Content uploaded by Siqi Liu
Author content
All content in this area was uploaded by Siqi Liu on Jun 07, 2022
Content may be subject to copyright.
JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 39, NO. 18, SEPTEMBER 15, 2021 5783
Tbit/s Multi-Dimensional Multiplexing
THz-Over-Fiber for 6G Wireless Communication
Hongqi Zhang , Lu Zhang , Member, IEEE,ShiweiWang , Zijie Lu , Zuomin Yang, Siqi Liu, Mengyao Qiao ,
Yuqian He , Xiaodan Pang , Senior Member, IEEE, Xianmin Zhang , and Xianbin Yu , Senior Member, IEEE
Abstract—Photonics-aided terahertz (THz) wireless systems
have been progressively developed to accommodate the forthcom-
ing wireless communication with extremely high data rates in
recent years. However, restrained by the obtainable signal-to-noise
ratio (SNR) and the dimensions explored in THz photonic wire-
less systems, achieving data rates of Tbit/s and beyond is still
challenging. In this paper, we present a multi-dimensional mul-
tiplexing Tbit/s THz-over-Fiber wireless communication system,
by efficiently benefiting from the multiplexing gain in both optical
wavelength and space domains. Enabled by a combined routine
of an optical frequency comb, a low inter-core crosstalk (IC-XT)
multi-core fiber and advanced digital signal processing, a line rate
of up to 1176 Gbit/s over a wireless distance of 10 m in the 350 GHz
band is experimentally demonstrated without any THz amplifiers,
resulting in a net data rate of up to 1059 Gbit/s. To the best of our
knowledge, this is the first time that beyond Tbit/s wireless data rate
is successfully achieved in the THz region above 300 GHz, making
a significant contribution to the development of THz-over-Fiber
systems for the sixth generation (6G) wireless communication.
Index Terms—Multi-core fiber, multi-dimensional multiplexing,
sixth generation, terahertz photonics, terahertz communication.
I. INTRODUCTION
OVER the past decades, the explosive growth of data traffic
has been well-witnessed in the wireless communication
networks. Currently, millimeter-wave band (30-300 GHz) has
Manuscript received April 9, 2021; revised June 8, 2021 and June 26, 2021;
accepted June 26, 2021. Date of publication June 30, 2021; date of current version
September 18, 2021. This work is supported by the National Key Research
and Development Program of China (2020YFB1805700), in part by the Natural
National Science Foundation of China under Grant 61771424, in part by the Nat-
ural Science Foundation of Zhejiang Province under Grant LZ18F010001 and
LQ21F010015, in part by the Swedish Research Council 2019-05197_VR, State
Key Laboratory of Advanced Optical Communication Systems and Networks of
Shanghai Jiao Tong University, in part by the Fundamental Research Funds for
the Central Universities 2020QNA5012 and Zhejiang Lab (no. 2020LC0AD01).
(Corresponding authors: Xianbin Yu; Lu Zhang.)
Hongqi Zhang, Shiwei Wang, Zijie Lu, Zuomin Yang, Siqi Liu,
Mengyao Qiao, Yuqian He, and Xianmin Zhang are with the College
of Information Science and Electronic Engineering, Zhejiang University,
Hangzhou 310027, China (e-mail: zhanghongqi@zju.edu.cn; wsw@zju.edu.cn;
luzijie@zju.edu.cn; yangzuomin@zju.edu.cn; 22031015@zju.edu.cn;
qiaomy@zju.edu.cn; heyuqian139@qq.com; zhangxm@zju.edu.cn).
Lu Zhang and Xianbin Yu are with the College of Information Science
and Electronic Engineering, Zhejiang University, Hangzhou 310027, Chi-
naZhejiang Lab, Hangzhou 310000, China (e-mail: zhanglu1993@zju.edu.cn;
xyu@zju.edu.cn).
Xiaodan Pang is with the Applied Physics Department, KTH Royal Institute
of Technology, Kista 164 40, Sweden (e-mail: xiaodan@kth.se).
Color versions of one or more figures in this article are available at https:
//doi.org/10.1109/JLT.2021.3093628.
Digital Object Identifier 10.1109/JLT.2021.3093628
been adopted in 5G communication networks, which is expected
to reach the data rates of 10 Gbit/s and continue boosting. Ac-
cording to the Cisco report [1], the number of Internet users are
predicted to exceed 5 billion by 2023. Besides, both wireless and
mobile data rates are anticipated to boost threefold from 2018
to 2023. To sustain this growth trend, wireless data transmission
with beyond 100 Gbit/s and eventually Tbit/s is of vital essence.
However, such high data rates are still difficult to be obtained
based on current millimeter-wave communications since the
total consecutive available bandwidth of that is less than 10
GHz. The terahertz (THz) band (0.3-10 THz) featuring very
large bandwidth is hereby recognized as a promising candidate
to accommodate the ever-increasing data traffic for 6G wireless
networks [2],[3].
In recent years, THz wireless communication systems have
evolved rapidly, and photonics-aided THz communication links
have so far achieved some impressive results, with a special ded-
ication in driving high capacity [4]–[14]. Benefit from the large
available bandwidth of optoelectronic devices and low propaga-
tion loss of optical fibers, photonics-aided THz communication
systems can not only enhance the data rates of wireless transmis-
sion, but also bridge the wireless frontends with optical access
networks seamlessly. For example, by utilizing a single optical
wavelength and spectrally efficient modulation formats, pho-
tonic wireless transmission of 100 Gbit/s at 280 GHz, 106 Gbit/s
at 400 GHz and 128 Gbit/s at 300 GHz over a distance of
0.5 m have been experimentally demonstrated [4]–[6]. To further
enable the data rates exceeding 100 Gbit/s, some multiplexing
techniques such as wavelength division multiplexing (WDM)
and polarization multiplexing (PDM) have been individually
employed in sub-THz/THz wireless communication systems.
For instance, a 100 GHz communication link with 108 Gbit/s
based on PDM [13], a data rate of 260 Gbit/s at 400 GHz
based on WDM method [14] have been achieved. Please note
WDM is usually translated to frequency division multiplexing
(FDM) in the wireless domain by photomixing-based frequency
conversion. These single dimensional multiplexing schemes
have indeed significantly contributed to the development of
high-speed THz communications, however not sufficient to
support wireless data rates approaching Tbit/s and beyond,
which is not yet achieved in the THz region above 300 GHz.
This is mainly limited by the obtainable signal-to-noise ratio
(SNR) after hybrid photonic wireless transmission, particularly
caused by atmospheric propagation loss in the THz frequency
region.
0733-8724 © 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.
5784 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 39, NO. 18, SEPTEMBER 15, 2021
Fig. 1. A typical application scenario based on THz-over-multicore-fiber
technology. BS: base station.
Here, we propose and experimentally demonstrate a multi-
dimensional multiplexing scheme aiming to deliver Tbit/s and
beyond data rates by exploring the multiplexing gain in both
optical wavelength and space domains. In our system, a cutting-
edge uni-traveling carrier photodiode (UTC-PD) [14] featur-
ing fast response time and high output saturation current is
employed for generating THz signals, and a high sensitivity
Schottky diode heterodyne mixer is used to down-convert the
THz signal. The transparency of heterodyne photo-mixing in the
generation scheme makes the dimensions available in the optical
domain beneficial for the wireless. Combining an optical fre-
quency comb (OFC), 16-ary quadrature amplitude modulation
format (16-QAM), a low inter-core crosstalk (IC-XT) multi-core
fiber and advanced digital signal processing (DSP) routine, we
demonstrate an aggregated net data rate of up to 1059 Gbit/s
over a wireless distance of 10 m in the 350 GHz band. The
signal bandwidth in each core is within that of THz receiver for
simultaneous reception, and the overall data rate beyond 1 Tbit/s
is the highest data rate in the THz band above 300 GHz, which is
expected to open a door towards Tbit/s for 6G wireless networks.
One of the typical application scenarios envisioned utilizing
THz-over-Fiber technology is shown in Fig. 1. A central office
is deployed to centrally implement such as baseband processing
and optical generation of signals, and then the optical signals are
coupled into multi-core fibers via fan-in modules. After trans-
mission through the multi-core fibers, the signals are extracted
by fan-out modules core-by-core, which can be then distributed
over the single mode fiber links and destined to endusers in
such as the remote and rural areas, urban areas, suburbs, and
so on. Here a multi-core fiber-based link is considered as the
ultra-high-speed infrastructure in future optical fiber networks
providing an additional space dimension infrastructure. In this
manner, the THz wireless and optical fiber access network
architecture are seamlessly connected, which can not only boost
the capacity of wireless access, but also provide great network
resilience and flexibility.
II. THZMULTI-DIMENSIONAL MULTIPLEXING WIRELESS
COMMUNICATION LINK
A. Experimental Setup
The conceptual multi-dimensional multiplexing THz-over-
Fiber wireless communication system is displayed in Fig. 2(a).
Here, a wavelength demultiplexer is employed to line-by-line
separate OFC frequency components for subsequent WDM
strategy. After data modulation, the modulated signals are group-
combined via wavelength multiplexers, which are then individ-
ually launched into each core of multi-core fiber by a fan-in
module. After the transmission through multi-core fiber and
extraction via a fan-out module, the output signals from each
core are then connected to a base station (BS) to wirelessly cover
one area, where an unmodulated light and a THz transmitter
are employed to implement photomixing-based opto-electronic
conversion and wireless radiation.
Fig. 2(b) depicts the experimental configuration of multi-
dimensional multiplexing THz-over-Fiber wireless transmission
with Tbit/s data rate. Firstly, we use an external cavity laser
(ECL1, <100 kHz linewidth) to generate a 1550 nm continuous
wave (CW) light. Then the light is launched into a polarization
controller (PC1) and a phase modulator (PM), where the PM
is driven by an amplified RF sinusoidal signal. The sinusoidal
signal is generated from a vector signal generator (Keysight,
E8267D) with an output power of 12 dBm at 15 GHz, and then
is amplified by a RF amplifier with 18 dB gain before driving the
PM. The PC1 is employed to adjust the polarization state of the
incident light from the ECL1, to maximize the polarization de-
pendent modulation efficiency of the PM. The optical spectra of
OFC after the PM is shown in Fig. 3(a), measured with a high res-
olution optical complex spectrum analyzer (APEX, AP2683B).
The resolution is set to 0.8 pm (equivalent to 100 MHz). The
center wavelength of the OFC is 1550 nm and the wavelength
space between the optical comb lines is 15 GHz. From the
Fig. 3(a), the carrier-to-noise ratio of the middle 3-line is the
largest, which is more than 40 dB. Therefore, three optical comb
lines, at 1549.88 nm, 1550 nm and 1550.13 nm, are selected
and filtered out by a programmable wavelength selective switch
(WSS, FINISAR, WaveShaper 4000A) for the WDM. After the
amplification by an Erbium-doped fiber amplifier (EDFA1), the
three selected optical comb lines are launched into an in-phase
and quadrature optical modulator (IQ-MOD) to implement the
complex digital baseband modulation (16-QAM), where the
PC2 is used to optimize the polarization state of three optical
carriers. Here an arbitrary waveform generator (AWG, Keysight,
M8195A) with 65 GSa/s sampling rates is employed to generate
a pseudo-random binary sequence (PRBS) with a word length
of 215-1 for user data, and its baud rates is adjustable. The output
voltage of the AWG is set to 80 mV, which is amplified by two
electrical amplifiers (RFA2, RFA3) with 26 dB gain for driving
the IQ-MOD. The three modulated optical signals are then
divided into seven copies by a 1 ×7 splitter and coupled into a
piece of 7-core fiber through a fan-in module. In the experiment,
a low IC-XT 7-core fiber with a length of 1 km is employed
to transmit high-speed communication signals. Compared with
standard single-core fiber transmission, this 7-core fiber can
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.
ZHANG et al.: TBIT/S MULTI-DIMENSIONAL MULTIPLEXING THZ-OVER-FIBER FOR 6G WIRELESS COMMUNICATION 5785
Fig. 2. (a) The conceptual multi-dimensional multiplexing THz-over-Fiber wireless communication system. (b) Experimental configuration of Tbit/s multi-
dimensional multiplexing photonic wireless transmission link. OFC: optical frequency comb; MOD: modulation; MCF: multi-core fiber; Tx: transmitter; Rx:
receiver; ECL: external cavity laser; PC: polarization controller; PM: phase modulator; RF: radio frequency; RFA: RF amplifier; WSS: wavelength selective
switch; EDFA: erbium-doped fiber amplifier; AWG: arbitrary waveform generator; IQ-MOD: in-phase and quadrature modulator; VOA: variable optical attenuator;
UTC-PD: uni-traveling carrier photodiode; DSO: digital storage oscilloscope; LO: local oscillator.
transmit multiple optical signals simultaneously in a single fiber
with space domain multiplexing, which significantly boost the
communication capacity.
The cross-section view of the 7-core fiber is shown in Fig. 4(a).
For the light beams centered at 1550 nm, the attenuation is
around 0.25 dB/km each core. The average core pitch and
cladding diameter of the 7-core fiber are 41.5 µm and 8 µm,
respectively. Besides, the chromatic dispersion and the IC-XT
are around 17.1 ps/nm/km and −50 dB/100 km, respectively. In
addition, the fan-in/fan-out modules shown in Fig. 4(b) hold a
crosstalk between adjacent cores of −55 dB and insertion loss
of around 1 dB, respectively, to establish a low-loss and reliable
connection between the splitter and the 7-core fiber. Data decor-
relations in the spatial channels are performed by the optical
delay and dispersion. After 1 km 7-core fiber transmission, the
output signal from each core is independently boosted by another
EDFA2, and filtered by an optical band-pass filter to suppress
the out-of-band amplified spontaneous emission (ASE) noise, as
well as polarization controlled. We perform this implementation
core by core in the experiment for all cores while all signals keep
running. Subsequently, the filtered optical signal is coupled with
a LO light centered at 1553.89 nm from the ECL2 (<100 kHz
linewidth) by an optical coupler. The output optical signal from
the coupler passes through the PC4 and a polarizer to align the
polarization to maximize the responsivity of a broadband UTC-
PD (NTT Electronics Corp. IOD-PMJ-13001). A polarization
maintaining variable attenuator (VOA) is employed to adjust
the input optical power of UTC-PD. Later, the combined optical
signals are fed into the UTC-PD to generate THz signals, whose
carrier frequency is determined by the frequency difference
between two optical carriers of ECL1 and ECL2.
Fig. 3(b) depicts the optical spectra at the input of the UTC-
PD. The center wavelengths of modulated and unmodulated
laser beams are 1550 nm and 1553.89 nm, respectively, with
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.
5786 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 39, NO. 18, SEPTEMBER 15, 2021
Fig. 3. (a) Optical spectra of optical frequency comb after the PM. (b) Optical spectra at the input of the UTC-PD.
Fig. 4. (a) The cross-section view of the 1 km 7-core fiber. (b) The fabricated
fan-in/fan-out modules.
a frequency difference of 350 GHz. The modulated optical path
consists of three wavelengths with an equal spacing of 15 GHz.
When two laser beams are fed into the UTC-PD and photo-
mixed, three signals with corresponding different frequencies
in the THz region are generated, the frequency of which are
located at 365 GHz, 350 GHz and 335 GHz, respectively, named
channel_1 (Ch_1), channel_2 (Ch_2) and channel_3 (Ch_3),
respectively. Afterwards, the THz signals from the UTC-PD are
radiated into a 10 m line-of-sight (LOS) wireless link via a horn
antenna. Within the free space, a pair of THz lenses with 25 dBi
gain are employed to collimate the THz beam to reduce the
propagation path loss.
At the receiver side, a Schottky mixer driven by a 12-order fre-
quency multiplied electrical LO signal is used to down-convert
the received THz signals into the intermediate frequency (IF)
domain. The LO signal is generated from an analog signal gener-
ator (Keysight, E8257D), the frequency and amplitude of which
are at 372.5 GHz and 0 dBm, respectively. After mixing with
the LO signal, three IF signals centered at 7.5 GHz, 22.5 GHz
and 37.5 GHz are generated. Subsequently, the IF signals are
amplified by an electrical amplifier (RFA4) with 22 dB gain, and
then sampled by a real-time digital storage oscilloscope (DSO,
Keysight, DSOZ594A) with 160 GSa/s sampling rates. Finally,
the three IF signals are demodulated separately through offline
DSP algorithms in the digital domain.
B. Digital Signal Processing (DSP) Routine
At the transmitter side, a pseudo-random binary sequence
(PRBS) with a word length of 215-1 is used as the user data.
Then, the 16-QAM modulation and pulse shaping are employed
to generate the transmitted samples. Here a squared root raised
cosine (SRRC) function is used for pulse shaping. At the receiver
side, the received signal is first down-converted to the baseband
with a digital mixer, which consists of a digital down-converter,
a digital current block and a low-pass filter [16]. Then, the IQ
imbalance of the baseband signal is compensated with Gram-
Schmidt algorithm [17]. After this, a time recovery algorithm,
which consists of the Gardner resampling and decimator, is
used to resample the signal to the original baud rate. Following
the time recovery, a linear equalization is employed based on
Multi-Modulus-Algorithm (MMA) [18]. Then the frequency
offset is compensated by Viterbi & Viterbi algorithm [19] and the
phase noise is compensated by a blind phase search method [20].
Finally, the error vector magnitude (EVM) performance between
the received signals and the baseband signals is evaluated, and
the bit-error-rate (BER) performance is then estimated based on
it [21][22]. In our experiment, the threshold of SD-FEC with
20% overhead is 2.7 ×10−2, and the threshold of HD-FEC with
7% overhead is 4.5 ×10−3. The parameter optimization in DSP
routine is shown in Fig. 5.
The optimization of DSP algorithms is performed step by step
following the DSP routine in the digital domain. The bandwidth
of a low-pass filter is firstly optimized in terms of transmission
BER performance. In the case of 14 Gbaud, a bandwidth of
14.7 GHz filter ensures the best performance, as shown in
Fig. 5(a). After this, we optimize the tap number of the linear
equalization to 70, as shown in Fig. 5(b). Subsequently, the
iteration number of the linear equalizer and the phase angle
resolution of phase noise compensator are also optimized, as
shown in Fig. 5(c) and Fig. 5(d). As the number of iterations
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.
ZHANG et al.: TBIT/S MULTI-DIMENSIONAL MULTIPLEXING THZ-OVER-FIBER FOR 6G WIRELESS COMMUNICATION 5787
Fig. 5. The optimization of parameters in digital signal processing. (a) Bandwidth optimization of low-pass filter, (b) optimization of number of linear equalization
taps, (c) optimization of linear equalizer iterations, (d) optimization of phase angle resolution.
increases and the phase angle resolution decreases, the BER
performance becomes better. However, the BER performance is
basically stable when the number of iteration exceeds 3 and the
phase angle resolution less than 0.0245 rad. When the number
of iterations further increases and the phase angle resolution is
further reduced, the complexity of the algorithm and the com-
puting time increases significantly. By considering this trade-off,
the iteration number and phase angle resolution are set as 3 and
0.0245 rad, respectively.
III. EXPERIMENTAL RESULTS AND DISCUSSIONS
A. BER Versus Optical Power
In the experiment, all 7-core duplicate the optical signals.
Three-wavelength is all modulated by 16-QAM baseband for
each core transmission. Fig. 6 presents the measured BER
performance of overall 21 channels after 1 km 7-core fiber and
10 m wireless transmission. The baud rates are set to 14 Gbaud.
From the Fig. 5, we can observe that the BER performance of
each individual channel becomes better as the optical power
increases, and for the same channel within different cores, the
BER performance is constant forming a cluster with ignorable
penalty. All the BER performance of 21 channels can reach be-
low the soft decision forward-error-correction (SD-FEC) thresh-
old with 20% overhead [23]. In particular, the BER performance
of for Ch_1 and Ch_2 in all cores can reach below the hard
decision forward-error-correction (HD-FEC) threshold with 7%
overhead [24], [25]. In this case, each core carries a data rate of
Fig. 6. The measured BER performance of overall 21 channels after 7-core
fiber and 10 m wireless transmission.
14 Gbaud ×4 bit/s/Hz ×3 channels =168 Gbit/s, resulting an
aggregate line data rate of 1176 Gbit/s (168 Gbit/s ×7) in the
7-core fiber. Taking into account the error-correction overhead, a
net data rate of 1059 Gbit/s after hybrid fiber wireless transmis-
sion is achieved. We can observe that the BER performance of
the Ch_1 is worse than that of the Ch_2, which is mainly caused
by frequency-dependent noise figure of IF amplifier (RFA4) and
correlative interference from neighbor channels. Besides, due to
the limitation of the Schottky mixer (∼40 GHz IF bandwidth),
the Ch_3 with modulation exceeds the 3 dB bandwidth range.
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.
5788 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 39, NO. 18, SEPTEMBER 15, 2021
Fig. 7. The measured 3-channel BER performance of core_5 versus baud rates
when the optical power is 13 dBm.
Consequently, the SNR of Ch_3 degrades dramatically, and it is
with the worst performance.
B. BER Versus Baud Rates
In the experiment, we also demonstrate 16-QAM THz wire-
less transmission at different baud rates. The optical power
launched into the UTC-PD is 13 dBm. The measured BER
performance of the core_5 for 3-channel, as an example, is
depicted in Fig. 7 with baud rates of 2 Gbaud, 4 Gbaud, 6 Gbaud,
8 Gbaud, 10 Gbaud, 12 Gbaud and 14 Gbaud. When the baud
rates are 14 Gbaud, the constellation of 3 channels in core_5
is shown on the right side of Fig. 7 and the BER values for 3
channels are 4.2 ×10−3,1.5×10−3and 6 ×10−3, respectively.
From the Fig. 7, we can see that the BER performance becomes
worse as the baud rate increases, this is due to the reduction
of electrical SNR in the case of fixed power, as well as less
guardband between neighbor channels. When the baud rates are
lower than 10 Gbaud, the BER performance of all 3-channel can
reach below the HD-FEC.
C. System Stability Test
In addition, we evaluate the stability of BER performance
in THz wireless communication link at the highest baud rates
(14 Gbaud) in the lab environment. The optical power is set
to 13 dBm. Fig. 8 shows the measured results for selective
3 channels - core_1, core_3 and core_5. For each channel of
core_1, core_3 and core_5, we run the link continuously over
1 hour, collecting 300 traces per channel and recording the
value of BER. From the Fig. 8, the BER performance of each
channel in core_1, core_3 and core_5 is maintained within a
small fluctuation range. We also see that 3 channels individually
possess stable performance forming a cluster over 300 measured
traces, while features different performance channel by channel,
as observed in Fig. 6. The measured results indicate that the THz
wireless communication systems can operate stably with data
rates exceeding Tbit/s over a long time.
Fig. 8. System stability of 3-channel in core_1, core_3 and core_5.
Fig. 9. The measured wireless distance dependent BER performance of
3-channel in core_1.
D. BER Versus Wireless Distance
Finally, we measure the BER performance at different wire-
less distances as well. The optical power and band rates are
set to 13 dBm and 14 Gbaud, respectively in the experiment.
Fig. 9 shows the measured BER performance of 3 channels over
core_1, as an example, and wireless transmission distance of 6 m,
8 m, 10 m, respectively. When the wireless distance is 10 m, the
constellation of 3 channels in core_1 is shown on the right side
of Fig. 9 and the BER values for 3 channels are 4.1 ×10−3,
1.4 ×10−3and 5.7 ×10−3, respectively. From the Fig. 9, we
can see the BER performance slightly degrades as the wireless
transmission distance increases from 6 m to 8 m. Meanwhile, the
BER performance variation can be also observed when distance
further increases to 10 m. This is mainly caused by the alignment
accuracy variation of the THz transmitter and receiver.
IV. CONCLUSION
In summary, we propose and experimentally demonstrate a
multi-dimensional multiplexing THz-over-Fiber wireless com-
munication system targeting an extremely large capacity of
Tbit/s. Based on the combination of the OFC for WDM, the
multi-core fiber for SDM, spectrally efficient modulation format
for amplitude and phase multiplexing, as well as the advanced
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.
ZHANG et al.: TBIT/S MULTI-DIMENSIONAL MULTIPLEXING THZ-OVER-FIBER FOR 6G WIRELESS COMMUNICATION 5789
DSP algorithm, successful hybrid transmission of a line data rate
of up to 1176 Gbit/s over 1 km fiber and a wireless distance of
10 m in the 350 GHz is achieved. Amongst 3 FDM channels,
the BER of one after transmission is below the SD-FEC and
two is below the the HD-FEC thresholds, resulting in a net data
rate of 1059 Gbit/s. This is the largest capacity ever-reported
in the THz region above 300 GHz, to our best knowledge. The
proposed THz communication system is expected to pave the
way towards the evolution for Tbit/s wireless communication,
and boost the evolution and implementation of the 6G wireless
communication networks.
ACKNOWLEDGMENT
The authors would like to thank Dr. B. Wei at the Training
Platform of Information and Microelectronic Engineering at the
Polytechnic Institute of Zhejiang University, for his help in the
experiment.
REFERENCES
[1] “Cisco visual networking index: Forecast and methodology, 2018–2023,”
Cisco, Mar. 2020. [Online]. Available: https://www.cisco.com/c/
en/us/solutions/ collateral/serviceprovider/visual- networking-index-
vni/complete-white-paper-c11-481360.html
[2] V. Petrov, A. Pyattaev, D. Moltchanov, and Y. Koucheryavy, “Terahertz
band communications: Applications, research challenges, and standardiza-
tion activities,” in Proc. 8th Int. Congr. Ultra Modern Telecommun. Control
Syst. Workshops (ICUMT), Lisbon, Portugal, Oct. 2016, pp. 183–190.
[3] I. F. Akyildiz, J. M. Jornet, and C. Han, “Terahertz band: Next frontier
for wireless communications,” Phys. Commun-Amst., vol. 12, pp. 16–32,
Sep. 2014.
[4] V. K. Chinni et al., “Single-channel 100 Gbit/s transmission using III–V
UTC-PDs for future IEEE 802.15.3d wireless links in the 300 GHz band,”
Electron. Lett., vol. 54, no. 10, pp. 638–640, May 2018.
[5] S. Jia et al., “0.4 THz photonic-wireless link with 106 Gb/s single channel
bitrate,” J. Light. Technol., vol. 36, no. 2, pp. 610–616, Jan. 2018.
[6] C. Castro et al., “32 GBd 16QAM wireless transmission in the 300
GHz band using a PIN diode for THz upconversion,” in Proc. Opt.
Fiber Commun. Conf. (OFC), San Diego, CA, USA, Mar. 2019,
Paper M4F. 5.
[7] T. Nagatsuma et al., “Real-time 100-Gbit/s QPSK transmission using
photonics-based 300-GHz-band wireless link,” in Proc. IEEE Int. Topical
Meeting Microw. Photon. (MWP), Long Beach, CA, USA, Nov. 2016,
pp. 27–30.
[8] X. Yu et al., “160 Gbit/s photonics wireless transmission in the 300–500
GHz band,” APL Photon., vol. 1, no. 8, Nov. 2016, Art. no. 081301.
[9] H. Hamada et al., “300-GHz, 100-Gb/s InP-HEMT wireless transceiver
using a 300-GHz fundamental mixer,” in Proc. IEEE/MTT-S Int. Microw.
Symp. – IMS, Philadelphia, PA, USA, Jun. 2018, pp. 1480–1483.
[10] G. Ducournau et al., “Ultrawide-bandwidth single-channel 0.4-THz wire-
less link combining broadband quasi-optic photomixer and coherent detec-
tion,”IEEE Trans. THz Sci. Technol., vol. 4, no. 3, pp. 328–337, Mar. 2014.
[11] H. Shams et al., “100 Gb/s multicarrier THz wireless transmission system
with high frequency stability based on a gain-switched laser comb source,”
IEEE Photon. J., vol. 7, no. 3, pp. 1–11, Jun. 2015.
[12] K. Liu et al., “100 Gbit/s THz photonic wireless transmission in the 350-
GHz band with extended reach,” IEEE Photon. Technol. Lett., vol. 30,
no. 11, pp. 1064–1067, Apr. 2018.
[13] X. Li, J. Yu, J. Zhang, Z. Dong, F. Li, and N. Chi, “A 400G optical wireless
integration delivery system,” Opt. Exp.., vol. 21, no. 16, pp. 18812–18819,
Aug. 2013.
[14] X. Pang et al., “260 Gbit/s photonic-wireless link in the THz band,” in
Proc. IEEE Photon. Conf., Waikoloa, HI, Oct. 2016, Paper PD1-2.
[15] H. Ito, T. Furuta, S. Kodama, and T. Ishibashi, “InP/InGaAs uni-travelling-
carrier photodiode with 310 GHz bandwidth,” Electron. Lett., vol. 36,
no. 21, pp. 1809–1810, Oct. 2000.
[16] X. Pang et al., “25 Gbit/s QPSK hybrid fiber-wireless transmission in the
W-band (75–110 GHz) with remote antenna unit for in-building wireless
networks,” IEEE Photon. J., vol. 4, no. 3, pp. 691–698, Apr. 2012.
[17] I. Fatadin, S. J. Savory, and D. Ives, “Compensation of quadrature imbal-
ance in an optical QPSK coherent receiver,” IEEE Photon. Technol. Lett.,
vol. 20, no. 20, pp. 1733–1735, Sep. 2008.
[18] P. J. Winzer, A. H. Gnauck, C. R. Doerr, M. Magarini, and L. L.
Buhl, “Spectrally efficient long-haul optical networking using 112-Gb/s
polarization-multiplexed 16-QAM,” J. Light. Technol., vol. 28, no. 4,
pp. 547–556, Sep. 2010.
[19] I. Fatadin and S. J. Savory, “Compensation of frequency offset for 16-
QAM optical coherent systems using QPSK partitioning,” IEEE Photon.
Technol. Lett., vol. 23, no. 17, pp. 1246–1248, Jun. 2011.
[20] A. J. Viterbi and A. M. Viterbi, “Nonlinear estimation of PSK-modulated
carrier phase with application to burst digital transmission,” IEEE Trans.
Inf. Theory, vol. 29, no. 4, pp. 543–551, Jul. 1983.
[21] R. Schmogrow et al., “Error vector magnitude as a performance measure
for advanced modulation formats,” IEEE Photon. Technol. Lett., vol. 24,
no. 1, pp. 61–63, Oct. 2012.
[22] R. Schmogrow et al., “Corrections to ‘Error vector magnitude as a
performance measure for advanced modulation formats,”’ IEEE Photon.
Technol. Lett., vol. 24, no. 23, pp. 2198–2198, Nov. 2012.
[23] D. Chang et al., “LDPC convolutional codes using layered decod-
ing algorithm for high speed coherent optical transmission,” in Proc.
OFC/NFOEC, Los Angeles, CA, USA, Mar. 2012, pp. 1–3.
[24] F. Chang, K. Onohara, and T. Mizuochi, “Forward error correction for 100
g transport networks,” IEEE Commun. Mag., vol. 48, no. 3, pp. S48–S55,
Mar. 2010.
[25] S. Koenig et al., “Wireless sub-THz communication system with high data
rate,” Nat. Photon., vol. 7, no. 12, pp. 977–981, Oct. 2013.
Hongqi Zhang received the B.S. degree from Zhengzhou University,
Zhengzhou, China, in 2019. He is currently working toward the Ph.D. degree
with the College of Information Science and Electronic Engineering, Zhejiang
University, China. His current research focuses on THz wireless communication
technologies.
Lu Zhang (Member, IEEE) received the bachelor’s degree from Southeast
University, Nanjing, China, in 2014 and the Ph.D. degree from Shanghai
Jiao Tong University, Shanghai, China, in 2019. He is currently a Research
Associate Professor with the College of Information Science and Electronic
Engineering, Zhejiang University, Hangzhou, China. From 2016 to 2017, he
was a Visiting Ph.D. Student with the KTH Royal Institute of Technology,
Stockholm, Sweden, sponsored by China Scholarship Council. Since 2018,
he has been a Visiting Research Engineer with the KTH Royal Institute of
Technology and Kista High-Speed Transmission Lab of RISE AB. His research
interests include ultrafast THz communications, fiber-optic communications,
digital signal processing algorithms for optical, and THz transmission systems.
Xiaodan Pang (Senior Member, IEEE) received the M.Sc. degree from the
KTH Royal Institute of Technology, Stockholm, Sweden, in 2010 and the PhD
degree from DTU Fotonik, Technical University of Denmark, Kongens Lyngby,
Denmark, in 2013. From October 2013 to March 2017, he was a Postdoc with the
RISE Research Institutes of Sweden (former ACREO Swedish ICT), and March
2017 to February 2018, he was a Researcher with the KTH Optical Networks
Lab (ONLab). From March 2018 to February 2020, he was a Staff Opto Engineer
and a Marie Curie Research Fellow with the R&D Team, Infinera Global HW,
Stockholm, Sweden. Since March 2020, he has been a Senior Researcher with
the Department of Applied Physics, KTH Royal Institute of Technology. He
is the PI of a Swedish Research Council Starting Grant, the EU H2020 Marie
Curie Individual Fellowship Project NEWMAN, and a Swedish SRA ICT-TNG
Postdoc project. He has authored and coauthored more than 190 publications
in journals and conferences. His research focuses on ultrafast communications
with millimeter-wave, terahertz wave, free-space optics, and fiber-optics. He has
been a TPC Member of more 20 conferences, including OFC, GLOBECOM,
OECC, ACP, and CLEO-PR. He is a Senior Member of OSA, and a Board
Member of IEEE Photonics Society Sweden Chapter.
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.
5790 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 39, NO. 18, SEPTEMBER 15, 2021
Xianmin Zhang received the B.S. and Ph.D. degrees in physical electronics and
optoelectronics from Zhejiang University, Hangzhou, China, in 1987 and 1992,
respectively. He was appointed as an Associate Professor of information and
electronic engineering with Zhejiang University in 1994 and a Full Professor
in 1999. From November 1996 to September 1997 and from October 1997
to September 1998, he was a Research Fellow with the University of Tokyo,
Tokyo, Japan, and Hokkaido University, Sapporo, Japan. In 2007, he spent two
months with the Research Laboratory of Electronics, Massachusetts Institute
of Technology, Cambridge, MA, USA, as a Visiting Research Fellow. He is
currently the Chair with the Department of Information Science and Electronic
Engineering, Zhejiang University. His research interests include microwave
photonics, and electromagnetic wave theory and applications.
Xianbin Yu (Senior Member, IEEE) received the M.Sc. degree from Tianjin
University, Tianjin, China, in 2002, and the Ph.D. degree from Zhejiang Uni-
versity, Hangzhou, China, in 2005. From 2005 to 2007, he was a Postdoctoral
Researcher with Tsinghua University, Beijing, China. Since November 2007,
he has been with DTU Fotonik, Technical University of Denmark, Kongens
Lyngby, Denmark, where he became an Assistant Professor in 2009 and was
promoted to Senior Researcher in 2013. He is currently a Research Professor
with Zhejiang University. He has coauthored two book chapters and more than
180 peer-reviewed international journal and conference papers in his research
interests, which include microwave photonics and optical fiber communications.
He has given more than 30 invited international conference presentations and
was a Session Chair/TPC Member for a number of international conferences. His
current research interests include mm-wave/THz photonics and its applications,
THz communications, ultrafast photonic RF signal processing, and high-speed
photonic wireless access technologies.
Authorized licensed use limited to: Zhejiang University. Downloaded on June 07,2022 at 02:26:58 UTC from IEEE Xplore. Restrictions apply.