Characterization of InN layers grown by high-pressure chemical vapor deposition

Article abstract of DOI:10.1063/1.2352797

Structural and optical properties of indium nitride (InN) layers grown by high-pressure chemical vapor deposition (HPCVD) on sapphire and GaN epilayers have been studied. HPCVD extends processing parameters beyond those accessible by molecular beam epitax

Full text of DOI:10.1063/1.2352797

Characterization of InN layers grown by high-pressure chemical vapor deposition
M. Alevli, G. Durkaya, A. Weerasekara, A. G. U. Perera, N. Dietz, W. Fenwick, V. Woods, and I. Ferguson
Citation: Applied Physics Letters 89, 112119 (2006); doi: 10.1063/1.2352797
View online: http://dx.doi.org/10.1063/1.2352797
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/89/11?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Effect of reactor pressure on the electrical and structural properties of InN epilayers grown by high-pressure
chemical vapor deposition
J. Vac. Sci. Technol. A 30, 031511 (2012); 10.1116/1.4705727
Optical characterization of InN layers grown by high-pressure chemical vapor deposition
J. Vac. Sci. Technol. A 26, 1023 (2008); 10.1116/1.2908736
The influence of substrate polarity on the structural quality of InN layers grown by high-pressure chemical vapor
deposition
Appl. Phys. Lett. 92, 041911 (2008); 10.1063/1.2840192
Optical and electrical properties of Al 1−x In x N films grown by plasma source molecular-beam epitaxy
Appl. Phys. Lett. 79, 632 (2001); 10.1063/1.1388883
GaN homoepitaxial layers grown by metalorganic chemical vapor deposition
Appl. Phys. Lett. 75, 1276 (1999); 10.1063/1.124666
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.211.208.19 On: Wed, 03 Dec 2014 09:26:56

APPLIED PHYSICS LETTERS 89, 112119 ͑2006͒
Characterization of InN layers grown by high-pressure chemical
vapor deposition
a͒
M. Alevli, G. Durkaya, A. Weerasekara, A. G. U. Perera, and N. Dietz
Department of Physics and Astronomy, Georgia State University, Atlanta, Georgia 30303
W. Fenwick, V. Woods, and I. Ferguson
School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332
͑
Received 4 May 2006; accepted 21 July 2006; published online 13 September 2006͒
Structural and optical properties of indium nitride ͑InN͒ layers grown by high-pressure chemical
vapor deposition ͑HPCVD͒ on sapphire and GaN epilayers have been studied. HPCVD extends
processing parameters beyond those accessible by molecular beam epitaxy and metal organic
chemical vapor deposition, enabling the growth of epitaxial InN layers at temperatures as high as
1150 K for reactor pressures around 15 bars, leading to vastly improved material properties. InN
layers grown on GaN͑0002͒ epilayers exhibit single-phase InN͑0002͒ x-ray diffraction peaks with
full width at half maximum ͑FWHM͒ around 430 arc sec. Optical characterization of the InN layers
+
19
−3
by infrared ͑IR͒ reflectance reveals free carrier concentrations in the low to mid-10 -cm and
optical dielectric function ␧ =5.8. The optical properties in the visible and near IR spectral ranges
ϱ
were analyzed by transmission spectroscopy, showing an absorption edge around 1.5 eV. The shift
of the absorption edge correlates with deviations in the InN stoichiometry, indicating that the
understanding and control of the point defect chemistry of InN is critical for improved material
properties. © 2006 American Institute of Physics. ͓DOI: 10.1063/1.2352797͔
In recent years, research on InN material optimization
has dramatically increased, largely due to the crucial impor-
tance of indium-rich group III-nitride alloys in novel device
structures for solid-state lighting, photovoltaics, spintronics,
or terahertz applications utilizing the large spectral tunability
and multifunctionality of group III-nitride alloys. However,
the integration of indium-rich group III-nitride layers into
The InN layers have been grown by HPCVD, a high-
pressure flow channel reactor with incorporated real-time op-
tical characterization capabilities in order to study and opti-
mize InN nucleation and growth. Ammonia ͑NH ͒ and
3
trimethyl indium ͑TMI͒ are employed as precursors in a
pulsed injection scheme utilizing pulse width, precursor
pulse separation, and cycle sequence time as control param-
eters in order to control the gas phase and surface chemistry
kinetics. The precursors are embedded in a high-pressure car-
rier gas stream and injected in the reactor utilizing a tempo-
rally controlled gas injection system. This approach enables
the precise control/suppression of gas phase reactions and the
tailoring of surface chemistry processes, which are moni-
tored in real time by UV absorption spectroscopy and prin-
cipal angle reflectance spectroscopy.
Ga1−xAl N alloys strongly depends on the existence of over-
x
lapping processing windows as well as the precise control of
the thermal decomposition pressures and stoichiometry of
indium-rich alloys at the optimum processing temperatures.
Presently, the most efficient growth of wide band gap
group III-nitride semiconductors are achieved by metal or-
ganic chemical vapor deposition ͑MOCVD͒ and molecular
beam epitaxy ͑MBE͒. However, the epitaxial growth of InN
under low-pressure conditions such as MOCVD or MBE is
problematic due to large thermal decomposition pressure at
the optimum growth temperature, creating conflicting mate-
rial properties due to point defect chemistry in InN, which at
The InN layers investigated here have been grown at
temperatures around 1150 K, a reactor pressure of 15 bars,
an ammonia to TMI precursor ratio of 600, and total gas
flows between 12 and 15 slm ͑standard liters per minute͒.
During the growth, the gas flow as well as the reactor pres-
sure are maintained constant at all times. The temperature
setting refers to the calibrated correlation of the analyzed
blackbody radiation as a function of the power setting of the
substrate heaters and is not corrected for the change in sur-
face emissivity during the growth. Details of the HPCVD
reactor, the growth configuration, as well as real-time optical
characterization techniques employed have been published
1
,2
present is not well understood. Surface stabilization data
have shown that InN can be grown at much higher tempera-
3
tures if stabilized at high nitrogen pressures, evoking the
development of a novel high-pressure chemical vapor depo-
sition ͑HPCVD͒ system at Georgia State University. This
approach allows us to control the vastly different partial
pressures of the constituents involved in the growth of
4
–9
indium-rich group III-nitride alloys. The combination of
HPCVD and real-time process monitoring/control has been
demonstrated to be a viable method to improve the InN ma-
terial properties, with InN growth temperatures as high as
4
,6–9
elsewhere.
The x-ray diffraction ͑XRD͒ measurements have been
carried out in a ␪–2␪ coupled geometry using Cu K␣ x rays
to evaluate the presence of secondary phases or polycrystal-
linity. A 1/2 deg slit is used on the incident beam optics and
a 1/4 deg slit on the diffracted beam optics for XRD. The
Raman spectra are observed in the backscattering configura-
tion. All Raman spectra are normalized to the peak intensity
1150 K for reactor pressures around 15 bars. This is a major
step towards the fabrication of indium-rich group III-nitride
heterostructures by providing a closer match to the process-
ing windows used for GaN–AlN alloys.
a͒Author to whom correspondence should be addressed; electronic mail:
ndietz@gsu.edu
of the E ͑high͒ phonon feature. Room temperature IR reflec-
2
tion measurements have been performed over the frequency
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
003-6951/2006/89͑11͒/112119/3/$23.00 89, 112119-1 © 2006 American Institute of Physics
31.211.208.19 On: Wed, 03 Dec 2014 09:26:56
0
1

112119-2
Alevli et al.
Appl. Phys. Lett. 89, 112119 ͑2006͒
FIG. 1. ͑Color online͒ XRD spectra for InN layers grown on sapphire sub-
strate ͑71U͒ and virtual GaN/sapphire substrate ͑76U͒.
FIG. 3. ͑Color online͒ IR reflectance spectrum and best fit for an InN layer
deposited on a virtual GaN/sapphire substrate ͑76U͒.
range of 200–8000 cm−1 ͑50–1.25 ␮m͒ by using a Perkin-
Elmer system 2000 fast Fourier transformed infrared spec-
trometer and Graesby optical reflection accessory setup. All
IR reflection spectra were taken under near normal
89–492 cm−1 is attributed to the scattering of light from
4
10
E ͑high͒ phonon mode which is most sensitive to strain,
2
−1
whereas the peak at 587 cm is assigned to A ͑LO͒ phonons
in InN. The peak positions are in good agreement with those
found in InN layers grown by MBE and MOCVD. Based on
1
͑
ϳ8 deg͒ incident light arrangement in order to minimize
anisotropy effects. Room temperature transmission measure-
ments were performed with a custom build near-infrared–
visible–UV spectrometer, which consists of a triple-grating
1
the allowed phonon mode peak positions for E ͑high͒ and
2
A ͑LO͒, our experimental peak positions are similar to the
1
1
/2 m monochromator with phase-sensitive signal detection
−1
calculated values of 483 and 588 cm by Davydov and
Klochikhin.
and processing. A photomultiplier and InGaSb or CdMnTe
photodiodes are utilized as detectors for the appropriate
wavelength regions.
Figure 1 shows the variation of XRD patterns on a loga-
rithmic scale for InN layers grown directly on sapphire and
GaN epilayers/sapphire substrates. The XRD peaks for
samples 76U and 71U are centered at 2−␪=31.357 and
1
Figure 3 depicts the IR reflectance spectrum for sample
7
6U and the analysis of optical properties in the IR region.
The dielectric functions of the InN and GaN layers are mod-
eled using Eq. ͑1͒ assuming two contributions, one of a Lor-
entz oscillator for phonons and the second one is the classi-
11,12
cal Drude model for the plasma permittivity,
31.243 deg, respectively. These peaks correspond to the dif-
fraction of the hexagonal phase InN͑0002͒ plane. The full
2
− ␻2
2
2
␻
2
TO
␻
LO
TO
P
width at half maximum ͑FWHM͒ of InN͑0002͒ peaks are
␧͑␻͒ = ␧ϱ
ͫ
1 +
+
ͬ
,
͑1͒
2
␻
− ␻ − i␻⌫ ␻ + i␻␥P
432 and 528 arc sec for the InN layers deposited on GaN/
sapphire and sapphire substrates, respectively. The slight
asymmetry on the left hand of the InN͑0002͒ peaks may
indicate the presence of a second phase in very close prox-
imity, which will need further studies to clarify the origin.
Figure 2 shows Raman spectra of InN samples grown on
sapphire substrate ͑71U͒ and GaN/sapphire substrate
where ␧ is the high frequency dielectric constant, ␻ and
ϱ
LO
␻TO are LO and TO frequencies of E phonon mode, ␻ is
1
P
the plasma frequency, and ⌫ and ␥ are the two correspond-
P
ing damping constants. Similarly, the dielectric functions for
the GaN epilayer and sapphire substrate are modeled as de-
1
3
scribed in Ref. 12. A matrix method is used to calculate the
multilayer stack reflection. The optical properties of the GaN
layer were measured and analyzed using a three-layer reflec-
tion model ͑air/GaN/sapphire͒ before the InN layer was de-
posited. The best fit parameters for the GaN film was ob-
tained using the nonlinear Levenberg-Marquardt fitting
͑
76UL͒, which were analyzed using excitation energy of
2
3
.33 eV. The two optical phonon modes of hexagonal InN at
00 K analyzed in Raman spectra are E ͑high͒ and A ͑LO͒.
2
1
The peak centered at wave numbers in the range of
1
4
algorithm.
Thereafter,
the
IR
reflection
of
InN/GaN/sapphire structure has been measured and the best
fit parameters for InN film have been obtained by using the
same fitting algorithm as above. The frequencies of the LO
and TO phonon modes were kept constant at 593 and
76 cm− during the fitting process. The free carrier concen-
1
5
trations were calculated by
2
P
␻
m ␧ ␧
eff ϱ 0
nc =
,
͑2͒
q2
with ␻ obtained from the fitting process. In Eq. ͑2͒, m is
P
eff
the effective mass of electron in InN, ␧ is permittivity in
0
vacuum, and q is the electron charge. The electron effective
1
5
mass was set as 0.09m , where m is the free electron mass.
FIG. 2. ͑Color online͒ Raman spectra for InN layers grown on sapphire
0
0
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
substrate ͑71U͒ and on a GaN epilayer ͑76U͒.
The change of the electron effective mass as a function of the
1
31.211.208.19 On: Wed, 03 Dec 2014 09:26:56

112119-3
Alevli et al.
Appl. Phys. Lett. 89, 112119 ͑2006͒
monia:TMI flow ratio also provided no direct correlation be-
tween free carrier concentration and InN stoichiometry de-
viations for the InN layers investigated. Further studies on
InN layers grown in an expanded process window are needed
to correlate the point defect chemistry, growth temperature,
and reactor pressure with the structural and optical properties
of InN.
In conclusion we have studied the structural and optical
properties of InN layers grown by HPCVD on quasi-lattice-
matched GaN epilayer/sapphire substrates and sapphire sub-
strates. The XRD analysis showed high-quality, single-phase
InN͑0002͒ peaks with hexagonal symmetry and a FWHM
around 430 arc sec for InN layers deposited on GaN epilay-
ers. The FWHM increases with the lattice mismatch to about
FIG. 4. ͑Color online͒ Transmission spectra for InN grown on sapphire
substrate ͑71U͒ and GaN epilayer/sapphire substrate ͑76U͒.
5
30 arc sec for InN deposited on sapphire substrates. Sharp
−
1
E ͑high͒ and A ͑LO͒ at 488 and 590 cm phonon modes are
2
1
free carrier concentration is not taken into account. The av-
observed. At present, the free carrier concentrations in these
erage of the values ␧ obtained from the fittings is 5.59 and
+19
−3
ϱ
InN layers are in the mid-10 -cm and carrier mobilities
is used for the free carrier calculation.
−2 −1 −1
around 430 cm
V
s , values that can be improved upon
The best fit approximation of the experimental data
shown in Fig. 3 reveals an InN layer thickness d=317 nm, a
further process optimization. The transmission spectra indi-
cate that the band gap of the InN layers grown by HPCVD is
well above 1 eV, which is contrary to results reported for
plasma-assisted MBE grown InN layers. Our results also in-
dicate that the observed absorption edge shift does not follow
the Moss-Burstein effect but is extremely sensitive to small
deviations in the InN stoichiometry, requiring a precise con-
trol of gas phase and surface chemistry processes.
−
1
plasma frequency ␻ =3461 cm , and a plasma damping
P
−
1
constant ␥ =239 cm . The carrier mobility ␮ is calculated
P
c
1
6
via the effective mass and the damping constant ␥P and
−
2
−1 −1
found to be ␮ =434 cm
V
s .
c
Figure 4 shows the room temperature transmission spec-
tra for samples 71U and 76U, taken at the same spots as the
XRD spectra shown in Fig. 1 were taken. Each spectrum is
corrected for the substrate and spectrometer response char-
acteristics. The calculated absorption spectra indicate an ab-
sorption edge around 1.5 eV with absorption structures
around 1.2 and 0.7 eV. Sample 76U exhibits characteristic
interference fringes due to the underlying GaN epilayer of
This work was supported by NASA Grant No. NAG8-
1686 and GSU-RPE.
1
V. Yu. Davydov and A. A. Klochikhin, Semiconductors 38, 861 ͑2004͒.
2
K. S. A. Butcher and T. L. Tansley, Superlattices Microstruct. 38, 1
3͑
J. MacChesney, P. M. Bridenbaugh, and P. B. O’Connor, Mater. Res. Bull.
, 783 ͑1970͒.
2005͒.
1789 nm thickness. The optical analysis on InN layers grown
under similar conditions but with slight variations in the am-
5
4
5
6
7
8
monia:TMI flow ratio ͑NH :TMI=600±30͒ showed that
3
N. Dietz, M. Strassburg, and V. Woods, J. Vac. Sci. Technol. A 23, 1221
͑2005͒.
these small deviations in the InN stoichiometry shift the ab-
sorption edge from 1.5 down to 1.1 eV, while the free carrier
concentration and mobility values obtained by IR reflectance
are almost unchanged. The XRD analysis showed single-
phase InN͑0002͒ peaks with a slight broadening to FWHM
values for ammonia:TMI flow ratios below 570.
N. Dietz, M. Alevli, H. Kang, M. Straßburg, V. Woods, I. T. Ferguson, C.
E. Moore, and B. H. Cardelino, Proc. SPIE 5912, 78 ͑2005͒.
B. H. Cardelino, C. E. Moore, C. A. Cardelino, and N. Dietz, Proc. SPIE
5
912, 86 ͑2005͒.
N. Dietz, M. Alevli, V. Woods, M. Strassburg, H. Kang, and I. T. Fergu-
son, Phys. Status Solidi B 242, 2985 ͑2005͒.
N. Dietz, in III-Nitrides Semiconductor Materials, edited by Z. C. Feng
The correlation of absorption edge shift and free carrier
concentration as obtained by IR reflectance does not support
9͑
Imperial College Press, 2006͒, Chap. 6, pp. 203–235.
1
7
V. Woods and N. Dietz, Mater. Sci. Eng., B 127, 239 ͑2006͒.
V. Y. Davydov, V. V. Emtsev, I. N. Goncharuk, A. N. Smirnov, V. D.
Petrikov, V. V. Mamutin, V. A. Vekshin, S. V. Ivanov, M. B. Smirnov, and
T. Inushima, Appl. Phys. Lett. 75, 3297 ͑1999͒.
the proposed Moss-Burstein effect as the leading cause for
the shift of the fundamental absorption edge to higher values
when the carrier concentration is increased. Additional ef-
10
1
8
11
J. S. Thakur, G. W. Auner, D. B. Haddad, R. Naik, and V. M. Naik, J.
Appl. Phys. 95, 4795 ͑2004͒.
fects such as stoichiometry deviations and the associated
point defect chemistry have to be considered to understand
the physical properties of InN.
1
2
Z. G. Hu, M. Strassburg, A. Weerasekara, N. Dietz, A. G. U. Perera, M. H.
Kane, A. Asghar, and I. T. Ferguson, Appl. Phys. Lett. 88, 061914 ͑2006͒.
C. C. Katsidis and D. I. Siapkas, Appl. Opt. 41, 3978 ͑2002͒.
Numerical Recipes in C: The Art of Scientific Computing, edited by W. H.
Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, 2nd ed.
͑Cambridge University Press, Cambridge, MA, 1992͒, p. 681.
T. Inushima, M. Higashiwaki, and T. Matsui, Phys. Rev. B 68, 235204
7
–9
1
1
3
4
As shown previously, ammonia:TMI flow ratios be-
low 500 cause an absorption edge shift down well below
0
.7 eV, while the free carrier concentration remains in the
+
19
−3
mid-10 -cm . For ammonia:TMI flow ratios below 500
the XRD analysis has shown also the presence of the
InN͑101͒ phase in addition to the InN͑0002͒ phase.
15
1
6͑
2003͒.
A. Kasic, M. Schubert, Y. Saito, Y. Nanishi, and G. Wagner, Phys. Rev. B
65, 115206 ͑2002͒.
7
No equivalent changes are found in the Raman spectra
that would provide a link between the appearance of the
InN͑101͒ peak in the XRD spectra and changes in the
E ͑high͒ and A ͑LO͒ Raman modes. The analysis of the
1
W. Walukiewicz, S. X. Li, J. Wu, K. M. Yu, J. W. Ager III, E. E. Haller,
H. Lu, and W. J. Schaff, J. Cryst. Growth 269, 119 ͑2004͒.
P. P.-T. Chen, K. S. A. Butcher, M. Wintrebert-Fouquet, R. Wuhrer, M. R.
Phillips, K. E. Prince, H. Timmers, S. K. Shrestha, and B. F. Usher, J.
Cryst. Growth 288, 241 ͑2006͒.
18
2
1
E ͑high͒ and A ͑LO͒ Raman modes as a function of the am-
2
1
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
31.211.208.19 On: Wed, 03 Dec 2014 09:26:56
1