High-surface-quality nanocrystalline InN layers deposited on GaN templates by RF sputtering

Article abstract of DOI:10.1002/pssa.200925636

We report a detailed study of the effect of deposition parameters on optical, structural, and morphological properties of InN films grown by reactive radio-frequency (RF) sputtering on GaN-on-sapphire templates in a pure nitrogen atmosphere. Deposition parameters under study are substrate temperature, RF power, and sputtering pressure. Wurtzite crystallographic structure with c-axis preferred growth orientation is confirmed by X-ray diffraction measurements. For the optimized deposition conditions, namely at a substrate temperature of 450 °C and RF power of 30 W, InN films present a root-mean-square surface roughness as low as ～0.4 nm, comparable to the underlying substrate. The apparent optical bandgap is estimated at 720 nm (1.7 eV) in all cases. However, the InN absorption band tail is strongly influenced by the sputtering pressure due to a change in the species of the plasma.

Full text of DOI:10.1002/pssa.200925636

Phys. Status Solidi A 208, No. 1, 65–69 (2011) / DOI 10.1002/pssa.200925636
a
p s s
www.pss-a.com
applications and materials science
High-surface-quality nanocrystalline
InN layers deposited on GaN templates by RF sputtering
,
1
1
1
2
*
Sirona Valdueza-Felip , Fernando B. Naranjo , Miguel Gonz a´ lez-Herr a´ ez , Lise Lahourcade ,
2
3
Eva Monroy , and Susana Fern a´ ndez
1
Grupo de Ingenier ´ı a Fot o´ nica, Departamento de Electr o´ nica, Escuela Polit e´ cnica Superior, Universidad de Alcal a´ , Campus
Universitario, 28871 Alcal a´ de Henares, Madrid, Spain
Equipe mixte CEA-CNRS-UJF «Nanophysique et Semiconducteurs», INAC/SP2M/PSC, CEA-Grenoble, 17 rue des Martyrs,
2
3
38054 Grenoble Cedex 9, France
Departamento de Energ ´ı as Renovables, Energ ´ı a Solar Fotovoltaica, Centro de Investigaciones Energ e´ ticas, Medioambientales
y Tecnol o´ gicas (CIEMAT), Avda. Complutense 22, 28040 Madrid, Spain
Received 22 December 2009, revised 16 July 2010, accepted 22 July 2010
Published online 26 August 2010
Keywords atomic force microscopy, InN, optical properties, sputtering, X-ray diffraction
*
Corresponding author: e-mail sirona.valdueza@depeca.uah.es, , Phone: þ34 91 885 6913, Fax: þ34 91 885 6591
We report a detailed study of the effect of deposition parameters ditions, namely at a substrate temperature of 450 8C and RF
on optical, structural, and morphological properties of InN films power of 30 W, InN films present a root-mean-square surface
grown by reactive radio-frequency (RF) sputtering on GaN-on- roughness as low as ꢀ0.4 nm, comparable to the underlying
sapphire templates in a pure nitrogen atmosphere. Deposition substrate. The apparent optical bandgap is estimated at 720 nm
parameters under study are substrate temperature, RF power, (1.7 eV) in all cases. However, the InN absorption band tail is
and sputtering pressure. Wurtzite crystallographic structure strongly influenced by the sputtering pressure due to a change in
with c-axis preferred growth orientation is confirmed by X-ray the species of the plasma.
diffraction measurements. For the optimized deposition con-
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1
Introduction Nanocrystalline InN is an interesting sputtering [10–13], and InN deposition on Si(111) substrates
material for application in the development of opto- has been recently investigated [11, 14, 15]. However, to the
electronic devices like high-efficiency low-cost solar cells, best of our knowledge, a systematic study of InN growth on
optical coatings, and opto-chemical sensors [1–3]. However, GaN-on-sapphire templates by RF sputtering has not been
InN has been less studied than other III-nitride semiconduc- reported so far. This study is indeed interesting in view of the
tor materials such as GaN and AlN due to its low dissociation application of nanocrystalline InN as a low-resistivity
temperature (ꢀ500 8C under vacuum) [4]. In order to reduce coating/contact layer.
the decomposition rate, InN should be deposited at low
In this work, we report the successful growth of InN
substrate temperatures. From this point of view, reactive layers on GaN-on-sapphire templates by reactive RF
radio-frequency (RF) sputtering holds an advantage com- sputtering. We discuss the effect of various deposition
pared to other well-known growth methods used for InN parameters, like substrate temperature, RF power, and
growth, like molecular-beam epitaxy (MBE), metalorganic sputtering pressure, on the crystal structure, surface
vapor phase epitaxy (MOVPE), pulsed laser deposition, and morphology, and optical properties of the InN films.
high-pressure chemical vapor deposition (a selection out of
many studies can be found in Refs. [5–8]). Among them,
2 Experimental The InN layers investigated in this
reactive sputtering is considered as a promising technique to work were deposited in a reactive RF sputtering system
synthesize nanocrystalline InN films for relatively low-cost equipped with a 2 in. confocal magnetron cathode (AJA
technology [9]. International, ATC ORION-3-HV). Substrates were com-
Previous work demonstrated the feasibility of obtaining mercial 3.5-mm thick GaN layers grown on sapphire by
InN layers on glass and sapphire substrates by using reactive MOVPE (GaN templates). Pure nitrogen (6N) was used as
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

a
p s s
S. Valdueza-Felip et al.: High-surface-quality nanocrystalline InN layers deposited on GaN templates
6
6
1
6
4
reactive gas, whereas the target was a pure indium disk
4N5). The distance between the substrate and the indium
target was fixed at 10.5 cm. The base pressure for the
deposition was of the order of 10 mTorr. The nitrogen flux
into the deposition chamber was fixed at 14 sccm. The
temperature of the substrate was measured using a K-type
thermocouple placed 2 mm below the sample holder. To
guarantee a uniform temperature, the sample was kept
(
1
ꢁ
5
12
1
0
8
6
4
2
3
0 min at the growth temperature before starting the
deposition. Reproducibility tests point to a temperature error
bar of ꢂ10 8C. Prior to the growth, the substrates were
chemically cleaned, loaded in the sputtering system and
outgassed in the deposition chamber at 600 8C for 30 min.
400 420 440 460 480 500 520 540 560
Substrate temperature (°C)
Figure 2 Dependence of rms surface roughness on substrate tem-
Results and discussion A typical X-ray diffraction perature for InN layers deposited at 40 W and 3.5 mTorr.
XRD) pattern of an InN film deposited on GaN-on-sapphire
template is shown in Fig. 1. Only reflections associated to
0001)-oriented InN are observed, together with reflections on growth rate and grain size in polycrystalline materials.
3
(
(
related to GaN-on-sapphire template. We can hence Therefore, a set of samples varying the substrate temperature
conclude that all samples under study present a wurtzite from 400 to 550 8C was deposited to evaluate its effect on the
crystallographic structure with c-axis preferred growth layer properties. For all samples, RF power and sputtering
orientation and no parasitic crystallographic orientations pressure were fixed at 40 W and 3.5 mTorr, respectively. The
within the resolution limit of the XRD system were observed. layer thickness is ꢀ150 nm.
From XRD measurements (u–2u scans) of InN on GaN
Figure 2 shows the evolution of root-mean-square (rms)
2
˚
templates, we deduce a c-lattice parameter of 5.74 A for all roughness, obtained from 2 ꢃ 2 mm atomic force micro-
InN samples under study, indicating a compressive strain of scopy (AFM) images measuring in the tapping mode, with
ꢁ
0.85%. This value is similar to that obtained by Shinoda the substrate temperature. The best result is obtained for InN
and Mutsukura [12] for InN samples deposited by RF films deposited at 450 8C, which presents an rms roughness
sputtering on sapphire and glass substrates, but it is slightly of ꢀ3.5 nm for a layer thickness ꢀ180 nm. The InN surface
˚
higher than the 5.70 A reported for fully relaxed MBE deterioration at higher temperatures (see Fig. 2) is attributed
samples [16].
to the increase of the InN desorption rate. On the other hand,
The influence of growth conditions on structural, for temperatures below 450 8C the InN surface degrades due
morphological, and optical qualities of InN films deposited to the decrease of the mobility of the adatoms.
2
on GaN-on-sapphire templates has been studied. In particu-
lar, we have analyzed the effect of parameters such as deposited at 450 and 550 8C, illustrating a strong increase of
Figure 3 shows the 2 ꢃ 2 mm AFM images of InN films
substrate temperature, RF power, and sputtering pressure.
the grain size with the substrate temperature. To our
knowledge, this is the first study of morphological quality
of InN films deposited by RF sputtering on GaN-on-sapphire
3.1 Influence of substrate temperature The templates. For comparison, Guo et al. [13] reported recently
mobility of the adsorbed species is controlled by the an rms surface roughness of 2.6 nm for InN deposited by RF
substrate temperature, which should hence have an influence sputtering on Si(111) at low substrate temperature
(
T ¼ 100 8C).
Structural properties have been further analyzed by XRD
measurements. The v-scan full width at half-maximum
FWHM) of the (0002) InN reflection peak remains around
7
5
3
1
1
1
1
1
0
0
0
0
(
30
40
50
60
70
2
θ (deg)
Figure 1 XRDpatternofanInNfilmdepositedat400 8C,30 Wand Figure 3 (online color at: www.pss-a.com) AFM micrograph of
2
3
.5 mTorr.
2 ꢃ 2 mm area of InN layers deposited at (a) 450 8C and (b) 550 8C.
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.pss-a.com

Original
Paper
Phys. Status Solidi A 208, No. 1 (2011)
67
1
.68, which is a factor of 3 narrower than measurements of
InN deposited on Si(111) under the same deposition
conditions [15]. Finally, u–2u XRD scans allow us to
estimate the grain size, D, from the Scherrer formula
0
:9l
D2u
D ¼
cosu;
(1)
where l, u, and D2u are the X-ray wavelength, Bragg Figure 4 (online color at: www.pss-a.com) AFM micrograph of
2
diffraction angle, and FWHM of the InN 2u reflection, 2 ꢃ 2 mm area of InN layers deposited at different RF powers:
respectively. A grain size of ꢀ33 nm is obtained indepen- (a) 30 W and (b) 45 W.
dently of the deposition conditions, which is consistent with
AFM observations and comparable to results reported by
Yamaguchi et al. [17] for InN deposited on GaN by
3
.3 Influence of N2 sputtering pressure The
MOVPE.
variation of the nitrogen pressure in the sputtering process
leads to a change in the partial composition of the different N
species within the N2 plasma, which influences the
morphological and optical properties of the material [20].
In order to evaluate the effect of this parameter on the layer
properties, a set of InN films with sputtering pressure varying
from 3.5 to 14 mTorr were deposited. For this set of samples,
the substrate temperature and RF power were fixed at 450 8C
and 30 W, respectively.
3
.2 Effect of RF power The increase of the RF
power leads to an increase of the deposition rate due to the
higher density of ionized atoms in the sputtering atmosphere.
Nevertheless, the energy of the active elements increases
with the RF power, thus increasing the damage that these
species can cause in the growing surface. To study the effect
of this parameter, we deposited a set of InN samples in which
the RF power was varied between 20 and 45 W, whereas the
substrate temperature and sputtering pressure were fixed at
50 8C and 3.5 mTorr, respectively.
Morphological properties of the material improve when
decreasing the RF power (see Table 1). The best result is
obtained for InN films deposited at low RF power (30 W),
which display an rms roughness of the same order as the
Figure 5 shows the evolution of the growth rate as well as
2
the rms surface roughness estimated on 2 ꢃ 2 mm scanning
4
areas as a function of the sputtering pressure. The layer
thickness of the samples is 130, 114, and 79 nm, for
sputtering pressures of 3.5, 7.0, and 14 mTorr, respectively.
The increase of N pressure causes a strong deterioration of
2
the InN surface. This effect is attributed to a change of the
nitrogen species in the plasma, which has been confirmed by
in situ monitorization of the spectral lines of the plasma
2
underlying substrate. Figure 4 shows the 2 ꢃ 2 mm AFM
images of InN films deposited at 30 and 45 W, illustrating a
strong increase of the grain height with the RF power. The
morphological deterioration for higher RF powers is
attributed to surface damage induced by the high energy of
the impinging In atoms [18, 19].
ꢄ
during deposition. An increase of the reactive nitrogen (N )/
N ratio and a decrease of the deposition rate were observed
when increasing the sputtering pressure. Both effects could
2
ꢄ
be related to the reaction of N with other impurities present
in the sputtering chamber and their incorporation into the
growing layer.
X-ray diffraction characterization confirms that the v-
scan FWHM of the (0002) reflection peak stays below 2.18,
independently of the RF power used in this study; being a
factor of 2 narrower than InN samples deposited on Si(111)
at the same conditions [15]. This enhancement in terms of
crystal quality is attributed to a better substrate/epilayer
interface, which could be deteriorated in the case of Si(111)
by the formation of amorphous Si N [9]. Finally, for all
6
x
y
80
samples under study a grain size of ꢀ29 nm is obtained.
4
Table 1 Morphological characterization of InN samples depos-
ited at different RF powers. Substrate temperature was fixed at
60
2
4
50 8C and sputtering pressure at 3.5 mTorr. The rms roughness of
the GaN-on-sapphire template is ꢀ0.3 nm.
4
0
0
RF power (W)
rms roughness (nm)
thickness (nm)
2
4
6
8
10
12
14
2
3
4
4
0
0
0
5
1
90
Pressure (mTorr)
0.4
3.5
10.5
115
180
125
Figure 5 Evolution of rms surface roughness and growth rate with
the sputtering pressure for InN samples deposited at 450 8C and
3
0 W.
www.pss-a.com
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

a
p s s
S. Valdueza-Felip et al.: High-surface-quality nanocrystalline InN layers deposited on GaN templates
6
8
The lowest rms surface roughness (ꢀ0.4 nm) is obtained index n (l), and absorption coefficient a (l). First-order
0
0
for InN samples deposited at low sputtering pressures Sellmeier dispersion formula was considered for refractive-
3.5 mTorr). For higher N pressures also a decrease of the index calculations in the transparent region. Besides, a
(
2
deposition rate is observed. Nevertheless, the deterioration sigmoidal approximation was used for the absorption-
of the surface is not attributed to the decrease of the growth coefficient estimation [21]. The estimated apparent optical
rate with the pressure, as it is supported by the low rms values bandgap energy is ꢀ720 nm (1.72 eV) for samples deposited
obtained in samples deposited at RF power of 20 W, with a at 3.5 mTorr. This bandgap value is similar to the apparent
growth rate of 60 nm/h (see Table 1).
optical bandgap obtained by other authors in InN samples
deposited by reactive sputtering [22], and it is much larger
than the well-established values of 0.7–1.0 eV reported for
high-quality single-crystalline InN grown by MBE or
MOVPE [9]. This large apparent optical bandgap is
attributed to the polycrystalline nature of the InN material
and due to the high free electron concentration of the layers
3.4 Characterization of optimized layers The
sheet resistance (R ) of InN layers was determined using
s
the four-point probe method; and resistivity values (r ) were
s
obtained from r ¼ R t, where t is the film thickness.
s
s
Optimized InN samples present
ꢀ
a
resistivity of
20
ꢁ3
(
time, from these simulations an absorption coefficient of
ꢀ10 cm ), as observed by other authors [23]. At the same
ꢁ
4
6.2 ꢃ 10 V cm (sheet resistance ꢀ48 V/sq), which is
consistent with typical resistivity values of polycrystalline
4
ꢁ1
a ¼ 3 ꢃ 10 cm and a refractive index of n ¼ 2.4 were
ꢁ
3
0
0
InN deposited by RF sputtering (ꢀ10 V cm [9]).
obtained at l ¼ 700 nm for InN samples deposited at
As an estimation of the concentration of electrically
active impurities, the free carrier concentration of the layers
was obtained by Hall-effect measurements, being in the
3
.5 mTorr. These values are in good agreement with
measurements previously reported [24]. For all calculations,
the GaN-on-sapphire template refractive index was assumed
to be 2.29.
2
0
ꢁ3
range of 10 cm . As Butcher and Tansley [22] discuss, if
all the oxygen was acting as a donor, the oxygen
concentration would be ꢀ6%. However, in InN material
the solubility of oxygen donor is limited to 0.3% [22]. Thus,
the origin of the measured carrier concentration should be
attributed not only to the existence of oxygen but also to other
impurity species incorporated mainly at the grain boundaries
of the InN layers.
For samples deposited at sputtering pressures above
.5 mTorr, a blue-shift of the apparent optical bandgap is
3
observed, as shown in the optical transmission measure-
ments of Fig. 6. This blue-shift of the bandgap energy when
increasing the working pressure, (and thus when changing
the partial composition of N species in the plasma [25]),
2
could be related to an increase of the density of impurities
and defects incorporated into the layer during the sputtering
process [26]. This phenomenon could be responsible of the
change in the absorption band edge, as can be observed in the
low-wavelength region of Fig. 6.
Regarding the optical properties of the InN films, room-
temperature optical transmittance was measured under
normal incidence with an optical spectrum analyzer (OSA)
in the visible/near-infrared spectral range. Figure 6 shows
optical transmission spectra of InN samples deposited at 3.5,
Regarding the oxygen contamination of the InN layers,
no peak related to the formation of the InON alloy has been
7
.0, and 14 mTorr, where the GaN-on-sapphire template
x
limits the maximum transmittance to ꢀ80%. Experimental
transmission results were compared with theoretical calcu-
lations using the transfer matrix method [21] in order to
observed near the (0002) InN reflection peak in the u–2u scan
obtained by XRD measurements (see Fig. 1). We can hence
conclude that there is not enough oxygen in the sputtered
samples to explain the increase of the apparent optical
obtain values of bandgap wavelength l , ordinary refractive
g
bandgap energy expected only due to the InON alloy
x
formation [22, 27].
1
00
8
6
4
2
0
0
0
0
0
4
Conclusions In this work, we have evaluated the
influence of deposition conditions, like substrate tempera-
ture, RF power, and nitrogen pressure, on the structural,
morphological, and optical properties of InN films deposited
on GaN-on-sapphire templates by reactive RF sputtering.
For all deposition conditions, wurtzite InN films with (0001)
preferred growth orientation have been deposited, as has
been confirmed by XRD measurements. Regarding the
morphological characterization of the samples, best results
have been obtained for substrate temperatures of 450 8C and
low RF powers (20–30 W). The sputtering pressure seems to
strongly influence the InN absorption edge due to a change of
the nitrogen species in the plasma. For the optimized
1
7
3
4 mTorr
mTorr
.5 mTorr
6
00
900
1200
1500
Wavelength (nm)
Figure 6 Optical transmission measurements of InN samples
deposited at sputtering pressures of 3.5, 7.0, and 14 mTorr on
GaN-on-sapphire templates. Substrate temperature and RF power deposition conditions, high surface quality InN films have
been achieved, with rms roughness as low as ꢀ0.4 nm,
were fixed at 450 8C and 30 W, respectively.
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.pss-a.com

Original
Paper
Phys. Status Solidi A 208, No. 1 (2011)
69
comparable to the surface roughness of the underlying [10] S. Inoue, T. Namazu, T. Suda, and K. Koterazawa, Vacuum
74, 443 (2004).
substrate. The optimized InN layers present resistivity values
ꢁ
4
[
[
[
11] Q. X. Guo, T. Tanaka, M. Nishio, H. Ogawa, X. D. Pu, and
W. Z. Shen, Appl. Phys. Lett. 86, 231913 (2005).
12] H. Shinoda and N. Mutsukura, Thin Solid Films 503, 8
of ꢀ6.2 ꢃ 10 V cm. The room-temperature optical band-
gap has been estimated through transmission measurements
ꢀ
720 nm (1.7 eV) for InN samples deposited at 3.5 mTorr of
pressure. These results indicate that reactive RF
sputtering is a promising deposition technique for obtaining
(
2005).
N
2
13] Q. Guo, M. Ogata, Y. Ding, T. Tanaka, and M. Nishio,
J. Cryst. Growth 311, 2783 (2009).
low-cost high-surface-quality nanocrystalline InN films on
GaN.
[
[
14] K. P. Biju and M. K. Jain, Appl. Surf. Sci. 254, 72595 (2008).
15] S. Valdueza-Felip, F. B. Naranjo, M. Gonz a´ lez-Herr a´ ez,
L. Lahourcade, E. Monroy, and S. Fern a´ ndez, J. Cryst.
Growth (2010); 10.1016/j.jcrysgro.2010.05.036.
Acknowledgements Partial financial support was provided
by Spanish Government Projects TEC2006-09990-C02-02/TCM, [16] W. Paszkowicz, R. Cerny, and S. Krukowski, Powder Diffr.
TEC2009-14423-C02-02, andHF2007-0065, andbyComunidadde
Madrid Project S-0505/AMB-0374.
18, 114 (2003).
[17] S. Yamaguchi, M. Kariya, S. Nitta, T. Takeuchi, C. Wetzel,
H. Amano, and I. Akasaki, J. Appl. Phys. 85, 7682 (1999).
[
18] Q. Guo, N. Shingai, Y. Mitsuishi, M. Nishio, and H. Ogawa,
Thin Solid Films 343, 527 (1999).
References
[
1] X. D. Pu, W. Z. Shen, Z. Q. Zhang, H. Ogawa, and Q. X. Guo, [19] H. Cheng, Y. Sun, and P. Hing, Thin Solid Films 434, 112
Appl. Phys. Lett. 88, 151904 (2006).
(2003).
2] S. Strite and H. Morko c¸ , J. Vac. Sci. Technol. B 10, 1237 [20] I. Sugimoto, S. Nakano, and H. Kuwano, J. Appl. Phys. 75,
[
(
1992).
7710 (1994).
[
[
3] O. Ambacher, J. Phys. D 31, 2653 (1998).
4] A. Koukitu, N. Takahashi, and H. Seki, Jpn. J. Appl. Phys. 36,
L1136 (1997).
[21] F. B. Naranjo, M. Gonz a´ lez-Herr a´ ez, H. Fern a´ ndez, J. Sol ´ı s,
and E. Monroy, Appl. Phys. Lett. 90, 091903 (2007).
[22] K. S. A. Butcher and T. L. Tansley, Superlatt. Microstruct.
38, 1 (2005).
[23] J. Wu, W. Walukiewicz, S. X. Li, R. Armitage, J. C. Ho, E. R.
Weber, E. E. Haller, Hai Lu, W. J. Schaff, A. Barcz, and R.
Jakiela, Appl. Phys. Lett. 84, 2805 (2004).
[24] R. T. Shamrell and C. Parman, Opt. Mater. 13, 289
(1999).
[
[
5] T. Yamaguchi, D. Muto, T. Araki, and Y. Nanishi, J. Cryst.
Growth 311, 2780 (2009).
6] O. Briot, S. Ruffenach, M. Moret, B. Gil, Ch. Giesen, M.
Heuken, S. Rushworth, T. Leese, and M. Succi, J. Cryst.
Growth 311, 2761 (2009).
[
7] K. Mitamura, T. Honke, J. Ohta, A. Kobayashi, H. Fujioka,
and M. Oshima, J. Cryst. Growth 311, 1316 (2009).
8] M. Alevli, G. Durkaya, A. Weerasekara, A. G. U. Perera, N.
[25] I. Sugimoto, S. Nakano, and H. Kuwano, J. Appl. Phys. 75, 12
(1994).
[
Dietz, W. Fenwick, V. Woods, and I. Ferguson, Appl. Phys. [26] N. Saito and Y. Igasaki, Appl. Surf. Sci. 169, 349 (2001).
Lett. 89, 112119 (2006).
9] A. G. Bhuiyan, A. Hashimoto, and A. Yamamoto, J. Appl.
Phys. 94, 2779 (2003).
[27] E. Trybus, G. Namkoong, W. Henderson, S. Burnham, W.
Alan Doolittle, M. Cheung, and A. Cartwright, J. Cryst.
Growth 288, 218 (2006).
[
www.pss-a.com
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim