Journal of Crystal Growth 548 (2020) 125848

Contents lists available at ScienceDirect

Journal of Crystal Growth

journal homepage: www.elsevier.com/locate/jcrysgro

GaAs epitaxial growth on R-plane sapphire substrate T

a b,⁎ b b b
Samir K. Saha , Rahul Kumar , Andrian Kuchuk , Hryhorii Stanchu , Yuriy I. Mazur ,
Shui-Qing Yuc, Gregory J. Salamoa,b
a
Department of Physics, University of Arkansas, Fayetteville, AR 72701, USA
b
Institute for Nanoscience and Engineering, University of Arkansas, Fayetteville, AR 72701, USA
c
Department of Electrical Engineering, University of Arkansas, Fayetteville, AR 72701, USA

ARTICLE INFO ABSTRACT

Communicated by H. Asahi Achieving epitaxial growth of III-As on r-plane sapphire would potentially allow the integration of both laser and
amplifier with corresponding RF electronics. Here we report on the growth of high-quality GaAs on an r-plane
Keywords:
sapphire substrate by molecular beam epitaxy. The epitaxial relationship between GaAs and r-plane sapphire is
A3. Molecular beam epitaxy
B1. Dissimilar materials observed and explained. GaAs on r-plane sapphire resulted in (1 1 1) orientation, similar to growth orientation
B1. Sapphire observed on c-plane sapphire. However, in comparison with growth of GaAs growth on a c-plane sapphire
B1. GaAs substrate, a stronger interaction is observed between GaAs on r-plane sapphire. The effect of growth temperature
A1. High resolution X-ray diffraction is also investigated for GaAs growth on r-plane sapphire. It is found that GaAs island size, density, and or­
A1. Twinning ientation can be tuned by varying the growth temperature. Finally, a thin AlAs nucleation layer on r-plane
sapphire has been introduced to study its effect on the growth of GaAs. The introduction of the AlAs nucleation
layer is found to enhance the wetting of GaAs but at the expense of introducing twinning defects.

1. Introduction GaAs may align with the rectangular r-plane of sapphire, while the
hexagonal nature of the sapphire substrate might force the epitaxial
GaAs and AlAs on sapphire are excellent candidates for microwave growth of GaAs to be of (1 1 1) orientation, or GaAs might take a totally
photonics, optoelectronics and electronics owing to a thermal expan­ different orientation. Here we report on an investigation of the epitaxial
sion coefficient match and a large refractive index contrast between the growth of III-As on r-plane (11̄02) sapphire.
two materials [1]. An additional attractive quality of the sapphire The epitaxial orientation and the correlation between film and
substrate is its large bandgap (~9 eV) resulting in excellent resistance substrate are certainly important and can have an impact on both the
to radiation damage, making it a candidate for space and nuclear-re­ optical and carrier transport properties of GaAs structures. To under­
lated applications [2]. In fact, r-plane (11̄02) sapphire, as a substrate, is stand the possible orientations for GaAs on r-plane sapphire, the atomic
an ideal platform for integrated RF electronics and already the choice arrangement of the r-plane of sapphire, the (1 1 1) plane and (1 1 0)
for silicon on sapphire (SOS) technology. While there are a few reports plane of GaAs are shown together in Fig. 1. The lattice mismatch has
of epitaxial growth of cubic Si(Ge) on both c and r plane sapphire [3–6] been calculated considering each different potential alignment shown.
and cubic III-As on c-plane sapphire [1,7–12], to the best of our For example, for the case of the GaAs (1 1 1) crystal plane the com­
knowledge, there are no reports about epitaxial growth of III-As on r- pressive strain for GaAs on sapphire is 35.2% and the tensile strain is
plane sapphire as a substrate. Achieving epitaxial growth of III-As on r- 16.2%, in the two perpendicular directions. On the other hand, for the
plane sapphire would potentially allow the integration of both laser and (1 1 0) crystal plane, it has compressive strain (10.7%) and tensile strain
amplifier with corresponding RF electronics. (16.2%).
Heteroepitaxy normally combines two different materials having a While the lattice mismatch is very important, the substrate surface
similar crystal structure and most often results in epitaxial films having energy also plays an important role in deciding the orientation and
the same orientation as of the substrate. However, for the hetero­ quality of the epitaxial growth of GaAs on sapphire. For example, for
epitaxial material system of GaAs on sapphire we have a combination of layer-by-layer growth of epitaxial GaAs material, the energy dynamics
dissimilar material systems. Consequently, it is not trivial to grow nor at the surface plays a role through the relationship:
predict the orientation and quality of the GaAs film. For example, for
the growth of GaAs on r-plane sapphire, the rectangular (1 1 0) plane of + (1)
s i e

⁎
Corresponding author.

https://doi.org/10.1016/j.jcrysgro.2020.125848
Received 3 August 2020; Received in revised form 19 August 2020; Accepted 21 August 2020
Available online 27 August 2020
0022-0248/ © 2020 Elsevier B.V. All rights reserved.
S.K. Saha, et al. Journal of Crystal Growth 548 (2020) 125848

Fig. 1. Atomic arrangement of (a) r-plane sapphire, (b) GaAs (1 1 0) plane, and (c) GaAs (1 1 1) plane.

where γs is substrate surface energy, γi is interface energy between microscopy (AFM) (Bruker, model number 3000 dimension III). A si­
substrate and film, and γe is film surface energy. The surface energy of licon tip with radius 10 nm was used in the AFM to image samples
relaxed c-plane and r-plane is 1.85 and 2.26 Jm−2 respectively [13]. under equilibrium conditions and using optimized feedback and force
Interestingly, the higher surface energy of r-plane sapphire over c-plane parameters. X-ray diffraction (XRD) was used to determine the epitaxial
sapphire favors a better film to substrate interaction in the case of GaAs orientation and crystal quality of the substrate and films. The XRD in­
growth on an r-plane sapphire substrate. With this difference in mind, strument was equipped with a multilayer focusing mirror and a Cu Kα1
we investigated the growth of GaAs on sapphire by (i) comparing source of radiation with wavelength 1.54 Å.
growth on c-plane and r-plane substrates, (ii) investigating the effect of
growth parameters on the material morphology, and (iii) probing the
3. Results and discussion
effect of an AlAs nucleation layer on the growth of GaAs on r-plane
sapphire.
All sapphire substrates, used in this work, possessed a well-defined
starting step-terrace surface. The importance of having a step-terrace
2. Experimental procedure and clean substrate surface for dissimilar material growth is shown
elsewhere [1,16]. These surfaces were achieved by heating as-received
A Riber-32 molecular beam epitaxy (MBE) chamber was used to substrates at 1200 °C for 6 h in atmospheric conditions. Fig. 2 shows the
grow all the samples. Substrates were back side coated with 1-µm thick representative AFM images of a c-plane and an r-plane sapphire sub­
Ti for efficient and uniform radiative heating from the heater [14]. strate. For the c-plane sapphire substrate, the average step height is
Substrates were introduced into the load lock chamber and heated at 0.22 nm, which corresponds to 1 monolayer (ML) of c-plane sapphire
200 °C for one hour to evaporate water vapor from the substrate sur­ with an average terrace width of 236 nm.
face. Afterward, substrates were annealed at 900 °C for 6 h in the de­ On the other hand, the average step height and terrace width of r-
gassing chamber to remove organic contaminants and then transferred plane sapphire was 0.76 nm and 198 nm, respectively. The difference is
to the MBE growth chamber, all under vacuum. In the growth chamber, also due to the unintentional miscut of the r-plane substrate (0.2°),
the surface of each substrate was exposed to an arsenic flux of 2 × 10−6 which was higher than the c-plane substrate (0.06°). The RHEED pat­
torr at 650 °C for half an hour [15], after which the substrate tem­ tern for both surfaces is shown in the inset of both images. Narrow
perature was set to the growth temperature. All temperatures men­ streaks, and the presence of Kikuchi lines, confirm the cleanliness of the
tioned in this work are the thermocouple temperature reading, which is substrate surfaces.
not in touch with the substrate so there can be as much as a 50 °C
temperature difference between the thermocouple reading and the ac­ (a) Comparison of GaAs growth on c and r-plane sapphire:
tual surface temperature during growth. The growth rates for GaAs and
AlAs were 0.75 and 0.2 ML/s, respectively. To study the growth in real- GaAs was grown directly on c-plane (sample: C-600) and r-plane
time, reflection high energy electron diffraction (RHEED) was used at sapphire (sample: R-600) after substrate preparation. 10 nm GaAs were
20 keV accelerating voltage and 1.5A cathode current at a glancing deposited on both substrates at 600 °C under identical growth condi­
angle of 1-2° to the substrate. tions. Thickness values were calibrated for homoepitaxial GaAs (1 0 0)
The surface morphology, after growth, for each sample was in­ growth. AFM surface morphologies of both samples are shown in
vestigated using the intermittent contact mode of atomic force Fig. 3(a) and (b). Both images show GaAs three-dimensional (3D)

2
S.K. Saha, et al. Journal of Crystal Growth 548 (2020) 125848

Fig. 2. AFM images of c-plane (a) and r-plane (b) sapphire substrates. Inset shows the substrate RHEED just before the growth.

Fig. 3(c) shows the GaAs (2 2 0) phi-scans of both samples. All GaAs
(2 2 0) peaks of C-600 are shown with arrows. GaAs peaks in R-600 are
very sharp, whereas peaks are broad in C-600. This sharp peak shows a
well-defined in-plane correlation in R-600 as opposed to C-600 where
the in-plane correlation with the substrate is very weak. Moreover, phi-
scan shows six-fold symmetry for C-600, indicating the presence of 60°
rotated twins. Six peaks corresponding to GaAs (2 2 0) can also be ob­
served in R-600, but these peaks are not exactly separated by 60°. This
shows that there are two different domains of [1 1 1] oriented crystals,
and these two domains are maintaining two different in-plane crystal­
lographic orientation relationships with the r-plane sapphire substrate.
The equal intensity of both types of peaks shows that both orientation
relationships are equally preferable. RHEED images after GaAs growth
from C-600 and R-600 are shown in the inset of Fig. 3(c). For C-600,
RHEED shows spots and rings. The spotty pattern indicates the 3-D
growth mode, whereas the ring pattern indicates the weak in-plane
correlation between substrate and GaAs which agrees with the XRD
result. On the other hand, for the r-plane (R-600), spots without a ring
pattern was observed. This again shows the better interaction and well-
defined GaAs orientation relationship of r-plane sapphire than with the
c-plane sapphire substrate. All results are consistent with expectations
based on Eq. (1).
XRD omega-2theta scans of both samples are shown in Fig. 4. For C-
600, only a GaAs (1 1 1) orientation is detected, whereas (1 1 1) and
Fig. 3. AFM Surface morphology on 2 μm × 2 μm scan area of 10 nm GaAs
growth on (a) c-plane (C-600) and (b) r-plane sapphire substrate (R-600); (c)
Phi-scans of both samples, arrows (red for R-600 and black for C-600) showing
GaAs (2 2 0) peaks, corresponding RHEED images after the growth are shown in
the inset of figure.

Table 1
Island height, density, and total volume on 2 μm × 2 μm AFM scan area.
Sample ID Average height (nm) Density (cm−2) Volume of 3D islands (m3)

C-600 85 4.25 × 108 4 × 10−20

R-600 40 4.35 × 109 4.3 × 10−20

islands, indicating a 3D growth mode. 3D island size, density, and vo­

lume of GaAs material deposited are listed in Table 1 for a 2 μm × 2 μm
scanned area. In both cases, 3D islands are observed. High lattice
mismatch and dissimilar crystal structure in both cases promote 3D
growth. Since the surface energy of r-plane sapphire is higher than the
c-plane sapphire, the interaction of GaAs on the r- plane should be
higher than the c-plane sapphire substrate, which increased the density
of GaAs islands in r-plane. Basically, diffusion and ripening are slower Fig. 4. XRD omega-2theta scan of C-600 and R-600. Sharp and intense peaks
on r-plane. are related to sapphire substrates.

3
S.K. Saha, et al. Journal of Crystal Growth 548 (2020) 125848

(1 1 0) orientations are detected for R-600. A higher angle tail in the Table 2
GaAs (1 1 1) peak for R-600 is also observed, perhaps due to the strain Growth temperature, GaAs island height and density for three GaAs/r-plane
difference in the two orientations of GaAs observed in phi-scan. sapphire samples.
Sample ID Growth temperature (°C) Height (nm) Density (/cm2)
(b) Effect of growth temperature on GaAs/r-plane sapphire:
R-500 500 9 NA
R-600 600 40 4.35 × 109
There are a number of parameters that can affect the surface mor­
R-650 650 52.6 3.4 × 109
phology and crystal quality of the grown materials, such as growth
temperature, growth rate, the ratio of V/III beam flux, and the substrate
surface [1,17]. To investigate the effect of growth temperature, four for a better in-plane correlation between epitaxial GaAs films and
different temperatures 500 °C (R-500), 600 °C (R-600), 650 °C (R-650) substrates.
and 700 °C (R-700) were used to grow 10 nm of GaAs on the r-plane
sapphire substrate. The study of growth temperature effects of GaAs (a) Introduction of AlAs layer:
growth on c-plane sapphire is discussed in our previous work [1]. For
GaAs on the r-plane, growth at 700 °C resulted in no detectable de­ Since GaAs does not wet the sapphire surface very well, we in­
position of GaAs. AFM, RHEED and X-ray photoelectron spectroscopy vestigated adding a thin AlAs buffer layer on r-plane sapphire before
(XPS) show no signature of GaAs or metallic Ga deposition. Apparently, adding GaAs. This was motivated by our earlier work that showed that
the growth temperature is too high for any appreciable sticking of ei­ a thin AlAs layer improved the chemical interaction and hence, the
ther Ga or As adatoms. However, GaAs was deposited, and the surface quality of subsequent GaAs thin films on c-plane sapphire [1]. There­
morphology and crystal quality were investigated, at 500 °C, 600 °C, fore, after adding a thin AlAs layer on sapphire, GaAs was grown under
and 650 °C. Fig. 5 shows the surface morphologies for these three three different arsenic flux conditions: 1 × 10−6 (RA-1), 3.5 × 10−6
samples. (RA-3.5) and 6 × 10−6 (RA-6). For these samples, the GaAs growth
R-500 (Fig. 5a) has the best surface coverage by GaAs islands. GaAs temperature was held at 600 °C. Fig. 7 shows the AFM surface
island size increases with increasing growth temperature while the is­ morphologies of these samples. For all three samples, highly dense 3D
land density decreases with GaAs growth temperature (see Table 2). islands are observed. The surface roughness of these samples is in­
These observations are consistent with ripening [18,19]. From the creased with increasing arsenic pressure, and the roughness values are
RHEED images, R-500 sample RHEED shows the ring pattern, which 2.53 nm (RA-1), 2.86 nm (RA-3.5), and 3.68 nm (RA-6). While com­
indicates a weak in-plane correlation between the film and the sub­ paring these surfaces to the samples where GaAs were directly grown
strate. With increasing temperature, however, for both R-600 and R- on sapphire (R-500, R-600, and R-650), we can see that adding a thin
650, the ring pattern vanishes, and a spotty pattern becomes prominent. AlAs layer does improve the wetting the sapphire substrate with GaAs,
Twinning can also be seen in RHEED images for R-600 and R-650. similar to our observation on c-plane sapphire 1. The Al-O (502 kJ/mol)
Omega-2theta scans of these three samples are shown in Fig. 6(a). bond is stronger compared to the Ga-O (374 kJ/mol) and the As-O
Only the (1 1 1) orientation was observed for R-500 and R-650. Mean­ (374 kJ/mol) bonds [20]. Stronger bonding at the interface will reduce
while, the R-600 sample shows the presence of two orientations: (1 1 1) the interface energy. Hence, improved wetting after introduction of
and (1 1 0). Fig. 6(b) shows the phi-scans of GaAs (2 2 0) plane for (1 1 1) AlAs is consistent with Eq. (1).
orientation. The orientation relationship remains unchanged with the Fig. 8(a) shows the omega-2theta scan of these three samples. Only
variation of growth temperature. It is also noticeable from the phi scan GaAs (1 1 1) orientation was observed. Here we observe a clear peak
that a higher temperature helps to improve the in-plane correlation. For beside the GaAs (1 1 1) peak towards the higher angle side instead of a
the low -temperature sample (R-500), the phi peaks are broad and tail that was observed for direct growth of GaAs on r-plane sapphire.
diffuse, which indicates the weak in-plane correlation, and this result is This peak corresponds to a high tensile strain for GaAs/AlAs material.
also in agreement with RHEED. The phi peaks for both samples (R-600, Fig. 8(b) shows the phi-scans of these samples and r plane sapphire.
R-650) are sharp and well defined, so the high temperature is necessary

Fig. 5. AFM and RHEED of (a) R-500 (b) R-600 (c) R-650.

4
S.K. Saha, et al. Journal of Crystal Growth 548 (2020) 125848

Fig. 6. (a) Omega-2theta scan, and (b) phi-scan of R-500, R-600 and R650.

Fig. 7. AFM surface morphology of (a) RA-1 (b) RA-3.5 (c) RA-6.

Fig. 8. (a) Omega-2theta scan and (b) phi-scan of RA-1, twin peaks are denoted by rectangles, angles between different types of peaks are shown in the figure. Pole
figure of the same sample is shown in the inset with sample tilt ranging from 0 to 70°.

In the zinc blende system, the phi-scan of the 220 reflection shows orientations are depicted in Fig. 9 by two hexagons O1 and O2. Atomic
three equally separated peaks, which represent three-fold symmetry. In distances are to the scale with small error. These two orientations of the
the presence of 60° twins, it shows six equally separated peaks. 110 axis of GaAs make an angle of approximately ± 14° with respect to
However, here we observe 12 different 220 peaks, which is unusual for the [1 0 0] in-plane direction of r-plane sapphire.
a cubic zinc blende system (shown in Fig. 8b). This is due to two major It is possible that initially, with smaller coverage of GaAs, only one
orientations, which are shifted by approximately ± 14° with respect to orientation of GaAs existed where [1 0 0] in-plane direction of sapphire
the 0006 reflection of the r-plane sapphire. Both orientations have their was aligned to [1 1 0] in-plane direction of GaAs. In other words, GaAs
60° twin peaks, which makes a total of 12 (2 2 0) peaks. The two hexagon was sitting on seed hexagon (shown by green dashed lines in

5
S.K. Saha, et al. Journal of Crystal Growth 548 (2020) 125848

Fig. 9. Arrangement of Al atoms in r plane sapphire (blue dots) and two possible in-plane orientations of GaAs separated by approximately 28°.

Fig. 9). As the coverage of GaAs increases, the strain also increases, and Acknowledgments
initial orientation split in two orientation O1 and O2 by a 14° twist to
minimize the strain. The authors acknowledge the financial support by the Institute for
Nanoscience and Engineering, University of Arkansas, and the project
4. Conclusion “Quantum Interfaces of Dissimilar Materials” funded by National
Science Foundation(NSF) (Grant No. 1809054). Additionally, Dr. Shui-
A highly dissimilar material system of GaAs on r-plane sapphire has Qing Yu would also like to acknowledge the NSF funding support
been grown. It has been compared with GaAs on c-plane sapphire. The ECCS1745143.
role of growth temperature and the addition of an AlAs nucleation layer
on GaAs epitaxial films were investigated. Due to the higher surface References
energy of r-plane sapphire, the interaction between GaAs films and r-
plane sapphire is better than c-plane sapphire. Two in-plane correla­ [1] S.K. Saha, R. Kumar, A.V. Kuchuk, M.Z. Alavijeh, Y. Maidaniuk, Y.I. Mazur, S.-
Q. Yu, G.J. Salamo, Cryst. Growth Des. 19 (2019) 5088.
tions are observed for GaAs films with r -plane sapphire, which is ~28° [2] S. Ebnesajjad, S. Ebnesajjad, in: Surf. Treat. Mater. Adhes. Bond, second ed.,
apart, irrespective of growth temperatures. The AlAs nucleation layer William Andrew Publishing, 2014, pp. 283–299.
enhances the wetting of GaAs films on r-plane sapphire substrates; [3] H.J. Kim, Y. Park, H. Bin Bae, S.H. Choi, Adv. Condens. Matter Phys. (2015,
2015,) 1.
however, twinning defects are also observed. We are optimistic that a [4] H.J. Kim, H. Bin Bae, Y. Park, K. Lee, S.H. Choi, J. Cryst. Growth 353 (2012) 124.
two-step growth method, with annealing, will result in a quality GaAs [5] Y. Park, G.C. King, S.H. Choi, J. Cryst. Growth 310 (310) (2008) 2724.
surface as demonstrated for GaAs on c-plane sapphire. [6] W. Dubbelday, K. Kavanagh, J. Cryst. Growth 222 (2001) 20.
[7] A. Sugimura, T. Hosoi, K. Ishibitsu, T. Kawamura, J. Cryst. Growth 77 (1986) 524.
[8] K. Kasai, K. Nakai, M. Ozeki, J. Appl. Phys. 60 (1986) 1.
CRediT authorship contribution statement [9] T.P. Humphreys, C.J. Miner, J.B. Posthill, K. Das, M.K. Summerville, R.J. Nemanich,
C.A. Sukow, N.R. Parikh, Appl. Phys. Lett. 54 (1989) 1687.
[10] T.P. Humphreys, C.J. Miner, N.R. Parikh, K. Das, M.K. Summerville, J.B. Posthill,
Samir K. Saha: Data curation, Methodology, Investigation. Rahul R.J. Nemanich, C.A. Sukow, MRS Proc. 144 (1988) 195.
Kumar: Investigation, Visualization, Writing - original draft. Andrian [11] T.F. Kuech, A. Segmüller, T.S. Kuan, M.S. Goorsky, J. Appl. Phys. 67 (1990) 6497.
[12] L.C. Chuang, M. Moewe, K.W. Ng, T.-T.-D. Tran, S. Crankshaw, R. Chen, W.S. Ko,
Kuchuk: Data curation. Hryhorii Stanchu: Formal analysis. Yuriy I. C. Chang-Hasnain, Appl. Phys. Lett. 98 (2011) 123101.
Mazur: Project administration. Shui-Qing Yu: Conceptualization, [13] T. Stirner, J. Sun, M. Aust, Phys. Procedia 32 (2012) 635.
Funding acquisition. Gregory J. Salamo: Writing - review & editing, [14] V. Petukhov, A. Bakin, I. Tsiaoussis, J. Rothman, S. Ivanov, J. Stoemenos, A. Waag,
J. Cryst. Growth 323 (2011) 111.
Conceptualization, Funding acquisition.
[15] C.J.M.T.P. Humphreys, N.R. Parikh, K. Das, J.B. Posthill, R.J. Nemanich,
M.K. Summerville, C.A. Sukow, J. Chem. Inf. Model. 53 (2013) 1689.
[16] H.J. Kim, A. Duzik, S.H. Choi, Acta Mater. 145 (2018) 1.
Declaration of Competing Interest
[17] L. Esposito, S. Bietti, A. Fedorov, R. Nötzel, S. Sanguinetti, Phys. Rev. Mater. 1
(2017) 024602.
The authors declare that they have no known competing financial [18] G. Madras, B.J. McCoy, J. Chem. Phys. 119 (2003) 1683.
[19] T.V. Western, R. Groot, Cryst. Growth Des. 18 (2018) 4952.
interests or personal relationships that could have appeared to influ­ [20] Y.-R. Luo, Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC
ence the work reported in this paper. Press, 2007.