вход по аккаунту



код для вставкиСкачать
Applied Physics Express
Related content
Generation of high-peak-power sub-nanosecond
650-nm-band optical pulses based on
semiconductor-laser-controlling technologies
To cite this article: Jui-Hung Hung et al 2017 Appl. Phys. Express 10 102701
View the article online for updates and enhancements.
- Generation of 650-Femtosecond Optical
Pulses from a Gain-Switched Laser Diode
Using Second-Harmonic Generation
Hiroshi Ohta, Seiji Nogiwa, Yoshikazu
Kawaguchi et al.
- 300 W Peak Power Picosecond Optical
Pulse Generation by Blue-Violet GaInN
Mode-Locked Laser Diode and
Semiconductor Optical Amplifier
Rintaro Koda, Tomoyuki Oki, Shunsuke
Kono et al.
- Generation of 10 GHz, 1.9 ps optical pulse
train using semiconductor optical amplifier
andsilica-based highly nonlinear fiber
S Y Yan, X P Xie, Z Q Hui et al.
This content was downloaded from IP address on 26/10/2017 at 07:38
Applied Physics Express 10, 102701 (2017)
Generation of high-peak-power sub-nanosecond 650-nm-band optical pulses
based on semiconductor-laser-controlling technologies
Jui-Hung Hung1*, Kazuo Sato1, Yi-Cheng Fang1, Lung-Han Peng2, Tomomi Nemoto3, and Hiroyuki Yokoyama1*
New Industry Creation Hatchery Center, Tohoku University, Sendai 980-8579, Japan
Graduate Institute of Photonics and Optoelectronics, National Taiwan University, Taipei 10617, Taiwan, Republic of China
Research Institute for Electronic Science, Hokkaido University, Sapporo 001-0020, Japan
Received August 10, 2017; accepted September 6, 2017; published online September 20, 2017
We have developed a method to generate sub-nanosecond 650-nm-band optical pulses. These pulses have a peak power of 40 W and a pulse
energy of 13 nJ at a 1-MHz repetition rate. This technology is intended for application in stimulated-emission-depletion microscopy. Our method is
based on the pulsed operation of a 1.3-µm-band semiconductor-laser optical amplifier and the second-harmonic generation of the optical pulses
after amplification by a Pr-doped fiber amplifier. The resultant peak power and pulse energy of the 650-nm-band optical pulses are two orders of
magnitude higher than those directly obtained from a laser diode. © 2017 The Japan Society of Applied Physics
ecently, super-resolution microscopy technologies
have been attracting considerable attention owing
to their ability to surpass the diffraction limit of
traditional light microscopy for studying biological specimen.1,2) Among several techniques to obtain super-resolution
images, stimulated-emission-depletion (STED) microscopy
is the only technology that does not require mathematical
reconstruction.1) In brief, a fluorescent area is made smaller
than the light diffraction limit by overlapping a donut-shaped
STED light beam to a Gaussian-shaped excitation laser beam.
In STED microscopy, an excitation laser excites fluorescent
molecules from their ground state to their fluorescent state
while a red-shifted STED light depletes the fluorescence
along the outer rim of the excitation spot via stimulated
emission depletion of excited molecules, leaving a fluorescent spot smaller than the diffraction limit. By using continuous-wave (CW) laser sources for STED light, the spatial
resolution of STED microscopy is determined by the intensity of STED light, ISTED, and can be approximated by
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ;
2NA 1 þ ISTED =Is
where λ denotes the wavelength of the excitation light, NA
is the numerical aperture of the lens, and Is is the saturation
intensity.1) Therefore, sufficient STED light intensity (and
thus light power) is required for significant improvement of
spatial resolution. In general, ∼1-W average power CW light
sources are required for STED microscopy. This high power,
however, causes photobleaching in fluorescent molecules.3,4)
An effective method to reduce the optical power of the
STED laser light is to use synchronized pulsed-mode excitation and STED de-excitation. By using this method, the
averaged STED laser power can be decreased by two orders
of magnitude compared to that using the CW STED laser
method.5) This is because the fluorescent signal should be
retrieved within the fluorescent lifetime of the excited molecules. As the lifetime is ∼1 ns, the STED light should take
effect within 1 ns, shortly after the excitation pulse. Thereafter, the STED light cannot provide efficient depletion.
Consequently, the fundamental criteria for an applicable
STED light source are the STED pulse intensity and timing.
A major difficulty for pulsed STED is the preparation of a
suitable compact and stable STED optical pulse source. To
this end, optical pulse generation through supercontinuum
(SC) generation6) and stimulated Raman scattering (SRS)7) in
optical fibers have been proposed and applied for pulsedmode STED microscopy. An SC pulse source is beneficial for
extracting the necessary wavelength optical pulses. However,
it needs broadband (∼20 nm) spectral extraction to obtain
sufficient pulse energy (≥1-nJ level). This broadband feature
is not ideal for efficient de-excitation in view of the STED
principle. SRS is good for generating discrete multiwavelength STED pulses. However, it requires higher-order
SRS to obtain red region pulses starting from incident
green optical pulses. This multistage nonlinear wavelength
conversion process makes final-stage optical pulses very
sensitive to the stability conditions of the incident pulses and
the operation environment.
A much simpler method utilizing the pulsed operation of a
red laser diode (LD) has been demonstrated for successful
pulsed-mode STED microscopy incorporating synchronized
single-photon pulsed excitation.8,9) The available red pulse
energy from this simple scheme, however, is just above
the lowest limit for inducing notable STED effects. Given
the significant advantages of LD devices, in this paper, we
present a new approach to generate 650-nm-band optical
pulses based on LD-controlling technologies while providing
a pulse peak power two orders of magnitude higher than that
directly obtained from a pulsed red LD.
Our concept is to produce optical pulses from a 1.3-µmband LD; these pulses are then synchronized to the optical
pulses for excitation of fluorescent molecules in bio-tissues.
Thereafter, the generated optical pulses are amplified by
a proper optical amplifier and, subsequently, converted to
second-harmonic (SH) optical pulses using an appropriate
nonlinear crystal. A schematic representation of the subnanosecond 650-nm-band optical pulse source is shown in
Fig. 1.
During initial experiments, we employed the gain-switching (GS) method to generate optical pulses from a distributedfeedback (DFB) LD [Fig. 1(b)]. Gain-switched laser diode
(GS-LD) technologies are well developed. We have recently
generated optical pulses with minimal durations of ≤5 ps10–14)
and have demonstrated application in two-photon microscopy.10,13,14) However, GS-LDs are not very suitable for
generating sub-nanosecond optical pulses because, in general,
the obtainable long pulse duration is ≤100 ps. This is not ideal
for pulsed-mode STED application owing to the mutual
© 2017 The Japan Society of Applied Physics
Appl. Phys. Express 10, 102701 (2017)
J.-H. Hung et al.
Fig. 1. Schematics of sub-nanosecond 650-nm-band optical pulse generation. (a) Setup for converting a 1.3-µm seed pulse to 650-nm SH pulses. PDFA: Prdoped fiber amplifier; FPC: fiber polarization controller; PPMgLN: periodic-poled MgO-doped lithium niobate; BPF: optical band-pass filter. (b) 1.3-µm seed
pulse generation via gain-switching operation of a DFB LD. (c) 1.3-µm seed pulse generation via pulsed operation of a semiconductor-laser optical amplifier.
Fig. 2. Oscilloscope waveform traces and optical spectra for optical pulses
generated by using the GS-LD configuration. (a) Self-triggered 1306.9-nm
GS-LD pulse (64-times averaged); (b) GS-LD pulse spectrum; (c) electricaltriggered 653.5-nm SH pulse (64-times averaged); (d) SH pulse spectrum.
Comparing (a) and (c) shows that the initial sharp pulse is enhanced in the
SH pulse because of the nonlinear SH conversion process.
timing jitters between the synchronized optical pulses
generated by the two GS-LDs. Usually, mutual timing jitter
ranges from several picoseconds to several tens of picosecond
when incorporating the jitter from the GS-LD optical pulse
generation process as well as the electronic instrumental jitter.
By accounting for a typical timing jitter value, in practice,
pulse duration in the sub-nanosecond range can be appropriate
for STED optical pulses as it allows proper overlap between
excitation and STED optical pulses. Although it is possible to
generate longer pulses by controlling the excitation electric
pulse duration and intensity, the obtainable optical pulses do
not show smooth temporal shapes. The second- and thirdpulse components, or the slow tail components, follow the
initial sharp pulse component and have a duration of several
tens of picoseconds (a feature partly shown in Fig. 2). Since
the STED effect mainly relies on this initial short pulse
component, the timing jitter problem cannot be solved.
Based on these considerations, we have developed a new
method to generate sub-nanosecond optical pulses utilizing a
semiconductor-laser optical amplifier (SOA) that was operated by electric pulse excitation under incident CW laser
light, as is shown in Fig. 1(c). Henceforth, we will refer to
this method as “gain-switched SOA” (GS-SOA). During the
experiments, a 1.3-µm DFB LD (Qphotonics QDFBLD1300-10) was operated under constant DC current excitation.
It was controlled to produce CW laser light having an
average power within a range of 1–10 mW. The CW laser
light was coupled into an SOA device (Thorlabs BOA1017S)
that consisted of an InP=InGaAsP multiple-quantum-well
(MQW) active-layer structure and has polarization-dependent
gain up to 23 dB. The SOA was driven by a homemade electric pulse generator, which could provide 850-ps pulses at a
1-MHz repetition rate with a maximal amplitude of 12 V.
This GS-SOA setup produced sub-nanosecond optical pulses
having a duration of 400–500 ps (FWHM), which depended
slightly on the incident CW laser light power.
Thereafter, the sub-nanosecond optical pulses were
amplified by a praseodymium-doped fiber amplifier (PDFA;
Fiberlabs AMP-FL8611-OB), which has a small signal gain
of >35 dB with a two-stage PDFA configuration. Subsequently, the amplified optical pulses were focused into a 10mm-long periodic-poled magnesium oxide-doped lithium
niobate (PPMgLN) crystal (fabricated by HC Photonics) to
generate 650-nm SH light. The PPMgLN crystal used had a
12.7-µm poling period and was operated at a temperature of
19 °C to satisfy the phase-matching condition. Both facets of
the crystal were antireflection coated (R < 0.5% at both 650nm and 1.3-µm bands) to minimize reflection losses. After the
generation of the SH light, a 14-nm-bandwidth optical bandpass filter (BPF) was used to remove the residual 1.3-µm
laser light. Optical spectra were measured by an interferometer-type optical spectrum analyzer (Advantest Q8387).
The spectral resolution width depends on the measured wavelength range; for the 1.3-µm band, the resolution is 0.005 nm,
while for the 650-nm band, the resolution is 0.001 nm. The
temporal waveforms were recorded by a combination of a
high-speed InGaAs photodetector (New Focus 1434, with a
25-GHz bandwidth) and a sampling oscilloscope (Agilent
86100C, with a 40-GHz bandwidth).
Initially, we tried to generate high-peak-power, 650-nm
light pulses starting from 1.3-µm GS-LD laser pulses. The
GS-LD pulses had a pulse energy of 3.5 pJ and a duration of
500 ps. Figures 2(a) and 2(b) show a temporal waveform and
an optical spectrum for the GS-LD pulses. In this measurement, the oscilloscope was triggered by the optoelectronic conversion of the optical pulses; thus, both the electronic timing
© 2017 The Japan Society of Applied Physics
Appl. Phys. Express 10, 102701 (2017)
J.-H. Hung et al.
Fig. 4. Oscilloscope waveform traces for optical pulses generated by
using the GS-SOA configuration under 1-mW incident power. (a) Nonaveraged GS-SOA pulse shape and (b) corresponding SH pulse shape. When
CW laser incident power decreased, the GS-SOA pulse largely fluctuated and
the fluctuation was more pronounced after the SHG process.
Fig. 3. Oscilloscope waveform traces and optical spectrum for optical
pulses generated by using the GS-SOA configuration. (a) Non-averaged
1307.1-nm GS-SOA pulse; (b) GS-SOA pulse spectrum; (c) non-averaged
653.6-nm SH pulse; (d) SH pulse spectrum. The resultant 653.6-nm SH pulse
has nearly 40-W peak power, which is two orders of magnitude higher than
that of the direct pulse generation from an LD.
jitter and gain-switching timing jitter were eliminated. The
total jitter value was estimated to be 15 ps (rms value). This
was done by measuring the GS-LD pulse through the trigger
signal provided by the electrical pulse generator. The optical
pulse waveform consists of a sharp first-pulse component and
the following lower intensity tail component. The duration of
the first sharp pulse is ∼30 ps, and the peak intensity is a factor
of 3 higher than that of the subsequent pulse components.
Figure 2(c) shows a temporal waveform of the SH 650-nm
optical pulses. The exact center wavelength is 653.5 nm, which
is half the value of the 1.3-µm-band LD oscillation wavelength. These SH pulses had a pulse energy of 5 nJ and a
duration of ∼300 ps. The peak intensity of the sharp first-pulse
component is significantly manifested after the SH conversion
process. It is found that the initial sharp first-pulse component
of the SH optical pulse involves nearly 40% of the entire
pulse energy. Therefore, the electronic timing jitter would
compromise the effectiveness of the STED pulse irradiation.
In contrast, the GS-SOA method generated smooth,
single-peak optical pulses with sub-nanosecond duration.
Figure 3(a) shows a typical temporal waveform of the GSSOA pulses. The CW incident laser-light power was set to
10 mW with careful control of its polarization. The GS-SOA
pulses had a pulse energy of 15 pJ and a duration of 460 ps
(FWHM). The peak power of the GS-SOA pulses was
evaluated at 30 mW; the pulse peak power was increased by a
factor of 3 through the SOA amplification. Note that the pulse
width is less than the that of the driving electrical pulse
(850 ps) in this setup. This is due to the dynamics of electric
carrier excitation and stimulated de-excitation in the SOA.
This behavior will be the subject of future studies.
The maximum extinction ratio (ER) of the GS-SOA pulse
under 10-mW incident power was evaluated at 57 dB (calculated under 50-nW background leak CW power and 30-mW
pulse peak power), which is sufficient for the following
optical amplification by the PDFA, given the input average
optical pulse power of 15 µW. After the PDFA amplification,
the 1.3-µm pulse energy was increased to 68 nJ (with an
average optical power of 68 mW) without any waveform
distortion. These amplified optical pulses were converted into
650-nm-band SH optical pulses by the PPMgLN crystal. The
peak wavelength of the SH pulse was 653.6 nm, and the
temporal waveform is shown in Fig. 3(c). The optical pulse
energy was 13 nJ, and the average optical power was 13 mW;
thus, the SH conversion efficiency was 19%. The pulse duration was decreased to 330 ps (FWHM) via a second-order
nonlinear optical conversion process. The peak power of the
SH optical pulses was evaluated at 40 W. This value is two
orders of magnitude higher than the maximum peak power of
the optical pulses directly obtained from a single-transversemode GS-LD. Therefore, this 653.6-nm optical pulse source
will be quite useful for synchronous pulsed-mode STED
microscopy, solving the marginal optical power problem of
direct GS-LD pulse generation for STED application.
In Fig. 3(d), the spectrum of the SH pulses exhibits
multiple peaks and is not as smooth as the GS-SOA spectrum
[Fig. 3(b)]. We confirmed that this is due to the self-phasemodulation (SPM) effect inside the PDFA induced by the
amplified 1.3-µm optical pulses. We observed two spectral
peaks for the 1.3-µm optical pulses at 1307.11 and 1307.14
nm (not shown here); these correspond to the SH light
spectrum in Fig. 3(d). When the gain of the PDFA decreased,
the spectrum returned to one peak.
Obtaining desirable GS-SOA pulses with high ER and
high average power for efficient amplification by the PDFA
requires proper adjustments of the polarization and the
incident power of the CW laser light. A precise polarization
leads to maximum amplification when the SOA is driven
by electrical pulses. It also leads to minimum background
leakage of the CW laser light power when the SOA is in its
off state. Besides, to obtain sufficient average power (>10
µW, in our case) from the GS-SOA output optical pulses, the
incident CW laser light power should be high. In practice, we
set the CW laser light power to 10 mW, the maximal value
for the present 1.3-µm LD. This resulted in an average optical
power of 15 µW for the GS-SOA output optical pulses.
Simultaneously, the average power of the CW laser light
leakage was 50 nW, which is low enough with respect to the
average power of the GS-SOA output optical pulses. It is also
important to consider suppression of the intensity fluctuation
in the optical pulse output when operating the GS-SOA.
When we set the incident CW laser light power to 1 mW, we
observed increased intensity fluctuation in the GS-SOA
output optical pulses. Moreover, the intensity fluctuation was
further enhanced in the corresponding SH pulse owing to the
intensity-dependent nonlinear wavelength conversion process. These features are shown in Figs. 4(a) and 4(b). For
10-mW CW laser light incidence, the mechanism of gain
saturation successfully suppressed the optical pulse intensity
© 2017 The Japan Society of Applied Physics
Appl. Phys. Express 10, 102701 (2017)
J.-H. Hung et al.
fluctuation. Therefore, we consider that a 10-mW CW laser
light incident power, with precise polarization control, is
optimal for generating sub-nanosecond optical pulses under
our current GS-SOA configuration.
In summary, we have proposed and demonstrated a novel
approach to generate 650-nm-band sub-nanosecond optical
pulses that have sufficient peak power for application in
pulsed-mode STED microscopy. The combination of a
1.3-µm GS-SOA sub-nanosecond optical pulse generator, a
Pr-doped optical fiber amplifier, and a PPMgLN SH converter has produced smooth-shaped 330-ps-duration, 653.6nm optical pulses having a peak power of 40 W at a 1-MHz
repetition rate. This optical pulse generation method can be
highly beneficial for STED-based super-resolution microscopy in combination with GS-LD-based picosecond-pulse
excitation light sources.10,13,14)
Acknowledgments This work was supported, in part, by “Brain Mapping
by Integrated Neurotechnologies for Disease Studies (Brain=MINDS)” from the
Japan Agency for Medical Research and Development (AMED) and the Research
Program “Dynamic Alliance for Open Innovation Bridging Human, Environment
and Materials” in the Network Joint Research Center for Materials and Devices.
1) S. W. Hell and J. Wichmann, Opt. Lett. 19, 780 (1994).
2) P. Bianchini, C. Peres, M. Oneto, S. Galiani, G. Vicidomini, and A.
Diaspro, Cell Tissue Res. 360, 143 (2015).
3) G. Donnert, C. Eggeling, and S. W. Hell, Nat. Methods 4, 81 (2007).
4) M. Bouzin, G. Chirico, L. D’Alfonso, L. Sironi, G. Soavi, G. Cerullo, B.
Campanini, and M. Collini, J. Phys. Chem. B 117, 16405 (2013).
5) K. I. Willig, B. Harke, R. Medda, and S. W. Hell, Nat. Methods 4, 915
6) D. Wildanger, E. Rittweger, L. Kastrup, and S. W. Hell, Opt. Express 16,
9614 (2008).
7) B. R. Rankin and S. W. Hell, Opt. Express 17, 15679 (2009).
8) S. Schrof, T. Staudt, E. Rittweger, N. Wittenmayer, T. Dresbach, J.
Engelhardt, and S. W. Hell, Opt. Express 19, 8066 (2011).
9) K. Otomo, Y.-C. Fang, J.-H. Hung, H. Yokoyama, and T. Nemoto,
10) H. Yokoyama, H. Guo, T. Yoda, K. Takashima, K. Sato, H. Taniguchi, and
H. Ito, Opt. Express 14, 3467 (2006).
11) S. Chen, A. Sato, T. Ito, M. Yoshita, H. Akiyama, and H. Yokoyama, Opt.
Express 20, 24843 (2012).
12) S. Chen, M. Yoshita, A. Sato, T. Ito, H. Akiyama, and H. Yokoyama, Opt.
Express 21, 10597 (2013).
13) Y. Kusama, Y. Tanushi, M. Yokoyama, R. Kawakami, T. Hibi, Y. Kozawa,
T. Nemoto, S. Sato, and H. Yokoyama, Opt. Express 22, 5746 (2014).
14) R. Kawakami, K. Sawada, Y. Kusama, Y.-C. Fang, S. Kanazawa, Y.
Kozawa, S. Sato, H. Yokoyama, and T. Nemoto, Biomed. Opt. Express 6,
891 (2015).
© 2017 The Japan Society of Applied Physics
Без категории
Размер файла
614 Кб
102701, apex
Пожаловаться на содержимое документа