ArticlePDF Available

Blue to yellow emission from (Ga,In)/GaN quantum wells grown on pixelated silicon substrate


Abstract and Figures

It is shown that substrate pixelisation before epitaxial growth can significantly impact the emission color of semiconductor heterostructures. The wavelength emission from InxGa1−xN/GaN quantum wells can be shifted from blue to yellow simply by reducing the mesa size from 90 × 90 µm2 to 10 × 10 µm2 of the patterned silicon used as the substrate. This color shift is mainly attributed to an increase of the quantum well thickness when the mesa size decreases. The color is also affected, in a lesser extent, by the trench width between the mesas. Cathodoluminescence hyperspectral imaging is used to map the wavelength emission of the InxGa1−xN/GaN quantum wells. Whatever the mesa size is, the wavelength emission is red-shifted at the mesa edges due to a larger quantum well thickness and In composition.
This content is subject to copyright. Terms and conditions apply.
Scientic Reports | (2020) 10:18919 | 
Blue to yellow emission
from (Ga,In)/GaN quantum
wells grown on pixelated silicon
Benjamin Damilano1*, Marc Portail1, Eric Frayssinet1, Virginie Brändli1, Florian Faure2,
Christophe Largeron2, David Cooper2, Guy Feuillet2 & Daniel Turover3
It is shown that substrate pixelisation before epitaxial growth can signicantly impact the emission
color of semiconductor heterostructures. The wavelength emission from InxGa1−xN/GaN quantum
wells can be shifted from blue to yellow simply by reducing the mesa size from 90 × 90 µm2 to 10 × 10
µm2 of the patterned silicon used as the substrate. This color shift is mainly attributed to an increase
of the quantum well thickness when the mesa size decreases. The color is also aected, in a lesser
extent, by the trench width between the mesas. Cathodoluminescence hyperspectral imaging is used
to map the wavelength emission of the InxGa1−xN/GaN quantum wells. Whatever the mesa size is, the
wavelength emission is red-shifted at the mesa edges due to a larger quantum well thickness and In
Group-III nitride semiconductors are the materials of choice for the fabrication of blue and white light emit-
ting diodes1. e active region is made of blue light emitting (Ga,In)N quantum wells (QWs) heterostructures
encapsulated in a yellow/red emitting luminophore which converts part of these blue photons to generate white
light. ese devices are now the most ecient white light sources for general lighting2. However, the color of
these lamps can only be weakly tuned. Fine control of the light emission over the visible range is only possible
if colors can be controlled separately. Ideally, an independent control of blue, green and red colors opens up
a wide color range. In principle, this can be done easily by mixing three dierent LEDs: for example (Ga,In)
N-based for blue and green, and (Al,Ga,In)P for red. However, with the aim of developing high resolution and
high luminance micro-displays3,4, this approach is no longer possible. Mixing the three basic blue/green/red
colors directly on the same wafer is therefore highly desirable and constitutes an important challenge. Eorts
towards this objective have been reported in the literature using for example nanowires grown by molecular
beam epitaxy5,6, light conversion by (Ga,In)N multiple quantum wells7, facets with dierent orientations8, and
local etching of red-emitting LED structures9.
Here, we propose the use of mesa-patterned silicon substrates to tune the light emission of (Ga,In)N quan-
tum wells. Mesa-patterned silicon substrates are known to be an ecient way to avoid crack formation in thick
GaN structures1013. is is achieved by stress relaxation at the mesa edges that compensate the tensile strain in
the GaN layers grown on silicon arising from the thermal expansion coecient mismatch between GaN and
silicon1013. Mesa-patterned silicon substrates have been successfully exploited for the fabrication of devices such
as ecient visible LEDs on silicon substrate14, laser diodes15, exciton–polariton laser16, LEDs incorporating a high
reectivity distributed light reector17, high electron mobility transistors18. Xu etal. reported an enhanced indium
incorporation in (Ga,In)N QWs at the corner of 340 × 340 µm2 mesas grown on patterned silicon substrate19 but
mesa-patterned silicon substrates have never been studied with the objective of modifying the emission color
as a function of the mesa size.
In order to determine the impact of substrate patterning, the micro-photoluminescence was measured at room
temperature at the center of 24 dierent mesas. e sample is divided into dierent regions with a surface area
of 4 × 6 mm2. Each region contains a square array of square mesas, each with a side length L and a trench width
W. ere are 6 dierent mesas sizes: L = 10, 20, 30, 40, 60, 90µm and 4 dierent trench widths W = 10, 13, 30,
Université Côte D’Azur, CNRS, CRHEA, Rue B. Gregory, Valbonne, France. Univ. Grenoble Alpes, CEA, LETI,
 *
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Scientic Reports | (2020) 10:18919 | 
60µm, totalizing 24 dierent regions. In the following the mesas are identied by LxWy, where x is the side
length and y the trench width in µms. e peak wavelength of the (Ga,In)N/GaN MQW extracted from a Gauss-
ian t of the PL spectra is shown in Fig.1a. Some of the spectra are given as an example in Fig.1b for dierent
mesa sizes and a trench width of 10µm. Note that the PL spectra are structured by oscillations coming from the
interference of light in the cavity formed by the nitride layers sandwiched by the air and the silicon substrate. e
main result shown by the Fig.1a is that the PL wavelength increases from ~ 430 to 440nm for the largest mesa
sizes of 60–90µm to ~ 570–580nm for the smallest mesa size of 10µm. is shows that simply by structuring
the substrate, dierent color emissions (at least from blue to green and yellow) can be monolithically integrated
on the same sample in one single growth run. e wavelength emission also increases when the trench width
increases however to a lesser extent than for the dependence observed as a function of the mesa size. is varia-
tion is the largest for the mesas with side length L = 20µm (Fig.1a): the peak PL wavelength shis from 494nm
(blue-green) to 544nm (green) when varying the trench width from 10 to 60µm. Some variation is found in the
data, for example the QW emission of L60W60 is at slightly shorter wavelength than the one of L90W60. is
is due to the non-perfect homogeneity of the QW on the 2-inch wafer during the growth (see Supplementary
Information, FigureS1).
e integrated photoluminescence intensity from the center of the mesas as a function of the peak wave-
length emission follows an asymmetric bell shape curve. It is maximum in the 470–500nm wavelength range
and decreases by a factor 2.5 at 430nm and by a factor 100 at 580nm. e PL signal of the mesa L10W60 is too
weak for a reliable estimation of the peak wavelength emission.
Cathodoluminescence (CL) mapping of the mesas was used to obtain spatially resolved spectra from each
mesa type, with a step size of 700nm. One example is shown in Fig.2, where a mesa L40W30 is investigated
Figure1. (a) Room temperature micro-photoluminescence peak wavelength from the InGaN/GaN multiple
quantum well at the center of the mesas with variable sizes and trench widths. e dashed line corresponds to
the emission at the center of the reference sample without patterns. (b) Examples of spectra with normalized
intensities obtained for variable mesa sizes L = 10, 20, and 90µm with a constant trench width of 10µm (mesas
L10W10, L20W10, and L90W10).
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Scientic Reports | (2020) 10:18919 | 
(scanning electron microcopy (SEM) image in Fig.2a). Figure2b shows the CL intensity map for 3 dierent wave-
lengths: 390nm, 475nm and 520nm (corresponding to the violet, blue and green colors). e spectra recorded
at 3 dierent positions on the mesas are shown in Fig.2d. Blue CL emission is observed at the center of the mesa
while the color is shiing towards the green at the edge of the mesa. Violet emission is observed on the inclined
facets located at the edges of the GaN mesa. e CL spectra extracted at the center of the mesa following the white
dotted line in Fig.2b are shown in Fig.3g. is prole indicates that the PL emission is rather homogeneous at
the center of the mesa and shis to longer wavelength at ~ 10µm from the mesa edge. e panchromatic image of
Fig.2c indicates a slightly weaker intensity at the edges of the GaN mesa. is decrease in the internal quantum
eciency is the typical evolution expected from InxGa1−xN/GaN MQWs when the wavelength increases20,21.
e CL mapping experiments have been repeated on dierent mesas and in order to reduce the time required
for these measurements, only line scan analysis at the center of the mesas were performed. e results are shown
in Fig.3a–h for mesas L20W10, L20W13, L20W30, L20W60, L40W10, L40W13, L40W30, and L40W60. As
reported above for µPL experiments, the peak CL wavelength at the mesa center shis towards longer wavelength
when the mesa size decreases and the trench width between mesas increases. For all the mesas investigated, the
peak CL wavelength at the mesa edges is red-shied compared to the value at the center of the mesa.
e origin of the large emission red-shi when the mesa size decreases and the trench width increases can be
explained. Firstly, the fundamental energy of the e1-hh1 excitonic transition of the InxGa1−xN/GaN quantum well
is given by the following expression22:
where Eg is the bandgap energy of InxGa1−xN, e1 and hh1 are the connement energies of the electron and hole,
Ry is the exciton binding energy, F is the internal electric eld, e is the electron charge, and h is the quantum
well thickness.
A red shi of the emission could be due to an increased Indium composition. Indeed, as the In composition
x increases the emission energy of the QW decreases because the bandgap of InxGa1−xN narrows and in addi-
tion the internal electric eld increases. Similarly, an increase of the QW thickness can cause a decrease of EQW,
mainly due to the term –eFh in (1). e strain state of the InxGa1−xN QWs can cause a change in the bandgap
energy Eg and also of the internal piezoelectric eld F. Indeed, strain relaxation at the mesa edges can cause a PL
Figure2. (a) Scanning electron microscopy image of a mesa with a side length of L = 40µm and a trench
width W = 30µm. (b) Cathodoluminescence map of the same mesa. e regions corresponding to a wavelength
of 390, 475, and 510nm (spectral width of 10nm) are highlighted in violet, blue, and green, respectively. (c)
Corresponding cathodoluminescence panchromatic image. (d) Characteristic examples of the spectra obtained
in three dierent regions of the mesa: center, edge, and side facet. e white scale bar in (a), (b) and (c) is 10µm.
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Scientic Reports | (2020) 10:18919 | 
blue-shi of the InxGa1−xN QWs due to a reduction of the internal piezoelectric eld. is last eect has been
shown to have a signicant impact for very small nanostructures: the PL emission of red-emitting InxGa1−xN QWs
is shied to green and blue when they are selectively etched to form nanowires with diameters of 150 and 50nm,
respectively9. In our work, the minimum mesa size is 10µm and therefore a strong impact of strain variation on
the InxGa1−xN bandgap and the piezoelectric eld at the center of the mesas is not expected.
As described above and shown in Fig.1b, the PL spectra are structured by interference fringes. e spacing of
these fringes is characteristic of the total nitride layer thickness grown on the silicon substrate23. At a wavelength
around 460nm, at the mesa center, this spacing is estimated to be 8.3nm and 6.3nm for the mesas L90W10
and L20W10, respectively. is clearly shows that the total thickness of the nitride layers is larger for L20W10
compared to L90W10.
In order to obtain direct and more quantitative data, TEM experiments were conducted to measure the
InxGa1−xN QW thicknesses. Figure4a shows cross section HAADF STEM images at the top of a mesa with a
side length of 20µm and a trench width of 60µm (L20W60) showing the InxGa1−xN quantum wells at dierent
positions. Figure4b shows a higher magnication HAADF STEM image at the mesa edge and Fig.4c shows a
corresponding EDX map. Figure4d shows a HAADF STEM image acquired 2µm away from the mesa edge and
Fig.4e the corresponding EDX map. ese experiments were repeated at other positions of the mesa L20W60
and also for mesa L20W10. e QW thicknesses were extracted from all these images and are shown in Fig.4f
as a function of the distance from the mesa edge. e InxGa1−xN QW thickness tends to a constant at the center
of the mesa of about 2.9nm and 4.5nm, for L20W10 and L20W60, respectively. Other thickness measurements
were performed by TEM on a mesa L10W10. e InxGa1−xN QW thickness at the center of this mesa is 5nm
(Fig.S2 of Supplementary Information). ese values are signicantly larger than the nominal InxGa1−xN QW
thickness of 2nm (on a planar substrate). Calculations of the e1-hh1 fundamental transition of the InxGa1−xN/
GaN quantum well were performed to determine whether the thickness variation can account for the wavelength
shi at the center of the mesas. is calculation was performed using the nominal indium content of 0.13 and
the other parameters (bandgap, piezoelectric constants,…) were taken from Ref.24 and summarized in TableS1
of Supplementary Information. e calculated QW emission wavelengths are 430, 460, 525, and 551nm for
QW thicknesses of 2, 2.9, 4.5, and 5nm, respectively. erefore, this QW thickness increase can account for a
large part the observed red-shi of the QW emission energy at the mesa center when the mesa size decreases.
EDX measurements were performed at the edge of mesas L20W10 and L10W10 and give an average In com-
position of 13.8 ± 2.0% and 18.6 ± 2.0%, respectively. is increase of the In composition can also explain the red-
shi of the photoluminescence emission of the mesas when their size decreases. e exact quantication of these
EDX measurements in such small QW structures is complicated. However, all the observations were made using
identical experimental conditions and quantication, such as the relative values can be compared. e dier-
ence in In concentration is also conrmed by the increase in intensity measured by HAADF in the InGaN wells.
Figure3. Room temperature cathodoluminescence spectra along a line in the middle of the mesas with a side
length L = 20µm (ad) or L = 40µm(eh).e width in between mesas is W = 10µm (a,e), 13µm (b,f), 30µm
(c,g), 60µm (d,h). e intensity (normalized) of the spectra is color coded from blue (0) to red (1). e doted
white line corresponds to a wavelength of 430nm,as obtained for the unpatterned substrate. e white scale bar
is 10µm in all gures.
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Scientic Reports | (2020) 10:18919 | 
e HAADF STEM images indicate a strong increase of the QW thicknesses above a few micrometers from
the mesa edges. is is again in good agreement with the wavelength redshis observed close to the mesa edges
for all the CL spectra shown in Fig.3. More specically for the mesas L20W10 and L20W60, the CL peak wave-
length of the InxGa1−xN/GaN MQW also starts to shi at a few micrometers from the mesa edge as shown in
Fig.3a,d. To evaluate the variation of the indium composition of the InxGa1−xN QWs the EDX measurements
that were performed on the mesa L20W60 are shown in Fig.4g. e quantitative In composition measurements
show an average value of 13.8 ± 2.0% at the edge and 11.6 ± 2.0% at a position of 2µm from the edge. Indeed,
this composition increase also contributes to the wavelength emission red-shi of the InxGa1−xN /GaN MQW
at the mesa edges.
e enhanced growth at the edges is similar to the observations made by Honda etal.12 and attributed to the
mass transport of chemical species from the SiO2 mask to the mesa edges. Honda etal. observed a strong GaN
thickness variation between the center and the edge of the mesas for a mesa spacing larger than 50µm, while it
decreases down to 10% when the spacing between mesas is 10µm and the mesa size 200µm. However, in our
case, growth occurs in the trenches between the mesas and the growth enhancement should be attributed to
another eect such as a perturbation of the critical layer in the gas phase close to the mesa edges25. e slightly
larger In incorporation at the mesa edges can be related to a stress relaxation eect by the free edges. Indeed,
the InxGa1−xN layer grown close to the mesa edge can be submitted to a lower compressive stress compared to
Figure4. (a) HAADF STEM cross section images taken in the top area of a mesa L20W60 (side length of
20µm and trench width of 60µm). (b) shows a higher magnication STEM image at the mesa edge and (c)
an EDX map showing the InGaN layers. (d) A higher magnication STEM image acquired 2µm away from
the mesa edge and (e) an EDX map showing the InGaN layers. (f) e InGaN quantum well thicknesses
extracted from the transmission electron microscopy imagesas a function of the position (0 corresponds to the
mesa edge) for mesa with a side length of 20µm and a trench width of 60µm (in red) or 10µm (in blue). (g)
Quantitative proles of the Indium content extracted from the EDX maps both at the edge and 2µm away from
the edge for the mesa L20W60.
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Scientic Reports | (2020) 10:18919 | 
the one at the center of the mesa and this reduction of stress can induce a larger In incorporation as shown for
InxGa1−xN quantum wells grown on relaxed (Ga,In)N layers26,27.
We have shown that the wavelength emission from an InxGa1−xN/GaN MQW grown on a micro-pixelated Si
substrate can be tuned from blue (430nm) to yellow (580nm) depending on the pixel size and the trench width
between the pixels. Decreasing the pixel size and increasing the trench width leads to a longer wavelength emis-
sion for mesas with a side length smaller than 40µm. is eect is mainly due to an increase of the QW thickness
as the mesa side length decreases. e emission is relatively homogeneous at the center of the mesa but a red
shi is observed near the mesa edges, again due to an increase of the QW thickness. e eciency of the yellow-
emitting mesas is much smaller than for blue and green. Further optimization is required to get blue, green and
red colors on the same wafer. e mesa sizes, pitches have to be nely tuned as a function of QW thickness and
indium composition. is approach should then become promising for the fabrication of monolithic full-color
micro-display emitters.
Patterned silicon substrate fabrication. e 2-inch Si(111) substrates were covered with a hard mask
made of a tri-layer SiO2-Al-SiO2 and subsequently patterned by direct laser photolithography using an Heidel-
berg MGP 1 machine. e Si mesas were etched using uorine (SF6) based inductively coupled plasma (ICP) to
provide high aspect ratio silicon etching.
e sample is divided into dierent regions with a surface area of 4 × 6 mm2. Each region contains a square
array of square mesas, each with a side length L and a trench width W. ere are 6 dierent mesas sizes: L = 10,
20, 30, 40, 60, 90µm and 4 dierent trench widths W = 10, 13, 30, 60µm, totalizing 24 dierent regions. e
trench depth in the silicon substrate is 10µm. SEM images of some of the mesas obtained aer growth are shown
in Fig.5.
Epitaxial growth. e samples were grown on the patterned Si(111) substrates as described above using
low-pressure metal organic vapor phase epitaxy (MOVPE) in a commercial close coupled 7 × 2-in. showerhead
reactor. e growth rates of the dierent layers are monitored insitu by a laser reectivity set-up. Trimethyla-
luminium, trimethylgallium (for GaN growth), triethylgallium (for InGaN growth) and trimethylindium are
used as group-III precursors while ammonia is used as group-V precursor. e structure of the samples com-
prises (starting from the substrate) a 220-nm thick AlN buer layer, a 2µm-thick GaN layer and a 5 periods
In0.13Ga0.87N (2nm)/GaN (12nm) multiple quantum well (MQW). More details about the nucleation procedure
on the silicon substrate can be found in Ref.28. e rst stage of the GaN layer is grown in a 3-dimensional
growth mode to reduce the threading dislocation density. On a separate sample grown without the MQW on
Figure5. Scanning electron microscopy images at a tilted view of 30° of mesas with variable sizes L = 10, 20 30,
40, and 60µm and a constant trench width W = 30µm (a) and mesas with a constant size L = 20µm and variable
trench widths W = 10, 13, 30, and 60µm (b). e scale bar is respectively 100 and 20µm for (a) and (b).
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Scientic Reports | (2020) 10:18919 | 
patterned silicon, a threading dislocation density of 1–2 × 109cm−2 was found emerging at the surface as meas-
ured by atomic force microscopy. e room temperature photoluminescence of such QW samples on un-pat-
terned silicon (this sample was grown in the same growth run than the sample on patterned silicon substrate)
is centred at 430nm.
Characterization. e surface of the samples was imaged by scanning electron microscopy. e micro-
photoluminescence (µPL) was measured at room temperature using the 244nm line of a frequency doubled
Argon laser with a power density of 10W/cm2. e laser spot diameter is ~ 2µm. Cathodoluminescence (CL)
was also performed at room temperature both in panchromatic and spatially and spectrally resolved modes.
High resolution scanning transmission electron microscopy (STEM) was performed using a probe corrected
FEI emis operated at 200kV. in specimens were prepared using focused ion beam (FIB) milling in a FEI
Strata 400 dual beam operated at 16kV to reduce specimen damage. e specimens were then cleaned using
2kV ions. On the dierent specimens, high-angle annular dark eld (HAADF) STEM was performed on the
same mesa structures looking down both the 0120 and 0110 zone axis. e indium content in the specimens
was evaluated by comparing the relative intensity of the HAADF signal which is sensitive to the Z number in
the specimens. In addition, energy dispersive X-ray spectroscopy (EDX) was performed in order to retrieve the
relative indium concentrations. Quantication of the In concentration was performed using the Cli-Lorimer
method29. Although it is dicult to quantify at these small dimensions using EDX, we assume that the rela-
tive concentrations in the same specimens that are measured using the same TEM settings, can be interpreted
Received: 5 May 2020; Accepted: 19 October 2020
1. Nakamura, S. Nobel Lecture: Background story of the invention of ecient blue InGaN light emitting diodes. Rev. Mod. Phys. 87,
1139–1151 (2015).
2. Narukawa, Y., Ichikawa, M., Sanga, D., Sano, M. & Mukai, T. White light emitting diodes with super-high luminous ecacy. J.
Phys. Appl. Phys. 43, 354002 (2010).
3. Jiang, H. X., Jin, S. X., Li, J., Shakya, J. & Lin, J. Y. III-nitride blue microdisplays. Appl. Phys. Lett. 78, 1303 (2001).
4. Templier, F. GaN-based emissive microdisplays: a very promising technology for compact, ultra-high brightness display systems:
GaN-based emissive microdisplays. J. Soc. Inf. Disp. 24, 669–675 (2016).
5. Kishino, K., Nagashima, K. & Yamano, K. Monolithic integration of InGaN-based nanocolumn light-emitting diodes with dierent
emission colors. Appl. Phys. Express 6, 012101 (2013).
6. Ra, Y.-H. et al. Full-color single nanowire pixels for projection displays. Nano Lett. 16, 4608–4615 (2016).
7. Damilano, B. et al. Metal organic vapor phase epitaxy of monolithic two-color light-emitting diodes using an InGaN-based light
converter. Appl. Phys. Express 6, 092105 (2013).
8. Funato, M. et al. Experimental and theoretical considerations of polarization eld direction in semipolar InGaN/GaN quantum
wells. Appl. Phys. Express 3, 071001 (2010).
9. Chung, K., Sui, J., Demory, B. & Ku, P.-C. Color mixing from monolithically integrated InGaN-based light-emitting diodes by
local strain engineering. Appl. Phys. Lett. 111, 041101 (2017).
10. Zamir, S., Meyler, B. & Salzman, J. ermal microcrack distribution control in GaN layers on Si substrates by lateral conned
epitaxy. Appl. Phys. Lett. 78, 288 (2001).
11. Dadgar, A. et al. Crack-Free InGaN/GaN Light Emitters on Si(111). Phys. Status Solidi A 188, 155–158 (2001).
12. Honda, Y., Kuroiwa, Y., Yamaguchi, M. & Sawaki, N. Growth of GaN free from cracks on a (111)Si substrate by selective metalor-
ganic vapor-phase epitaxy. Appl. Phys. Lett. 80, 222–224 (2002).
13. Hossain, T. et al. Stress distribution of 12 μm thick crack free continuous GaN on patterned Si(110) substrate. Phys. Status Solidi
C 10, 425–428 (2013).
14. Tao, X. et al. Performance enhancement of yellow InGaN-based multiple-quantum-well light-emitting diodes grown on Si sub-
strates by optimizing the InGaN/GaN superlattice interlayer. Opt. Mater. Express 8, 1221 (2018).
15. Sun, Y. et al. Room-temperature continuous-wave electrically pumped InGaN/GaN quantum well blue laser diode directly grown
on Si. Light Sci. Appl. 7, (2018).
16. Zuniga-Perez, J. et al. Patterned silicon substrates: a common platform for room temperature GaN and ZnO polariton lasers. Appl.
Phys. Lett. 104, 241113 (2014).
17. Damilano, B. et al. Growth of nitride-based light emitting diodes with a high-reectivity distributed Bragg reector on mesa-
patterned silicon substrate: growth of nitride-based LEDs with a high-reectivity DBR. Phys. Status Solidi A 212, 2297–2301 (2015).
18. Comyn, R. et al. AlGaN/GaN/AlGaN DH-HEMTs grown on a patterned silicon substrate. Phys. Status Solidi A 1700642 (2017).
https :// 0642.
19. Xu, J. et al. Cathodoluminescence study of InGaN/GaN quantum-well LED structures grown on a Si substrate. J. Electron. Mater.
36, 1144–1148 (2007).
20. Mukai, T. Recent progress in group-III nitride light-emitting diodes. IEEE J. Sel. Top. Quantum Electron. 8, 264–270 (2002).
21. Krames, M. R. et al. Status and future of high-power light-emitting diodes for solid-state lighting. J. Disp. Technol. 3, 160–175
22. Gil, B. Physics of Wurtzite nitrides and oxides: passport to devices. (Springer, 2014).
23. Hums, C. et al. Fabry-Perot eects in InGaN∕GaN heterostructures on Si-substrate. J. Appl. Phys. 101, 033113 (2007).
24. Vurgaman, I. & Meyer, J. R. Band parameters for nitrogen-containing semiconductors. J. Appl. Phys. 94, 3675 (2003).
25. Hossain, T. GaN based structures on patterned silicon substrate: Stress and strain analysis. (Nice-Sophia Antipolis, 2012).
26. Even, A. et al. Enhanced In incorporation in full InGaN heterostructure grown on relaxed InGaN pseudo-substrate. Appl. Phys.
Lett. 110, 262103 (2017).
27. Pasayat, S. S. et al. Fabrication of relaxed InGaN pseudo-substrates composed of micron-sized pattern arrays with high ll factors
using porous GaN. Semicond. Sci. Technol. 34, 115020 (2019).
28. Frayssinet, E., Cordier, Y., Schenk, H. P. D. & Bavard, A. Growth of thick GaN layers on 4-in. and 6-in. silicon (111) by metal-
organic vapor phase epitaxy. Phys. Status Solidi C 8, 1479–1482 (2011).
29. Cli, G. & Lorimer, G. W. e quantitative analysis of thin specimens. J. Microsc. 103, 203–207 (1975).
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Scientic Reports | (2020) 10:18919 | 
is work was partly funded by the French National Research Agencey (ANR) within the project no. 2011
RMNP 018 01 “MOSAIC.” e work done on the NanoCharacterisation PlatForm (PFNC), was supported by
the “Recherches Technologiques de Base” Program of the French Ministry of Research. e authors acknowledge
the support from the OPTITEC, MINALOGIC, and MONT BLANC INDUSTRIES competitiveness clusters.
e authors would like to thank Julien Brault, Mathieu Leroux and Jean Massies for the critical reading of the
Author contributions
B.D. conceived the experimental work with the contribution of G.F. and D.T., participated to the MOVPE growth
of the samples and carried out the photoluminescence experiments. E.F. carried out the MOVPE growth of the
samples. M.P. carried out the cathodoluminescence experiments. V.B. carried out the scanning electron micros-
copy experiments. F.F. and C.L. prepared the patterned silicon substrates. D.C. carried out the transmission
electron microscopy experiments. B.D. wrote the main manuscript text with the contribution of D.C., M.P., and
C.L. All the authors discussed the results and commented on the manuscript during the manuscript writing.
Competing interests
e authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https :// 8-020-76031 -3.
Correspondence and requests for materials should be addressed to B.D.
Reprints and permissions information is available at
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access is article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.
© e Author(s) 2020
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Current laser-based display and lighting applications are invariably using blue laser diodes (LDs) grown on free-standing GaN substrates, which are costly and smaller in size compared with other substrate materials.1–3 Utilizing less expensive and large-diameter Si substrates for hetero-epitaxial growth of indium gallium nitride/gallium nitride (InGaN/GaN) multiple quantum well (MQW) structure can substantially reduce the cost of blue LDs and boost their applications. To obtain a high crystalline quality crack-free GaN thin film on Si for the subsequent growth of a blue laser structure, a hand-shaking structure was formed by inserting Al-composition step down-graded AlN/AlxGa1−xN buffer layers between GaN and Si substrate. Thermal degradation in InGaN/GaN blue MQWs was successfully suppressed with indium-rich clusters eliminated by introducing hydrogen during the growth of GaN quantum barriers (QBs) and lowering the growth temperature for the p-type AlGaN/GaN superlattice optical cladding layer. A continuous-wave (CW) electrically pumped InGaN/GaN quantum well (QW) blue (450 nm) LD grown on Si was successfully demonstrated at room temperature (RT) with a threshold current density of 7.8 kA/cm2.
Fully or partially relaxed micron-sized InGaN patterns with fill factors up to 69% were demonstrated via porosification of the underlying GaN: Si layer. The impact of the porosification etch conditions and the pattern geometry on the degree of InGaN relaxation were studied and monitored via high resolution x-ray diffraction reciprocal space maps. Additionally, a 45 nm redshift in the photoluminescence emission from In x Ga 1− x N/In y Ga 1− y N multi-quantum wells (MQWs) regrown on bi-axially relaxed InGaN buffer layers was observed when compared to a co-loaded reference sample grown on GaN. The longer emission wavelength was associated with higher indium incorporation into the InGaN layers deposited on the InGaN base layers with a lattice constant larger than GaN, due to the reduced lattice mismatch between MQW and InGaN base layer, also called compositional pulling effect.
A specially designed InGaN/GaN superlattice (SL) interlayer was inserted between n-GaN and a multiple quantum well to enhance the performance of yellow light-emitting diodes (LEDs) grown on Si (111). The number of SL periods was determined to be the key to enhancing the external quantum efficiency and reducing forward voltage. Our results show that more SLs could suppress nonradiative recombination by eliminating micron-scale indium-rich clusters and could promote hole injection with increased V-pit size. However, too many SLs reduce the effective luminescence area and lead to many voids formed in the p-type layer. We demonstrate that 32 is the optimum number of SLs for yellow InGaN/GaN LEDs, obtaining a high light output power of 63 mW with a dominant wavelength of 568 nm, and a low forward voltage of 2.38 V at 200 mA (20 A/cm²).
In the present work, a Si (111) substrate has been patterned with 120-μm wide square mesas separated by 5-μm deep grooves. An Al0.29Ga0.71N/GaN HEMT structure was grown by molecular beam epitaxy with a 2-μm thick Al0.15Ga0.85N buffer on a strain-mitigating stack previously developed for GaN and low Al-content AlGaN buffers. Surface inspections confirm the absence of cracks and roughness modification compared to planar growth. Interestingly, contrary to planar growth, X-ray diffraction shows that plastic strain relaxation was inhibited within the 150-nm GaN channel grown on top of the Al0.15Ga0.85N buffer. Devices were fabricated in order to assess the electrical properties of the grown films. C–V and TLM measurements reveal the presence of a 2DEG with a density of 7 × 10¹² cm⁻² and a sheet resistance around 450 Ω □⁻¹. Round geometry transistors were measured up to 200–V drain bias. Compared to devices previously fabricated on planar structures with Al0.05Ga0.95N buffers, an order of magnitude lower off-state leakage currents were obtained thanks to the larger Al content in the buffer.
Additive color mixing across the visible spectrum was demonstrated from an InGaN based light-emitting diode (LED) pixel comprising red, green, and blue subpixels monolithically integrated and enabled by local strain engineering. The device was fabricated using a top-down approach on a metal-organic chemical vapor deposition-grown sample consisting of a typical LED epitaxial stack. The three color subpixels were defined in a single lithographic step. The device was characterized for its electrical properties and emission spectra under an uncooled condition, which is desirable in practical applications. The color mixing was controlled by pulse-width modulation, and the degree of color control was also characterized.
The impact of a relaxed InGaN pseudosubstrate on indium incorporation in a full InGaN heterostructure was investigated. Three types of InGaN pseudosubstrates were tested with different a lattice parameters ranging from 3.190 to 3.205 Å, that is to say, greater than that of a GaN template on sapphire. Samples were loaded together in the growth chamber in order to apply exactly the same growth conditions. The effect of the photoluminescence(PL) emission redshift was observed on InyGa1-yN buffer layers and also on InxGa1-xN/InyGa1-yN multiple quantum wells(MQWs). It was found that these pseudosubstrates have the ability to improve the indium incorporation rate, with an increasing effect as the a lattice parameter increases. A strong PL emission redshift was observed in InxGa1-xN/InyGa1-yN MQWs as a function of the increasing a lattice parameter of the InGaN pseudosubstrate, compared to a reference grown on a GaN template. It has been shown that green and amber emissions can be easily reached. A redshift of up to 42 nm was detected between various InGaN pseudosubstrate samples and up to 62 nm compared to a conventional structure emerged from a GaN buffer on the sapphire substrate. The average QW width less than 3 nm indicates a higher In content. The reduced compressive strain originating from the relaxed InGaN substrate allows the reduction in the compositional pulling effect and consequently enables an enhanced In incorporation rate.
High-brightness GaN-based emissive microdisplays can be fabricated with different approaches that are listed and described. They consist either of hybridizing a GaN LED array on a CMOS circuit or building a monolithic component on a single substrate. Using the hybridization approach, two types of 10-μm pixel pitch GaN microdisplay prototypes were developed: (1) directly driven, 300 × 252 pixels and (2) active-matrix, 873 × 500 pixels. Brightness as high as 1 × 10⁶ and 1 × 10⁷ cd/m² for blue and green arrays, respectively, were reached. GaN-based emissive microdisplays are suitable for augmented reality systems or head-up displays, but some challenges remain before they can be put in production.
Multi-color single InGaN/GaN dot-in-nanowire light emitting diodes (LEDs) were fabricated on the same substrate using selective area epitaxy. It is observed that the structural and optical properties of InGaN/GaN quantum dots depend critically on nanowire diameters. Photoluminescence emission of single InGaN/GaN dot-in-nanowire structures exhibits a consistent blueshift with increasing nanowire diameter. This is explained by the significantly enhanced indium (In) incorporation for nanowires with small diameters, due to the more dominant contribution for In incorporation from the lateral diffusion of In adatoms. Single InGaN/GaN nanowire LEDs with emission wavelengths across nearly the entire visible spectral were demonstrated on a single chip by varying the nanowire diameters. Such nanowire LEDs also exhibit superior electrical performance, with a turn-on voltage ~ 2V and negligible leakage current under reverse bias. The monolithic integration of full-color LEDs on a single chip, coupled with the capacity to tune light emission characteristics at the single nanowire level, provides an unprecedented approach to realize ultra-small and efficient projection display, smart lighting, and on-chip spectrometer.