Nanocrystal-mediated crystallization of silicon and germanium nanowires in organic solvents: The role of catalysis and solid-phase seeding

Article abstract of DOI:10.1002/anie.200601078

(Figure Presented) The top seed: Various metal nanocrystals are studied as crystallization seeds for silicon and germanium nanowires. The wires are grown by decomposing silane or germane reactants in supercritical organic solvents. Co and Ni nanocrystals

Full text of DOI:10.1002/anie.200601078

Communications
and holes in both silicon and germanium and poses a serious
contamination problem for nanowire integration with silicon
CMOSs. Surprisingly few gold alternatives have been inves-
tigated for nanowire seeding. Si nanowires have been seeded
by chemical vapor deposition (CVD) with Ti particles[9] and
Ga droplets,[10] and in organic solvents, Si and Ge nanowires
were synthesized using Ni nanocrystals.[11,12] Gold alternatives
have been explored more extensively for other semiconduc-
tors: Sn for ZnO wires,[13] various metal films for vapor-grown
tin oxide nanowires,[14] and low-melting metals such as In and
Bi for solution synthesis of Group II–VI,[15,16] III-V[17–19] and
Ge nanowires.[20] Herein, many different nanocrystals—Co,
Ni, CuS, Mn, Ir, MnPt3, Fe2O3, and FePt—are explored as
seeds for Si and Ge nanowire synthesis to develop a more
general understanding of the role of the seed particles in the
nanowire growth process.
Si and Ge nanowires were grown by decomposing silanes
or germanes in high temperature (450–5008C), high pressure
(10.3MPa) supercritical toluene. [21,22] Under these reaction
conditions, Au nanocrystals seed nanowires by the “super-
critical fluid–liquid–solid” (SFLS) mechanism in which nano-
wires evolve from a liquid Au:Si (or Au:Ge) eutectic. In
combination with Si and Ge, many of the seed materials
studied herein do not form liquid eutectics until reaching
temperatures well above 5008C, and are not expected to work
for “VLS”-like growth. However, they form solid alloys
below 5008C, perhaps making solid-phase nanowire seeding
possible. Figure 1 shows TEM images of the various colloidal
nanocrystals used to seed Si and Ge nanowires; their size
distributions had standard deviations less than 20% about
mean diameters ranging between 4.2 and 10.2 nm. Figure 2
shows the reaction products. All of the nanocrystals seeded Si
and Ge nanowires from monophenylsilane (MPS) and
diphenylgermane (DPG), but with varying success (summar-
Nanowires
DOI: 10.1002/anie.200601078
Nanocrystal-Mediated Crystallization of Silicon
and Germanium Nanowires in Organic Solvents:
The Role of Catalysis and Solid-Phase Seeding**
Hsing-Yu Tuan, Doh C. Lee, and Brian A. Korgel*
The size-dependent properties, large surface-area-to-volume
ratios, dispersibility in solvents, and mechanical flexibility of
semiconductor nanowires make them an exciting class of
materials.[1–3] Silicon nanowires in particular can be both n and
p-doped and interface well with an insulating oxide and
conducting metal silicide and these nanowires might be
applied in silicon CMOS (complementary metal oxide semi-
conductor) circuits or with organic compounds on flexible
plastic substrates.[4,5] Germanium is similar to silicon in many
respects, but with a higher carrier mobility. Gold-seeded
vapor–liquid–solid (VLS) growth is a common route to silicon
and germanium nanowire synthesis at low growth temper-
atures ( ꢀ 3608C).[1,3,6–8] Unfortunately, gold traps electrons
[*] H.-Y. Tuan, D. C. Lee, Prof. B. A. Korgel
Department of Chemical Engineering
Texas Materials Institute
Center for Nano- and Molecular Science and Technology
The University of Texas
Austin, TX 78712-1062 (USA)
Fax: (+1)512-471-7060
E-mail: korgel@mail.che.utexas.edu
[**] We acknowledge the National Science Foundation, the Robert A.
Welch Foundation, the Advanced Materials Research Center in
collaboration with International SEMATECH, the Advanced Proc-
essing and Prototyping Center (DARPA: HR0011-06-1-0005), and
the Office of Naval Research (N00014-05-1-0857) for financial
support of this research. We are grateful to Felice Shieh, Cindy
Stowell, and Ali Ghezelbash for Co, Ir, and CuS nanocrystals, and to
J. P. Zhou for TEM assistance.
Figure 1. TEM images of metal nanocrystals studied as Si and Ge
nanowire seeds: a) Co, b) Ni, c) CuS, d) Mn, e) Ir, f) MnPt3, g) Fe2O3
and h) FePt. Insets: magnified images, scale bars 2 nm.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
5184
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 5184 –5187

Angewandte
Chemie
shorter lengths (3–10 mm) and Fe2O3
nanocrystals produced high quality Ge
nanowires with relatively high yield.
All of the nanocrystals produced
nanowires at temperatures signifi-
cantly below their bulk eutectic tem-
peratures (summarized in Table 1).
The small size of the seed nanocrystals
reduces the eutectic temperature;
however, a drop of nearly 3508C is
unlikely.[12] These nanocrystals most
likely promote nanowire crystalliza-
tion from solid-phase seeds. Solid-
phase metal-seeded nanowire growth
has also been proposed in other sys-
tems: Si (Ti,[9] Ni[11]), Ge (Fe,[24] Ni[12]),
GaAs (Au),[25] InAs (Au),[26] and ZnSe
(Au,[27] Fe[28]) nanowires. Solid-phase
seeding is a viable growth mechanism,
provided that the seed particles are
small enough for rapid saturation by
solid-state diffusion and the semicon-
ductor has a high solid solubility in the
metal. Co, Ni, Fe, and Cu all form
Figure 2. SEM images of a–h) Si and i–p) Ge nanowires synthesized in supercritical toluene from
MPS(150 m m, 5008C, 10.3 MPa) and DPG (80 mm, 4608C, 10.3 MPa), respectively.
ized in Table 1). In general, straight nanowires are crystalline
with few extended defects, whereas curly wires are usually
amorphous or polycrystalline.[23] Co nanocrystals gave the
highest yield of straight, long (> 10 mm) Si and Ge nanowires
(Figure 3). Ni nanocrystals also produced crystalline Si and
Ge nanowires with good yield. CuS nanocrystals produced
straight crystalline Si nanowires with good yield but slightly
alloys with Si and Ge at the growth temperatures and energy
dispersive spectroscopy (EDS) analysis of the seed particles
found at the tips of many nanowires revealed silicide and
germanides (see Supporting Information for representative
data). In the particles located at the tips of nanowires seeded
with CuS and Fe2O3 nanocrystals, S and O were not observed
in significant quantities, indicating that wires probably
crystallize from copper silicide
and iron germanide phases. Per-
Table 1: Summary of seed nanocrystal composition and selected properties and reaction conditions.
haps CuS and Fe2O3 are first
reduced to Cu and Fe metal in the
reaction mixture, prior to nanowire
growth. Particles found at the tips
of FePt-seeded nanowires showed
all three elements—Fe, Pt, and
Ge—indicating that nanowire
growth probably originates from a
ternary Fe:Pt:Ge phase.
Nanowire Seed
Nanocrystal
Reaction
Eutectic
Product description
diameter [nm] temp. [8C][a] temp. [8C][h]
Si
Co
9.7
5.6
11.7
4.7
4.2
5.1
500[b]
460
500
500
500
500
500
500
460[c]
460
1200
964
Long straight wires
Long straight wires
Si
Ni
Si
Si
Si
CuS
Mn
Ir
802[d]
1040
1470
N/A
1200[e]
N/A
817
Good yield, shorter wires
Crystalline wires, low yield
Amorphous particles & curly wires
Low yield, mostly curly wires
Amorphous particles & few wires
Amorphous particles
Long straight wires
Long straight wires
Short curly wires
Crystalline wires, low yield
Wires & amorphous particles
Straight wires with rough surfaces
Long straight wires
Si
MnPt3
S
i 2O3 10F.e2
Si
FePt
4.3
9.7
5.6
Si and Ge nanowires seeded
with Ni, Co, Fe2O3, and CuS show
two predominant growth direc-
tions—h111i and h110i—in nearly
Ge
Ge
Ge
Ge
Ge
Ge
Ge
Ge
Co
Ni
CuS11.7
Mn
Ir
MnPt3
Fe2O3 10.2
FePt 4.3
762
[f]
460
644
4.7
4.2
5.1
460
460
460
460
460
720
N/A
N/A
838[g]
N/A
equal
proportions
(Figure 3).
Approximately 5% of the nano-
wires also had h112i-oriented
growth. This result is different
from that for Au-seeded Si and
Ge nanowires in organic solvents,
which show preferential growth in
either the h111i (Si) or h110i (Ge)
directions, with only a small pro-
portion of h110i (or h111i) and
h112i orientations.[22,29] Perhaps
the difference in nanowire growth
direction relates to the solid-phase
Wires & amorphous particles
[a] The optimal synthesis temperature for Si nanowires was approximately 408C higher than for Ge
nanowires in all cases, indicating slower Si nanowire growth kinetics. [b] The optimum reaction
temperature for cobalt-seeded Si nanowires from MPS was approximately 508C higher than the optimal
synthesis temperature (4508C) using Au seeds. [c] The optimum reaction temperature for cobalt-seeded
Ge nanowires from DPG is approximately 808C higher than the optimal synthesis temperature (3808C)
using Au seeds. [d] Value based on Cu:Si phase diagram, because S appears to outgas from the seed
particles. [e] Value based on Fe:Si phase diagram, because iron silicide appears to be the seed
composition. [f] Value based on Cu:Ge phase diagram, because Sappears to outgas from the seed
particles. [g] Value based on Fe:Ge phase diagram, because iron germanocide appears to be the seed
composition. [h] Ref. [42].
Angew. Chem. Int. Ed. 2006, 45, 5184 –5187
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5185

Communications
Figure 3. HRTEM images of Si (a–c) and Ge (d–f) nanowires seeded
by Co (a–c), Fe2O3 (d–e), and CuS(f) showing h110i, h111i and h112i
growth directions. Approximately 5% of the sample contained h112i
oriented nanowires, usually with longitudinal {111} twins, as in (c)
and (f). The 0.326 nm lattice spacing agrees with the (111) d-spacing
for bulk Ge (0.327 nm).
Figure 4. Cobalt-nanocrystal-seeded Si nanowires. HRSEM images of
Si nanowires synthesized in supercritical toluene (10.3 MPa) from
a) MPS(500 8C), b) trisilane (4008C), and c) octylsilane (5008C).
HRSEM images of Si nanowires synthesized in supercritical toluene
(10.3 MPa) from trisilane at d) 3508C, e) 4008C, and f) 4508C. g–
i) TEM images of Si nanowires synthesized in supercritical toluene
(10.3 MPa) with g) trisilane (4008C), h) trisilane (4508C) and i) octylsi-
lane (5008C). The nanowires in (h) and (i) are coated with an
amorphous shell, as shown more clearly in the low-resolution TEM
images in the insets in (h) and (i). The shell material in (h) is
amorphous Si and in (i) it is amorphous 3:1 C:Si (by EDS).
nanowire seeding process. However, it is noteworthy that
regardless of the metal used to seed the nanowires (i.e., Au,
Ni, or Co.), h112i-oriented nanowires always tend to have
longitudinal {111} twins, as in the wires in Figure 3c and f.
In VLS growth, the metal seed dissolves the semiconduc-
tor and recrystallizes it as a nanowire and has a passive role in
the precursor decomposition chemistry.[3,6,22] Au does not
catalyze reactant decomposition, but simply helps drive
crystallization (Au could be called a “crystallization catalyst”,
but not a catalyst for reactant decomposition). In contrast to
Au, the transition metals Ni, Co, and Fe, are catalysts for
molecular decomposition reactions. For carbon nanotube
(CNT) growth, Fe, Co, and Ni are used as catalysts that
enhance reactant decomposition at the seed surface, while
also promoting “crystallization” of the nanostructure—for
CNTs that is graphitization and tube growth.[30–32] Au-seeded
CVD nanowire growth typically uses very reactive precursors,
such as silane, and there is generally no need to use catalytic
seed metals to promote reactant decomposition. However,
sidewall deposition can occur with reactive precursors and
lead to substantial diameter tapering over the length of the
wire.[33,34] Catalytic seeds might be able to lower the growth
temperature and help prevent sidewall deposition. In Au-
seeded, solution-phase nanowire growth, many organosilane
and organogermane reactants, such as octylsilane, are too
unreactive to promote nanowire formation at typical reaction
temperatures.[11,22] Trisilane on the other hand is too reactive
and forms amorphous Si particles.[22] These Si reactants
require a catalytic seed to promote nanowire formation.
Figure 4 shows TEM and SEM images of Si nanowires
seeded by Co nanocrystals in supercritical toluene from MPS,
trisilane, and octylsilane—all three reactants gave nanowires.
The cobalt nanocrystals enhance octylsilane decomposition
enough to promote nanowire formation; however, the nano-
wire quality is only marginal as the nanowires are coated with
an amorphous layer with significant carbon content (Fig-
ure 4i, 3:1 C:Si by EDS in shell). The cobalt nanocrystals
were found to promote Si nanowire formation using trisilane
at temperatures as low as 3508C, although the majority
product at this low temperature was amorphous Si colloids
(Figure 4d). Cobalt enhances heterogeneous trisilane decom-
position relative to homogeneous particle formation. Si
nanowires produced with trisilane at 4008C had very little
sidewall deposition and few particulates; slightly higher
temperature (4508C) gave significant sidewall deposition
(Figure 4 h).
Of the nanocrystals studied, Co gave the highest yield and
quality of both Si and Ge nanowires, rivaling Au-seeded
reactions (See Supporting Information for an example of Au-
seeded Si nanowires). All growth temperatures were well
below the bulk eutectic temperatures, indicating solid-phase
seeding. Solid-phase seeding can probably occur for any
semiconductor with a high solubility in the seed metal;
however, the growth temperature must be sufficiently high for
fast saturation by solid-state diffusion, which can occur
relatively fast in nanometer-diameter seed particles. Cobalt
was also found to catalyze silane decomposition to promote Si
nanowire growth from octylsilane and trisilane, similar to past
studies on Ni nanocrystals.[11,12] In gas-phase reactions, the use
of catalytic transition-metal seeds might enable lower temper-
ature reactions for less sidewall deposition and better
diameter control. As more seed materials are studied, the
solid-phase growth method may become as prevalent as
liquid-eutectic seeding, and the use of catalytic seed particles
might serve as a general approach for improved diameter
control, as it has for carbon nanotube growth.
5186
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 5184 –5187

Angewandte
Chemie
[12] H.-Y. Tuan, D. C. Lee, T. Hanrath, B. A. Korgel, Chem. Mater.
Experimental Section
2005, 17, 5705.
Anhydrous toluene (99.8%, Sigma-Aldrich), monophenylsilane
(MPS, 97%, Sigma-Aldrich), N-octylsilane ((C8H17)H3Si, Gelest),
trisilane (Si3H8, Voltaix), and diphenylgermane (DPG, Gelest) were
used as received. Co,[35] Ni,[36] CuS,[37] Ir,[38] Mn (Supporting Informa-
tion), MnPt3, Fe2O3,[39] and FePt[40] nanocrystals were prepared by
arrested precipitation as described in the literature.
[13] Y. Ding, P. X. Gao, Z. L. Wang, J. Am. Chem. Soc. 2004, 126,
2066.
[14] P. Nguyen, H. T. Ng, M. Meyyappan, Adv. Mater. 2005, 17, 1773.
[15] H. Yu, J. B. Li, R. A. Loomis, P. C. Gibbons, L. W. Wang, W. E.
Buhro, J. Am. Chem. Soc. 2003, 125, 16168.
[16] J. W. Grebinski, K. L. Hull, J. Zhang, T. H. Kosel, M. Kuno,
Chem. Mater. 2004, 16, 5260.
Nanowires were synthesized in semibatch 1 mL or 10 mL Ti
grade-2 reactor connected to a high-pressure liquid chromatography
(HPLC) pump for pressure control and reactant injection. The
reactor was heated in a brass block with temperature monitored by a
thermocouple.[22,32,41] A small piece of clean Si wafer was placed in the
reactor as a deposition substrate to gather reaction product. The
reactor was sealed in a nitrogen-filled glove box, connected to the
HPLC pump, and then filled with anhydrous oxygen-free toluene.
The reactor was pressurized to 3.4 MPa with toluene and then heated
to the reaction temperature. 500 mL of 150 mm MPS in toluene with
Si:nanocrystal mole ratios of 50:1 (Co), 100:1 (Ni), 150:1 (CuS), 50:1
(Mn), 50:1 (Ir), 50:1 (MnPt3), 50:1 (Fe2O3), and 50:1 (FePt), were
[17] T. J. Trentler, K. M. Hickman, S. C. Goel, A. M. Viano, P. C.
Gibbons, W. E. Buhro, Science 1995, 270, 1791.
[18] S. P. Ahrenkiel, O. I. Micic, A. Miedaner, C. J. Curtis, J. M.
Nedeljkovic, A. J. Nozik, Nano Lett. 2003, 3, 833.
[19] D. D. Fanfair, B. A. Korgel, Cryst. Growth Des. 2005, 5, 1971.
[20] X. Lu, D. D. Fanfair, K. P. Johnston, B. A. Korgel, J. Am. Chem.
Soc. 2005, 127, 15718.
[21] For a recent review, see P. S. Shah, T. Hanrath, K. P. Johnston,
B. A. Korgel, J. Phys. Chem. B 2004, 108, 9574.
[22] D. C. Lee, T. Hanrath, B. A. Korgel, Angew. Chem. 2005, 117,
3639; Angew. Chem. Int. Ed. 2005, 44, 3573.
[23] X. Lu, T. Hanrath, K. P. Johnston, B. A. Korgel, Nano Lett. 2003,
3, 93 .
[24] S. Mathur, H. Shen, V. Sivakov, U. Werner, Chem. Mater. 2004,
16, 2449.
[25] A. I. Persson, M. W. Larsson, S. Stenstroem, B. J. Ohlsson, L.
Samuelson, L. R. Wallenberg, Nature Mater. 2004, 3, 677.
[26] K. A. Dick, K. Deppert, T. Mrtensson, B. Mandl, L. Samuelson,
W. Seifert, Nano Lett. 2005, 5, 761.
[27] Y. Ohno, T. Shirahama, S. Takeda, A. Ishizumi, Y. Kanemitsu,
Appl. Phys. Lett. 2005, 87, 043105.
[28] A. Colli, S. Hofmann, A. C. Ferrari, C. Ducati, F. Martelli, S.
Rubini, S. Cabrini, A. Franciosi, J. Robertson, Appl. Phys. Lett.
2005, 86, 153103.
[29] T. Hanrath, B. A. Korgel, Small 2005, 1, 717.
[30] H. J. Dai, J. Kong, C. W. Zhou, N. Franklin, T. Tombler, A.
Cassell, S. S. Fan, M. Chapline, J. Phys. Chem. B 1999, 103,
11246.
[31] D. C. Lee, F. V. Mikulec, B. A. Korgel, J. Am. Chem. Soc. 2004,
126, 4951.
[32] D. C. Lee, B. A. Korgel, Mol. Simul. 2005, 31, 637.
[33] H. Adhikari, A. F. Marshall, C. E. D. Chidsey, P. C. McIntyre,
Nano Lett. 2006, 6, 318.
injected from a six-way valve injection loop (Valco) at 0.4 mLminꢁ1
.
For Ge nanowires, 500 mL of 80 mm DPG in toluene were injected
with Ge:nanocrystal mole ratios of 100:1 (Co), 100:1 (Ni), 100:1
(CuS), 50:1 (Mn), 20:1 (Ir), 100:1 (MnPt3), 200:1 (Fe2O3), and 200:1
(FePt). The reaction pressure was increased to 10.3MPa with
additional toluene. After 10 min, the reactor was immersed in ice-
water and cooled to room temperature. The reactor was opened and
the deposition substrate removed and stored under nitrogen until the
product was characterized by SEM and TEM.
Nanowires were imaged by HRSEM on the deposition substrate
without further purification using a LEO 1530 field-emission SEM at
1–3kV accelerating voltage with working distances ranging between
2–5 mm. For HRTEM and STEM imaging, nanowires were either
drop-cast from chloroform dispersions or dry transferred by scratch-
ing the deposition substrate onto 200-mesh lacey carbon-coated
copper or nickel grids (Electron Microscope Sciences). Images were
acquired using 200 kV accelerating voltage on a JEOL 2010F
equipped with an Oxford INCA ED spectrometer.
Received: March 20, 2006
Revised: May 1, 2006
Published online: July 5, 2006
[34] E. Tutuc, S. Guha, J. O. Chu, Appl. Phys. Lett. 2006, 88, 043113.
[35] V. F. Puntes, D. Zanchet, C. K. Erdonmez, A. P. Alivisatos, J.
Am. Chem. Soc. 2002, 124, 12874.
[36] C. B. Murray, S. Sun, H. Doyle, T. Betley, MRS Bull. 2001, 26,
985.
Keywords: germanium · nanocrystals · nanowires · silicon ·
supercritical fluids
.
[37] A. Ghezelbash, B. A. Korgel, Langmuir 2005, 21, 9451.
[38] C. A. Stowell, B. A. Korgel, Nano Lett. 2005, 5, 1203.
[39] T. Hyeon, S. S. Lee, J. Park, Y. Chung, H. B. Na, J. Am. Chem.
Soc. 2001, 123, 12798.
[40] S. Sun, C. B. Murray, D. Weller, L. Folks, A. Moser, Science 2000,
287, 1989.
[41] T. Hanrath, B. A. Korgel, J. Am. Chem. Soc. 2002, 124, 1424.
[42] Binary Alloy Phase Diagrams, 2nd Ed. (Eds.: B. M. Thaddeus, O.
Hiroaki, P. R. Subramanian, K. Linda), ASM International,
Materials Park, OH, 1990.
[1] J. D. Holmes, K. P. Johnston, R. C. Doty, B. A. Korgel, Science
2000, 287, 1471.
[2] B. A. Korgel, Science 2004, 303, 1308.
[3] For a recent review, see M. Law, J. Goldberger, P. Yang, Annu.
Rev. Mater. Res. 2004, 34, 83 .
[4] Y. Cui, C. M. Lieber, Science 2001, 291, 851.
[5] M. C. McAlpine, R. S. Friedman, S. Jin, K. H. Lin, W. U. Wang,
C. M. Lieber, Nano Lett. 2003, 3, 1531.
[6] J. T. Hu, T. W. Odom, C. M. Lieber, Acc. Chem. Res. 1999, 32,
435.
[7] T. Hanrath, B. A. Korgel, Adv. Mater. 2003, 15, 437.
[8] D. W. Wang, H. J. Dai, Angew. Chem. 2002, 114, 4783; Angew.
Chem. Int. Ed. 2002, 41, 4783.
[9] T. I. Kamins, R. S. Williams, D. P. Basile, T. Hesjedal, J. S. Harris,
J. Appl. Phys. 2001, 89, 1008.
[10] M. K. Sunkara, S. Sharma, R. Miranda, G. Lian, E. C. Dickey,
Appl. Phys. Lett. 2001, 79, 1546.
[11] H.-Y. Tuan, D. C. Lee, T. Hanrath, B. A. Korgel, Nano Lett. 2005,
5, 681.
Angew. Chem. Int. Ed. 2006, 45, 5184 –5187
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5187