High-quality AlInN for high index contrast Bragg mirrors lattice matched to GaN

J.-F. Carlin, and M. Ilegems

Citation: Appl. Phys. Lett. 83, 668 (2003); doi: 10.1063/1.1596733

View online: http://dx.doi.org/10.1063/1.1596733
View Table of Contents: http://aip.scitation.org/toc/apl/83/4
Published by the American Institute of Physics

Articles you may be interested in

Crack-free fully epitaxial nitride microcavity using highly reflective AlInNGaN Bragg mirrors
Applied Physics Letters 86, 031107 (2005); 10.1063/1.1849851

Band parameters for nitrogen-containing semiconductors

Journal of Applied Physics 94, 3675 (2003); 10.1063/1.1600519

Molecular beam epitaxy of InAlN lattice-matched to GaN with homogeneous composition using ammonia as
nitrogen source
Applied Physics Letters 100, 072107 (2012); 10.1063/1.3686922

Strain compensation in AlInN/GaN multilayers on GaN substrates: Application to the realization of defect-free
Bragg reflectors
Applied Physics Letters 98, 181111 (2011); 10.1063/1.3586767

CW lasing of current injection blue GaN-based vertical cavity surface emitting laser
Applied Physics Letters 92, 141102 (2008); 10.1063/1.2908034

Progress in AlInNGaN Bragg reflectors: Application to a microcavity light emitting diode

Journal of Applied Physics 97, 084505 (2005); 10.1063/1.1872197
APPLIED PHYSICS LETTERS VOLUME 83, NUMBER 4 28 JULY 2003

High-quality AlInN for high index contrast Bragg mirrors lattice matched
to GaN
J.-F. Carlina) and M. Ilegems
Institute of Quantum Electronics and Photonics, Swiss Federal Institute of Technology/Ecole Polytechnique
Federale, CH 1015 Lausanne EPFL, Switzerland
Received 10 March 2003; accepted 28 May 2003
We report on the growth by metalorganic vapor phase epitaxy of high-quality Al1x Inx N layers and
AlInN/GaN Bragg mirrors near lattice matched to GaN. Layers are grown on a GaN buffer layer
with no cracks over full 2 in. sapphire wafers. The index contrast relative to GaN is around 7% for
wavelengths ranging from 950 to 450 nm. We demonstrate the growth of a crack-free, 20 pairs
Al0.84In0.16N/GaN distributed Bragg reflector centered at 515 nm with an over 90% reflectivity and
a 35 nm stop band. The growth of high quality AlInN lattice matched to GaN may represent a
breakthrough in GaN-based optoelectronics which is presently limited by the lack of a
high-index-contrast and high-band gap lattice-matched material. 2003 American Institute of
Physics. DOI: 10.1063/1.1596733

AlInN materials hold great potential for GaN-based op- GaN buffer layer. AlInN was deposited between 800 and
toelectronics. Alloys with indium content between 14% and 850 C and at 50 to 75 mbar pressure using N2 carrier gas.
22% which are within a 0.5% lattice mismatch to GaN, Lower growth temperatures led to lower crystalline quality
would be of special interest if they prove to exhibit a suffi- as revealed by high-resolution x-ray diffraction HRXD
ciently high band gap and index contrast with GaN. Indeed, 0002 scans. Higher growth temperatures resulted in a de-
AlGaN is presently the standard material for optical engi- creased indium incorporation so that near-lattice matched al-
neering of GaN-based devices, but the requirement of loys could no longer be obtained. Deposition rates ranged
achieving a high index contrast while at the same time avoid- between 0.6 and 0.2 m/h. During the DBR runs, the growth
ing the generation of cracks due to the lattice mismatch to was interrupted at each interface. GaN was deposited at
GaN are contradictory. As a consequence, for nitride-based 1050 C using H2 and N2 carrier gas. No degradation of
laser diodes AlGaN waveguide cladding layers with hardly AlInN could be detected due to this thermal cycling as
more than 10% Al content1 0.25% lattice mismatch and an shown on Fig. 1 which compares 0002 HRXD rocking
index contrast that does not exceed 2% are used. Distributed curves of a 0.5 m Al0.84In0.16N layer with that of a 20 pairs
Bragg reflectors DBRs also suffer from this issue;2 over Al0.84In0.16N/GaN DBR centered at 515 nm wavelength. The
50% Al content can be used in AlGaN/GaN DBRs with no HRXD scans were performed without a slit on the detector,
cracks,3,4 but in this case the whole structure relaxes to an in this case the diffracted intensity is integrated over a 5
average in-plane lattice parameter. It means that GaN/GaInN detector angle and the full width at half maximum FWHM
multiple quantum well active layers grown on top of such of the peaks are influenced by both composition fluctuations
DBRs are no longer lattice matched and strain relaxation and c-axis tilt. The DBR superlattice satellites are not re-
issues may arise in the active zone. Thus, most authors that solved on the DBR sample as their spacing is too narrow, and
demonstrate AlGaN/GaN DBRs in devices5mainly reso-
nant cavity light emitting diodesprefer to keep Al contents
below 30%, at the price of a reduced optical stop band.
AlInN is not yet a well-known material mainly because
the growth is difficult as AlInN is subject to phase
separation.6 There still subsist a considerable uncertainty
concerning the band gap of AlInN lattice matched to GaN as
values ranging from 2.8 to 4.2 eV were reported by different
groups.79
We demonstrate in this letter the growth of
Al0.84In0.16N/GaN DBRs near lattice matched to GaN. These
DBRs are optically equivalent to state-of-the-art
Al0.6Ga0.4N/GaN mirrors and avoid the issues related to
strain. Layers were grown in an AIXTRON 200/4 RF-S met-
alorganic vapor phase epitaxy system, on 2 in. c-plane sap-
phire substrates. The growth was initiated by a low-
temperature GaN nucleation layer followed by a 1-m-thick
FIG. 1. 0002 x-ray diffraction rocking curves of a 20 pairs AlInN/GaN
a
Electronic mail: [email protected] DBR and of a single 0.5 m AlInN layer grown on GaN buffer layers.

0003-6951/2003/83(4)/668/3/$20.00 668 2003 American Institute of Physics

Appl. Phys. Lett., Vol. 83, No. 4, 28 July 2003 J.-F. Carlin and M. Ilegems 669

FIG. 2. Evolution of the reflectivity at 950 nm wavelength during the

growth of an AlInN/GaN DBR matched to the measurement wavelength. FIG. 3. AlInN/GaN optical index contrast versus AlInN indium content,
The inset shows the index contrast calculated from the period-to-period calculated from in situ reflectivity experiments 950 nm wavelength and
increase of the reflectivity signal. from ex situ analysis of shorter wavelengths DBRs 455 and 515 nm.

the x-ray scan rather reflects de quality of the bulk materials. an estimated 0.4% compressive strain. On more mismatched
The single-layer and the DBR sample show identical high samples, cracks appeared earlier, and in some cases only the
crystalline quality, with 360 arcsec FWHM for the first period could be taken into account for index contrast
Al0.84In0.16N peak, nearly as narrow as the 340 arcsec evaluation.
FWHM GaN peak. Figure 3 summarizes the index contrast measured on dif-
We evaluated the optical index contrast between AlInN ferent samples, and presents the dependence of n/n as a
and GaN, n/n(nAlInN nGaN)/nGaN, by recording the function of the indium content as estimated from HRXD
reflectivity of the layers in situ during the growth of a few 0002 measurements. Open symbols are the in situ measure-
periods of a DBR whose center wavelength matched that of ments described earlier and thus represent n/n at
the measurement wavelength. Our experimental setup con- 950 nm and at growth temperature. The other two data
sists of a LUXTRON TR-100 using a 950 nm wavelength points correspond to ex situ analysis of the blue-green DBRs
source under normal incidence, which allows for an absolute tuned at 455 and 515 nm presented later. We note that the
reflectivity measurement. Figure 2 shows the evolution of index contrast is not much dependent on wavelength within
reflectivity during a typical run; the growth of the GaN this range. The experimental data are well fitted by a linear
buffer layer is stopped when its maximum reflectivity is dependence with indium content within the 6%21% ex-
reached around 26%, then AlInN is grown during the nega- plored range, according to
tive slope of the reflectivity signal, followed by GaN during
the positive slope. Lets note R i for the reflectivity value n
Al1x Inx N/GaN 0.1270.35x. 2
after deposition of the ith DBR period. R i increases with the n
number of periods starting from the very first period. This
Extrapolation of Eq. 2 to zero indium content gives a
already indicates that AlInN has a lower optical index than
12.7% index contrast for AlN/GaN, in agreement with lit-
GaN, otherwise reflections at the AlInN/GaN and GaN/
erature values.11
AlInN interfaces would be in antiphase with the GaN/air and
sapphire/GaN reflections, leading to a decrease of R i during
the first periods. As reflections at all interfaces are in phase,
the well-known formulas for DBRs reflectivity10 can be used
and we can calculate the optical index contrast from the pe-
riod to period increase in reflectivity using

n
n
i 1 1 R i 1 R i1
1 R i 1 R i1
. 1

Of course, this result is only valid if no parasitic effects,

such as absorption, appearance of cracks or development of
surface roughness, decrease the reflectivity. We can insure
that the index contrast given by Eq. 1 is correct by plotting
n/n as a function of the number of periods as shown in the
inset of Fig. 2. Any parasitic effect will manifests itself by a
decrease of n/n. In the case of the run shown in Fig. 2, a
marked decrease occurs at the seventh period, and indeed
further examination of the sample revealed the presence of FIG. 4. Index contrast vs lattice mismatch to GaN: comparison between
cracks. This sample was still quite near lattice matched, with AlInN/GaN and AlGaN/GaN materials systems.
670 Appl. Phys. Lett., Vol. 83, No. 4, 28 July 2003 J.-F. Carlin and M. Ilegems

tometer in double-reflection mode. These data were fitted

with a standard transfer matrix model dotted lines on Fig. 5
to extract the index contrast values previously shown in Fig.
3. The ten periods sample shows a maximum reflectivity of
76% with a 41 nm FWHM stop band. The dip in the stop-
band is due to the GaN cavity between the DBR and the
sapphire substrate. For the 20 periods sample, the reflectivity
reaches over 90% with a 35 nm FWHM stop band. For com-
parison, Nakada et al.3 reported 70% and 83% reflectivity
for, respectively, 10 and 20 periods Al0.6Ga0.4N/GaN DBRs
relaxed on GaN.
Hall measurements showed a residual donor density of
71017 cm3 in the Al0.84In0.16N layer. This value repre-
sents an upper-limit estimate as a bidimensional electron gas
may be present at the AlInN/GaN interface. Preliminary re-
flectivity data also indicate the presence of an optical transi-
tion around 4.2 eV, in agreement with the value reported for
the Al0.84In0.16N band gap measured on layers deposited by
plasma source molecular beam epitaxy9 and sputtering.7
This work was performed in the framework of the Swiss
National Center of Competence Quantum Photonics pro-
gram. The authors thank H. Martinez-Fraiz and R. Sanjines
of the Institute for Complex Materials at EPFL for the reflec-
tivity measurements.
FIG. 5. Experimental full lines and simulated dotted lines reflectivity
spectra of AlInN/GaN distributed Bragg reflectors.
1
T. Asano, K. Yanashima, T. Asatsuma, T. Hino, T. Yamaguchi, S. Tomiya,
K. Funato, T. Kobayashi, and M. Ikeda, Phys. Status Solidi A 176, 23
The advantage of using near lattice matched AlInN as 1999.
the low-index material is evident from Fig. 4, where the 2
T. Someya and Y. Arakawa, Appl. Phys. Lett. 73, 3653 1998; R. Langer,
AlInN/GaN index contrast is plotted as a function of lattice A. Barski, J. Simon, N. T. Pelekanos, O. Konovalov, R. Andre, and L. S.
Dang, ibid. 74, 3610 1999; K. E. Waldrip, J. Han, J. J. Figiel, H. Zhou,
mismatch to GaN and compared with that of the AlGaN/GaN E. Makarona, and A. V. Nurmikko, ibid. 78, 3205 2001.
material system.11 A lattice mismatch that lies within 0.5% 3
N. Nakada, H. Ishikawa, T. Egawa, T. Jimbo, and M. Umeno, J. Cryst.
is sufficient to avoid relaxation in blue DBR applications. In Growth 237, 961 2002.
4
this case the maximum index contrast achievable with F. Natali, D. Byrne, A. Dussaigne, N. Grandjean, J. Massies, and B. Dami-
lano, Appl. Phys. Lett. 82, 499 2003.
AlGaN/GaN is about 3%, while up to 8% is obtained with 5
N. Nakada, M. Nakaji, H. Ishikawa, T. Egawa, M. Umeno, and T. Jimbo,
AlInN/GaN. The gain is even more dramatic when consider- Appl. Phys. Lett. 76, 1804 2000; M. Diagne, Y. He, H. Zhou, E. Maka-
ing laser applications where the lattice mismatch is rather rona, A. V. Nurmikko, J. Han, K. E. Waldrip, J. J. Figiel, T. Takeuchi, and
limited to 0.25%. M. Krames, ibid. 79, 3720 2001; S. Fernandez, F. B. Naranjo, F. Calle,
M. A. Sanchez-Garcia, E. Calleja, and P. Vennegues, Phys. Status Solidi A
To demonstrate the interest of the AlInN/GaN system, 192, 389 2002.
we show in Fig. 5 the reflectivity spectra of two AlInN/GaN 6
T. Matsuoka, Appl. Phys. Lett. 71, 105 1997.
7
DBRs having stop bands in the visible wavelength range. T. Peng, J. Piprek, G. Qiu, J. O. Olowolafe, K. M. Unruh, C. P. Swann,
and E. F. Schubert, Appl. Phys. Lett. 71, 2439 1997.
Sample A is a ten period DBR centered at 455 nm, sample B 8
S. Yamaguchi, M. Kariya, S. Nitta, H. Kato, T. Takeuchi, C. Wetzel, H.
has 20 periods and is centered at 515 nm. Careful examina- Amano, and I. Akasaki, J. Cryst. Growth 195, 309 1998.
tions by phase contrast microscopy just after the growth 9
M. J. Lukitsch, Y. V. Danylyuk, V. M. Naik, C. Huang, G. W. Auner, L.
showed that both samples were completely crack-free over Rimai, and R. Naik, Appl. Phys. Lett. 79, 632 2001.
10
H. Angus Macleod, Thin-film Optical Filters Hilger, Bristol, 1985, p.
the full 2 in. area. Some cracksabout 10however, ap- 165.
peared after some weeks and rather careless manipulation. 11
D. Brunner, H. Angerer, E. Bustarret, F. Freudenberg, R. Hopler, R. Dim-
The measurements were performed with a Cary 500 reflec- itrov, O. Ambacher, and M. Stutzmann, J. Appl. Phys. 82, 5090 1997.