Modification of the growth mode of Ge on Si by buried Ge islands

Article abstract of DOI:10.1063/1.126762

Photoluminescence experiments on double Ge layers were performed to give deep insights on the growth mode of Ge on Si in the presence of buried 4.5 monolayers of Ge islands. The critical coverage of the island formation and the wetting layer thickness were confirmed to be reduced in the second Ge layer. In addition, a drastic increase of the island density as well as a shape transition were observed by atomic force microscopy. These modifications of the growth mode are explained in terms of the surface strain induced by the buried Ge islands and the reduction of the nucleation barrier due to the alloying.

Full text of DOI:10.1063/1.126762

Modification of the growth mode of Ge on Si by buried Ge islands
N. Usami, Y. Araki, Y. Ito, M. Miura, and Y. Shiraki
Citation: Applied Physics Letters 76, 3723 (2000); doi: 10.1063/1.126762
View online: http://dx.doi.org/10.1063/1.126762
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/76/25?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Localized Si enrichment in coherent self-assembled Ge islands grown by molecular beam epitaxy on (001)Si
single crystal
J. Appl. Phys. 113, 033513 (2013); 10.1063/1.4775772
Positioning of self-assembled Ge islands on stripe-patterned Si(001) substrates
J. Appl. Phys. 93, 6258 (2003); 10.1063/1.1566455
The effect of strain field seeding on the epitaxial growth of Ge islands on Si(001)
Appl. Phys. Lett. 78, 1658 (2001); 10.1063/1.1352660
Influence of molecular hydrogen on Ge island nucleation on Si(001)
J. Appl. Phys. 88, 5113 (2000); 10.1063/1.1317245
Photoluminescence study of the initial stages of island formation for Ge pyramids/domes and hut clusters on
Si(001)
Appl. Phys. Lett. 75, 1905 (1999); 10.1063/1.124867
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.121.247.60 On: Sat, 08 Nov 2014 02:53:26

APPLIED PHYSICS LETTERS
VOLUME 76, NUMBER 25
19 JUNE 2000
Modification of the growth mode of Ge on Si by buried Ge islands
N. Usami,a) Y. Araki, Y. Ito, M. Miura, and Y. Shiraki
Research Center for Advanced Science and Technology (RCAST), The University of Tokyo, 4-6-1 Komaba,
Meguro-ku, Tokyo 153-8904, Japan
͑Received 10 January 2000; accepted for publication 25 April 2000͒
Photoluminescence experiments on double Ge layers were performed to give deep insights on the
growth mode of Ge on Si in the presence of buried 4.5 monolayers of Ge islands. The critical
coverage of the island formation and the wetting layer thickness were confirmed to be reduced in the
second Ge layer. In addition, a drastic increase of the island density as well as a shape transition
were observed by atomic force microscopy. These modifications of the growth mode are explained
in terms of the surface strain induced by the buried Ge islands and the reduction of the nucleation
barrier due to the alloying. © 2000 American Institute of Physics. ͓S0003-6951͑00͒03725-6͔
Self-assembled Ge islands on Si͑100͒ have been inten-
sively investigated as the basis of future electronic and opti-
cal devices.1–6 The island is known to develop during epi-
taxial growth in the Stranski–Krastanow growth mode;
where beyond a critical thickness, a nucleation of islands on
the two-dimensional wetting layer becomes energetically fa-
vorable since the gain of strain energy due to the partial
strain relaxation overcompensates the increased surface en-
ergy. Owing to the simple in situ growth approach without
lithography, the self-assembly is now regarded as an attrac-
tive and promising route to the fabrication of the small is-
lands which might act as quantum dots. However, to estab-
lish a method to achieve sufficiently uniform island sizes
with regular spatial distribution still remains a critical issue.
This should be solved since a well defined size with little
dispersion is generally required for any practical applica-
tions.
As a possible way to obtain the size uniformity, Tersoff
et al. have pointed out that the growth of multilayer arrays of
coherently strained islands is useful.7 Based on a simple
model of the surface strain, the arrangement of islands was
predicted to become more uniform with stacking the islands
regardless the initial condition, and atomic force microscope
͑AFM͒ images of SiGe/Si superlattices were given as experi-
mental evidence.
coverage of the first layer was fixed to be 4.5 monolayers
͑ML͒, which is larger than the critical coverage of the island
formation of 3.7 ML.1 Therefore, three-dimensional Ge is-
lands with leaving 3.0 ML of the wetting layer are expected
in the first layer. The coverage of the second Ge layer was
systematically changed from 0.6 to 9.0 ML. As a sensitive
tool to investigate the growth mode, PL spectroscopy was
performed. In Ge/Si system, PL spectroscopy has been al-
ready demonstrated to be a powerful tool to monitor the
growth mode changeover to the island formation at high
growth temperatures1,2 as well as the different island phases
such as domes, pyramids, and hut clusters grown at low
temperatures.8 The samples for PL measurements were
capped with 300 nm of Si. PL spectra were recorded using a
standard lock-in technique with a liquid-nitrogen-cooled Ge
photodetector ͑North-Coast EO-817L͒. The samples were
mounted on a cold finger of a closed-cycle cryostat. An ar-
gon ion laser was used as an excitation source. For some
samples, AFM measurements were performed ex situ by in-
terrupting the epitaxial growth at heterointerfaces.
Figure 1 shows the PL results of a series of samples
where the Ge coverage of the second layer was systemati-
In this letter, we report on the growth mode of Ge on
Si͑100͒ with buried Ge islands to give deep insight on the
initial stage of stacking process. By using photolumines-
cence ͑PL͒ spectroscopy, we demonstrate that the critical
coverage of the island formation and the wetting layer thick-
ness are reduced in the second Ge layer. These changes are
shown to be accompanied by morphological changes such as
a drastic increase of the island density and a shape transition.
The samples were grown by gas-source molecular beam
epitaxy ͑MBE͒ ͑Daido-Hoxan VCE S2020͒ using disilane
(Si2H6) and germane (GeH4) as source gases at fixed growth
temperature of 700 °C. The substrates were p-type Si͑100͒
with resistivity of 10–20 ⍀ cm. The samples contain double
Ge layers separated by a 144 Å Si spacer layer. The Ge
FIG. 1. PL spectra of a series of samples where the Ge coverage of the
second layer was systematically changed while that of the first layer and the
Si spacer thickness were fixed at 4.5 ML and 144 Å, respectively.
a͒
Present address: IMR, Tohoku University, 2-1-1 Katahira, Aoba-ku,
Sendai 980-8577, Japan; electronic mail: usa@imr.jp
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2000/76(25)/3723/3/$17.00 3723 © 2000 American Institute of Physics
158.121.247.60 On: Sat, 08 Nov 2014 02:53:26

3724
Appl. Phys. Lett., Vol. 76, No. 25, 19 June 2000
Usami et al.
FIG. 3. 2 ␮mϫ2 ␮m AFM images of ͑a͒ the first Ge layer and ͑b͒ the
second Ge layer separated by a 144 Å Si layer. It is noted that the Ge
coverage is 4.5 ML for the both samples.
FIG. 2. PL peak energies from the wetting layer ͑solid circles͒, NP2, as a
function of the Ge coverage of the second layer. For comparison, the NP
energies of a single Ge layer ͑open circiles͒ in Si are also plotted as a
function of the Ge coverage.
ness in a five-fold stack of 6.5 ML of Ge. On the other hand,
this energy relationship is seen to be broken at lower and
higher Ge coverages. It should be noted that NP2 does not
reach the band gap of unstrained Si even if the Ge coverage
is extrapolated to 0. This would be explained by the reduc-
tion of the band gap of the Si spacer layer due to the strain.
In other words, this supports that the modification of the
growth mode of the second Ge layer is brought by the strain
induced by the buried Ge islands.
As shown in Fig. 1, PL from the islands, L1 and L2 , can
be clearly resolved even if the Ge coverage of each layer is
same. AFM observations revealed that the PL feature from
the islands is closely related to morphological changes. Fig-
ure 3 shows 2 ␮mϫ2 ␮m AFM images of ͑a͒ the first Ge
layer and ͑b͒ the second Ge layer separated by a 144 Å Si
layer. It is seen that the density of the islands in the second
layer is drastically increased and about four times larger than
that in the first layer. As a result, the averaged volume of the
island in the second layer is much smaller than that in the
first layer. This would be the reason for the blueshift of L2
compared to L1 and the dominance of ‘‘pyramids’’ which
are known to be stable phases for smaller volume.10
cally changed measured at around 25 K. The PL lines at
around 1130 and 1095 meV originate from the Si substrates.
At lower energies of the Si-related peaks, several additional
lines can be identified. Based on our previous publication
where we investigated PL from Ge embedded in Si,1 we
assigned these peaks as no phonon ͑NP͒ and transverse op-
tical ͑TO͒ phonon lines from the wetting layer and PL from
the Ge islands ͑L͒. PL from the wetting layer is clearly seen
to consist of two ͑NP,TO͒ pairs. The pair at lower energies
labeled NP1 and TO1 is considered to come from the first Ge
layer since they show no peak shift with changing Ge cov-
erage of the second layer. This confirms that the electronic
coupling between the wetting layers of double Ge is negli-
gible, showing that PL spectral feature related with the wet-
ting layer of the second Ge layer reflects the modification of
the growth mode. PL from the Ge islands ͑L͒ are also di-
vided into two components; L1 and L2 . A striking feature is
that L2 , which seems to be related with the second Ge layer,
can be clearly identified at 1.8 ML. This amount is much
smaller than the critical coverage of the island formation of a
single Ge layer, showing that the critical coverage of the
island formation is drastically decreased in the presence of
the buried Ge islands.
In Fig. 2, the PL peak energies from the second wetting
layer, NP2, are plotted against the Ge coverage of the second
layer. For comparison, the NP energies of a single Ge layer
in Si are also plotted as a function of the Ge coverage. The
energy dip at 3.7 ML and following blueshift was previously
interpreted as evidence for the island formation.1,2 That is,
after the nucleation of the islands, Ge atoms at the topmost
layer are considered to be incorporated to the islands, result-
ing in the reduction of the wetting layer thickness. This
mechanism is not directly applied to the stacked Ge islands
since NP2 continuously shifts to lower energies after the ap-
pearance of L2 at 1.8 ML. This suggests that a part of addi-
tional Ge atoms contribute to the increase of the wetting
layer thickness even after the nucleation of the islands takes
place.
It is noted that the total volume of the islands in the
second layer estimated from Fig. 3 showed only 8% increase
compared with that in the first layer, which could be com-
pensated with the decrease of the wetting layer thickness.
Therefore, the growth rate of Ge seems to be not greatly
affected by the presence of the buried islands. The small
amount of excess Ge atoms for the islands formation is un-
like to bring the observed drastic increase of the island den-
sity. The observed growth mode could be explained in terms
of the surface strain induced by the buried Ge islands. It is
well known that the Si lattice on top of the islands becomes
an energetically preferable site for the successive Ge growth
due to smaller lattice mismatch, resulting in the vertical or-
dering of the islands.11 In fact, the vertical ordering of
‘‘domes’’ were observed in our samples by cross-sectional
transmission electron microscopy. The additional appearance
of pyramids would be due to the interaction of the strain
induced by neighboring islands, leading to the formation of
the preferable sites not only on top of the islands but also
between the islands.7 Another possibility is the reduction of
the nucleation barrier in the second layer since Ge atoms
segregating to the growth front is predicted to drastically
At an intermediate Ge coverage, NP2 lies at higher en-
ergies than NP of the single Ge layer, showing that the wet-
ting layer thickness of the second layer is reduced. This re-
sult is consistent with the recent report of Schmidt and Eberl9
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
where they clarified the reduction of the wetting layer thick-
reduce it due to the contribution of the free energy of entropy
158.121.247.60 On: Sat, 08 Nov 2014 02:53:26

Appl. Phys. Lett., Vol. 76, No. 25, 19 June 2000
Usami et al.
3725
mixing.12 The interface mixing due to the Ge surface segre-
gation is less probable since gas-source MBE was used for
the sample growth where is suppressed by the hydrogen.13
However, an alloying during the capping process14 must be
considered since the averaged height of domes is around 200
Å, which is even larger than that of the spacer thickness of
144 Å. In fact, a preliminary experiment showed that to cap
Ge islands with a thin Si layer brings enhanced nucleation.15
Further study to clarify the mechanism of the growth mode
modification is in progress.
In summary, we investigated the growth mode of Ge on
Si͑100͒ with buried Ge islands by using PL spectroscopy and
AFM. The critical coverage of the island formation and the
wetting layer thickness were found to be reduced in the pres-
ence of the buried Ge islands. The morphological changes
such as a drastic increase of the island density and a shape
transition were also observed.
the Promotion of Science ͑JSPS͒, and the Core Research for
Evolutional Science and Technology ͑CREST͒ of the Japan
Science and Technology Corporation ͑JST͒.
1 H. Sunamura, N. Usami, Y. Shiraki, and S. Fukatsu, Appl. Phys. Lett. 66,
3024 ͑1995͒.
2
¨
P. Schittenhelm, M. Gail, J. Brunner, J. F. Nutzel, and G. Abstreiter, Appl.
Phys. Lett. 67, 1292 ͑1995͒.
3 G. Abstreiter, P. Schittenhelm, C. Engel, E. Silveira, A. Zrenner, D.
¨
Meertens, and W. Jager, Semicond. Sci. Technol. 11, 1521 ͑1996͒.
4 C. S. Peng, Q. Huang, W. Q. Cheng, J. M. Zhou, Y. H. Zhang, T. T.
Scheng, and C. H. Tung, Phys. Rev. B 57, 8805 ͑1998͒.
5 P. Boucaud, V. Le Thanh, S. Sauvage, D. Debarre, and D. Bouchier, Appl.
Phys. Lett. 74, 401 ͑1999͒.
6 L. P. Rokhinson, D. C. Tsui, J. L. Benton, and Y. H. Xie, Appl. Phys. Lett.
75, 2413 ͑1999͒.
7 J. Tersoff, C. Teichert, and M. G. Lagally, Phys. Rev. Lett. 76, 1675
͑1996͒.
8 O. Schmidt, C. Lange, and K. Eberl, Appl. Phys. Lett. 75, 1905 ͑1999͒.
9 O. G. Schmidt and K. Eberl, Appl. Phys. Lett. 74, 1272 ͑1999͒.
10 F. M. Ross, J. Tersoff, and R. M. Tromp, Phys. Rev. Lett. 80, 984 ͑1998͒.
11 Q. Xie, A. Madhukar, P. Chen, and N. P. Kobayashi, Phys. Rev. Lett. 75,
2542 ͑1995͒.
The authors would like to acknowledge K. Kawaguchi
and H. Takamiya for their cooperation in MBE growth, and
S. Ohtake for his technical assistance. This work was in part
supported by a Grant-in-Aid for Scientific Research from the
Ministry of Education, Science, Sports, and Culture, Interna-
tional Priority Collaboration Program of Japan Society for
12 J. Tersoff, Phys. Rev. Lett. 81, 3183 ͑1998͒.
13 S. Fukatsu, H. Yoshida, N. Usami, A. Fujiwara, Y. Takahashi, Y. Shiraki,
and R. Ito, Thin Solid Films 222, 1 ͑1992͒.
14
´
J. M. Garcıa, G. Medeiros-Ribeiro, K. Schmidt, T. Ngo, J. L. Feng, A.
Lorke, J. Kotthaus, and P. M. Petroff, Appl. Phys. Lett. 71, 2014 ͑1997͒.
15 N. Usami, M. Miura, and Y. Shiraki ͑unpublished͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.121.247.60 On: Sat, 08 Nov 2014 02:53:26