RESEARCH ARTICLE

www.lpr-journal.org

Monolithically Integrated Ultra-Low Threshold

GeSn-on-Insulator Laser Using Rapid Melting Growth
Zhi Liu, Melvina Chen, Xiangquan Liu, Qinxin Huang, Yupeng Zhu, Yuhua Zuo,
Zoran Ikonic, Buwen Cheng, Donguk Nam,* and Jun Zheng*

including photonics-based artifi-

A low-threshold, monolithically integrated laser on Si is considered a crucial cial intelligence[1] and quantum
missing ingredient in realizing efficient fully functional photonic-integrated computing.[2] Despite the technolog-
circuits (PICs). Owing to its compatibility with complementary ical maturity of most Si photonics
metal-semiconductor-oxide (CMOS) processes, direct bandgap GeSn alloy components,[3] the development of scal-
able and practical Si photonics lasers
has recently been studied intensively in hopes of making GeSn lasers the remains elusive owing to the funda-
mainstream technology for PICs. However, the inevitable formation of mental issue of the indirect bandgap
harmful defects in GeSn directly grown on Si has thus far required the use of nature of group IV materials. Hetero-
non-monolithic approaches such as wafer bonding to obtain high-quality geneous integration of direct bandgap
GeSn layers, preventing the realization of practical, low-threshold GeSn III-V materials and lasers onto a pre-
fabricated Si photonics chip has been
lasers. Here, ultra-low threshold lasing in a monolithically-grown,
extensively used to provide photons to
nearly-defect-free GeSn single-crystal layer is demonstrated. The rapid melting the chip,[4] but this method hinders
growth method used in this study allows the fabrication of a compact, the realization of a fully monolithic,
integrated laser that simultaneously achieves an ideal GeSn gain medium laser-integrated Si photonics platform.
with built-in tensile strain and an excellent GeSn optical cavity on an Among the potential routes to-
insulating layer. The measured threshold is ≈0.52 kW cm−2 under the optical wards monolithic Si photonics lasers,
GeSn alloys have emerged as one
pumping scheme at 10 K, which is the lowest among all the reported GeSn of the most promising candidates
lasers. This work provides a new solution for building a truly for CMOS-compatible laser.[5] Since
CMOS-compatible, monolithic laser that can complete the device library of the first successful demonstration
the mature Si photonics foundries. of GeSn lasers on Si,[6] there have
been relentless efforts in improv-
ing the laser performance.[7–9] Early
theoretical studies predicted that the
1. Introduction laser threshold can be reduced by increasing the Sn content
A monolithically integrated Si photonics platform has recently and thereby the directness of the material,[10,11] motivating re-
become a cost-effective, scalable solution for various applications searchers to experimentally achieve ever-higher Sn contents.

Z. Liu, X. Liu, Q. Huang, Y. Zhu, Y. Zuo, B. Cheng, J. Zheng M. Chen

Key Laboratory of Optoelectronic Materials and Devices Global Device Engineering
Institute of Semiconductors GlobalFoundries Singapore Pte. Ltd
Chinese Academy of Sciences 60 Woodlands Industrial Park D Street 2, Singapore 738406, Singapore
Beijing 100083, China
Z. Ikonic
E-mail: [email protected]
School of Electronic and Electrical Engineering
Z. Liu, X. Liu, Q. Huang, Y. Zhu, Y. Zuo, B. Cheng, J. Zheng University of Leeds
Center of Materials Science and Optoelectronics Engineering Leeds LS2 9JT, UK
University of Chinese Academy of Sciences
D. Nam
Beijing 100049, China
Department of Mechanical Engineering
M. Chen KAIST
School of Electrical and Electronic Engineering 291 Daehak-ro, Yuseong-gu, Daejeon 34141, South Korea
Nanyang Technological University E-mail: [email protected]
50 Nanyang Avenue, Singapore 639798, Singapore

The ORCID identification number(s) for the author(s) of this article

can be found under https://doi.org/10.1002/lpor.202401077
DOI: 10.1002/lpor.202401077

Laser Photonics Rev. 2024, 2401077 2401077 (1 of 6) © 2024 Wiley-VCH GmbH

 18638899, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lpor.202401077 by Korea Advanced Institute Of, Wiley Online Library on [27/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.lpr-journal.org

Contrary to the prediction, however, the threshold was steadily sation of GeSn during the following processes (top two panels).
increased at higher Sn contents,[11] possibly due to the increased Subsequently, rapid thermal annealing (RTA) is performed for
defect density in GeSn with higher Sn contents[12] —another im- 1 s at 920 °C to melt the amorphous GeSn strip. During the
portant consideration for laser threshold.[13] cooling down, liquified GeSn undergoes lateral liquid phase epi-
Recently, following another theoretical prediction,[14] a re- taxy with the single crystal Si seed, forming single crystal GeSn
search team utilized tensile strain to achieve direct bandgap in with Sn segregation at the end of the strip (third-from-the-top
GeSn with only 5.4 at.% Sn and reported the threshold value as panel). The necking effect, commonly occurring in nanoscale se-
low as 0.8 kW cm−2 .[11] Tensile strain complements the Sn alloy- lective epitaxial growth, forces the threading dislocation to termi-
ing for achieving an improved directness,[15] removing the ne- nate at sidewalls, allowing the GeSnOI strip to be nearly defect-
cessity of increasing the Sn content. Also, strain-induced valence free. Moreover, the thermal expansion mismatch between GeSn
band splitting lowers the density of states and the carrier den- and the Si substrate leads to the accumulation of ≈0.15% tensile
sity required for achieving population inversion,[14] enabling a strain in GeSn, making an ideal GeSnOI gain medium. More
reduced laser threshold. However, due to the harmful residual details about the RMG method can be found in our previous
compressive strain in GeSn grown on the Ge buffer layer, achiev- work[16,17] and in Experimental section. Lastly, FIB is used to form
ing tensile strain requires complicated fabrication processes such the two end facets (bottom panel).
as wafer bonding,[11] preventing monolithic integration of Si pho-
tonics lasers.
In this work, by harnessing the rapid melting growth (RMG) 2.2. Material Characterization
method that is a fully monolithic approach, we demonstrate ten-
sile strained GeSn-on-insulator (GeSnOI) lasers operating at an Cross-sectional transmission electron microscopy (TEM) was
ultra-low threshold. The RMG method used in this study al- employed to study the crystal quality, structure, and Sn distribu-
lows obtaining a nearly-defect-free, tensile strained GeSn gain tion of the sample prepared by the RMG method. Figure 2a shows
medium with ≈9.6 at.% Sn on an Si3 N4 insulating layer without a typical TEM image near the interface between the single crystal
requiring any complex fabrication procedures. The formation of GeSn and the Sn segregation areas. The entire length of the strip
two end facets in GeSn enabled by focused ion beam (FIB) results is ≈120 μm. Figure 2b shows a selected area electron diffraction
in a high-quality Fabry–Pérot (F–P) cavity on insulator, providing (SAED) image of the single crystal GeSn region, which is ≈4 μm
strong optical confinement. With an optical pumping at 10 K, the away from the Sn segregation area. The orientation of the single
GeSnOI laser achieves a laser threshold of 0.52 kW cm−2 , which crystal GeSn is determined to be (001), which is consistent with
is the lowest value among all of the demonstrated GeSn lasers. the (001) Si substrate, confirming that the GeSnOI strip is crystal-
Rigorous theoretical calculations show a significant net gain im- ized from the Si seed. Based on the distance between the diffrac-
provement in our RMG-enabled GeSn gain medium compared to tion spots in the SAED image, the lattice constants of the single
a compressive strained, epitaxially grown GeSn counterpart. This crystal GeSn along the <002> and <220> directions are calcu-
work marks a major milestone in the development of monolith- lated as 5.752 ± 0.005 Å and 5.761 ± 0.005 Å, respectively, corre-
ically integrated Si photonics lasers toward realizing fully func- sponding to the Sn-content of 9.7 ± 0.5%. The lattice constant
tional Si photonics platforms. difference between the <002> and <220> directions suggests
that the GeSn strip is under the tensile strain of ≈0.15%. A sharp
2. Results GeSn/SiO2 interface and a perfect atomic arrangement of GeSn
shown in the high-resolution TEM further provide evidence of
2.1. Design of Monolithic GeSnOI Lasers the excellent quality of the RMG-based GeSn layer (Figure 2c).
The compositional profile along the strip is estimated from the
Figure 1a shows the schematic illustration of the F–P GeSnOI energy dispersive X-ray spectroscopy (EDX) under the TEM mode
laser enabled by the RMG method. The F–P laser is pumped op- (Figure 2d). The Sn content in the single crystal GeSn near the Sn
tically by a 1550-nm pulsed laser. The photons emitted from the segregation region is ≈20%. Due to the destructive sampling and
GeSn gain medium are confined between the two end facets of limited measurement range of TEM, the Sn content along the en-
the strip, allowing a small fraction of the scattered emission to be tire 10-μm GeSn gain medium was estimated nondestructively by
measured perpendicularly above the sample. Figure 1b presents micro-Raman scattering. The Sn content is verified to gradually
the scanning electron microscopy (SEM) image of the GeSnOI vary from ≈12.6% to ≈7.4% (see Note S2, Supporting Informa-
lasers with two end facets formed by FIB (see Experimental sec- tion for Raman measurement). Supporting The average Sn con-
tion for more details on the device fabrication). The length of the tent in the GeSnOI laser is estimated as 9.6%. It is noteworthy
F–P cavity is ≈10 μm. Figure 1c shows the top view (left) and the that this Sn content, combined with ≈0.15% tensile strain, is suf-
cross-sectional view (right) of the simulated electric field distribu- ficient in achieving direct bandgap GeSn.[15]
tions of the fundamental TEM00 mode in the laser cavity. A sim-
ulated optical quality factor was ≈1000 (See Note S1, Supporting
Information for more details on Optical cavity design). 2.3. Lasing Characteristics in RMG-Based GeSnOI Lasers at 10 K
Figure 1d shows the growth and fabrication processes of the
RMG-based GeSnOI laser. First, an amorphous GeSn strip on We conducted photoluminescence (PL) measurements to char-
a Si3 N4 layer is prepared with one end of the strip in contact acterize the lasing behaviors of RMG-based GeSnOI lasers (see
with the Si seed layer. The entire structure is then covered with Experimental Section for more details on the PL measurement
a SiO2 insulating layer to prevent the evaporation and conden- setup and See Note S3 (Supporting Information) for more details

Laser Photonics Rev. 2024, 2401077 2401077 (2 of 6) © 2024 Wiley-VCH GmbH

 18638899, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lpor.202401077 by Korea Advanced Institute Of, Wiley Online Library on [27/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.lpr-journal.org

Figure 1. Design of monolithic GeSnOI lasers. a) Schematic illustration of a GeSnOI laser. Under the 1550 nm pump laser illumination, laser emission
is scattered and collected from both facets of the GeSnOI Fabry–Pérot (F–P) cavity. b) Scanning electron microscope (SEM) image of the GeSnOI lasers.
The two end facets of the lasers are formed by focused ion beam (FIB). Scale bar, 5 μm. c) Electric field simulation of the top (left) and cross-sectional
view (right) of fundamental TEM00 mode in the laser cavity. Scale bar, 1 μm (left) and 200 nm (right). d) Schematics of the growth and fabrication
processes of GeSnOI laser by the rapid melting growth (RMG). One end of amorphous GeSn strip on Si3 N4 is in contact with Si seed window. SiO2 is
used to cover the entire structure to prevent evaporation and condensation in the subsequent processes. Upon rapid thermal annealing (RTA) of 920 °C
for 1 s, GeSn is melted and on cooling down, GeSn undergoes lateral liquid phase epitaxy using Si seed to form single crystal GeSn and Sn segregation
at the end of the strip. Threading dislocations are terminated at the sidewalls due to necking effect and nearly defect-free single crystal GeSn strip is
obtained. Lastly, FIB is used to form the end facets of the cavity.

on PL spectra without F–P cavities). Laser devices with different sion. The lasing peak increases sharply in intensity at increased
cavity lengths show clear single-mode lasing, confirming the ef- pump power densities. The emission from the laser device is
fectiveness of FIB in obtaining smooth facets for an F–P optical highly polarized along the width of the F–P cavity, providing fur-
cavity. (See Note S4, Supporting Information) Figure 3a shows ther evidence of lasing (See Note S5, Supporting Information for
the emission spectra of a GeSnOI laser with increasing pump more details on emission polarization measurement). The inset
power densities. The sample temperature is fixed at 10 K. At a low to Figure 3a shows the light-in-light-out (L-L) curve for the inte-
pumping power density of 0.3 kW cm−2 (green), only broad spon- grated output intensity, presenting a clear threshold behavior. A
taneous emission was observed. As the pumping power density double-logarithmic plot of the same data with an S-shape curve
is increased to 1.0 kW cm−2 (blue), a strong lasing peak emerges signifies clear evidence of lasing (See Note S6, Figure S7, Sup-
at ≈2200 nm, dominating the background spontaneous emis- porting Information). The full-width at half-maximum (FWHM)

Laser Photonics Rev. 2024, 2401077 2401077 (3 of 6) © 2024 Wiley-VCH GmbH

 18638899, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lpor.202401077 by Korea Advanced Institute Of, Wiley Online Library on [27/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.lpr-journal.org

Figure 2. Material characterization. a) Cross-sectional transmission electron microscope (TEM) image near GeSn single crystal and Sn segregation
interface. Scale bar, 500 nm. b) Selected area electron diffraction (SAED) diffraction pattern of GeSn single crystal. The pattern is taken at 4 μm away
from the Sn segregation area. <002> and <220> diffraction directions are indicated. Scale bar, 10 nm−1 . c) High-resolution TEM image of sharp
GeSn/SiO2 interface and perfect GeSn atomic arrangement further showing the excellent quality of RMG-based GeSn layer. Scale bar, 10 nm. d) Energy
dispersive X-ray spectroscopy (EDX) compositional profile along the strip near GeSn/Sn segregation interface.

of the emission peak is presented in the same inset. Near the pressive strain (typical strain of epitaxial GeSn layer grown on Ge
threshold, the sudden reduction in the FWHM occurs due to the buffer)[1] at an injection carrier density of 5 × 1016 cm−3 , which
abrupt transition from broad spontaneous emission to sharp las- corresponds to 6.5 kW cm−2 pump power density. The net gain is
ing peak. From the L-L curve, the lasing threshold is extracted as the resulting gain of an interplay between direct interband transi-
0.52 kW cm−2 , which is the lowest value among all GeSn lasers tions, inter-valence band absorption (IVBA), free carrier absorp-
reported thus far. (See Note S7, Supporting Information for the tion (FCA), and indirect transitions.[18] At 10 K, losses from afore-
comparison of state-of-the-art GeSn lasers.) It is noteworthy that mentioned absorptions at wavelength range of interest are neg-
this exceptionally low threshold is achieved in an as-grown GeSn ligible, and the net gain is dominated by the interband gain. In
layer, signaling the importance of the defect-free nature of our the case of –0.5% epitaxial compressive strain (blue), interband
material prepared by the RMG method. gain is nonexistent, and therefore no net gain is observed. For
To understand the effect of the tensile strain on the laser the 0% strain relaxed case (red), a peak net gain of ≈1060 cm−1
threshold, we conducted theoretical simulations using 8-band is obtained at ≈2130 nm. With 0.15% tensile strain as found in
k∙p model. (See Note S8, Supporting Information for more de- our RMG-based GeSnOI lasers (black), the peak net gain is in-
tails on theoretical simulations) Figure 3b presents the net gain creased by nearly three times to 2670 cm−1 , while the wavelength
spectra calculated for 9.6 at.% Sn content GeSn at 10 K for three for the peak net gain is redshifted to ≈2190 nm owing to the
different strain values: 0.15% tensile strain (strain accumulated strain-induced bandgap reduction. It is worth noting that such
by the RMG method), 0% (strain relaxed case), and –0.5% com- beneficial tensile strain can only be achieved in our RMG method

Figure 3. 10 K lasing characteristics and gain modeling in RMG-based GeSn laser. a) Power-dependent photoluminescence spectra. Inset: Light-in light-
out curve (left axis) and FWHM (right axis) as a function of pump power density. Clear threshold behavior is observed in the L-L curve, coincident with
the sudden reduction in FWHM. b) Theoretical gain simulation for GeSn with 9.6 at.% Sn content as a function of energy at 10 K with 5 × 1016 cm−3
injection carrier densities. The net gain is given as 0.15% tensile strain (black, RMG-based GeSn), 0% (red, strain relaxed), and –0.5% compressive
strain (blue, typical epitaxy). The 0.15% tensile strain achieved in RMG-based GeSn has nearly three times increase in optical net gain, enabling low
GeSn laser threshold.

Laser Photonics Rev. 2024, 2401077 2401077 (4 of 6) © 2024 Wiley-VCH GmbH

 18638899, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lpor.202401077 by Korea Advanced Institute Of, Wiley Online Library on [27/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.lpr-journal.org

Figure 4. Temperature-dependent study of lasing characteristics from GeSn laser grown by RMG method. a) Emission spectra at a fixed power density
of 6.5 kW cm−2 for various temperatures. Lasing is sustained until 53 K. The spectrum at 84 K is enlarged for clarity. b) Light-in light-out curve taken at
various temperatures. Clear threshold behavior is evident for temperatures up to 53 K.

because in the typical GeSn epitaxy by chemical vapor deposition confinement in the gain medium, which is another unique fea-
(CVD), it is inevitable to have compressive strain in GeSn due ture of the RMG method. Ultra-low threshold of 0.52 kW cm−2 at
to the lattice mismatch between GeSn and underlying Ge buffer 10 K was achieved from our RMG-based GeSnOI lasers. In con-
layer. clusion, the RMG method enables the realization of truly mono-
lithic, high-performance on-chip lasers with ideal GeSn gain me-
dia, potentially completing the missing piece of the monolithi-
2.4. Temperature-Dependent Lasing Characteristics in cally integrated Si photonics platform.
RMG-Based GeSnOI lasers

We conducted temperature-dependent PL to investigate the ef-

fect of temperature on lasing behavior. Figure 4a shows emission 4. Experimental Section
spectra at four different temperatures. The pump power density Material Growth and Device Fabrication: 2 inch n− -Si (001) substrates
is fixed at 6.5 kW cm−2 . A distinct lasing peak can be observed for with a resistivity of 1–10 Ω cm were loaded into a low pressure chem-
temperatures up to 53 K, while at 84 K, only broad spontaneous ical vapor deposition (LPCVD) chamber to grow a 200 nm high quality
emission remains. The lasing peak intensity decreases with in- Si3 N4 layer. The Si3 N4 layer was then patterned by photolithography and
creasing temperature due to the decreased net gain.[19,20] The L- dry etching to open seed windows in which the Si substrate was fully ex-
posed. Natural oxide layer of the seed windows was removed by diluted HF
L curves measured at the four temperatures are also shown in
solution to guarantee good contact between GeSn layer and Si substrate in
Figure 4b. Clear threshold behaviors are evident for temperatures seed windows. In a magnetron sputterrer, a 600 nm GeSn amorphous film
of 10 K (black), 28 K (magenta), and 53 K (blue). In contrast, the was deposited by the sputtering of Ge and Sn targets in an Ar atmosphere
linearity of the plot shown for 84 K (green) indicates the mate- at room temperature with a growth pressure of 3 mTorr and substrate ro-
rial gain is insufficient to overcome the prevailing losses at ele- tation of 20 rpm. The initial Sn-content in the amorphous GeSn film was
vated temperature, resulting in the quenching of lasing emission, ≈25%. Next, the GeSn film was patterned into strips with 1.2 μm widths
and various lengths by dry etching. 1.5 μm-thick SiO2 was deposited by
which is a typical behavior in GeSn lasers.[6,8,11,21,22]
plasma enhancement chemical vapor deposition (PECVD) at 300 °C to
suppress agglomeration and evaporation of the molten GeSn strips. The
wafer was heated to 920 °C for 1 s with a ramp up rate of 100 °C s−1 by
3. Conclusion rapid thermal annealing (RTA). GeSn laser with F–P resonance cavity was
fabricated at the end of GeSn strip (close to metal Sn strip) by focused ion
In summary, we have demonstrated ultra-low threshold GeSnOI beam (FIB) (FEI Helios G4 CX). The length of the GeSn lasers is ≈10 μm.
laser using a defect-free and tensile-strained GeSn layer grown Material Characterization: The structure, Sn distribution, crystal orien-
tation, and strain of the GeSn strips were measured by TEM, EDX, SAED,
by the RMG method. The threading dislocations induced by the
and micro-Raman scattering. The TEM specimen was prepared by a dual-
large lattice mismatch between GeSn and Si are terminated at beam FIB system (FEI Helios Nanolab 600). The TEM images of the spec-
the narrow necking region, resulting in the defect-free GeSnOI imen were obtained by using a transmission electron microscope (FEI
layer. It should be noted that such defect-free GeSn layers can- Tecnai G2 F20 S-Twin). SAED and EDX were performed under the TEM
not be obtained even by removing the defective Ge/GeSn inter- instrument. The micro-Raman scattering was measured by Raman spec-
facial layer via wafer bonding and chemical-mechanical polish- troscopy (LabRAM HR800, HORIBA Jobin Yvon) at room temperature with
ing (CMP).[23] The tensile strain accumulated during the growth, a 488 nm line Ar+ laser and a Si photodetector. The spot size of the laser
focused on the sample was ≈1.5 μm.
which is only possible via the RMG method, plays a crucial role Photoluminescence Measurement: The sample was mounted into a
in enhancing the net gain, as supported by our theoretical mod- close-cycle helium cryostat operating between 4 and 350 K. A 1550 nm
eling. The large refractive index contrast between the GeSn gain pulsed laser with 5 ns pulse width and 1 MHz repetition rate was used
medium and the underlying insulating layer allows strong optical to optically pump the sample. The laser was focused by a ×15 reflec-

Laser Photonics Rev. 2024, 2401077 2401077 (5 of 6) © 2024 Wiley-VCH GmbH

 18638899, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/lpor.202401077 by Korea Advanced Institute Of, Wiley Online Library on [27/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.lpr-journal.org

tive objective onto a spot size of 25 μm. Emission from the sample [4] C. Xiang, W. Jin, O. Terra, B. Dong, H. Wang, L. Wu, J. Guo, T. J. J.
was collected by the same objective lens perpendicular to the sam- Morin, E. Hughes, J. Peters, Q.-X. Ji, A. Feshali, M. Paniccia, K. J. J.
ple surface and the signal was guided to an FTIR spectrometer with Vahala, J. E. E. Bowers, Nature 2023, 620, 78.
an extended liquid nitrogen-cooled InGaAs detector of a 2.4 μm cut- [5] K. P. Homewood, M. A. Lourenco, Nat. Photonics 2015, 9, 78.
off wavelength. The resolution of the FTIR used in data collection was [6] S. Wirths, R. Geiger, N. von den Driesch, G. Mussler, T. Stoica, S.
1 cm−1 . Mantl, Z. Ikonic, M. Luysberg, S. Chiussi, J. M. Hartmann, H. Sigg, J.
Faist, D. Buca, D. Gruetzmacher, Nat. Photonics 2015, 9, 88.
[7] Y. Kim, H.-J. Joo, M. Chen, B. Son, D. Burt, X. Shi, L. Zhang, Z. Ikonic,
Supporting Information C. S. Tan, D. Nam, Adv. Sci. 2023, 10, 202207611.
[8] D. Buca, A. Bjelajac, D. Spirito, O. Concepcion, M. Gromovyi, E.
Supporting Information is available from the Wiley Online Library or from Sakat, X. Lafosse, L. Ferlazzo, N. von den Driesch, Z. Ikonic, D.
the author. Grutzmacher, G. Capellini, M. El Kurdi, Adv. Opt. Mater. 2022, 10,
2201024.
[9] S. Amoah, S. Ojo, T. Huong, G. Abernathy, Y. Zhou, W. Du, J.
Margetis, J. Tolle, B. Li, S.-Q. Yu, Ieee, Conference on Lasers and
Acknowledgements Electro-Optics (CLEO), Conference on Lasers and Electro-Optics
Z.L. and M.C. contributed equally to this work. This work was sup- ELECTR NETWORK, 2021.
ported by the National Key Research and Development Program of China [10] D. Sukhdeo, Y. Kim, S. Gupta, K. Saraswat, B. Dutt, D. Nam, IEEE
(2018YFB2200103); National Natural Science Foundation (62250010, Electron Device Lett. 2016, 37, 1307.
62090054, 62274160); Strategic leading science and technology project, [11] A. Elbaz, D. Buca, N. von den Driesch, K. Pantzas, G. Patriarche,
CAS (XDB43020100); Youth Innovation Promotion Association Chinese N. Zerounian, E. Herth, X. Checoury, S. Sauvage, I. Sagnes, A. Foti,
Academy of Sciences (2021111). R. Ossikovski, J.-M. Hartmann, F. Boeuf, Z. Ikonic, P. Boucaud, D.
Gruetzmacher, M. El Kurdi, Nat. Photonics 2020, 14, 375.
[12] O. Moutanabbir, S. Assali, X. Gong, E. O’Reilly, C. A. Broderick, B.
Marzban, J. Witzens, W. Du, S. Q. Yu, A. Chelnokov, D. Buca, D. Nam,
Conflict of Interest Appl. Phys. Lett. 2021, 118, 110502.
The authors declare no conflict of interest. [13] D. S. Sukhdeo, S. Gupta, K. C. Saraswat, B. Dutt, D. Nam, Opt. Com-
mun. 2016, 364, 233.
[14] D. Rainko, Z. Ikonic, A. Elbaz, N. von den Driesch, D. Stange, E.
Herth, P. Boucaud, M. El Kurdi, D. Gruetzmacher, D. Buca, Sci. Rep.
Data Availability Statement 2019, 9, 259.
[15] S. Gupta, B. Magyari-Koepe, Y. Nishi, K. C. Saraswat, J. Appl. Phys.
The data that support the findings of this study are available from the cor-
2013, 113, 073707.
responding author upon reasonable request.
[16] Q. Huang, X. Liu, J. Zheng, Y. Yang, D. Zhang, Y. Pang, J. Cui, Z. Liu,
Y. Zuo, B. Cheng, J. Lumin. 2023, 255, 119623.
[17] Z. Liu, H. Cong, F. Yang, C. Li, J. Zheng, C. Xue, Y. Zuo, B. Cheng, Q.
Keywords Wang, Sci. Rep. 2016, 6, 38386.
[18] T. Quang Minh, J. Chretien, M. Bertrand, L. Casiez, A. Chelnokov, V.
GeSn alloys, lasers, rapid melting growth method, Si photonics Reboud, N. Pauc, V. Calvo, Phys. Rev. B 2020, 102, 155203.
[19] Y. Kim, S. Assali, D. Burt, Y. Jung, H.-J. Joo, M. Chen, Z. Ikonic, O.
Received: July 10, 2024 Moutanabbir, D. Nam, Adv. Opt. Mater. 2022, 10, 2101213.
Revised: November 25, 2024 [20] H.-J. Joo, Y. Kim, D. Burt, Y. Jung, L. Zhang, M. Chen, S. J. Parluhutan,
Published online: D.-H. Kang, C. Lee, S. Assali, Z. Ikonic, O. Moutanabbir, Y.-H. Cho,
C. S. Tan, D. Nam, Appl. Phys. Lett. 2021, 119, 201101.
[21] A. Bjelajac, M. Gromovyi, E. Sakat, B. Wang, G. Patriarche, N. Pauc, V.
Calvo, P. Boucaud, F. Boeuf, A. Chelnokov, V. Reboud, M. Frauenrath,
[1] W. Bogaerts, D. Pérez, J. Capmany, D. A. B. Miller, A. Melloni, Nature J. Hartmann, M. El Kurdi, Opt. Express 2022, 30, 3954.
2020, 586, 207. [22] A. Elbaz, R. Arefin, E. Sakat, B. Wang, E. Herth, G. Patriarche, A.
[2] E. Pelucchi, G. Fagas, I. Aharonovich, D. Englund, E. Figueroa, Q. Foti, R. Ossikovski, S. Sauvage, X. Checoury, K. Pantzas, I. Sagnes,
Gong, H. Hannes, J. Liu, C.-Y. Lu, N. Matsuda, J.-W. Pan, F. Schreck, J. Chretien, L. Casiez, M. Bertrand, V. Calvo, N. Pauc, A. Chelnokov,
F. Sciarrino, C. Silberhorn, J. Wang, K. D. Jons, Nat. Rev. Phys. 2022, P. Boucaud, F. Boeuf, V. Reboud, J.-M. Hartmann, M. El Kurdi, ACS
4, 194. Photonics 2020, 7, 2713.
[3] S. Y. Siew, B. Li, F. Gao, H. Y. Zheng, W. Zhang, P. Guo, S. W. Xie, A. [23] J. R. Jain, D.-S. Ly-Gagnon, K. C. Balram, J. S. White, M. L.
Song, B. Dong, L. W. Luo, C. Li, X. Luo, G. Q. Lo, J. Lightwave Technol. Brongersma, D. A. B. Miller, R. T. Howe, Opt. Mater. Express 2011,
2021, 39, 4374. 1, 1121.

Laser Photonics Rev. 2024, 2401077 2401077 (6 of 6) © 2024 Wiley-VCH GmbH