Optical gain from the direct gap transition of Ge-on-Si at
room temperature
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Jifeng Liu et al. “Optical gain from the direct gap transition of Geon-Si at room temperature.” Group IV Photonics, 2009. GFP '09.
6th IEEE International Conference on. 2009. 262-264. ©
Copyright 2010 IEEE
As Published
http://dx.doi.org/10.1109/GROUP4.2009.5338364
Publisher
Version
Final published version
Accessed
Thu May 26 08:46:24 EDT 2016
Citable Link
http://hdl.handle.net/1721.1/59838
Terms of Use
Article is made available in accordance with the publisher's policy
and may be subject to US copyright law. Please refer to the
publisher's site for terms of use.
Detailed Terms
FD2
4:30 PM – 4:45 PM
Optical Gain from the Direct Gap Transition of Ge-on-Si at Room Temperature
Jifeng Liu, Xiaochen Sun, Lionel C. Kimerling, and Jurgen Michel
Microphotonics Center, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA
Abstract
We report direct band gap optical gain of tensile strained
n+ epitaxial Ge-on-Si at room temperature, which confirms
that band-engineered Ge-on-Si is a promising gain medium
for monolithic optical amplifiers and lasers on Si.
1. Introduction
Lasers on silicon are one of the most crucial components
for silicon-based electronic-photonic integration [1,2].
Epitaxial Ge-on-Si is a particularly interesting candidate
due to its pseudo-direct band gap behavior [3] and its
compatibility with advanced electronic devices on Si [4].
Integrated photonic devices such as waveguide-coupled
photodetectors [5-7] and electro-absorption modulators [8,9]
have already been demonstrated based on the direct band
gap transition of Ge. If a high performance Ge-on-Si light
source exists, all active photonic devices on Si can be
realized using Ge, which greatly simplifies the monolithic
electronic-photonic integration process. Our theoretical
analysis has shown that Ge can be band-engineered by
tensile strain and n-type doping to achieve efficient light
emission and optical gain from its direct gap transition [10].
Indeed, direct gap photoluminescence (PL) [11,12] and
electroluminescence (EL) [13] at room temperature have
already been demonstrated from these band engineered Geon-Si materials. Here we report experimental observation of
optical gain in epitaxial tensile strained n+ Ge-on-Si at room
temperature. The optical gain has been observed in the
wavelength range of 1600-1608 nm near the direct band
gap of band-engineered Ge mesas on Si, and a maximum
gain coefficient of Ȗ=56±25cm-1 was observed at 1605 nm.
This gain coefficient is much greater than the waveguide
loss in Si photonics even with the most conservative
estimate and can be applied to optical gain devices on-chip.
These results demonstrate the potential of band-engineered
Ge as an optical gain medium for monolithically integrated
lasers and optical amplifiers on Si.
2. Experiment
Ge mesas with in-situ phosphorous doping were
selectively grown on Si by ultra-high vacuum chemical
vapor deposition (UHVCVD) using a SiO2 mask layer. A
50 nm Ge buffer layer was directly grown on Si at a low
temperature of 360°C to kinetically suppress islanding, then
the growth temperature was ramped up to 650°C to grow
the rest of the Ge layer. Details about this two-step growth
method were reported earlier [14]. The samples were
annealed by rapid thermal annealing at 780°C for 30 sec to
reduce the defect density. The Ge layer was nearly fully
relaxed at the growth and annealing temperatures, and
tensile strain was accumulated upon cooling to room
temperature due to the large thermal expansion coefficient
difference between Ge and Si [15, 16]. The thermally
induced tensile strain in the Ge layer is 0.23%, and the
corresponding direct band gaps from the maxima of light
(lh) and heavy (hh) hole bands to the minimum of the ī
valley are EgΓ (lh) =0.767 eV and E gΓ (hh) =0.782 eV,
respectively [16]. There are two major advantages in using
selectively grown Ge mesas to investigate optical gain
compared to blanket Ge films on Si: (1) the threading
dislocation density is lower than blanket Ge films [14] so
that non-radiative recombination is reduced; (2) The SiO2
mask layer for the selective growth naturally provides
carrier confinement in the lateral direction to enhance the
injected carrier concentration in the selectively grown Ge
mesa upon optical pumping.
Fig. 1 Schematic drawing of the pump-probe measurement setup. A lockin method is used to detect the transmitted probe laser signal only.
To investigate the optical bleaching and gain of n+
tensile-strained Ge-on-Si, a pump-probe measurement was
performed using the setup schematically shown in the inset
of Fig. 1. A 1480 nm continuous wave (CW) pump laser
and a tunable probe laser with an output wavelength range
of 1510-1640 nm were coupled into a single mode lenstipped optical fiber through a wavelength division
multiplexing (WDM) coupler. The light was incident on the
front surface of the sample. A selectively grown Ge mesa
with an area of 500 μm2 and an active doping level of
n=1.0×1019 cm-3 was used in this study. The active doping
concentration was determined by Hall Effect measurements.
The Ge film grown on the backside of the Si wafer during
UHVCVD process was removed to simplify data analysis.
The Ge layer on the front side was 870 nm thick. An
InGaAs detector with an integration sphere was placed at
the backside of the sample to collect the transmitted signal.
In order to accurately measure the transmittance of the
probe laser through the n+ Ge-on-Si mesa, the probe laser
was internally modulated at 500 Hz and a lock-in approach
was adopted to record the transmitted probe laser signal
only. This method achieves ±0.25% accuracy in the
transmittance measurement of the probe laser, and to be
more conservative we adopted an accuracy of ±0.4% to
reflect the upper limit of the transmittance measurement
error in the data analysis. Compared to the commonly used
1
978-1-4244-4403-8/09/$25.00 ©2009 IEEE
262
pulse laser pumping in pump-probe measurements that
emphasizes transient behavior, the optical bleaching
measured under CW pumping in this case reflects the
properties of the band-engineered Ge-on-Si sample at a
steady-state injection level, which is more relevant to the
real operation of laser devices.
change in transmittance below the direct band gap is
dominated by the increase in free carrier absorption.
3. Results and Discussions
Fig. 3 Comparison of (a) absorption spectra, and (b) refractive index
spectra of bulk Ge with the tensile strained n+ Ge-on-Si mesa sample.
Fig. 2 Transmittance spectra of a 500 μm2 Ge-on-Si mesa sample with
n=1.0×1019 cm-3 under 0 and 100 mW optical pumping.
The transmittance spectra of the sample under 0 and 100
mW optical pumping are shown in Fig. 2. The effective
pump power density at 100 mW is estimated to be ~7.0
kW/cm2 considering coupling loss at the fiber connectors
and the effective absorption of the pump laser by the thin
Ge layer. Without optical pumping, the transmittance starts
to decrease dramatically at photon energies >0.767 eV due
to the onset of the direct gap absorption ( E gΓ (lh) =0.767 eV).
Upon optical pumping, the transmittance of the Ge film is
expected to change due to two reasons. On one hand, the
direct band gap absorption is decreased due to the band
filling by optically injected carriers and the start of
population inversion. On the other hand, the optically
injected free carriers in the conduction and valence bands
increase the free carrier absorption and the transmittance
tends to decrease. Overall, the change in transmittance upon
optical pumping at photon energies greater than the direct
band gap reflects the competition between the decrease in
the direct band gap absorption and the increase in the free
carrier absorption. For photon energies below the direct
band gap, the free carrier absorption is expected to
dominate upon optical pumping due to the lack of bleaching
mechanism from the direct gap transition. Experimentally,
we observe that the transmittance at photon energies >0.767
eV increases while the transmittance at photon energies
<0.767 eV slightly decreases upon optical pumping (see Fig.
5b). Since 0.767 eV corresponds to the onset of the direct
gap transition between the light hole valence band and the
conduction band at the ī valley, the experimental results
indicate that the change in transmittance at photon energies
above the direct band gap is dominated by the decrease in
the direct band gap absorption due to band filling, while the
To derive the absorption spectra of the sample under 0
and 100 mW optical pumping from the transmittance data
in Fig. 2, we use transfer matrix method [17] and KramersKronig relation to solve both the real refractive index (nr)
and the absorption coefficient (Į) deterministically using an
iterative self-consistent regression approach. We substitute
the refractive index nr(Ȝ) of bulk Ge [3] into the transfer
matrix equation as a starting point to solve Į(Ȝ) from the
transmittance data, then we use the newly derived Į(Ȝ) to
obtain a corrected nr(Ȝ) through Kramers-Kronig relation.
This process is iterated back and forth until self-consistency
is obtained for both the transfer matrix equation of
transmittance and the Kramers-Kronig equation. As an
example, the derived absorption spectrum and refractive
index spectrum of the tensile strained n+ Ge mesa on Si
without optical pumping are compared with those of bulk
Ge in Fig. 3. The absorption edge is significantly red
shifted due to the band gap shrinkage induced by the tensile
strain [15,16]. As a result, the refractive index is also
notably modified compared to bulk Ge due to the KramersKronig relation. Such a self-consistency approach
guarantees a more accurate solution to the absorption
coefficients than simply assuming bulk Ge refractive index
for tensile strained Ge in the transfer matrix calculation. In
fact, the absorption spectrum derived here is in good
agreement with a recent report on the absorption of tensile
strained Ge-on-Si films [18].
Figure 4 shows the derived absorption spectra of the n+
Ge mesa sample under 0 and 100 mW optical pumping. The
absorption coefficients at photon energies >0.767 eV
(Ȝ<1617 nm) decreases significantly upon optical pumping.
Especially, negative absorption coefficients corresponding
to optical gain are observed in the wavelength range of
1600-1608 nm, as shown in the inset of Fig. 4. The shape of
the gain spectrum near the direct band edge of Ge
resembles those of III-V semiconductor materials [17]. The
maximum gain coefficient observed is Ȗ=-Į=56±25cm-1 at
1605 nm. The error bars given here reflect the upper limit
of the transmittance measurement error of ±0.4%. Even
2
978-1-4244-4403-8/09/$25.00 ©2009 IEEE
263
with the most conservative estimate, a gain coefficient of
>25 cm-1 at 1605 nm, which is equivalent to >100 dB cm-1,
is guaranteed. This gain coefficient is much greater than the
waveguide loss in Si photonics (typically<10 dB cm-1) and
can be applied to optical gain devices on-chip. Further
improvement in gain coefficients is expected with higher ntype doping level and is supported by the trend in PL
intensity enhancement [17, 18].
Fig. 4 Absorption spectra of the n+ Ge mesa sample under 0 and 100 mW
optical pumping. Negative absorption coefficients corresponding to optical
gain are observed in the wavelength range of 1600-1608 nm, as shown in
the inset. The error bars reflect the upper limit of the transmittance
measurement error of ±0.4%.
[5] D. Ahn, C. Y. Hong, J. F. Liu, M. Beals, W. Giziewicz, L. C.
Kimerling and J. Michel, Opt. Express 15, 3916-3921 (2007).
[6] G. Masini, G. Capellini, J. Witzens and C. Gunn, 4th IEEE
International Conference on Group IV Photonics (IEEE Cat. No.
07EX1662C), Tokyo, Japan, 19-21 Sept. 2007, pp. 28-30.
[7] T. Yin, R. Cohen, M. Morse, G. Sarid, Y. Chetrit, D. Rubin, and M. J.
Paniccia, Opt. Express 15, 13965-13971 (2007)
[8] J. E. Roth, O. Fidaner, R. K. Schaevitz, Y. H. Kuo, T. I. Kamins, J. S.
Harris, and D. A. B. Miller Opt. Express. 15, 5851-5859 (2007)
[9] J. F. Liu, M. Beals, A. Pomerene, S. Bernardis, R. Sun, J. Cheng, L. C.
Kimerling and J. Michel, Nature Photonics. 2, 433 (2008).
[10] J. F. Liu, X. C. Sun, D. Pan, X. X. Wang, L. C. Kimerling, J. Michel
and T. L. Koch, Optics Express 15, 11272-11277 (2007)
[11] J. F. Liu, X. C. Sun, P. Becla, L. C. Kimerling and J. Michel, in 5th
IEEE International Conference on Group IV Photonic (Sept. 2008,
Sorrento, Italy), IEEE Catalog Number: CFP08GFP-CDR, pp. 16-18
[12] X. C. Sun, J. F. Liu, L. C. Kimerling, J. Michel and T. L. Koch, in
2008 OSA Conference on Integrated Photonics and Nanophotonics
research and Application (Optical Society of America, 2008), paper IMC5
[13] X. C. Sun, J. F. Liu, L. C. Kimerling, and J. Michel, Opt. Lett. 34,
1198 (2009).
[14] H. Luan, D. R. Lim, K. K. Lee, K. M. Chen, J. G. Sandland, K. Wada,
and L. C. Kimerling, Appl. Phys. Lett. 75, 2909-2911 (1999)
[15] Y. Ishikawa, K. Wada, D. D. Cannon, J. F. Liu, H. C. Luan and L. C.
Kimerling, Appl. Phys. Lett. 82, 2044-2046 (2003).
[16] J. F. Liu, D. D. Cannon, K. Wada, Y. Ishikawa, D. T. Danielson, S.
Jongthammanurak, J. Michel and L. C. Kimerling, Phys. Rev B 70,
155309 (2004)
[17] S. L. Chuang, Physics of optoelectronic devices, (Wiley, New York,
1995).
[18] V. Sorianello, A. Perna, L. Colace, G. Assanto, H. C. Luan, and L. C.
Kimerling, Appl. Phys. Lett. 93, 111115 (2008).
4. Conclusion
In conclusion, we report the first observation of optical
gain in epitaxial Ge-on-Si at room temperature by using
tensile strain and n-type doping for band engineering. The
optical gain has been observed in the wavelength range of
1600-1608 nm near the direct band gap of band-engineered
Ge mesa on Si, and a maximum gain coefficient of
-1
γ = 56 ± 25 cm was observed at 1605 nm. This gain
coefficient is much greater than the waveguide loss in Si
photonics even with the most conservative estimate and can
be applied to optical gain devices on-chip. These results
demonstrate the potential of band-engineered Ge as an
optical gain medium for monolithically integrated lasers on
Si.
Acknowledgements
This work is supported by the Si-based Laser Initiative of
the Multidisciplinary University Research Initiative (MURI)
sponsored by the Air Force Office of Scientific Research
(AFOSR) and supervised by Dr. Gernot Pomrenke.
References
[1] D. J. Lockwood and L. Pavesi, Silicon Photonics (Springer-Verlag,
Berlin, 2004), pp 1-50.
[2] L. C. Kimerling, D. Ahn, A. B. Apsel, M. Beals, D. Carothers, Y-K.
Chen, T. Conway, D. M. Gill, M. Grove, C-Y Hong, M. Lipson, J. Liu, J.
Michel, D. Pan, S. S. Patel, A. T. Pomerene, M. Rasras, D. K. Sparacin, KY. Tu, A. E. White, and C. W. Wong, Proc. SPIE 6125, 612502 (2006).
[3] Physics of Group IV Elements and III–V Compounds, LandoltBörnstein: Numerical Data and Functional Relationships in Science and
Technology, edited by O. Madelung, (Springer, Berlin, 1982), vol. 17a.
[4] M. J. Lee and E. A. Fitzgerald, J. Appl. Phys. 97, 011101 (2005).
3
978-1-4244-4403-8/09/$25.00 ©2009 IEEE
264