. These OA pro-
cesses have proved to be size- and surface-dependent, which
clearly indicates that the size and surface properties of the
nanoseeds will determine the subsequent crystal growth
process. This finding may serve as a new and general principle
for the further investigation of crystal growth processes.
Classically, Ostwald ripening has been used extensively to
describe and explain the growth of large crystals at the expense
of smaller ones, driven by surface energy reduction.3 Usually,
the products grown by Ostwald ripening are nearly spherical
morphologies and thermodynamically stable because of the
minimization of the overall surface energy. However, the gro-
wth of particles strongly depends on the structure of the
material, the properties of the solution, capping ligands, the
temperature, and the interface between the particles and the
surrounding environment. Recent works have shown that
Ostwald ripening does not adequately describe the crystal gro-
wth for many systems, especially for the anisotropic growth of
1D nanorods or nanowires and 2D nanodisks or nanoplates.4
Well-defined nanocrystals have attracted considerable
attention in recent years due to their unique size- and
shape-dependent properties and potential applications in
assembling advanced materials and devices.1 One of the main
interesting and notable advances in this field is the asse-
mbly of nanocrystal building blocks into single-crystal 1D
nanowires, 2D nanosheets, and other complex architectures
directly via an oriented-attachment (OA) growth process.2
While being effective in generating novel nanostructures, this
OA process may also provide a deeper understanding
of the growth process of nano- and microcrystals and
*To whom correspondence should be addressed. E-mail: wangxun@
mail.tsinghua.edu.cn.
(1) (a) Bruchez, M.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P.
Science. 1998, 281, 2013–2016. (b) El-Sayed, M. A. Acc. Chem. Res. 2004, 37,
326–333. (c) Li, J. J.; Wang, Y. A.; Guo, W. Z.; Keay, J. C.; Mishima, T. D.;
Johnson, M. B.; Peng, X. G. J. Am. Chem. Soc. 2003, 125, 12567–12575. (d)
Hyeon, T.; Lee, S. S.; Park, J.; Chung, Y.; Na, H. B. J. Am. Chem. Soc. 2001,
123, 12798–12801. (e) Collier, C. P.; Vossmeyer, T.; Heath, J. R. Annu. Rev.
Phys. Chem. 1998, 49, 371–404. (f ) Guo, Q.; Kim, S. J.; Kar, M.; Shafarman,
W. N.; Birkmire, R. W.; Stach, E. A.; Agrawal, R.; Hillhouse, H. W. Nano
Lett. 2008, 8, 2982–2987. (g) Alivistos, A. P.; Gu, W.; Larabell, C. Annu. Rev.
Biomed. Eng. 2005, 7, 55–76. (h) Murray, C. B.; Sun, S.; Doyle, H.; Betley, T.
Mater. Res. Soc. Bull. 2001, 26, 985–991. (i) Wang, X.; Zhuang, J; Peng, Q.;
Li, Y. D. Nature 2005, 437, 121–124. (j) Shevchenko, E. V.; Ringler, M.;
Schwemer, A.; Talapin, D. V.; Klar, T. A.; Rogach, A. L.; Feldmann, J.;
Alivisatos, A. P. J. Am. Chem. Soc. 2008, 130, 3274–3275.
(3) Lifshitz, I. M.; Slyozov, V. V. J. Phys. Chem. Solids 1961, 19, 35–50.
(4) (a) Pradhan, N.; Efrima, S. J. Phys. Chem. B 2004, 108, 11964–11970.
(b) Nguyen, T. D.; Mrabet, D.; Do, T. O. J. Phys. Chem. C 2008, 112, 15226–
15235. (c) Lilly, G. D.; Lee, J.; Sun, Kai.; Tang, Z. Y.; Kim, K. S.; Kotov, N.
A. J. Phys. Chem. C 2008, 112, 370–377.
(2) (a) Yong, K. T.; Sahoo, Y.; Zeng, H.; Swihart, M. T.; Minter, J. R.;
Prasad, P. N. Chem. Mater. 2007, 19, 4108–4110. (b) Tang, H.; Chang, J. C.;
Shan, Y.; Lee, S. T. J. Phys. Chem. B 2008, 112, 4016–4021. (c) Liu, B.; Zeng,
H. C. J. Am. Chem. Soc. 2004, 126, 8124–8125.
r
2009 American Chemical Society
Published on Web 05/04/2009
pubs.acs.org/IC

5118 Inorganic Chemistry, Vol. 48, No. 12, 2009
Shen et al.
As a helpful complement, the view “oriented attachment”,
demonstrated by Penn and Banfield5 for the first time,
has been highlighted in many recent works. The growth of
gold nanowires,6 ZnO nanorods,7 ZnS nanorods,8 PbSe
nanowires and nanorings,9 CdTe 1D nanowires,10 and 2D
nanosheets11 are all well-explained by the OA mechanism,
which provides an efficient method for the preparation of
nanocrystals with unique morphologies and controllable sizes.
For the controlled OA process of nanocrystals into 1D or
higher-dimensional nanostructures, surface capping ligands
usually play a critical role. Polleux et al.12 demonstrated a
ligand-directed assembly of nanoparticles into anisotropic
TiO2 nanocrystals. Tang et al.10 reported a similar process
for CdTe nanowires. They demonstrated that the removal of
excess stabilizer is the key step in the preparation of nanowires
from nanoparticles. But current results show that the orienta-
tion is maintained for all nanowires produced, regardless of the
initial particle size. In this article, we report the solvothermal-
based synthesis of crystalline Cu2S nanorods and nonodisks
using the OA growth process of Cu nanoseeds. Size effects of
the Cu nanoseeds in morphology evolution during the OA
process were investigated for the first time. At the same time,
the OA growth characteristics determined by the surface
ligand were also discussed and compared.
colloid solution cooled to room temperature, 36 μL of DT was
injected to get ∼5 nm Cu-DT nanoseeds. After the system was
cooled to room temperature, 30 mL of ethanol was added to
precipitate Cu-DT nanoseeds with centrifugation at 10 000 rpm
for 5 min. The precipitate was redispersed in toluene for the
subsequent synthesis.
Nanorods and Rodlike Superlattice of Nanodisks. A total
of 1 mL of OLA was added to 5 mL of toluene containing 9.6 mg
of Cu nanoseeds with stirring. Then, the whole mixture was
transferred into a Teflon-lined autoclave of 10 mL capacity and
heated at 140∼200 °C. The precipitate was washed with ethanol
and separated by centrifugation several times to remove residual
impurities.
Characterization. The size, morphology, and superstruc-
ture of the nanocrystals were probed by transmission electron
microscopy (TEM; JEOL JEM 1200EX working at 100 kV),
high-resolution transmission electron microscopy (HRTEM;
Tecnai G2 F20 S-Twin working at 200 kV), and scanning
electron microscopy (SEM; JEOL JSM-6300F working at
15 kV), using a microscope equipped with an X-ray energy
dispersive spectrometer (EDS). TEM samples were prepared as
follows: A small amount of desired product was dispersed in
toluene. One drop of the resulting suspension was deposited on
carbon-coated Cu grids, and the solvent was then evaporated
at room temperature in the air. UV-vis absorption spectra were
measured on a Hitachi U-3010 spectrophotometer with a 1-cm
quartz cuvette. Solutions were prepared typically at a concen-
tration of 0.1 mg/mL in toluene. FT-IR spectroscopy was
performed using a Nicolet 360 spectrograph with the pressed
KBr pellet technique. X-ray photoelectron spectra (XPS) were
recorded on a PHI Quantera SXM spectrometer using mono-
chromatic Al KR X-ray sources (1486.6 eV) at 2.0 kV and
20 mA. All XPS data were acquired at a nominal photoelectron
takeoff angle of 45°. The binding energies of the peaks obtained
were made with reference to the binding energy of the C1s
line, set at 284.8 eV. The spectrum was fitted using an 80% linear
combination of Gaussian-Lorentzian profiles. Powder X-ray
diffraction (XRD) measurement of the products was carried
out on a Bruker D8 Advance X-ray powder diffractometer
Experimental Section
Materials.
Dioctadecyldimethylammonium
bromide
(DODA, 99%) was purchased from Acros. The rest of the
chemicals were purchased from a Beijing chemical reagent
company. All chemicals were of analytical grade and were used
as received without further purification. Deionized water was
used throughout.
The 2 and 3 nm Cu nanoseeds. In a typical synthesis, 30 mg
of cupric acetate and 120 mg of DODA were added into 10 mL
of toluene, and the mixture was heated up to boiling under
continuous magnetic stirring to form a dark green solution. An
aqueous NaBH4 solution (36 μL, 9.4M) was added under
vigorous stirring. The dark green solution turned dark brown
within a minute. A total of 36 μL of dodecanethiol (DT) was
added, and the mixture was stirred at room temperature for
30 min (for ∼2 nm Cu nanoseeds) and 2 h (for ∼3 nm Cu
nanoseeds). In order to prevent the oxidation of copper, all of
the synthetic steps were carried out in a nitrogen atmosphere.
Ethanol (30 mL) was added to precipitate Cu nanoseeds coated
by DT (Cu-DT) with centrifugation at 10 000 rpm for 5 min.
The precipitate was redispersed in toluene for the subsequent
synthesis.
1
˚
with Cu KR radiation (λ = 1.5418 A). H NMR spectra were
recorded using a JEOL ECA-600 spectrometer. Solutions
were prepared in toluene-d8, and measurements were conducted
at 25 °C.
Results and Ddiscussion
Synthesis of Cu-DT Nanoseeds with Diameters of 2, 3,
and 5 nm. The size and size distribution of highly crystal-
line Cu nanoseeds is the key for the subsequent investi-
gation of the OA process. In previous reports, many
methods have been developed for the synthesis of mon-
idisperse Cu nanoparticles including microemulsion
techniques,13 the polyol method,14 metal salt reduc-
tion,15 electrochemical reduction,16 and the solvothermal
method.17 However, few methods have been established
to control the size and size distribution effectively,
The 5 nm Cu nanoseeds15. A total of 30 mg of cupric acetate
and 0.2 mL of oleic acid (OLA) were dissolved in 5 mL of toluene
by heating to boiling under a nitrogen atmosphere. A total of
5 mL of N2H4 and 1 mL of OLA dissolved in 5 mL of toluene
were injected, and boiling was continued for 2 h. Before the
(5) (a) Penn, R. L.; Banfield, J. F. Geochim. Cosmochim. Acta 1999, 63,
1549–1557. (b) Banfield, J. F.; Welch, S. A.; Zhang, H. Z.; Ebert, T. T.; Penn,
R. L. Science 2000, 289, 751–754.
(13) (a) Lisiecki, I.; Pileni, M. P. J. Am. Chem. Soc. 1993, 115, 3887–3896.
(b) Lisiecki, I.; Bjorling, M.; Motte, L.; Ninham, B.; Pileni, M. P. Langmuir
1995, 11, 2385–2392. (c) Pileni, M. P. J. Phys. Chem. C 2007, 111, 9019–9038.
(14) Park, B. K.; Jeong, S.; Kim, D.; Moon, J.; Lim, S.; Kim, J. S. J.
Colloid Interface Sci. 2007, 311, 417–424.
(15) (a) Jana, N. R.; Peng, X. G. J. Am. Chem. Soc. 2003, 125, 14280–
14281. (b) Wu, S. H.; Chen, D. H. J. Colloid Interface Sci. 2004, 273, 165–
169.
(6) Halder, A.; Ravishankar, N. Adv. Mater. 2007, 19, 1854–1858.
(7) Pacholski, C.; Kornowski, A.; Weller, H. Angew. Chem., Int. Ed. 2002,
41, 1188–1191.
(8) Yu, J. H.; Joo, J.; Park, H. M.; Baik, S.-I.; Kim, Y. W.; Kim, S. C.;
Hyeon, T. J. Am. Chem. Soc. 2005, 127, 5662–5670.
(9) Cho, K. S.; Talapin, D. V.; Gaschler, W.; Murray, C. B. J. Am. Chem.
Soc. 2005, 127, 7140–7147.
(10) Tang, Z. Y.; Kotov, N. A.; Giersig, M. Science 2002, 297, 237–240.
(11) Tang, Z. Y.; Zhang, Z. L.; Wang, Y.; Glotzer, S. C.; Kotov, N. A.
Science 2006, 314, 274–278.
(16) Han, W. K.; Choi, J. W.; Hwang, G. H.; Hong, S. J.; Lee, J. S.; Kang,
S. G. Appl. Surf. Sci. 2006, 252, 2832–2838.
(12) Polleux, J.; Pinna, N.; Antonietti, M.; Niederberger, M. Adv. Mater.
2004, 16, 436–439.
(17) Mott, D.; Galkowski, J.; Wang, L. Y.; Luo, J.; Zhong, C. J. Langmuir
2007, 23, 5740–5745.

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5119
Figure 1. TEM images of Cu nanoseeds with different sizes: (a) 2.2 ( 0.1 nm, (b) 3.4 ( 0.1 nm, and (c) 5.2 ( 0.2 nm. The insets are the HRTEM images and
the corresponding FFT patterns. (d) Corresponding UV-vis absorption spectra.
because the agglomeration and chaotic growth of Cu
particles are often substantial, which leads to the irregular
shape and big size. Here, two single-phase approaches
were used to precisely control the diameter of Cu nano-
seeds. First, the synthesis of 1-4 nm Cu nanoseeds
was carried out in a boiling toluene solution containing
cationic surfactant (DODA) by using NaBH4 as a redu-
cing reagent. The as-prepared Cu nanocrystals were
then coated with DT through ligand exchange. Second,
according to a procedure previously developed by Peng
and Jana,15a Cu nanoseeds about 5 nm in diameter were
synthesized in a boiling toluene solution with a fatty acid
(OLA) as the surfactant and hydrazine as the reducing
reagent. The as-prepared Cu nanocrystals were also
coated with DT through ligand exchange.
Figure 1 illustrates the representative TEM images of
the obtained Cu nanoseeds. The Cu nanoseeds are all
well-dispersed and have a highly uniform size. The lattice
spacing in the HRTEM images (inset, Figure 1) of 0.206-
0.209 nm calculated by FFT patterns is consistent with
the distance of (111) lattice spacing in Cu, revealing the
cubic and single-crystalline nature of the nanoseeds.
Statistical analysis of the size of the nanoseeds, obtained
from counting ∼100 nanoseeds from different parts of the
grids, indicates that the nanoseeds have an average size of
2.2 ( 0.1 nm, 3.4 ( 0.1 nm, and 5.2 ( 0.2 nm, respectively.
Figure 1d shows the UV-visible absorption spectra of the
obtained Cu nanoseeds. The surface plasmon peak of
small copper nanoseeds (<4 nm) is strongly broad-
ening due to the predominant quantum size effects.
When the diameter is larger than 4 nm (5.2 nm), the
spectrum exhibits an absorption band at around
∼575 nm, which is the typical surface plasmon band of
copper nanocrystals.18
The size control of Cu nanoseeds was achieved by using
DODA or OLA as a capping agent during the formation
of the Cu nucleus. The interaction between the outer
organic shell and the inner Cu core of Cu-DODA is
stronger than that of Cu-OLA, which permits the remain-
ing monomers in the solution to continuously nucleate
and, consequently, results in samples full of smaller
(18) Chen, T.Yu .; Chen, S. F.; Sheu, H. S.; Yeh, C. S. J. Phys. Chem. B
2002, 106, 9717–9722.

5120 Inorganic Chemistry, Vol. 48, No. 12, 2009
Shen et al.
Table 1. Mode Assignments of DT, OLA, DODA, (Cu-OLA)-DT, and
(Cu-DODA)-DT
peak
assignmenta DT OLA DODA (Cu-OLA)-DT (Cu-DODA)-DT
ν
ν
sym(C-H) 2852 2854 2849
asym(C-H) 2922 2925 2916
2852
2932
1465
1377
720
2852
2923
1465
1377
721
δs(C-H)
1465 1465 1469
1376 1377 1374
Fr(CH2)n
ν(S-H)
ν(C-S)
ν(CdO)
ν(CdC)
721 723
2570
654
719
652
653
1709
1412
ν(C-N+
)
1079
1055
982
893
ν(CH3-N+
)
1489
a ν = stretching, sym =symmetric, asym =antisymmetric, δs
methylene scissoring, Fr = rocking.
=
Figure 2. FTIR spectra of (a) DT, (b) OLA, (c) DODA, (d) (Cu-OLA)-
DT, and (e) (Cu-DODA)-DT.
Figure 3. XPS spectra of 2, 3, and 5 nm Cu nanoseeds.
particles. The role of DODA and OLA is to control the
growth of the Cu nucleus to different extents, and the final
stabilizer is DT. DT was chosen as the ultimate stabilizer
because it is not only a strong stabilizer capable of
preventing the Cu nanoseeds from oxidation and aggre-
gation, but also the sulfur source for the preparation of
copper sulfide. DODA and OLA are weaker ligands
compared to DT, and the ligand exchange can be
achieved straight-away by adding DT into the system
containing the as-synthesized Cu nanocrystals coated by
DODA or OLA.
More evidence can be obtained from the FTIR spectra
of the Cu nanocrystals after ligand exchange by DT,
and various peak assignments have been compared in
Table 1. (Cu-OLA)-DT and (Cu-DODA)-DT repre-
sent Cu nanocrystals synthesized in the presence of
OLA and DODA, which are finally exchanged by DT,
respectively. Figure 2d indicates that, after ligand ex-
change, the CdO stretching vibration at 1709 cm-1
disappeared. The CH3-N+ stretching mode at 1489
cm-1 and the stretching modes of C-N+ at 1079 cm-1
,
982 cm-1, and 893 cm-1 were also absent after ligand

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5121
Figure 4. (a, b) HRTEM images of Cu2S nanorods (2.2 nm Cu nanoseeds, 150 °C for 12), (c) HRTEM image of Cu2S nanorods (3.4 nm Cu nanoseeds,
150 °C for 12), (d) TEM image of Cu2S particles (5.2 nm Cu nanoseeds, 140 °C for 12 h). Insets are FFT patterns.
exchange (Figure 2e). The stretching vibration of C-S at
600-700 cm-1 can be found in Figure 2d and e. This
indicates that almost all of the OLA and DODA were
exchanged by DT. The absence of S-H stretching
vibrations at 2550-2600 cm-1 in (Cu-DODA)-DT and
(Cu-OLA)-DT implies that S-H is destroyed and S
anchors on the surface of Cu nanoseeds with strong
interaction.
Detailed information about the surface molecular
and electronic structure of the obtained Cu nanoseeds
was provided by the XPS analysis. Figure 3 shows the
XPS spectra of Cu2p, O1s, C1s, and S2p of the copper
surface covered with a monolayer of DT. For all samples,
the XPS spectra are similar. The Cu2p3/2 and Cu2p1/2
are detected at 932.8 and 952.6 eV, respectively. The
absence of a series of satellite structures of Cu2p3/2
on the high-binding-energy side (934∼943 eV) pre-
cludes significant oxidation of the copper nanoseeds in
the form of Cu2+ 19
. The insignificant O1s peaks at
∼531ev indicate that the surface of copper nanoseeds is
slightly oxidized in the form of Cu2O. The intensity of O1s
increases slightly with the decrease of Cu size due to
the dramatic increasing in surface activity. A single
S2p doublet at 162.1 eV (S2p3/2) is also observed. The
binding energies of these doublets are consistent with
thiolate species bonded to the surfaces of metal nano-
crystals.20 These clearly indicate that DT absorbed on
the surface of Cu nanoseeds through chemisorption,
which prevents the oxidation of the copper surface
effectively.
(19) Ghijsen, J.; Tjeng, L. H.; Elp, J. v.; Eskes, H.; Westerink, J.;
Sawatzky, G. A. Phys. Rev. B 1988, 38, 11322–11330.
(20) (a) Castner, D. G.; Hinds, K.; Grainger, D. W. Langmuir 1996, 12,
5083–5086. (b) Vollmer, S.; Witte, G.; Woll, C. Langmuir 2001, 17, 7560–
7565.

5122 Inorganic Chemistry, Vol. 48, No. 12, 2009
Shen et al.
Figure 5. (a) TEM image of rodlike Cu2S superlattice (2.2 nm Cu nanoseeds, 180 °C for 16 h), (b) HRTEM images of unstacked Cu2S nanodisks, and (c)
the edge of rodlike Cu2S superlattice. (d) Cu2S nanodisks (3.4 nm Cu nanoseeds, 180 °C for 16 h). (e) Cu2S particles (5.2 nm Cu nanoseeds, 180 °C for 16 h).
i-iii are the FFT patterns of corresponding regions.
Size Effects in OA Growth Process of Cu Nanoseeds.
TEM investigations were performed to monitor the
structural development of Cu nanoseeds. As shown in
Figure 4a, the products synthesized from 2.2 nm Cu
nanoseeds at 150 °C are composed of many local ordered
serpentine nanorods. The HRTEM images (Figure 4b
and Figure SI-1a,b in the Supporting Information) show
that the nanorods preferentially align with their long
axes parallel to each other into separate close-packed
“stacks”, which is similar to the nematic phase observed

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5123
Figure 6. TEM and HRTEM images of (a, b) nanocrystal chains (3.4 nm Cu nanoseeds, 140 °C for 8 h), (c, d) disks (3.4 nm Cu nanoseeds, 180 °C for 8 h),
and (e, f) particles (5.2 nm Cu nanoseeds, 140 °C for 12 h).
for rodlike colloids in high-volume fractions.21 The
nanorods are single-crystalline Cu2S in nature, about
∼30-100 nm in longitudinal size and 2.2 nm in lateral
size. The latter is very close to the diameter of Cu
nanoseeds. A similar result has been observed by Tang
and co-workers.10 They reported the self-organization
of single CdTe nanoparticles into nanowires controlled
(21) Gelbart, W. M.; Ben-Shaul, A. J. Phys. Chem. 1996, 100, 13169–
13189.

5124 Inorganic Chemistry, Vol. 48, No. 12, 2009
Shen et al.
and a diameter of 400 nm were obtained from 2.2 nm
Cu nanoseeds (Figure SI-2a in the Supporting Infor-
mation). The TEM and HRTEM images reveal that
the as-obtained three-dimensional superlattices are
composed of uniform nanodisks with a diameter of
6.5 nm and a thickness of 2.3 nm (Figure 5b), which
stack face-to-face perpendicular to the substrate, and
the distance between two adjacent nanodisks is about
2.1 nm (Figure 5c). The distance of 2.1 nm, which is
shorter than twice the length of the alkyl chains of
DT (1.77 nm),22 suggests that the DT chains are inter-
digitated during the superlattice formation. In addition
to the rodlike superstructures, a few unassembled nano-
disks remain to be isolated. Furthermore, the edgewise
nanodisks are almost never observed isolated but are
usually stacked in lines and made up of a few lying
nanodisks sharing their c axis (Figure 5b). The plane
distances of edgewise nanodisks and lying nanodisks
were estimated to be 0.34, 0.24, and 0.199 nm, respec-
tively, which correspond to the (002), (102), and (110)
planes of hexagonal Cu2S nanocrystals. The EDS spec-
trum (Figure SI-2b in the Supporting Information) con-
firms that the nanodisks consist of sulfur and copper.
Surprisingly, no rodlike superlattices formed, and
only unstacked nanodisks (Figure 5d) and irregular
particles (Figure 5e) can be obtained by using 3.4 and
5.2 nm Cu nanoseeds, respectively. The above results
further confirm the size effects in the growth of Cu
nanoseeds.
In fact, the presence of defects in the nanorods and
the identical diameter of the nanorods and nanoseeds
are characteristic of the OA process. In order to further
clarify the formation mechanism of nanorods, nano-
disks, and particles, detailed morphological and struc-
tural analyses for the early stage of nanorod, nanodisk,
and particle formation with different sizes of Cu nano-
seeds are shown in Figure 6. As can be seen from
Figure 6a, most nanoseeds are self-organized into neck-
lace chains at 140 °C after 8 h from 3.4 nm Cu nanoseeds.
The HRTEM image shows that the multimer chain was
formed by the alignment of 3.4 nm Cu nanoseeds along
the (111) direction. This indication confirms that growth
occurred not by monomer deposition but by a coales-
cence mechanism. It is also worth noting that the multi-
mer chain displays some dislocations (white arrows) and
branches (black arrows) in the regions where coalescence
occurs. Figure 6c and d show the TEM and HRTEM
images of nanodisks at 180 °C for 8 h from 3.4 nm Cu
nanoseeds. It can be seen that at this stage the disks are
approximately trigonal, suggesting that, in the early
stage of reaction, the nanodisks were not grown and
crystallized well. Figure 6e and f illustrate the typical
morphologies of the agglomerate by the disordered
coalescence of 5.2 nm Cu nanoseeds at 140 °C for
12 h, which finally recrystallized into irregular particles
(Figure 5e).
Figure 7. (a) XRD patterns of (i) 3.4 nm Cu nanoseeds, (ii) Cu nano-
crystal chains (140 °C), (iii) Cu2S nanorods (150 °C), and (iv) Cu2S
nanodisks (180 °C). The chains, nanorods, and nanodisks are all prepared
by using the 3.4 nm Cu nanoseeds. (b) UV-vis absorption spectra of Cu
chains, Cu2S nanorods, and Cu2S nanodisks.
by strong dipole-dipole interactions and demonstrated
that the diameters of the nanowires are virtually identical
to the diameters of precursor nanoparticles. As expected,
nanorods of about 3.4 nm in width were obtained for
3.4 nm Cu nanoseeds grown under the same conditions.
However, more defects (white arrows) and branches
(black arrows) can be observed obviously in the nanor-
ods, which makes the rods look intermittent (Figure 4c
and Figure SI-1c in the Supporting Information). When
the size of the Cu nanoseeds increases up to 5.2 nm,
the nanorod is never so much as observed, and only
irregular particles with a diameter of 10-20 nm can be
obtained (Figure 4d). There are still many small nano-
particles distributed around the bigger particles. This
suggests that aggregates of smaller Cu nanoseeds are
more ordered and have fewer defects than bigger ones.
An important conclusion can be made on the basis of the
above results that the morphology of the final Cu2S
nanocrystals was determined by the size of the initial Cu
nanoseeds.
XRD measurements (Figure 7a) reveal the occurrence
of structural and componential transitions along with the
morphology evolution. For the nanoseeds with diameters
of 3.4 nm, XRD measurement indicates that they are
Cu with a cubic phase, which is very consistent with
When the growth temperature was elevated (180-
200 °C), rodlike superlattices with a length of 2-4 μm
(22) Taleb, A.; Petit, C.; Pileni, M. P. Chem. Mater. 1997, 9, 950–959.

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5125
Figure 8. Scheme of size effects in the oriented attachment growth process of Cu-DT nanoseeds.
the HRTEM result. Due to the low intensity of O1s
(Figure 3), the characteristic peaks of Cu2O were not
detected in XRD measurements. After an OA process at a
temperature of 140 °C, necklace structures can be ob-
tained, combining with the HRTEM analysis, which can
be indexed to a cubic Cu phase, though the characteristic
peaks of Cu2S emerge in XRD. After treatment at ele-
vated temperatures (150-180 °C), the obtained nanorods
or nanodisks can be indexed as hexagonal Cu2S (JCPDS
card No.84-0208), indicating a complete transformation
of Cu into Cu2S. For 2.2 and 5.2 nm nanoseeds, similar
componential transitions can be observed. The UV-vis
adsorption spectra also monitored the optical behavior
from Cu-DT to Cu2S (Figure 7b). The Cu nanochains
exhibit the typical adsorption peak of elongated Cu
nanocrystals (564 nm).23 Cu2S nanorods and nanodisks
exhibit significant blue-shift peaks (228 and 231 nm,
respectively) in the UV-vis absorption spectrum com-
pared to that of the bulk phase (1022 nm),24 which shows
a quantum size effect. The similar adsorption peaks of
Cu2S nanodisks and nanorods were possibly induced by
the identical diameter of nanodisks and nanorods. When
the XPS, HRTEM, and XRD results are combined, it can
be concluded that DT as the sulfur source of sulfide
anchored on the surface of Cu nanoseeds by chemisorp-
tion, and then the Cu-DT nanoseeds self-organized into
1D, 2D, or 3D aggregates, which is determined by the size
and surface properties of the nanoseeds, and gradually
transferred into Cu2S through controlling solvothermal
conditions.
The total interaction potential (WT) is expressed as
the sum of WL and Wvdw
:
WT ¼ Wvdw þ WL
ð1Þ
If the repulsion is weaker than the attraction (WT < 0),
nanoseeds aggregate. If the repulsion is stronger than the
attraction (WT > 0), nanoseeds disperse.
The energy of the steric repulsion (WL) between two
Cu-DT nanoseeds can be expressed by26
ꢀ
ꢁ
100Rδ2
ðC -RÞπσ3thiol
-πðC -RÞ
WL≈
kT exp
ð2Þ
δ
where δ is the thickness of the thiol monolayer, T is the
temperature, σthiol is the diameter of the area occupied by
the thiol on the particle surface, R is the diameter of the
nanoseed and C is the distance between the centers of the
two nanoseeds. From eq 2, we can see that the higher
packing density (i.e., smaller σthiol) of DT should correlate
with a stronger steric repulsion. On the basis of the above
analysis, DT may be dominant and adhere on the (111)
facets of Cu nanoseeds. Passivation of the (111) facets
lowers the energy of these facets. When the Cu nanoseeds
are transferred from the nucleation solution into the
solvothermal system, the facets coated with less DT be-
come “defects”, which provide the junctures for the attach-
ment of adjacent nanoseeds driven by van der Waals
interaction. Generally, surface energy increases dramati-
cally with the decrease of nanoseed diameter due to the
obvious increase in the surface-to-volume ratio, which will
increase the adsorption of DT on other facets. This is why
smaller nanoseeds with less “defects” tend to attach into
chain structures. Comparatively, there are many more
“defects” distributed on the surface of bigger Cu nano-
seeds, and van der Waals forces are known to increase
with crystal size, so 2D or 3D structures are easy to obtain.
Increasing the solvothermal treatment temperature
According to the Derjaguin-Landau-Verwey-Over-
beek model,25 the van der Waals interaction (Wvdw
)
and ligand steric repulsion (WL), which originated
from the steric repulsion of surface-adsorbed DT, are
believed to be the dominant forces for the formation
of 1D, 2D, and 3D aggregates of Cu-DT nanoseeds.
(23) Lisiecki, I.; Billoudet, F.; Pileni, M. P. J. Phys. Chem. 1996, 100,
4160–4166.
(26) (a) de Gennes, P. G. Adv. Colloid Interface Sci. 1987, 27, 189–209. (b)
Korgel, B. A.; Fullam, S.; Connolly, S.; Fitzmaurice, D. J. Phys. Chem. B
1998, 102, 8379–8388. (c) Zhang, S. H.; Leem, G.; Srisombat, L. T.; Lee, R. J.
Am. Chem. Soc. 2008, 130, 113–120.
(24) Haram, S. K.; Mahadeshwar, A. R.; Dixit, S. G. J. Phys. Chem. 1996,
100, 5868–5873.
(25) Sinyagin, A. Y.; Belov, A.; Tang, Z. Y.; Kotov, N. A. J. Phys. Chem.
B 2006, 110, 7500–7507.

5126 Inorganic Chemistry, Vol. 48, No. 12, 2009
Shen et al.
Figure 9. TEMand HRTEMimagesofCu2S nanocrystalspreparedatdifferentDTconcentrations using3.4nmCu nanoseeds at180°C. (a-c)3Drodlike
superlattices (0.03% DT). Inset is FFT pattern. (d, e) 2D superlattices (0.12% DT).
quickens the decompositionofcopper thiolate; inaddition,
DT adhered on the (111) facets will be desorbed, which
leads to more “defects” appearing. This explains why the
smaller nanoseeds can also attach to 2D disks at higher
temperatures(Figure8). Furthermore, for the samesizeCu
nanoseeds, more “defects” on the surface will result in
nonuniform nanocrysals, which can be attested to by the
following self-assembly behaviors of Cu2S nanodisks.
Surface-Determined OA Process of Cu Nanoseeds.
The formation of Cu2S nanodisks considerably enhances
the long-range dipolar forces compared with isolated
Cu nanocrystals, leading to an anisotropic organization.
However, the self-assembly of nanocrystals depends
to a great extent on homogenization of the building
blocks. The uniformity of the unstacked nanodisks
(Figure 5d) can be ameliorated by introducing a correct

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5127
amount of DT to reduce the “defects” (decrease of σthiol
)
during the solvothermal treatment of 3.4 nm Cu
nanoseeds. Figure 9 displays the TEM photographs
of the Cu2S nanocrystals from 3.4 nm Cu nanoseeds
at different DT concentrations. When the concentra-
tion of DT is 0.03%, three-dimensional rodlike super-
lattices with a length of 2-4 μm and a diameter of
400-500 nm can be obtained as shown in Figure 9a-c.
The obtained 3D superlattices are composed of face-
to-face stacked nanodisks with a diameter of 13 nm
and a thickness of 3.5 nm (Figure 9b, c). The distance
between two adjacent nanodisks is also about 2 nm. The
similarity in distances shows that the energy balance
is basically the same as in the 3D superlattices of nano-
disks grown from 2.2 nm Cu nanoseeds. It can be seen
that almost all of the nanodisks self-assemble into super-
lattices.
A different close-packed 2D superstructure (Figure 9d)
was obtained in the case of adding 0.12% DT. The
building blocks in these superlattices are 3.3 nm nano-
crystals (Figure 9e), which are similar in size to the initial
nanoseeds (3.4 nm). This result indicates that, once there
was excess DT in the solution, Cu nanoseeds did not
orient into nanorods or nanodisks but experienced a
Figure 10. 1H NMR spectra of Cu2S nanocrystals in (a) 3D rodlike
superlattices and (b) 2D superlattices. Inset: 1H NMR spectrum of free
dodecanethiol in toluene-d8. The hydrogen numbering is depicted in the
upper-left corner.
Figure 11. TEM images of Cu2S nanodisks prepared in different solvents using 2.2 nm Cu nanoseeds at 180 °C: (a) toluene, (b) n-octane, (c) cyclohexane,
and (d) hexane.

5128 Inorganic Chemistry, Vol. 48, No. 12, 2009
Shen et al.
digestive ripening procedure,27 during which the growth
was inhibited and narrow size distribution, allowing the
nanocrystals to order into a 2D or 3D close-packed array,
was achieved.
with a uniform diameter and the nanodisks are close-
packed. When n-octane was used to replace toluene as the
solvent, rodlike superlattices could also be obtained, but
the assembly of nanodisks was loose and there were many
more unstacked nanodisks depositing on the substrate.
When cyclohexane and hexane were used, nanodisks
were almost unstacked and the size distributions of the
disks were wide. The solubility of DT in the solvents
increases with the decrease of solvents’ polarity (toluene
> n-octane > cyclohexane > hexane). More surface
“defects” lead to the formation of irregular disks and
decrease the assembly rate. However, there is actually
still no model which could quantitatively evaluate this
effect.
1
A comparison of the H NMR spectra of the Cu2S
nanocrystals in 3D rodlike superlattices and 2D planar
superlattices reveals the different capping degrees of DT.
The 1H NMR spectrum of the nanocrystals in 3D rodlike
superlattices shows some broadened peaks and changes in
chemical shifts compared to the free dodecanethiol mo-
lecules (inset of Figure 10). In fact, resonances of H1-H3
methylene groups do not appear in the spectrum of
nanocrystals in 2D planar superlattices, whereas the
resonances assigned to the H12 methylene protons remain
largely unchanged. This is representative for chemisorbed
functional groups on nanocrystal surfaces.28 Because the
H1 atom is closest to the nanocrystal/sulfur interface, it is
thus expected to give the broadest peak. Comparatively,
the terminal methyl hydrogen (e.g., H12 in Cu-DT)
experiencing the slowest spin relaxation will have the
highest orientation degree of freedom. Thus, the bound
dodecanethiol extends into the solvent with a relatively
high degree of mobility, which provides the steric stabi-
lization necessary to construct different superstructures.
The increase in peak width from curve a to curve b in
Figure 10 suggests stronger dipolar interactions and thus
a more ordered chain arrangement in the latter.
Similarly, for the smaller Cu nanoseeds (2.2 nm), which
can be used to prepare rodlike superlattices, when the
other reaction parameters are kept unchanged (180 °C,
16 h), decreasing the density of DT (increase of σthiol) will
result in the assembled superlattices becoming unstacked
nanodisks. Figure 11 illustrates the TEM images of Cu2S
nanodisks synthesized in different solvents by using
2.2 nm Cu nanoseeds. It can be seen that, when toluene
is used as a solvent, the rodlike superlattice is very thick
Conclusion
In summary, Cu nanoseeds with narrow size distribu-
tions have been synthesized by single-phase approaches.
Cu2S nanorods and nanodisks, as well as their assemblies
and particles, have been synthesized by using the obtained Cu
nanoseeds under solvothermal conditions. Size effects have
been found in the OA growth process of Cu nanoseeds with
different sizes, which induced versatile morphologies of final
nanocrystals. The surface density of DT also plays an
important role in determining the shape and the assembly
of final nanocrystals. The present work provides a control-
lable way to synthesize complex two- and three-dimensional
nanodevices, which will have potential applications in elec-
tronics, biotechnology, and energy.
Acknowledgment. The authors thank Prof. Yadong
Li for his help. This work was supported by NSFC
(20725102), the Foundation for the Author of National
Excellent Doctoral Dissertation of P. R. China, the
Program for New Century Excellent Talents of the Chi-
nese Ministry of Education, the Fok Ying Tung Educa-
tion Foundation (111012), and the State Key Project of
Fundamental Research for Nanoscience and Nanotech-
nology (2006CB932301).
(27) (a) Prasad, B. L. V.; Stoeva, S. I.; Sorensen, C. M.; Klabunde, K. J.
Chem. Mater. 2003, 15, 935–942. (b) Stoeva, S.; Klabunde, K. J.; Sorensen,
C. M.; Dragieva, I. J. Am. Chem. Soc. 2002, 124, 2305–2311. (c) Smetana, A.
B.; Klabunde, K. J.; Sorensen, C. M. J. Colloid Interface Sci. 2005, 284, 521–
526. (d) Hines, M. A.; Scholes, G. D. Adv. Mater. 2003, 15, 1844–1849.
(28) (a) Chaki, N. K.; Vijayamohanan, K. P. J. Phys. Chem. B 2005, 109,
2552–2558. (b) Templeton, A. C.; Cliffel, D. E.; Murray, R. W. J. Am. Chem.
Soc. 1999, 121, 7081–7089. (c) Guo, R.; Song, Y.; Wang, G.; Murray, R. W.
J. Am. Chem. Soc. 2005, 127, 2752–2757. (d) Wang, W.; Murray, R. W.
Langmuir 2005, 21, 7015–7022.
Supporting Information Available: HRTEM images of Cu2S
nanorods, SEM image of rodlike superlattice on silicon-wafer
surface, and EDS spectrum obtained during SEM operation.
This material is available free of charge via the Internet at http://
pubs.acs.org.