Size effects of nano-pattern in Si(1 1 1) substrate on the selective growth behavior of GaN nanowires by MOCVD

Article abstract of DOI:10.1016/j.materresbull.2011.11.039

We report on the size effects of nano-patterned Si(1 1 1) substrates on the selective growth of GaN nanowires (NWs) using metal organic chemical vapor deposition. The nano-patterns on Si(1 1 1) substrates were fabricated by etching process of Au nano-droplets. The size of nano-patterns fabricated on Si(1 1 1) substrates were corresponding to size of Au nano-droplets, and the diameter of GaN NWs grown on nano-patterns was similar to the size of nano-pattern. Dense and well-oriented GaN NWs were grown on Si(1 1 1) substrates corresponding to the nano-patterns with an average diameter of about 50 nm. However, only a few GaN bulk grains, and mixed phase of a few NWs and bulk crystal of GaN were grown on the nano-patterned Si(1 1 1) having too small and large diameter, respectively, compare to the nano-patterns with diameter of 50 nm. Our results suggested that the selective growth of GaN NWs is strongly affected by the size of nano-patterns and its related mechanism.

Full text of DOI:10.1016/j.materresbull.2011.11.039

Materials Research Bulletin 47 (2012) 836–842
Contents lists available at SciVerse ScienceDirect
Materials Research Bulletin
journal homepage: www.elsevier.com/locate/matresbu
Size effects of nano-pattern in Si(1 1 1) substrate on the selective growth
behavior of GaN nanowires by MOCVD
*
Ji-Hyeon Park, R. Navamathavan, Cheul-Ro Lee
Semiconductor Materials and Process Laboratory, School of Advanced Materials Engineering, Research Center for Advanced Materials Development (RCAMD),
Chonbuk National University, Jeonju 561-756, South Korea
A R T I C L E I N F O
A B S T R A C T
Article history:
We report on the size effects of nano-patterned Si(1 1 1) substrates on the selective growth of GaN
nanowires (NWs) using metal organic chemical vapor deposition. The nano-patterns on Si(1 1 1)
substrates were fabricated by etching process of Au nano-droplets. The size of nano-patterns fabricated
on Si(1 1 1) substrates were corresponding to size of Au nano-droplets, and the diameter of GaN NWs
grown on nano-patterns was similar to the size of nano-pattern. Dense and well-oriented GaN NWs were
grown on Si(1 1 1) substrates corresponding to the nano-patterns with an average diameter of about
50 nm. However, only a few GaN bulk grains, and mixed phase of a few NWs and bulk crystal of GaN were
grown on the nano-patterned Si(1 1 1) having too small and large diameter, respectively, compare to the
nano-patterns with diameter of 50 nm. Our results suggested that the selective growth of GaN NWs is
strongly affected by the size of nano-patterns and its related mechanism.
Received 15 March 2011
Received in revised form 24 October 2011
Accepted 16 November 2011
Available online 25 November 2011
Keywords:
A. Semiconductors
A. Nitrides
A. Nanostructures
B. Crystal growth
B. Vapor deposition
ß 2011 Elsevier Ltd. All rights reserved.
1. Introduction
strong luminescence was observed in such structures of GaN NWs
grown by MBE indicates that a potential way to improve the
The self-assembled 1-dimensional (1D) nanocrystals, for
example, nanowires (NWs), nanorods, nanopillars, and nanocol-
umns are expected to be the key elements in realizing nanometer-
scale electrical and optical devices in nano-electronics and nano-
optics [1–4]. Nanometer-scale III-nitride 1D structures have great
prospects in fundamental physical science and novel nano-
technological applications. There is a growing interest on the
fabrication of GaN NWs for potential applications in optoelec-
tronic and nanoscale electronic devices in which they take
advantage of the unique high-aspect ratio structure of NWs [5–7].
A variety of techniques have been utilized to synthesize GaN NWs
which include laser ablation, chemical vapor deposition (CVD),
hydride vapor-phase epitaxy (HVPE), molecular beam epitaxy
(MBE), and metalorganic chemical vapor deposition (MOCVD) [8–
17]. The preparation of GaN NWs is typically based on the vapor
liquid solid (VLS) growth mechanism and involves the use of
catalysts, such as Au, Ni, Fe, and In [18–25]. Considerable effort has
been devoted to the synthesis of vertically aligned with high dense
GaN NWs. The GaN NWs with high crystal quality can be achieved
by MBE using an external catalyst on Si(1 1 1) substrate. Very
efficiency of group III nitride-based devices. The growth of catalyst
free GaN NWs has been reported using selective growth on
patterned surfaces, in which the substrate preparations were
performed before the NW growth [6,26–28]. However, there were
no reports available based on the nano-patterns on Si(1 1 1) via Au
etching process. Therefore, we made an attempt to synthesis GaN
NWs on the nano-patterned Si(1 1 1) substrate and the influence
of the nano-pattern size effect on the growth behavior of the
resultant NWs. We believe that this method is very efficient for the
production of high quality and aligned GaN NWs without the
usage of any catalyst.
In this study, we examine the effect of nano-pattern size on
the selective growth behavior of GaN NWs grown on nano-
patterned Si(1 1 1) substrate by MOCVD. The experiments were
conducted using different size of nano-patterns fabricated on
Si(1 1 1) substrates. The size of Au nano-droplets and its related
nano-patterns was controlled by Au sputtering time. The
formation of Au nano-droplets, fabricated nano-patterns, and
grown GaN NWs were studied by using field emission scanning
electron microscopy (FE-SEM) and photoluminescence (PL)
spectroscopy. On the basis of experimental results, a model
has been established to describe the growth mechanism of GaN
NWs on the nano-patterned Si(1 1 1) substrates with the
variation of the size.
* Corresponding author. Tel.: +82 63 270 2304; fax: +82 63 270 2305.
E-mail address: crlee7@chonbuk.ac.kr (C.-R. Lee).
0025-5408/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2011.11.039

J.-H. Park et al. / Materials Research Bulletin 47 (2012) 836–842
837
Fig. 1. Schematic diagram illustrating the process flow to grow GaN NWs on nano-patterned Si(1 1 1) substrate which consists of 6 steps. (a) Deposition of the Au thin film on
Si(1 1 1) substrate. (b) The formation of Au nano-droplets to synthesize nano-patterns. (c) The oxidation of Si(1 1 1) substrate. (d) The etching of Au nano-droplets on the
Si(1 1 1) substrate. (e) The pre-deposition of Al in nano-patterns. (f) The growth GaN NWs on nano-patterned Si(1 1 1) substrate.
2. Experimental
Si(1 1 1) substrate was immersed in Au etchant (KI:I2:DI
water = 4 g:1 g:50 ml) for 10 min to remove Au nano-droplets
and were over flowed with DI water for 5 min (Fig. 1(d)). To grow
GaN NWs, Si(1 1 1) substrate with SiO2 nano-patterns was loaded
in MOCVD chamber. And then trimethylaluminum (TMAl) was
supplied for 5 s at 1020 8C to deposit Al on the nano-patterns
(Fig. 1(e)) and followed by the growth of GaN NWs, as shown in
Fig. 1(f). The Al source incorporated into the nano-patterns
provides the initiation of nucleation during the growth of GaN
NWs. The precursors for Ga and N were trimethylgallium (TMGa)
and ammonium (NH3), respectively. Hydrogen (H2) was used as
the carrier gas. GaN NWs on SiO2 nano-patterns/Si(1 1 1)
substrates were grown at 740 8C for 60 min at working pressure
of 600 Torr. The flow rates of TMGa and NH3 were of 0.2 SCCM and
1.5 SLM, respectively.
GaN NW arrays were grown on nano-patterned Si(1 1 1)
substrates by using MOCVD with a horizontal quartz reactor. The
growth of GaN NWs is typically based on the VLS growth
mechanism [29] which involves the use of nano-patterns. Fig. 1
shows schematic diagram illustrating the process for fabrication
of nano-patterns on Si(1 1 1) surface and the growth of GaN NWs
on it. Initially the Si(1 1 1) substrates were cleaned by ultrasonic
cleaning in acetone and methyl alcohol for 5 min. After that they
were over flowed with DI water for 5 min and were cleaned by HF
(2%) for 3 min. The Au thin film was deposited on Si(1 1 1)
substrates by an ion sputtering system using gold target
(Fig. 1(a)). In this present study, in order to investigate the effect
of nano-pattern size on the growth behavior of GaN NWs, we
deposited Au thin film with different sputtering durations. Then,
the Si(1 1 1) substrate was annealed at 650 8C (at a working
pressure of 600 Torr) under hydrogen ambient for 10 min in order
to form the Au nano-droplets as shown in Fig. 1(b). After this
process, the substrate was loaded in to the furnace to form SiO2
layer by oxidizing at 800 8C for 10 min (Fig. 1(c)). Then the
The morphology of the Au nano-droplets, nano-patterns and
the grown GaN NWs were analyzed by field emission scanning
electron microscopy (FE-SEM, Hitachi S-7400, Hitachi, Japan). The
optical properties of GaN NWs were characterized by temperature-
dependent photoluminescence (PL) analysis using a 325 nm He–Cd
laser.

838
J.-H. Park et al. / Materials Research Bulletin 47 (2012) 836–842
Fig. 2. The top-view FE-SEM images of the Au nano-droplets annealed on Si(1 1 1) substrate with different Au sputtering time; (a) 10, (b) 30, (c) 50 and (d) 120 s.
3. Results and discussion
growth of GaN NWs. These nano-patterns on the Si(1 1 1) substrate
act as nucleation centers for growing GaN NWs. Fig. 3(a)–(d) shows
the top-view FE-SEM images of the nano-patterns obtained by the
etching of Au on Si(1 1 1) substrates with the different Au
sputtering time of 10, 30, 50 and 120 s, respectively. The average
diameters of nano-patterns were 10 Æ 2, 25 Æ 5, 50 Æ 8 and
70 Æ 10 nm corresponding to the Au sputtering time of 10, 30, 50
and 120 s, respectively. The size of nano-patterns almost corresponds
to the Au nano-droplet size that was fabricated by different Au
sputtering time.
Fig. 4(a)–(d) shows the tilt-view FE-SEM images of GaN NWs
grown on nano-patterned Si(1 1 1) substrates with the different
diameters of 10, 25, 50 and 70 nm, respectively. These results
strongly suggested that the GaN NWs growth occurred only at
nano-patterned Si(1 1 1) substrates. The reason why the GaN NWs
or its bulks were formed only on the surface of nano-patterned
Si(1 1 1) substrate is due to the predeposited Al. The Al
predeposition is not only adsorbing GaN on the oxidized SiO2
surface but also provides the efficient nucleation seed for growing
GaN NWs in nano-patterned region. As shown in Fig. 4(a), only a
few GaN bulk grain were observed on the nano-patterned Si(1 1 1)
substrate having the diameter of about 10 nm. In this case, the GaN
NWs were not formed due to very limited sizes of the nano-
patterns. This is because; Ga and N may not nucleate in the too
small nano-patterns on the Si(1 1 1) substrates to grow as the NWs.
In this present study, we focused on the influence of nano-
pattern size on the growth of GaN NWs by keeping other
experimental parameters constant. Fig. 2(a)–(d) shows the top-
view FE-SEM images of the formed Au nano-droplets on Si(1 1 1)
substrates on the different Au sputtering time of 10, 30, 50 and
120 s, respectively. Spherical shaped Au nano-droplets were
formed after the annealing process. The average diameters of
the Au nano-droplets were of 20 Æ 3, 45 Æ 5, 70 Æ 5 and
100 Æ 10 nm corresponding to the different Au sputtering time of
10, 30, 50 and 120 s, respectively. The diameters of the Au nano-
droplets were proportional to the Au sputter time. These results
clearly indicate that the diameter of the Au nano-droplets was
determined by Au sputtering time.
In the early stage of the GaN NWs growth, Au catalyst is known
to play an important role to nucleate the NW growth through VLS
process. At the same time, the Au catalyst assisted NW growth
resulted in a severe contamination which may affect the physical
and chemical properties of the finally grown GaN NWs. During the
VLS growth, the diffusion of Au catalyst into GaN NWs may cause
stacking defects or a twinning plane which can affect the quality of
the NWs [30,31]. Therefore, to avoid unnecessary existence of Au
catalyst in GaN NWs, we successfully etched the Au nano-droplets
on Si(1 1 1) substrates to fabricate clear nano-patterns before the

J.-H. Park et al. / Materials Research Bulletin 47 (2012) 836–842
839
Fig. 3. The top-view FE-SEM images of the nano-patterned Si(1 1 1) substrate with different diameters of; (a) 10, (b) 25, (c) 50 and (d) 70 nm.
However, as the nano-pattern size was slightly increased, initiation
of GaN NWs nucleation occurred and relatively many NWs were
formed and then lengths terminated at certain level as shown in
Fig. 4(b). On the other hand, we were able to grow very high
density with almost vertically aligned GaN NWs corresponding to
the nano-pattern size of 50 nm as shown in Fig. 4(c). Very smooth
surface morphology of GaN NWs without any poly crystals was
obtained, the substrate surface was very clear. The average
diameter and length of GaN NWs grown on nano-patterned
Si(1 1 1) were of 150 Æ 50 nm and 1–2 mm, respectively. The
diameter of GaN NWs was similar to the size of nano-pattern. Inset
of Fig. 4(c) shows the typical alignment of a single GaN NW on the
Si(1 1 1) substrate. At this condition, the formation of GaN bulk was
disappeared and strongly favorable for the growth GaN NWs alone
when compared to the nano-patterns with the diameter less than that
of 50 nm, as in the cases of Fig. 4(a) and (b). The GaN NWs grown on
nano-patterned Si(1 1 1) substrate with the Au sputtering time of
120 s is shown in Fig. 4(d). The GaN NWs grew in random directions
with non-uniform diameters, which is resulted from the very large
diameter of nano-patterns (70 nm) on Si(1 1 1) substrates. In this
case, we observed only a few GaN NWs and many GaN bulk crystals.
We believe that the diameter of GaN NWs is strongly affected by the
diameter of nano-patterns. From this investigation, we can conclude
that the best condition to grow GaN NWs is the nano-pattern with the
diameter of 50 nm on the Si(1 1 1) substrate.
The optical properties of GaN NWs are useful in understand-
ing the physics of nanostructures and their potential applications
in the field of optoelectronics. Fig. 5 shows the room-tempera-
ture PL spectra of the GaN NWs grown with the different
diameters of nano-pattern on Si(1 1 1) substrates. The PL
spectrum of GaN is dominated by the band-edge emission at
362.54 nm. The narrow band emission from the PL spectrum at
room temperature indicates that the NWs have good crystal
quality. The defect related yellow luminescence peak was not
observed, which indicates that as-synthesized GaN NWs have
good optical quality. The PL emission intensity monotonically
increased with increasing nano-pattern size and the maximum
PL intensity was observed corresponding to the nano-patterns
with a diameter of 50 nm. For this case, the full width at half
maximum (FWHM) was of 174 meV. At the same time, the optical
characteristics of the GaN NWs grown were improved with the Al
pre-deposition compared to our previous reports [32,33]. This
may be due to the use of Al pre-deposition can be utilized to
suppress the defect related Ga vacancies of the GaN NWs. The
pre-deposited Al atoms on the Si(1 1 1) substrates can be filled
the deep level defects during the nucleation and growth the GaN
NWs. We could observe the Al have a good influence on the
growth of GaN NWs on Si(1 1 1) as buffer layer. The PL data
showed that the GaN NWs exhibited a good optical result which
leads to the device oriented application.

840
J.-H. Park et al. / Materials Research Bulletin 47 (2012) 836–842
Fig. 4. Tilt-view FE-SEM images of GaN NWs grown on nano-patterned Si(1 1 1) substrate with different diameters of; (a) 10, (b) 25, (c) 50 and (d) 70 nm. Inset of (c) showing a
single GaN NW on Si(1 1 1) substrate.
The fabrication of GaN NWs on nano-patterned Si(1 1 1)
substrates can also be explained more conveniently using the
following mechanism schematically illustrated as shown in Fig. 6.
When the size of nano-pattern was too small (of the order of
10 nm) to initiate the NW nucleation, the adsorption of Ga and N
was barely happened. Even if the Ga and N were absorbed in a
Diameter of nano-patterns
10 nm
25 nm
50 nm
70 nm
nano-pattern, we could not observe GaN NWs rather it would
result in GaN bulk crystals. However, the size of nano-pattern is
most appropriate to accommodate the Ga and
N adatoms
(corresponding to 50 nm), nucleation is readily occurred and we
can expect well aligned growth of GaN NWs, as shown in Fig. 6(b).
When the size of nano-pattern was too big (greater than 50 nm),
plenty amount of adsorption of Ga and N was more easily occurred
in the nano-patterns region. Such kind of nucleation will lead to
formation of NWs with larger diameter, at the same time the NW
growth direction cannot be confined. Therefore, the resultant GaN
NWs were observed to be of bulk crystal having random direction
with non-uniform diameters, as depicted in Fig. 6(c). We found
that the growth of GaN NWs was strongly affected by the size of
nano-patterns. As we observed carefully in this present investiga-
tion, too small and too large diameters of nano-patterns were not
suitable to obtain high quality GaN NWs. The optimum size of
nano-patterns was needed to grow high dense with vertically
aligned GaN NWs.
300
350
400
450
500
550
600
Wavelength (nm)
Fig. 5. Room-temperature PL spectrum of GaN NWs grown on nano-patterned.
Si(1 1 1) substrate with different diameters.

J.-H. Park et al. / Materials Research Bulletin 47 (2012) 836–842
841
Fig. 6. Schematic illustration showing the possible mechanism for the formation of GaN NWs on the (a) too small, (b) optimum and (c) too large sizes of the nano-patterns on
Si(1 1 1) substrates.
References
4. Conclusions
In summary, we have selectively grown high quality GaN NW
arrays on nano-patterned Si(1 1 1) substrate by MOCVD
technique with the variation of pattern size. The different
growth behaviors of GaN NWs were observed on different sizes
of the nano-patterned Si(1 1 1) substrates. Nano-pattern size can
be conveniently altered by varying the Au sputtering time of 10,
30, 50 and 120 s. There was no trace of any catalytic metal or
alloy cluster at top end of the grown GaN NWs because the Au
was completely etched by Au etchant before growing NWs.
Based on the experimental observation, we can conclude that
nano-patterns having the diameter of 50 nm is most favorable
for the GaN NWs growth. The selective growth of GaN NWs is
very sensitive to the diameter of nano-patterns; the lateral
dimensions of GaN NWs were correlated with the initial size of
Au nano-droplets and the related diameter of the fabricated
nano-patterns. These results suggested that the suitable
diameter of the fabricated nano-pattern on Si(1 1 1) for selective
area growth of GaN NWs having excellent characteristics is most
essential in order to apply it for electronic and optoelectronic
applications.
[1] A.M. Morales, C.M. Lieber, Science 279 (1998) 208–211.
[2] L.W. Tu, C.L. Hsiao, T.W. Chi, I. Lo, K.Y. Hsieh, Appl. Phys. Lett. 82 (2003) 1601–
1603.
[3] S. Dhara, A. Datta, C.T. Wu, Z.H. Lan, K.H. Chen, Y.L. Wang, Appl. Phys. Lett. 84
(2004) 3486–3488.
[4] M. He, P. Zhou, S.N. Mohammad, G.L. Harish, J.B. Halpern, R. Jacob, W.L. Sarney, L.S.
Riab, J. Cryst. Growth 231 (2001) 357–365.
[5] G.T. Wang, A.A. Talin, D.J. Werder, J.R. Creighton, E. Lai, R.J. Anderson, I. Arslan,
Nanotechnology 17 (2006) 5773–5780.
[6] R. Koester, J.S. Hwang, C. Durand, D.L.S. Dang, J. Eymery, Nanotechnology 21
(2010), 015602-1-9.
[7] F.M. Ross, Rep. Prog. Phys. 73 (2010), 114501-1-21.
[8] J.R. Kim, H.M. So, J.W. Park, J.J. Kim, J. Kim, C.J. Lee, S.C. Lyu, Appl. Phys. Lett. 80
(2002) 3548–3550.
[9] G. Seryogin, I. Shalish, W. Moberlychan, V. Narayananmurthi, Nanotechnology 16
(2005) 2342–2353.
[10] Y.K. Byeun, K.S. Han, S.C. Choi, J. Ceram. Process. Res. 6 (2005) 197–200.
[11] R. Calarco, M. Marso, T. Richter, A.I. Aykanat, R. Meijers, A.V. Hart, T. Stoica, H.
Luth, Nano Lett. 5 (2005) 981–984.
[12] Y.H. Kim, J.Y. Lee, S.H. Lee, J.E. Oh, H.S. Lee, Appl. Phys. A 80 (2005) 1635–1639.
[13] R. Navamathavan, Y.H. Ra, K.Y. Song, D.W. Kim, C.R. Lee, Curr. Appl. Phys. 11
(2011) 77–81.
[14] T. Kuykendall, P.J. Pauzauskie, Y. Zhang, J. Goldberger, D. Sirbuly, J. Denlinger, P.D.
Yang, Nano Lett. 3 (2003) 1063–1066.
[15] F. Qian, Y. Li, S. Gradecak, D. Wang, C.J. Barrelet, C.M. Lieber, Nano Lett. 4 (2004)
1975–1979.
[16] Y.H. Ra, R. Navamathavan, J.H. Park, K.Y. Song, Y.M. Lee, D.W. Kim, B.B. Jun, C.R.
Lee, Jpn. J. Appl. Phys. 49 (2010), 091003-1-5.
[17] T. Kuykendall, P.J. Pauzauskie, Y. Zhang, J. Goldberger, D. Sirbuly, J. Denlinger, P.D.
Yang, Nat. Mater. 3 (2004) 524–528.
Acknowledgments
[18] M. Yoshizawa, A. Kikuchi, N. Fujita, K. Kushi, H. Sasamoto, K. Kishino, J. Cryst.
Growth 189/190 (1998) 138–141.
[19] C.N.R. Rao, F.L. Deepak, G. Gundiah, A. Govindaraj, Prog. Solid State Chem. 31
(2003) 5–147.
[20] Q. Li, G.T. Wang, Appl. Phys. Lett. 93 (2008), 043119-1-3.
[21] D.H. Kuo, W.H. Wu, J. Electrochem. Soc. 156 (2009) K1–K3.
[22] S. Biswas, S. Kar, T. Choshal, V.D. Ashok, S. Chakrabarti, S. Chaudhuri, Mater. Res.
Bull. 42 (2007) 428–436.
[23] R.A. Burke, D.R. Lamborn, X. Weng, J.M. Redwing, J. Cryst. Growth 311 (2009)
3409–3416.
This research was supported by Basic Science Research Program
through the National Research Foundation of Korea (NRF) funded
by (NRL.R0A-2008-000-20031-0) of Korean Government
(MOEHRD), and the Human Resource Development (Proj.No:
201040100660) of the Korea Institute of Energy Technology
Evaluation and Planning (KETEP) grant funded by the Korea
Government Ministry of Knowledge Economy.

842
J.-H. Park et al. / Materials Research Bulletin 47 (2012) 836–842
[24] W. Han, S. Fan, Q. Li, Y. Hu, Science 277 (1997) 1287–1289.
[25] X. Duan, C.M. Lieber, J. Am. Chem. Soc. 122 (2000) 188–189.
[26] P. Deb, H. Kim, V. Rawat, M. Oliver, S. Kim, M. Marshall, E. Stach, T. Sands, Nano
Lett. 5 (2005) 1847–1851.
[30] S. Kang, B.K. Kang, S.W. Kim, D.H. Yoon, Cryst. Growth Des. 10 (2010) 2581–2584.
[31] D. Tham, C.Y. Nam, J.E. Fischer, Adv. Funct. Mater. 16 (2006) 1197–1202.
[32] E.S. Jang, Y.H. Ra, Y.M. Lee, S.H. Yun, D.W. Kim, R. Navamathavan, J.S. Kim, I.H. Lee,
C.R. Lee, Jpn. J. Appl. Phys. 48 (2009), 090011-1-4.
[27] X. Wang, X. Sun, M. Fairchild, S.D. Harsee, Appl. Phys. Lett. 89 (2006), 233115-1-3.
[28] S.D. Harsee, X. Sun, X. Wang, Nano Lett. 6 (2006) 1808–1811.
[29] R.S. Wagner, W.C. Ellis, Appl. Phys. Lett. 4 (1964) 89–91.
[33] Y.H. Ra, R. Navamathavan, Y.M. Lee, D.W. Kim, J.S. Kim, I.H. Lee, C.R. Lee, J. Cryst.
Growth 312 (2010) 770–774.