Vertically aligned, catalyst-free InP nanowires grown by metalorganic chemical vapor deposition

Article abstract of DOI:10.1063/1.2131182

Vertically-aligned InP nanowires are grown by metalorganic chemical vapor deposition (MOCVD) without the use of a deposited metal catalyst. A surface reconstruction induces indium droplets to form on the surface and thus act as nucleation sites for nanowire growth. Vertical growth from the InP(111)B substrate along with transmission electron microscopy (TEM) analysis indicate epitaxial growth from the substrate in the [111]B direction. A uniform cross section along the longitudinal axis can be achieved by optimizing the input VIII ratio. Small variations in the diameter and length are seen under optimal growth conditions.

Full text of DOI:10.1063/1.2131182

Vertically aligned, catalyst-free InP nanowires grown by metalorganic
chemical vapor deposition
Clint J. Novotny and Paul K. Yu
Citation: Appl. Phys. Lett. 87, 203111 (2005); doi: 10.1063/1.2131182
View online: http://dx.doi.org/10.1063/1.2131182
View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v87/i20
Published by the AIP Publishing LLC.
Additional information on Appl. Phys. Lett.
Journal Homepage: http://apl.aip.org/
Journal Information: http://apl.aip.org/about/about_the_journal
Top downloads: http://apl.aip.org/features/most_downloaded
Information for Authors: http://apl.aip.org/authors
Downloaded 28 Aug 2013 to 131.211.208.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

APPLIED PHYSICS LETTERS 87, 203111 ͑2005͒
Vertically aligned, catalyst-free InP nanowires grown by metalorganic
chemical vapor deposition
Clint J. Novotny and Paul K. L. Yu
Department of Electrical and Computer Engineering, University of California, San Diego, California 92093
͑
Received 20 July 2005; accepted 20 September 2005; published online 10 November 2005͒
Vertically-aligned InP nanowires are grown by metalorganic chemical vapor deposition ͑MOCVD͒
without the use of a deposited metal catalyst. A surface reconstruction induces indium droplets to
form on the surface and thus act as nucleation sites for nanowire growth. Vertical growth from the
InP͑111͒B substrate along with transmission electron microscopy ͑TEM͒ analysis indicate epitaxial
growth from the substrate in the ͓111͔B direction. A uniform cross section along the longitudinal
axis can be achieved by optimizing the input V/III ratio. Small variations in the diameter and length
are seen under optimal growth conditions. © 2005 American Institute of Physics.
͓
DOI: 10.1063/1.2131182͔
Semiconductor nanowires ͑NWs͒ have recently gained
much attention due to their promising electrical and optical
properties. High-quality NWs have been grown from a vari-
site for the NWs. In this experiment, the substrate is placed
into the reactor without any such catalyst or patterning. The
InP NWs nucleate and grow from In droplets that form on
the surface. We have found several conditions that affect this
nucleation.
1
–4
ety of materials, including III-V and II-VI alloys,
5
,6
7,8
9,10
silicon, nitrides, and oxides.
Most of this research is
11
attributed to the vapor-liquid-solid ͑VLS͒ technique. This
technique relies on the use of metal nanoparticles, such as Au
or Co, which can result in the unintentional incorporation
into the otherwise pure crystalline NWs. According to the
VLS theory, this metal catalyst must become supersaturated
with the desired material. The NW then precipitates out at
the liquid-solid interface. The diameter of the NW is con-
trolled by the Au nanoparticle. Other theories have been pre-
sented by groups such as Lars Samuelson’s, who suggest that
First, the initial annealing of the substrate is a key step in
the self-catalytic growth of the NWs. During the anneal, the
substrate can undergo a surface reconstruction that provides
an In-rich surface. As the impinging TMIn decomposes on
this surface, an excess of In adatoms is created due to the
lack of available P dangling bonds. This excess causes a
liquid phase separation, thus providing the droplets needed
for nucleation of the wires. Studies on InP͑001͒ show that
the surface changes from a P-rich ͑2ϫ1͒ reconstruction to
an In-rich ␦͑2ϫ4͒ reconstruction with increasing annealing
1
2,13
a liquid alloy is not required for NW growth.
So-called self-catalytic growth has been seen in several
1
7
temperature. Similarly, on InP͑111͒A, the surface trans-
1
4–16
material systems.
The group III metal acts as the cata-
ͱ ͱ
forms from a ͑ 3ϫ 3͒R30° P-rich surface to a ͑2ϫ2͒ In-
1
8
lyst, although instead of remaining neutral, it is actually in-
volved in the reaction. In this work, we focus on the growth
of NWs without any patterning or foreign material. Here,
excess indium collects on the surface, leading to a liquid
phase separation. Liquid In droplets thus act as nucleation
sites for the NWs to grow, removing any possibility of con-
tamination from a metal catalyst. The annealing conditions
as well as the growth conditions, such as temperature and
input V/III ratio, are examined.
rich surface upon heating. To our knowledge, there have
been no investigations made on the ͑111͒B surface. However,
we have used these trends to understand our observed results.
To look at the surface reconstruction, samples were
grown under identical conditions with and without the pre-
growth anneal. On the ͑100͒ substrate without the anneal,
almost zero wire growth was observed. Repeating the experi-
ment with annealing led to a high density of self-catalyzed
NWs on the entire surface of the substrate. Similarly, a
͑111͒B substrate was used without annealing and again no
wires were visible. With annealing, a high density of verti-
cally aligned, self-catalyzed wires formed. Further scanning
tunneling microscopy ͑STM͒ or x-ray photoelectron spec-
troscopy ͑XPS͒ studies are required to study the effects of
The NWs are grown on InP͑111͒B substrates in a hori-
zontal MOCVD reactor at a pressure of 100 Torr. Trimeth-
ylindium ͑TMIn͒ and phosphine ͑PH ͒ were used as the
3
group III and V sources, respectively, with hydrogen as the
carrier gas. The input V/III ratios are adjusted by keeping
the TMIn flow constant at 0.30 ␮mol/min and by adjusting
annealing InP͑111͒B under a flow of PH to exactly deter-
3
the PH flow to attain the desired V/III ratio. The substrates
mine the type of reconstruction that occurs under these
conditions.
3
are annealed at temperatures between 500 and 625 °C under
PH for 5–10 min, then cooled down to the growth tempera-
Second, the growth conditions affect the nucleation of
3
ture, ranging from 350 to 450 °C. The NWs nucleate and
grow when the TMIn is introduced into the system, usually
for 10 min. Sample characterization was carried out using an
FEI XL 30 Environmental Scanning Electron Microscope
the NWs. At the low growth temperatures used, PH does not
3
1
9,20
fully decompose.
Therefore, the V/III ratio at the surface
is much lower than the input V/III ratio. The impinging
TMIn lands on the reconstructed In-rich surface, thus leading
to an excess of In adatoms which form droplets. The pres-
ence of a liquid surface reduces the activation barrier for PH3
decomposition, thus providing a nucleation site for the NWs.
As the growth temperature increases from 400 to 450 °C,
͑
SEM͒, operating at 20 kV, and a Phillips CM20 Transmis-
sion Electron Microscope, operating at 200 kV.
For typical VLS growth, Au colloids are dispersed onto
the substrate surface before growth to act as the nucleation
0003-6951/2005/87͑20͒/203111/3/$22.50
87, 203111-1
© 2005 American Institute of Physics
Downloaded 28 Aug 2013 to 131.211.208.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

203111-2
C. J. Novotny and P. K. L. Yu
Appl. Phys. Lett. 87, 203111 ͑2005͒
FIG. 1. SEM of samples tilted at 45° with an input V/III of ͑a͒ 2, ͑b͒ 25,
and ͑c͒ 125. The scale bar is 2 ␮m in ͑a͒ and 1 ␮m in ͑b͒ and ͑c͒.
FIG. 2. SEM images of sample tilted at ͑a͒ 45° and ͑b͒ 90°, with scales bars
of 1 and 2 ␮m, respectively. Diameter variation of sample is shown in ͑c͒.
the amount of pyrolyzed PH increases. Too large of a flux of
22
3
5
–60 nm. Understanding the conditions that lead to
PH suppresses the formation of the droplets on the surface
3
catalyst-free growth is therefore very important when at-
tempting to grow metal-catalyzed NWs only.
and leads to 3D island growth on the surface. Thus, the tem-
perature and the input V/III ratio must be optimized to grow
the NWs.
Once the In droplet forms and becomes supersaturated,
the NW begins to form at the liquid-solid interface. The input
V/III ratio has a tremendous impact on this growth. At low
V/III ratios, the wires become very tapered. At higher V/III
ratios, the wires are nontapered, longer, and thinner. These
properties are all related to the In droplet.
In general, we see that for increasing V/III, the diameter
decreases and the length of the wires increases, as seen in
Fig. 3. These trends are explained by the consumption of
material by the In droplet. If the P flux is too low ͑i.e., low
V/III͒, the droplet will grow in size and thus the diameter
increases. Likewise, the growth rate of the wires will be af-
fected by the amount of P. If the volume of the droplet in-
creases, more P is required to supersaturate the droplet. Thus,
the growth rate slows as the droplet size increases.
TEM studies were carried out to perform structural
analysis of the NWs. The vertical growth of the wires off the
substrate implies that the NWs grow epitaxially from the
substrate in the ͓111͔B direction, as seen in Fig. 1. TEM
analysis on the wires from Fig. 1͑c͒ further confirms this
growth direction, as seen in the inset of Fig. 4. The NWs are
single crystalline with rectangular blocks of rotational twin
structures along the axis of the wire, which arise from the
small energetic differences for hexagonal or cubic stacking
Keeping the TMIn flow constant, the PH flow is varied
3
to produce input V/III ratios ranging from 2 to 250, with a
growth temperature of 350 °C. For an input V/III of 2,
shown in Fig. 1͑a͒, the wires are “reverse tapered.” Previ-
ously published works on NWs grown with MOCVD have
shown a different type of tapering due to epitaxial growth
1
,21
along the side walls of the NWs, where the wire is thinner
at the top and wider near the substrate. In this case, the wires
show the opposite trend, due to the use of an In droplet and
not a foreign material. In fact, the term self-catalytic growth
is misleading. A catalyst by definition reduces the activation
energy for a reaction but remains unchanged. Here, the In
takes place in the reaction at the liquid-solid interface. Both
the In from the impinging gas as well as the In adatoms
migrating from the surface easily dissolve into the droplet. If
the flux of P atoms to the droplet surface is not high enough
to keep up with the NW growth rate, then the droplet will
grow in size. We see this very clearly for a low V/III ratio.
As the V/III ratio is increased to 25 and then 125, the taper-
ing decreases to ϳ10 nm/␮m and then to an immeasurable
amount, as shown in Figs. 1͑b͒ and 1͑c͒, respectively.
22,23
sequences in the ͗111͘ directions.
In conclusion, we have shown that single crystalline InP
NWs can be grown without the use of metal catalysts. The
The dimensions of the NWs are thus very dependent on
the V/III ratio as well as the growth temperature. Wires
grown at a V/III of 25 at 450 °C show a high density of
9
2
approximately 10 wires/cm , as shown in Fig. 2͑a͒. A cross-
sectional view of the sample seen in Fig. 2͑b͒ shows that the
wires are very uniform in length. Here, the average diameter
of the wire was 36.7 nm with an average length of 1.9 ␮m.
The diameters range from 20 to 50 nm, as seen in Fig. 2͑c͒.
In comparison, previously grown vertically aligned InP NWs
using MOCVD and Au catalysts of 20 nm had a range of
FIG. 3. Comparison of diameter and growth rates of NWs as the V/III is
varied. Growth temperature is kept constant at 380 °C.
Downloaded 28 Aug 2013 to 131.211.208.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

203111-3
C. J. Novotny and P. K. L. Yu
Appl. Phys. Lett. 87, 203111 ͑2005͒
1
K. Hiruma, M. Yazawa, T. Katsuyama, K. Ogawa, K. Haraguchi,
M. Koguchi, and H. Kakibayashi, J. Appl. Phys. 77, 447 ͑1995͒.
X. Duan and C. M. Lieber, Adv. Mater. ͑Weinheim, Ger.͒ 12, 298 ͑2000͒.
B. J. Ohlsson, M. T. Björk, M. H. Magnusson, K. Deppert, L. Samuelson,
and L. R. Wallenberg, Appl. Phys. Lett. 79, 3335 ͑2001͒.
2
3
4
M. S. Gudiksen, J. Wang, and C. M. Lieber, J. Phys. Chem. B 105, 4062
5͑
2001͒.
Y. Cui, X. Duan, J. T. Hu, and C. M. Lieber, J. Phys. Chem. B 104, 5213
6͑
2000͒.
Y. Cui, L. J. Lauhon, M. S. Gudiksen, J. Wang, and C. M. Lieber, Appl.
Phys. Lett. 78, 2214 ͑2001͒.
7
8
X. Duan and C. M. Lieber, J. Am. Chem. Soc. 122, 188 ͑2000͒.
T. Kuykendall, P. Pauzauskie, S. K. Lee, Y. F. Zhang, J. Goldberger, and
P. Yang, Nano Lett. 3, 1063 ͑2003͒.
9
M. H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber,
R. Russo, and P. Yang, Science 292, 1897 ͑2001͒.
1
0
P. Yang, H. Yan, S. Mao, R. Russo, J. Johnson, R. Saykally, N. Morris,
J. Pham, R. R. He, and H. J. Choi, Adv. Funct. Mater. 12, 323 ͑2002͒.
R. S. Wagner, Whisker Technology ͑Wiley, New York, 1970͒.
A. Persson, M. W. Larsson, S. Stenstrom, B. J. Ohlsson, L. Samuelson,
and L. Reine Wallenberg, Nat. Mater. 3, 677 ͑2004͒.
K. A. Dick, K. Deppert, T. Martensson, B. Mandl, L. Samuelson, and
W. Seifert, Nano Lett. 5, 761 ͑2005͒.
1
1
FIG. 4. TEM image of NW, with a scale bar of 100 nm, and its diffraction
pattern, indicating growth along the ͓111͔B direction.
12
1
1
1
3
4
5
use of an initial annealing step causes a surface reconstruc-
tion necessary for the formation of indium droplets needed
for VLS growth. Vertically aligned NWs of uniform length
and a small distribution in diameter and length can be
achieved. The lack of a metal catalyst ensures a pure NW
free from unintentional incorporation of the catalyst. We be-
lieve that by taking advantage of their vertical alignment,
these NWs will be very useful for integrated nanoscale
devices.
E. Stach, P. Pauzauskie, T. Kuykendall, J. Goldberger, R. He, and P. Yang,
Nano Lett. 3, 865 ͑2003͒.
P. J. Poole, J. Lefebvre, and J. Fraser, Appl. Phys. Lett. 83, 2055 ͑2003͒.
Y. Zhang, H. Jia, and D. Yu, J. Phys. D 37, 413 ͑2004͒.
D. C. Law, Y. Sun, C. H. Li, G. Chen, S. B. Visbeck, and R. F. Hicks,
Phys. Rev. B 66, 045314 ͑2002͒.
C. H. Li, Y. Sun, D. C. Law, S. B. Visbeck, and R. F. Hicks, Phys. Rev. B
68, 085320 ͑2003͒.
16
17
1
1
2
2
8
9
0
1
C. A. Larsen, N. I. Buchan, and G. B. Stringfellow, J. Cryst. Growth 85,
1
48 ͑1987͒.
N. I. Buchan, C. A. Larsen, and G. B. Stringfellow, J. Cryst. Growth 92,
05 ͑1988͒.
K. Haraguchi, T. Katsuyama, K. Hiruma, and K. Ogawa, Appl. Phys. Lett.
0, 745 ͑1992͒.
This work was supported by the National Science Foun-
dation under Program Nos. ECS0307247 and ECS0403589.
The authors would like to thank Professor Deli Wang from
UCSD for introducing NW growth to us and for all his as-
sistance. They would also like to thank Dr. Wen-An Chiou
from UC Irvine for assisting with TEM imaging, Mr. Art
Clawson from UCSD for assisting with TEM analysis, and
Mr. David Aplin from UCSD for helpful discussions.
6
6
22
S. Bhunia, T. Kawamura, S. Fujikawa, K. Tokushima, and Y. Watanabe,
Physica E ͑Amsterdam͒ 21, 583 ͑2004͒.
3
2
U. Krishnamachari, M. Borgstrom, B. J. Ohlsson, N. Panev, L. Samuelson,
W. Seifert, M. W. Larsson, and L. R. Wallenberg, Appl. Phys. Lett. 85,
2077 ͑2004͒.
Downloaded 28 Aug 2013 to 131.211.208.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions