JOURNAL OF NANOSCIENCE AND NANOTECHNOLOGY
Raman Spectroscopy and Structure of Crystalline
Gallium Phosphide Nanowires
Qihua Xiong,a, b R. Gupta,a K. W. Adu,a E. C. Dickey,b G. D. Lian,b D. Tham,c J. E. Fischer,c and P. C. Eklunda, b, ¤
a
Department of Physics, Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A.
b
Department of Material Science and Engineering, Pennsylvania State University,
University Park, Pennsylvania 16802, U.S.A.
c
Department of Material Science and Engineering and LRSM, University of Pennsylvania,
Philadelphia, Pennsylvania 19104, U.S.A.
RESEARCH ARTICLE
Gallium phosphide nanowires with a most probable diameter of ¹20.0 nm and more than 10 m m
in length have been synthesized by pulsed laser vaporization of a heated GaP/5% Au target. The
morphology and microstructure of GaP nanowires have been investigated by scanning electron
microscopy and transmission electron microscopy. Twins have been observed along the crystalline
nanowires, which have a growth direction of [111]. Raman scattering shows a 4 cmƒ1 downshift of
the longitudinal optical phonon peak in the nanowire with respect to the bulk; the transverse optical
phonon frequency and line width are, however, the same as in the bulk. The quantum conŽ nement
model Ž rst proposed by Richter et al. cannot explain the observed behavior of the Raman modes.
Keywords: Pulsed Laser Vaporization, Gallium Phosphide Nanowires, Raman Scattering.
Successful synthesis of long silicon nanowires by pulsed
laser vaporization (PLV)1 has driven recent attempts to
synthesize compound semiconducting nanowires such as
GaAs, GaN, InP, etc.,2 and to understand the physical
properties of these quasi-one-dimensional systems.3 Semiconducting nanowires with diameters, d, in the range
10 nm < d < 50 nm and lengths exceeding several
micrometers have intriguing optical and electronic properties that may be distinct from those of bulk semiconductors
and may Ž nd applications in future optoelectronic devices.
Gallium phosphide (GaP) is a very important semiconductor material because of its wide indirect band gap
(¹2.3 eV) and potential applications in optoelectronics.4
Some theoretical calculations have predicted that fundamental physical properties, such as the band gap
of the semiconductor, will change because of quantum
conŽ nement.5 For Si nanowires, phonon quantum conŽ nement effects have been observed by Raman scattering6
and electron conŽ nement by photoluminescence.7 Several methods have been reported recently to produce GaP
nanowires, such as carbon nanotube template-conŽ ned
growth,8 oxide-assisted growth with excimer laser ablation
of a GaP and Ga2 O3 mixture,9 vapor-liquid-solid (VLS) 10
growth with pulsed laser vaporization (PLV) of a GaP target and monodisperse Au nanoparticles deposited on a silicon substrate.11
Here we report on the synthesis of highly crystalline
GaP nanowires by PLV of a GaP/5% Au target and studies
¤
Author to whom correspondence should be addressed.
J. Nanosci. Nanotech. 2003, Vol. 3, No. 4
of phonon conŽ nement by Raman scattering. Our electron microscopy data indicate that the growth proceeds by
the VLS mechanism.10 The morphology and structure of
the nanowires have been investigated by scanning electron
microscopy (SEM) and transmission electron microscopy
(TEM).
Our PLV apparatus is shown schematically in Figure 1.
A pulsed Nd:YAG laser (Telescopic Series SL803, Spectron Laser Systems) with a pulse width of 15 ns and a
repetition rate of 10 Hz was used to vaporize the Au/GaP
target. Both the fundamental (1064 nm at 850 mJ/pulse,
delayed by ¹40 ns) and Ž rst harmonic (532 nm at
450 mJ/pulse) lines were present. The target was made by
pressing GaP (99.99%) powder (Aldrich Inc.) mixed with
5 mol% gold (99.96%) powder (Alfa-Aesar Inc.). The target was centered in a ¹1-m-long quartz tube with an inner
diameter of 25 mm (Fig. 1). A tube furnace was used
to control the target temperature at ¹870 ž C, which was
monitored by an inconel-sheathed thermocouple (type K)
placed just above the target. The nanowires were found on
the wall of the quartz tube ¹30 cm behind the target as a
result of a small Ar/5% H2  ow (100 sccm) in the inner
tube. After the system was pumped down to 20 mTorr
and  ushed several times with Ar gas, the temperature
was increased to 850–890 ž C and the Ar/5% H2  ow was
initiated at a pressure of 100 torr. The laser beam was
focused on the target to a spot size of ¹2 mm and rastered
slowly across the target surface. The ablation or vaporization process was carried out for 2 h. The laser was
then shut down and the power to the furnace turned off,
© 2003 by American ScientiŽ c Publishers 1533-4880/2003/04/335/339/$17.00+.25
doi:10.1166/jnn.2003.208
335
Xiong et al./Raman Spectroscopy of GaP Nanowires
J. Nanosci. Nanotech. 2003, 3, 335–339
Fig. 1. Schematic diagram of the pulsed laser vaporization system. The
gas  ow was Ž rst introduced into the outer quartz tube to allow the gas
to be preheated before it enters the inner quartz tube.
RESEARCH ARTICLE
allowing the system to cool slowly to room temperature
in ¹3 h under  owing Ar/H2 .
The as-grown GaP nanowires exhibit a light yellow
color. The sample was Ž rst dispersed in ethanol with
an ultrasonic horn (Misonix Inc., XL2010) operated at
100 W, and then the suspension was spin-coated onto silicon substrates. A Leica LEO 400 SEM was used to study
the morphology of the GaP nanowires. SEM images were
collected with an acceleration voltage of 5 kV; a thin Au
overlayer was deposited on the nanowires to decrease the
charge accumulation. TEM images were collected with
a FEI-Philips 420ST TEM operated at 120 kV. Highresolution transmission electron microscopy (HRTEM)
images, energy-dispersive X-ray (EDX) spectroscopy, and
scanning transmission electron microscopy (STEM) data
were collected on a JOEL 2010F microscope operated
at 200 kV. The sample for these studies was prepared
by depositing a drop of the nanowire/ethanol suspension on the copper grid. We have used TEM to determine the diameter distribution and growth direction of our
nanowires. The instrument magniŽ cation was calibrated
with the use of a standard grating replica, and the rotation
was calibrated with a -MoO3 platelet. Raman spectra
were collected with a JY-Horiba T64000 micro-Raman
spectrometer in a backscattering conŽ guration at room
temperature. The sample was illuminated with 514.5-nm
radiation with a laser spot diameter of ¹1 Œm, with a
100 £ objective on an Olympus BX40 confocal microscope. The power at the sample was about 10 mW. For
Raman measurements, a thin Ž lm of tangled nanowires
was prepared by putting a few drops of the suspension
on a piece of indium foil and allowing the ethanol to
evaporate.
Figure 2 shows a typical SEM image of the as-prepared
GaP nanowires. The wires appear to be straight and entangled with each other. Most of the wires are longer than
10 Œm. Figure 3 shows a bright-Ž eld TEM image of GaP
nanowires. In the VLS mechanism, a small metal particle
serves as a solvent for the vapor reactant (GaP). The Ž lament grows from the surface of the particle and can be
viewed as a solid phase coexisting with a liquid pseudobinary alloy (e.g., Au-GaP) in the droplet.1–2 Inset (1) to
Figure 3 shows a small solidiŽ ed metal particle at the end
336
Fig. 2. SEM image of as-prepared GaP nanowires on a silicon substrate. A thin layer of Au was coated to decrease the charge accumulation.
The inset shows an individual nanowire with a length of about 10 Œm.
of a nanowire, consistent with the VLS mechanism. We
have used STEM (Fig. 4a) to map the Au concentration
along the nanowire from the droplet. It was found that the
droplet has an abrupt change in Au concentration. Only a
Fig. 3. (a) Bright-Ž eld TEM image of an individual GaP nanowire.
Inset (1): TEM image of the tip of the nanowire. Inset (2): Selected-area
diffraction (SAD) pattern from this nanowire. This pattern was indexed
N zone axis from
as two superimposed diffraction patterns along the [110]
twins, as shown by solid and dashed arrows, respectively. (b) Diameter
distribution of the GaP nanowires calculated from several TEM images.
The solid curve is a Ž t to a log-normal function.
Xiong et al./Raman Spectroscopy of GaP Nanowires
J. Nanosci. Nanotech. 2003, 3, 335–339
(a)
B
A
RESEARCH ARTICLE
10 nm
Fig. 5. HRTEM image of a GaP nanowire with a diameter of about
25 nm. No appreciable amorphous layers were found outside of the
nanowires. This is also supported by EDX data (Fig. 4c).
(b)
Cu
Au
P
Cu
Ga
C
O
Au
Cu
Ga
Au
0
5
Au
1100
E n e r gy ( k e V )
15
(c)
Ga
P
Ga
Cu
Cu
CO
0
Cu
2
4
6
8
E n e r gy ( k e V )
Ga
10
12
Fig. 4. (a) A STEM image of a nanowire with the tip. (b) An energydispersive X-ray spectrum obtained at the spot A (a) in the droplet.
(c) The same spectrum taken at spot B (a) on the nanowire. The C and
Cu are from the grids.
trace of Au was found within 1 nm of the droplet in one
wire (out of Ž ve). We believe this is due to Au diffusion
during the growth. No Au was found in the body of the
wires; this is desirable for future investigation of transport
and optical properties. We also checked the composition
of the droplet by energy dispersive X-ray spectroscopy
(EDX) (Fig. 4b). It was found that the droplet contains
mainly Au, Ga, and P, which support the VLS mechanism.
Figure 4c shows an EDX spectrum of a nanowire; it contains Ga and P. The tiny amount of oxygen might come
from the solvent or surface adsorption contamination from
the atmosphere, since no appreciable amorphous layer was
observed on the surface of the wires in HRTEM (Fig. 5).
The TEM images (Fig. 3a) also show some segmental
contrast features (arrow), which are attributed to twins.
Inset (2) to Figure 3a shows a selected-area diffraction (SAD) pattern from an individual GaP nanowire.
The growth direction has been determined to be “111”,
as shown in Fig. 3a and 5. GaP exhibits a zincblende
N
structure with space group F 43m.
The diffraction patN ”
tern in Fig. 3a was indexed as two superimposed “110
zone axis patterns that have a ¹110ž (or 70ž ) rotation
with respect to each other. This can be attributed to
twins with two equivalent “111” directions along the
wire axis. Twin boundaries have also been observed
in Si and Ge nanowires,1 although the origin of the
twins in nanowires is still not understood.1 This type
of rotational twin with a “111” twin axis has also been
observed in GaAs and InAs whiskers grown by metalorganic vapor phase epitaxy (MOVPE). 12 It has been suggested that this defect strongly depends on the growth
conditions12 (e.g., temperature, pressure, and gas  ow).
The diameter distribution, as shown in Figure 3b, has
been calculated from several TEM images. The distribution has been Ž tted to a log-normal distribution
function (log-normal function: AExpƒ4ln4x=x0 5=w52 ,
where x0 is the peak position in nanometers and w
is the full width at half-maximum in nanometers); the
peak in Fig. 3b is located at 20.3 (§ 108) nm with
the log-normal tail extending to larger diameters (¹70–
80 nm). It should be noted that the growth conditions had not been optimized to produce smaller-diameter
nanowires or a tighter diameter distribution. This diameter
distribution can be controlled to some extent by the
growth conditions, such as gas pressure and choice of gas
(nitrogen, argon, or helium).13 Figure 5 shows an HRTEM
image of a GaP nanowire with a diameter of about 25 nm.
The image shows that the wire is highly crystalline and
337
Xiong et al./Raman Spectroscopy of GaP Nanowires
RESEARCH ARTICLE
that no appreciable oxide layer was found, which is supported by EDX data (Fig. 4c).
Raman scattering is a powerful tool for characterizing the quality of materials. It is sensitive to composition, presence of disorder, impurities, and the size of
the system under investigation. The downshifting and
broadening of Raman bands in nanostructured semiconductors have been qualitatively explained in terms of a
phonon conŽ nement model of Richter.14 The Ž rst quantitative attempt to understand these effects was made by
Fauchet and Campbell,15 who used Richter’s phenomenological phonon conŽ nement model.14 The physics underlying the phonon conŽ nement effect can be understood as
follows: In a nanowire, because of the dimensional constraint along the directions perpendicular to the wire axis,
a range of phonon wave vectors, 0 < q < 1=d, is required
to describe the Raman-active phonon in the nanowire.
Since for most semiconductors, the optical-mode frequencies in the bulk decrease away from the Brillouin zone
center, the Raman-active mode in the nanowire broadens
to lower frequency and downshifts. This effect is larger
in smaller-diameter wires. Thus, conŽ nement leads to a
breakdown in the q D 0 Raman selection rule applied to
crystalline bulk materials.
Figure 6 shows a typical Raman spectrum of our GaP
nanowires; polycrystalline bulk GaP is measured under
the same optical conditions; and the spectrum is included
for comparison. Each spectrum consists of two peaks that
we identify with Ž rst-order scattering from longitudinal
optical (LO) and transverse optical (TO) phonons. We
presume that in the nanowires we have —4LO5 > —4TO5,
as in the bulk. Therefore, for GaP nanowires, the LO
phonon mode is seen to be downshifted by ¹4 cmƒ1 with
respect to its position in the bulk. Also, the nanowire LO
mode exhibits a noticeable asymmetric broadening of the
Fig. 6. Raman spectra of GaP nanowires and polycrystal GaP materials. For nanowires, the Ž rst-order LO phonon band is downshifted and
broadened by ¹4 cmƒ1 with respect to the bulk, whereas the Ž rst-order
TO band remains the same as that observed in bulk GaP.
338
J. Nanosci. Nanotech. 2003, 3, 335–339
LO band not seen in the bulk material. However, the TO
phonon peak in the nanowire remains almost the same
as that observed in the bulk GaP. It should be also noted
that the full width at half-maximum (FWHM) of the TO
band of our nanowires is the same as that of the bulk.
This is additional proof of the structural quality of these
nanowires. The observations of the LO and TO modes
are notably different from that reported by Shi et al. for
oxide-assisted growth of GaP nanowires.9 In the Raman
measurements of Shi et al., the LO and TO bands in the
nanowire downshifted by 8 and 5 cmƒ1 , respectively, and
both broadened. They attributed this to combined effects
of conŽ nement, planar defects, and stress within the GaP
nanowires arising because of lattice distortion.
We have carried out a simple calculation of the Raman
lineshape for our nanowires, using the model of Richter14
extended by Fauchet and Campbell.15 The Raman intensity for the “conŽ ned” TO or LO phonon in a nanowire
of radius r0 is written as
INW 4—5 D
Z
0
q max
—C 4q5 —2
2 q dq
6— ƒ —4q572 C 4â =252
(1)
where —4q5 is the phonon dispersion of the LO or
TO Raman mode, â is the natural linewidth, and the
wave vector q is perpendicular to the wire axis “111”.
C4q5 is the Fourier coefŽ cient of the phonon conŽ nement function and depends on r0 . We assume that the
phonon dispersion in the nanowires is the same as that
of bulk GaP, and we take an approximate isotropic dispersion given by —4q5 D 6A C B cos4q=4 =a5571=2 , where
A D 1371670 cmƒ1 , B D 211155 cmƒ1 for the LO branch
and A D 1261025 cmƒ1 , B D 71200 cmƒ1 for the TO
branch Ž t to the neutron scattering data.16 In Figure 7 we
Fig. 7. Calculated LO (a) and TO (b) Raman band line shape in the
Fauchet-Campbell model as a function of wire diameter d (d D ˆ represents the bulk GaP).
Xiong et al./Raman Spectroscopy of GaP Nanowires
J. Nanosci. Nanotech. 2003, 3, 335–339
and wire lengths of about 10 Œm are typical. The wires
grow by the VLS mechanism. TEM images and the
selected-area diffraction pattern show that some of the
nanowires are twinned along the [111] growth direction. An asymmetrical broadening of only the Ž rst-order
longitudinal optical phonon is observed; the transverse
optical phonon is unaffected. This observation cannot
be accounted for by the phonon conŽ nement model of
Richter-Fauchet-Campbell, and an alternative explanation
is needed to account for the experimentally observed
behavior of the Raman bands in GaP nanowire.
Acknowledgments: This work was primarily supported
by NSF-NIRT (grant DMR-0103585) and the NSFMRSEC program at U-Penn (grant DMR00-79909).
Q. Xiong acknowledges support from NSF-NIRT (DMR0103585). R. Gupta was supported by UPenn-MRSEC
(DMR00-79909).
RESEARCH ARTICLE
show the calculated Raman lineshape for both the LO
and TO modes for nanowires with different diameters.
There are two points of signiŽ cance in Figure 7. First,
it would be very difŽ cult to observe any conŽ nement
effects on the LO or TO Raman bands for nanowires
with d ¶ 10 nm. Second, the conŽ nement effect for
the LO mode at d D 5 nm is stronger than for the TO
mode. This stems from the stronger LO phonon dispersion. Returning to our experimental data (Fig. 6), we see
that only the LO mode is appreciably different from its
counterpart in the bulk. If this shift were a simple consequence of phonon conŽ nement, then we would expect
a similar shift in the TO mode. Secondly, the RichterFauchet-Campbell model predicts almost no shift for 20nm-diameter GaP nanowires, which is the most probable
diameter for our sample obtained from a log-normal Ž t
to the measured distribution. We do not expect the smalldiameter wires (<4 nm) in the sample to cause only the
LO nanowire band downshift. Therefore we have not carried out the above calculation incorporating the diameter
distribution of the nanowires. In an earlier work on bulk
crystalline GaP doped with sulfur or tellurium impurities
by Galtier and Martinez,17 it was found that in the presence of impurities only the LO mode downshifts. However, the downshifts observed in their experiments were
small (¹1 cmƒ1 ), even for degenerately doped GaP. In
our experiments the downshifting of the LO band is large,
¹4 cmƒ1 . We also believe that if the downshifting is due
to the screening of the LO mode in the presence of impurities, then it should lead to broadening of both the LO
and TO modes in the nanowires. In our experiments, we
observe that the line widths of the TO mode in bulk and
nanowires are nearly the same. In a recent work, Mahan
et al.18 have shown that the long-range dipolar interactions in a polar semiconductor responsible for the splitting
of the LO-TO modes in the bulk give rise to a different
splitting in the nanowire. Their study shows that the LOTO splitting depends on the aspect ratio of the sample;
it is a “shape” effect rather than a “size” effect. Preliminary analysis seems to indicate that our data are consistent with this model. However, it is important to point
out that further polarization-dependent measurements on
single nanowires are needed to see if the character of the
LO (TO) mode in a nanowire is different from that in
the bulk. The FWHM of the LO mode in the nanowire
is much larger than that in the bulk. Part of the reason for this is that a slightly lower frequency surface
mode is present and overlaps the LO band. A paper on
these surface modes will be submitted.19 Further work is
in progress to understand the surface-mode properties in
GaP nanowires.
In summary, highly crystalline GaP nanowires have
been successfully synthesized by pulsed laser ablation of
a target made of (GaP)0095 Au0005 powders. The nanowire
diameter distribution is well described by a log-normal
distribution with a most probable diameter of ¹20.0 nm,
References and Notes
1. A. M. Morales and C. M. Lieber, Science 279, 208 (1996).
2. X. Duan Duan, J. Wang, and C. M. Lieber, Appl. Phys. Lett.
76, 1116 (2000); X. Duan and C. M. Lieber, J. Am. Chem. Soc.
122, 188 (2000); X. Duan and C. M. Lieber, Adv. Mater. 12, 298
(2000).
3. X. Duan, Y. Huang, J. Wang, and C. M. Lieber, Nature 409, 66
(2001); Y. Cui, X. Duan, J. Hu, and C. M. Lieber, J. Phys. Chem.
B 104, 5213 (2000); J. Wang, M. S. Gudiksen, X. Duan, Y. Cui,
and C. M. Lieber, Science 293, 1455 (2001).
4. K. S. Jones, S. J. Pearton, and H. Kanber, Mater. Res. Symp. Proc.
300, 101 (1993).
5. A. J. Read, R. J. Needs, K. J. Nash, L. T. Canham, P. D. J.
Calcott, and A. Qteish, Phys. Rev. Lett. 69, 1232 (1992); J. Behren,
T. Buuren, M. Zacharias, E. H. Chimowitz, and P. M. Fachert, Solid
State Commun. 105, 317 (1998).
6. R. P. Wang, G. W. Zhou, Y. L. Liu, S. H. Pan, H. Z. Zhang, D. P.
Yu, and Z. Zhang, Phys. Rev. B 61, 16827 (2000).
7. D. P. Yu, Z. G. Bai, J. J. Wang, Y. H. Zou, W. Qian, J. S. Fu, H. Z.
Zhang, Y. Ding, G. C. Xiong, L. P. You, J. Xu, and S. Q. Feng,
Phys. Rev. B 59, 2498 (1999).
8. C. C. Tang and S. S. Fan, Adv. Mater. 12, 1346 (2000).
9. W. S. Shi, Y. F. Zheng, N. Wang, C. S. Lee, and S. T. Lee, J. Vac.
Sci. Technol., B 19, 1115 (2001).
10. R. S. Wagner and W. C. Ellis, Appl. Phys. Lett. 4, 89 (1964).
11. M. S. Gudiksen and C. M. Lieber, J. Am. Chem. Soc. 122, 8801
(2000).
12. M. Koguchi, H. Kakibayashi, M. Yazawa, K Hiruma, and T. Katsuyama, Jpn. J. Appl. Phys. 31, Pt. 1, 2061 (1992); K. Hiruma,
M. Yazawa, K. Ogawa, K. Haraguvhi, M. Koguchi, and H. Kakibayashi, J. Appl. Phys. 77, 447 (1995).
13. Y. F. Zhang, Y. H. Tang, H. Y. Peng, N. Wang, C. S. Lee, I. Bello,
and S. T. Lee, Appl. Phys. Lett. 75, 1842 (1999).
14. H. Richter, Z. P. Wang, and L. Ley, Solid State Commun. 39, 625
(1981).
15. L. H. Campbell and P. M. Fauchet, Solid State Commun. 58, 739
(1986).
16. J. L. Yarnell, J. L. Warren, R. G. Wenzel, and P. J. Dean, Neutron
Inelastic Scattering 1, 301 (1968).
17. P. Galtier and G. Martinez, Phys. Rev. B 38, 105542 (1975).
18. G. D. Mahan, R. Gupta, Q. Xiong, C. K. Adu, P. C. Eklund, Phys.
Rev. B, in press (2003).
19. R. Gupta, Q. Xiong, and P. C. Eklund, manuscript in preparation.
Received: 18 September 2002. Revised/Accepted: 10 March 2003.
339