783 MHz fundamental repetition rate all-fiber ring laser mode-locked by carbon nanotubes

Abstract

We demonstrate a 783 MHz fundamental repetition rate mode-locked Er-doped all-fiber ring laser with a pulse width of 623 fs. By using carbon nanotubes saturable absorber, a relatively low self-starting pump threshold of 108 mW is achieved. The laser has a very compact footprint less than 10 cm × 10 cm, benefiting from the all-active-fiber cavity design. The robust mode-locking is confirmed by the low relative intensity noise and a long-term stability test. We propose a new scheme for generating high repetition rate femtosecond optical pulses from a compact and stable all-active-fiber ring oscillator.

Full text

783 MHz fundamental repetition rate all-fiber ring laser mode-locked by carbon
nanotubes
Maolin Dai
1,2
, Bowen Liu
1,2
, Yifan Ma
1,2
, Takuma Shirahata
1,2
, Ruoao Yang
3
, Zhigang Zhang
3
, Sze Yun Set
1,2*
, and
Shinji Yamashita
1,2*
1
Department of Electrical Engineering and Information Systems, The University of Tokyo, Bunkyo-ku, Tokyo 113-8656, Japan
2
Research Center for Advanced Science and Technology, The University of Tokyo, Meguro-ku, Tokyo 153-8904, Japan
3
State Key Laboratory of Advanced Optical Communication Systems and Networks, School of Electronics, Peking University, Beijing 100871, People’s
Republic of China
*
E-mail: set@cntp.t.u-tokyo.ac.jp;syama@cntp.t.u-tokyo.ac.jp
Received May 30, 2024; accepted June 5, 2024; published online June 18, 2024
We demonstrate a 783 MHz fundamental repetition rate mode-locked Er-doped all-fiber ring laser with a pulse width of 623 fs. By using carbon
nanotubes saturable absorber, a relatively low self-starting pump threshold of 108 mW is achieved. The laser has a very compact footprint less
than 10 cm ×10 cm, benefiting from the all-active-fiber cavity design. The robust mode-locking is confirmed by the low relative intensity noise and
a long-term stability test. We propose a new scheme for generating high repetition rate femtosecond optical pulses from a compact and stable all-
active-fiber ring oscillator. ©2024 The Author(s). Published on behalf of The Japan Society of Applied Physics by IOP Publishing Ltd
Owing to the widespread applications in both scien-
tific research and industries such as frequency comb-
based spectroscopy,
1–3)
laser processing,
4,5)
and
bio-imaging,
6)
high repetition rate fiber lasers have been
developed intensively and many breakthroughs have been
achieved. Techniques such as active mode-locking,
7)
pas-
sively harmonic mode-locking,
8,9)
dissipative four-wave
mixing
10)
and mode filtering
11)
could achieve a very high
repetition rate, however, they exhibit a higher instability in
terms of output performance compared to fundamentally
passive mode-locking.
12)
Previous works have shown that
ultrahigh repetition rates can be obtained by using short fiber
Fabry–Perot resonators.
13–15)
Ring lasers can get rid of
standing waves and spatial hole burning effect that always
happen in linear lasers, however, it is challenging for
realizing high repetition rate ring lasers.
Recently, high repetition rates have been successfully
achieved from Yb-doped fiber ring lasers.
16–19)
Compared to
Yb-doped silica fibers, Er-doped silica fibers have lower gain
efficiency,therefore,manyofthereportedhighrepetitionrate
fiber lasers in 1.5 μmbandarebasedonhighlyEr-dopedorEr/
Yb co-doped phosphate glass fibers.
20–22)
In 2016, J. Zhang et al.
demonstrated a 517 MHz ring fiber laser mode-locked by
nonlinear polarization evolution (NPE) using commercialized
highly Er-doped silica fiber.
23)
In 2023, using the same oscillator
design, the cavity length was further decreased to 22.5 cm,
corresponding to a repetition rate of 895 MHz.
24)
However,
those lasers contain the free space components which require
fine alignment. All-fiber configuration would better benefitthe
compactness and robustness of the high repetition rate fiber
lasers. In 2017, 384 MHz repetition rate was achieved by an all-
fiber NPE ring laser using silica gain fiber.
25)
However, all-fiber
NPE lasers with higher repetition rate are hard to achieve owing
to the limited space for polarization management.
Reports have shown that using slow saturable absorber
(SA) such as carbon nanotubes (CNT) can realize high
repetition rate mode-locking with femtosecond pulse
width.
26–28)
Benefiting from its low saturation intensity,
self-started femtosecond pulses could be achieved in all-fiber
laser cavities under a relatively low pump power. By using
CNT-SA, high repetition rates of over 300 MHz have been
realized.
26–28)
We summarize the development of Er-doped
fiber ring lasers delivering high repetition rates as Table I.
Here, we report an Er-doped all-fiber ring laser mode-
locked by CNT-SA that delivers 783 MHz fundamental
repetition rate and 623 fs pulse width at the center wave-
length of 1562 nm. The laser oscillator has a very compact
size with footprint of less than 10 cm ×10 cm, where only
24 cm Er-doped silica fiber, an all-in-one integrated device
and CNT-SA coated on fiber ferrule are used in the laser
cavity, making it a robust all-active-fiber configuration. Turn-
key mode-locking operation can be achieved under a low
self-starting pump threshold of 108 mW. To the best of our
knowledge, 783 MHz is the highest fundamental repetition
rate from all-fiber ring lasers to date.
The schematic diagram and the image of the laser
oscillator are shown in Figs. 1(a) and 1(b), respectively. To
shorten the cavity length as much as possible, compact
devices should be used. Here we use an all-in-one device
called polarization insensitive tap/iso/wavelength division
multiplexer (PI-TIWDM) with active fiber pigtails to inte-
grate the functions of 10% output coupler, isolator, and 980/
1550 nm WDM. A 976 nm laser diode (LD, 1999CHB, 3SP
Technologies) is used as the pump source for the oscillator.
24 cm highly Er-doped fiber (EDF, Er80-8/125, LIEKKI)
with a core absorption of 80 ± 8 dB m
−1
at 1530 nm and an
anomalous dispersion of −20 ps
2
km
−1
is used for providing
gain in 1.5 μm for laser oscillation. The net dispersion in the
cavity is calculated as −4800 fs
2
. It is noted that no passive
fiber is used in the laser cavity. The CNT-SA is prepared as
follows: First, the commercialized high-pressure carbon
monoxide (HiPco)-synthesized single-wall CNT powders
(NanoIntegris) with diameters of 0.8–1.2 nm are fully dis-
persed into the N,N-dimethylformamide (DMF) solvent using
an ultrasonic homogenizer (UH-600, SMT), then the CNT
solution is directly sprayed onto the surface of the ferrule
using a high-pressure spraying gun. At the same time, a dryer
is used to evaporate the DMF solvent. After a certain time of
Content from this work may be used under the terms of the Creative Commons Attribution 4.0 license. Any further distribution of this
work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
062001-1 ©2024 The Author(s). Published on behalf of
The Japan Society of Applied Physics by IOP Publishing Ltd
Applied Physics Express 17, 062001 (2024) LETTER
https://doi.org/10.35848/1882-0786/ad548f

spraying, the CNT film is formed on the fiber facet to act as
the SA. Owing to the absorption of the gain fiber, the
nonlinear transmission of the CNT-SA cannot be measured
directly. We fabricate another CNT-SA on passive fiber
under the same condition, and measure its saturable absorp-
tion property, shown in Fig. 1(c). The measurement setup and
the fitting equation can be found in Ref. 29. Results show that
the CNT-SA has a modulation depth (MD) of 1.9% when
peak intensity increases to 150 MW cm
−2
, and the saturation
intensity is measured as 22.8 MW cm
−2
. The low saturation
intensity is highly benefit for the low-threshold self-starting
of the proposed mode-locked laser.
The optical output is directly measured by an optical
spectrum analyzer (OSA, AQ6370D, YOKOGAWA) and an
autocorrelator (FR-103XL, Femtochrome). The optical-to-
electrical conversion is achieved by a high-speed photodetector
with 25 GHz bandwidth (PD, 1414, NEWFOCUS). The con-
verted pulse train is measured by an oscilloscope (DSA91304A,
Keysight) and an electrical spectrum analyzer (ESA, RSA3045,
RIGOL). The relative intensity noise (RIN) is analyzed by a
RIN analyzer (PNA1, Thorlabs). The output power is recorded
by a power meter (PM100USB, Thorlabs), and the stability of
repetition rate is measured by a frequency counter (53181A,
Agilent).
Turn-key, self-started mode-locking operation is achieved
when the pump power reaches 108 mW. We slightly turn up
the pump power to 150 mW while keeping its single-pulse
mode-locking, the average output power is measured as 0.7 mW,
and the output performance is plotted in Fig. 2.Figure2(a)
shows the optical spectra under a 0.1 nm resolution bandwidth
(RBW), both in log scale and linear scale, where the latter one is
well fitted by the sech
2
curve. The laser has a wide emission
spectrum from 1545 to 1575 nm with a center wavelength λ
c
of
1562 nm and a full width at half maximum (FWHM) λ
FWHM
of
5.6 nm. No obvious Kelly sidebands are observed owing to the
low peak power under such a high repetition rate. In Fig. 2(b),
the corresponding autocorrelation (AC) trace is well fitted by the
sech
2
curve, indicating the pulse is soliton shaped, which is
expected to be obtained from an oscillator with anomalous
dispersion. The pulse width is measured as 623 fs, and the time-
bandwidth product (TBP) is calculated as 0.43, showing the
pulse is slightly chirped. The pulse could be further compressed
outside cavity. The laser owns an all-anomalous-dispersion
cavity; shorter pulse width could be obtained if we properly
manage the cavity dispersion.
Figure 2(c) shows oscillogram of the laser pulse train with
a roundtrip time of 1.3 ns. As shown in Fig. 2(d), the RF
spectrum with a RBW of 300 Hz indicates the laser has a
fundamental repetition rate f
c
of 783 MHz with a high signal-
to-noise ratio (SNR) of 70 dB. The inset shows the higher-
order harmonics up to 4 GHz with a 10 kHz RBW. In the
broader RF spectrum, there is no other noisy frequency,
showing the laser purely operates at the mode-locking
regime. The intensity decline of the harmonic frequencies
may result from the limited bandwidth of the ESA (4.5 GHz).
In our attempt for laser mode-locking, we find the mode-
locking is highly dependent on the deposited CNT-SA. When
thick CNT film (darker color in experimental observation) is
deposited onto the ferrule, corresponding to a higher linear
loss as well as high MD,
30)
the laser operates in the Q-
switching mode with higher possibility. When the laser
operates at Q-switch regime, the CNT-SA is easy to be burnt
owing to the high pulse energy. We gradually decrease the
thickness of the CNT layer (decrease the MD), and finally the
mode-locking is achieved. Our finding is consistent with
Table I. Comparisons between high fundamental repetition rate Er-doped fiber ring lasers.
References Configuration Fundamental repetition rate Self-starting pump power Mechanism SNR (RBW)
27 All-fiber 300 MHz Not mentioned CNT 70 dB (1 kHz)
28 All-fiber 358 MHz 60 mW CNT 73 dB (1 Hz)
25 All-fiber 374 MHz 620 mW NPE 78 dB (1 kHz)
26 All-fiber 447 MHz N/A CNT N/A
22 All-fiber 500 MHz 650 mW NPE 85 dB (2 kHz)
23 Non-all-fiber 517 MHz 2 W NPE 50 dB (30 kHz)
24 Non-all-fiber 895 MHz N/A NPE 75 dB (30 kHz)
This work All-fiber 783 MHz 108 mW CNT 70 dB (300 Hz)
Fig. 1. (a) The schematic diagram of the proposed fiber laser; (b) the image
of the proposed fiber laser; (c) nonlinear transmission curve of the CNT-SA.
062001-2 ©2024 The Author(s). Published on behalf of
The Japan Society of Applied Physics by IOP Publishing Ltd
Appl. Phys. Express 17, 062001 (2024) M. Dai et al.

previous works reporting that small MD of SA should be
used for laser mode-locking in high repetition regimes.
15,31)
The relationship between the output power and pump power
as well as the corresponding operation regimes is plotted in
Fig. 3. In the process of increasing pump power, the laser
experiences continuous-wave (CW) regime, Q-switching regime,
and finally mode-locking regime when the pump power reaches
22 mW, 39 mW, and 108 mW, respectively. Under the pump
power of 108 mW, the intracavity peak intensity is estimated
around 25 MW cm
−2
, the effective absorption of the CNT-SA is
around 1% corresponding to Fig. 1(c). Once the pump power is
higher than 230 mW, multi-pulse mode-locking is observed.
Under the boundary state, the maximum output power under
single-pulse mode-locking is measured as 1.21 mW. Although
the output power is at mW level, it can be further amplified by
an external amplification system.
We have noted that the threshold pump power is relatively
higher than that of the CNT-mode-locked fiber lasers with
lower repetition rates,
28)
the main reason may be the less net
gain of the laser cavity and small MD of the CNT-SA. The
short active fiber needs more power to pump to reach the
saturation intensity, and the loss induced by the coupling
between fiber and TIWDM will cause an extra pump
consumption. The small MD would induce a higher self-
starting threshold.
32)
However, compared to lasers based on
NPE scheme, the mode-locking pump power threshold is
much lower. For examples, the thresholds are 650 mW in
Ref. 22, 620 mW in Ref. 25, and 2 W in Ref. 23, respectively.
To evaluate the amplitude stability of the pulse train, the
RIN of the free-running laser is measured. The RIN power
spectral density (PSD) in the Fourier frequency range from
100 Hz to 1 MHz is shown in Fig. 4. The RIN-PSD maintains
flat until 200 kHz, and then meets a hump located around
250 kHz, which could be the resonant relaxation oscillation
frequency. The relaxation oscillation is caused by the finite
upper-state lifetime of Er
3+
ions. Below the relaxation
oscillation frequency of 250 kHz, the noise is introduced
from amplified spontaneous emission (ASE) noise and pump
noise. Above 250 kHz, the RIN-PSD drops dramatically and
goes to the quantum noise limit.
33)
Across the frequency
range of [100 Hz–1 MHz], the integrated root mean square
(rms) RIN is measured as 0.058%, which is comparable to
that of the lasers with lower repetition rate,
27)
showing a good
output stability. Due to the limitation of equipment, we are
unable to measure the free-running timing jitter.
The free-running laser operates stably without any polar-
ization controllers owing to the ultrashort cavity. The 6 h
operation stability of the laser is evaluated under room
temperature (22 ± 1 °C) without active protection stabiliza-
tion. In the test period, seen from Fig. 5(a), there is no
obvious change in the optical spectrum for both center
wavelength and the border region, showing the robust
mode-locking under free-running mode. We inject a high
pump power and record the fluctuation of the laser’s output
power. As shown in Fig. 5(b), under the pump power of
200 mW, the laser’s output power is measured as 1.01 mW,
the total power variation is less than 14 μW in 6 h, and the
rms fluctuation is calculated as 0.22%. The periodic pulsation
may result from the temperature change caused by the air
conditioner. The repetition rate drift and the corresponding
Allan deviation are shown in Figs. 5(c) and 5(d),
Fig. 3. The relationship between the output power with injected pump
power and the corresponding operation regimes.
Fig. 4. RIN-PSD and the integrated rms RIN of the laser output in the
Fourier frequency of [100 Hz–1 MHz].
(a)
(b)
(c)
(d)
Fig. 2. Output performance of the proposed laser when the pump power is
150 mW: (a) optical spectrum; (b) AC trace; (c) pulse train and (d) RF
spectrum.
062001-3 ©2024 The Author(s). Published on behalf of
The Japan Society of Applied Physics by IOP Publishing Ltd
Appl. Phys. Express 17, 062001 (2024) M. Dai et al.

respectively. A frequency drift of less than 2 kHz with a
mean frequency of 783.019 MHz and a standard deviation
(SD) of 514.653 Hz is obtained. The small jumping changes
at around 2, 3 and 5.5 h in the frequency drift spectrum may
be caused by environmental disturbance, such as opening/
closing the door, or the vibration induced by people walking
near the laser. The Allan deviation plot shows that, within the
average time of 1000 s, the deviation is significantly de-
creased with a slope of τ
−0.92
. The minimum deviation of
107 mHz is achieved at 2000 s gate time. After this point, the
deviation turns to increase, attributed to the environmental
disturbance. Although the free-running laser shows good
stability and robustness, we will further consider fully
stabilizing it for subsequent high-precision applications.
In conclusion, we have demonstrated a compact and robust
CNT-mode-locked Er-doped all-fiber laser delivering 623 fs
pulses at center wavelength of 1562 nm, with a fundamental
repetition rate up to 783 MHz. To the best of our knowledge, this
is the highest fundamental repetition rate reported from all-fiber
ring lasers. The all-active-fiber cavity design shortens the cavity
length as much as possible, and a CNT-SA on fiber ferrule is
adopted for laser mode-locking with low-threshold self-starting.
The repetition rate can be boosted to over GHz using such a
compact laser configuration by further shortening the fiber length
and decreasing the device size.
Acknowledgments This research is funded by Japan Society for the
Promotion of Science (22H00209, 23H00174); and Core Research for Evolutional
Science and Technology (JPMJCR1872). We also thank Mr. Hideru Sato and Mr.
Raymond Chen for their kind personal donation to this research work.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not
publicly available at this time but may be obtained from the
authors upon reasonable request.
1) T. Fortier and E. Baumann, Commun. Phys. 2, 153 (2019).
2) S. Xing, D. M.-B. Lesko, T. Umeki, A. J. Lind, N. Hoghooghi, T.-H. Wu,
and S. A. Diddams, APL Photonics 6, 086110 (2021).
3) N. Picque and T. W. Haensch, Nat. Photonics 13, 146 (2019).
4) H. Kalayciog, P. Elahi, O. Akcaalan, and F. O. Ilday, IEEE J. Sel. Top.
Quantum Electron. 24, 8800312 (2018).
5) K. Sugioka and Y. Cheng, Light-Sci. Appl. 3, e149 (2014).
6) N. Ji, J. C. Magee, and E. Betzig, Nat. Methods 5, 197 (2008).
7) J. Qin, R. Dai, Y. Li, Y. Meng, Y. Xu, S. Zhu, and F. Wang, Opt. Express
26, 25769 (2018).
8) C. M. Wu and N. K. Dutta, IEEE J. Quantum Electron. 36, 145 (2000).
9) Z. Zhao, L. Jin, S. Y. Set, and S. Yamashita, Opt. Lett. 46, 3621 (2021).
10) S.-S. Jyu, L.-G. Yang, C.-Y. Wong, C.-H. Yeh, C.-W. Chow, H.-K. Tsang,
and Y. Lai, IEEE Photonics J. 5, 1502107 (2013).
11) X. Cao, J. Zhou, Z. Cheng, S. Li, and Y. Feng, Laser Photonics Rev. 17,
2300537 (2023).
12) R. V. Gumenyuk, D. A. Korobko, and I. O. Zolotovskii, Opt. Lett. 45, 184
(2020).
13) A. Martinez and S. Yamashita, Appl. Phys. Lett. 101, 041118 (2012).
14) A. Martinez and S. Yamashita, Opt. Express 19, 6155 (2011).
15) X. Gao, Z. Zhao, Z. Cong, G. Gao, A. Zhang, H. Guo, G. Yao, and Z. Liu,
Opt. Express 29, 9021 (2021).
16) G. Liu, X. Jiang, A. Wang, G. Chang, F. Kaertner, and Z. Zhang, Opt.
Express 26, 26003 (2018).
17) C. Li, Y. Ma, X. Gao, F. Niu, T. Jiang, A. Wang, and Z. Zhang, Appl. Opt.
54, 8350 (2015).
18) Y. Wang, R. Yang, Z. Chen, D. Pan, B. Luo, Z. Zhang, and J. Chen,
Technical Digest SeriesFrontiers in Optics +Laser Science 2023 (FiO, LS)
2023 (Optica Publishing Group) FD2.5, 10.1364/FIO.2023.FD2.5.
19) R. Yang, M. Zhao, X. Jin, Q. Li, Z. Chen, A. Wang, and Z. Zhang, Optica 9,
874 (2022).
20) S. Yamashita, Y. Inoue, K. Hsu, T. Kotake, H. Yaguchi, D. Tanaka,
M. Jablonski, and S. Y. Set, IEEE Photonics Technol. Lett. 17, 750 (2005).
21) R. Thapa, N. Dan, J. Zong, and A. Chavez-Pirson, Opt. Lett. 39, 1418
(2014).
22) T. Yang, H. Huang, X. Yuan, X. Wei, X. He, S. Mo, H. Deng, and Z. Yang,
Appl. Phys. Express 6, 052702 (2013).
23) J. Zhang, Z. Kong, Y. Liu, A. Wang, and Z. Zhang, Photonics Res. 4,27
(2016).
24) Z. Chen, R. Yang, D. Pan, Y. Wang, B. Luo, J. Chen, and Z. Zhang,
Technical Digest SeriesFrontiers in Optics +Laser Science 2023 (FiO, LS)
2023 (Optica Publishing Group) JTu4A.70, 10.1364/FIO.2023.JTu4A.70.
25) W. Du, H. Xia, H. Li, C. Liu, P. Wang, and Y. Liu, Appl. Opt. 56, 2504
(2017).
26) J. W. Nicholson and D. J. DiGiovanni, IEEE Photonics Technol. Lett. 20,
2123 (2008).
27) C. Kim et al., Opt. Express 23, 26234 (2015).
28) J. Sotor, G. Sobon, J. Jagiello, L. Lipinska, and K. M. Abramski, J. Appl.
Phys. 117, 133103 (2015).
29) M. Dai, B. Liu, G. Ye, T. Shirahata, Y. Ma, N. Yamaguchi, S. Y. Set, and
S. Yamashita, Opt. Laser Technol. 176, 111002 (2024).
30) G. Sobon, A. Duzynska, M. Swiniarski, J. Judek, J. Sotor, and M. Zdrojek,
Sci. Rep. 7, 45491 (2017).
31) Y. Ma et al., Opt. Express 27, 14487 (2019).
32) S. Liu, Y. Chen, L. Huang, T. Cao, X. Qin, H. Ning, J. Yan, K. Hu, Z. Guo,
and J. Peng, Opt. Lett. 46, 2376 (2021).
33) J. Kim and Y. Song, Adv. Opt. Photonics 8, 465 (2016).
Fig. 5. Operation stability of the free-running fiber laser in 6 h: (a) optical
spectrum; (b) output power fluctuation under 200 mW pump power; (c)
repetition rate drift; (d) Allan deviation corresponding to (c).
062001-4 ©2024 The Author(s). Published on behalf of
The Japan Society of Applied Physics by IOP Publishing Ltd
Appl. Phys. Express 17, 062001 (2024) M. Dai et al.