808
IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS, VOL. 14, NO. 3, MAY/JUNE 2008
20 ps Transition Time All-Optical SOA-Based
Flip-Flop Used for Photonic 10 Gb/s Switching
Operation Without Any Bit Loss
Antonio Malacarne, Jing Wang, Yuancheng Zhang, Abhirup Das Barman, Gianluca Berrettini,
Luca Potı̀, Member, IEEE, and Antonella Bogoni
Abstract—A novel scheme for integrable ultrafast all-optical flipflop is demonstrated. Transition times as low as 20 ps with a contrast ratio higher than 17.5 dB have been experimentally measured.
All-optical switching operation in a 2×2 spatial and wavelength
preserving switch is reported with a power penalty of about 1 dB.
The proposed solution exploits the fast falling edge provided by a
semiconductor optical amplifier (SOA) based optical flip-flop. Numerical investigations already demonstrated high extinction ratios
(>40 dB) and low switching energies (15.6 fJ) for integrated optical flip-flop. On the other hand, slow rising times, due to the cavity
length, intrinsically limit such configurations. By using SOA-based
logic gates, two flip-flop outputs are combined in a new bistable
signal. Both the new rising and falling edges are related to the
primary flip-flop falling edge. This way it is possible to eliminate
the intrinsic slow rising time that limits the flip-flop configuration
based on the coupled ring lasers, without excessively increasing
the complexity of the structure and maintaining a reasonably high
contrast ratio. Furthermore, the noise on the high level has been
improved due to the regenerative properties of the logic gates based
on cross-gain modulation and cross-phase modulation in a single
nonlinear SOA. Finally, flip-flop output has been used to drive a
2×2 all-optical spatial and wavelength preserving switch based
on SOAs. For cross/bar switch configurations, 10 Gb/s error-free
operation has been obtained without bit loss.
Index Terms—All-optical flip-flop, cross-gain modulation
(XGM), optical bistable devices, optical packet switching (OPS),
optical signal processing (OSP), semiconductor optical amplifier
(SOA).
Manuscript received November 2, 2007, revised December 11, 2007. This
work was supported by the BONE project (“Building the Future Optical Network in Europe”), a Network of Excellence funded by the European Commission
through the 7th ICT-Framework Programme.
A. Malacarne and G. Berrettini are with the Scuola Superiore Sant’Anna,
Centre of Excellence for Information and Communication Engineering
(CEIIC), Pisa 1-56124, Italy (e-mail: antonio.malacarne@cnit.it; abhirup_rp@
yahoo.com; gianluca.berrettini@cnit.it).
J. Wang and Y. Zhang are with the Department of Electronic Engineering,
Tsinghua University, Beijing 100084, China. They are also with the Photonics Networks National Laboratory, Consorzio Nazionale Interuniversitario
per le Telecomunicazioni (CNIT), 1-56124 Pisa, Italy (e-mail: jwangl983@
gmail.com; zhang-yc06@mails.tsinghua.edu.cn).
A. D. Barman and L. Potı̀ are with the Scuola Superiore Sant’Anna, Centre
of Excellence for Information and Communication Engineering (CEIIC), Pisa
1-56124, Italy. They are also with the Photonic Networks National Laboratory,
Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), 156124 Pisa, Italy (e-mail: antonio.malacarne@cnit.it; abhirup_rp@yahoo.com;
gianluca.berrettini@cnit.it).
A. Bogoni is with the Integrated Research Center for Photonic Networks
and Technologies, Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), 1-56124 Pisa, Italy (e-mail: gianluca.meloni@cnit.it; antonella.
bogoni@cnit.it).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/JSTQE.2008.918654
I. INTRODUCTION
LL-OPTICAL packet switching seems to be the most
promising way to take advantage of fiber bandwidth to
increase routers forwarding capacity, being able to achieve very
high data rate operations. All-optical flip-flops have been widely
investigated mainly because they can be exploited in all-optical
packet switches [1]–[5], where switching, routing, and forwarding are directly carried out in the optical domain. Some examples concerning optical packet switches are shown in [1], [3], [4],
and [5], where an optical flip-flop stores the switch control information and drives the switching operation. Former solutions for
all-optical flip-flops have been demonstrated exploiting discrete
devices [1] or erbium-doped fiber properties [6] that suffer from
slow switching times and high set/reset input powers. Several
integrated or integrable solutions [7], [8] present a switching
energy in the femtojoule range and switching times of tens of
picoseconds at the expenses of poor contrast ratios. On the other
hand in [9], an integrated scheme exhibiting a very high contrast
ratio value but with transition times in the nanosecond range is
reported. In any case, a tradeoff between contrast ratio and edges
speed must be found as a function of the flip-flop application.
Microresonators-based bistable element has been demonstrated [10] presenting high optical operating power, picojoule
switching energies and microsecond switching times, theoretically reducible down to the order of tens of picosecond. Making
a comparison with electronics, recent large-scale integration
(LSI) circuits [11] show switching energies of 1 fJ even though
with slower switching speeds.
In [1], a solution based on coupled ring lasers is proposed.
This solution offers a certain number of advantages: it can
provide high contrast ratios between states; there is no difference in the mechanisms for switching from state 1 to state 2
and vice versa, making symmetric set and reset operations;
it presents a large input light wavelength range and a controllable switching threshold. Moreover, considering an integrated version of this kind of flip-flop, through numerical analysis, a switching energy in the femtojoule range has been
demonstrated.
Here, we analyze the aforementioned scheme in order to investigate its intrinsic limitations and find out a possible solution
to overcome these drawbacks. As a matter of fact in an all-optical
flip-flop based on gain quenching/saturation as the flip-flop presented in [1], the main problem consists in a slow rising time
due to the laser building-up time, which can be analytically
and experimentally estimated to be five to ten times the cavity
A
1077-260X/$25.00 © 2008 IEEE
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.
MALACARNE et al.: 20 PS TRANSITION TIME ALL-OPTICAL SOA-BASED FLIP-FLOP USED
propagation time. Hence, in the case of discrete-device-based
implementation, the rising time can be higher than hundreds of
nanoseconds, whereas for integrated version it can be reduced
down to tens of picoseconds. On the other hand, the falling time
depends on the semiconductor optical amplifier (SOA) dynamics, and consequently, can be reduced down to tens of picoseconds, considering set and reset pulses with very fast edges.
In this paper, we present a solution able to exploit the fast
falling edge of the coupled-ring-laser-based flip-flops overcoming the limitation due to their slow rising edges. The use of two
bistable elements that have one fast and one slow transition,
together to create a bistable element with two fast transitions,
has some history [12]. In particular, our solution requires two
identical replicas of the aforementioned flip-flops, and it provides a photonic processing of the two output signals in order
to produce a new bistable output signal with very fast transition
times. The photonic processing is based on two identical logic
gates exploiting cross-gain modulation (XGM) and cross-phase
modulation (XPM) in SOA.
Finally, an SOA-based photonic switching operation is
demonstrated as well, utilizing the fast-dynamics flip-flop to
store the switching decision information and to drive a 2×2
spatial and wavelength preserving all-optical switch. Bit error
rate (BER) measurements show a low power penalty of about
1 dB, making this scheme promising in terms of scalability.
Moreover, flip-flop and switch schemes are completely SOAbased, giving the possibility for a whole integrated version.
This paper is organized as follows. In Section II, the advantages and the issues of the optical flip-flop scheme realized
through two coupled SOA-based ring lasers are experimentally
and analytically investigated. Section III describes the proposed
solution to overcome the aforementioned limitations and to carry
out a fast-dynamics all-optical flip-flop. Section IV reports an
experiment of all-optical 10Gb/s data stream switching operation driven by the proposed flip-flop, and finally, in Section V,
the conclusions are drawn.
II. ADVANTAGES AND LIMITATIONS OF A FLIP-FLOP
EXPLOITING SOA-BASED COUPLED RING LASERS
All-optical flip-flops exploiting SOA-based coupled ring
lasers have been well documented [1], [13]. Here, we experimentally and analytically investigate their advantages and limitations. We consider the setup shown in Fig. 1.
The flip-flop consists of two coupled ring lasers emitting at
two different wavelengths (λ1 = 1550 nm and λ2 = 1560 nm).
In each ring, an SOA acts as the gain element, a 0.25 nm bandpass filter (BPF) is used as the wavelength selective element, and
an isolator makes the light propagation unidirectional. Both the
SOAs are polarization insensitive multiquantum well (MQW)
structures with a small-signal gain of 31 dB, a saturation power
of 13 dBm, and a peak of amplified spontaneous emission (ASE)
noise at 1547 nm. The system can have two states. In “state 1,”
light from ring 1 suppresses lasing in ring 2, reaching cavity
2 through the 50/50 coupler and saturating the SOA 2 gain. In
this state, the optical flip-flop output 1 emits continuous-wave
(CW) light at wavelength λ1 . Conversely, in “state 2,” light from
809
Fig. 1. Experimental setup of the all-optical flip-flop; SOA: semiconductor
optical amplifier; BPF: bandpass filter.
ring 2 suppresses lasing in ring 1 (saturating SOA 1 gain). In this
case, the optical flip-flop output 2 emits CW light at wavelength
λ2 . To dynamically change state, lasing in the dominant cavity
can be switched OFF by injecting external pulsed light with a
wavelength different from λ1 and λ2 (λIN = 1554.5 nm). In
Fig. 2, experimental measurements of the two states optical
spectra are investigated and a graph of the output power of
both the ring lasers versus the CW input power injected into
each cavity is reported. In all considered cases, we obtained an
output contrast ratio higher than 40 dB.
By injecting two regular sequences of pulses into the set and
reset ports, we demonstrate a dynamic flip-flop operation, as
shown in Fig. 3.
When the flip-flop switches from one state to the other, the
transitions speed is determined by the switching OFF and switching ON time of the two states. We experimentally observed that
falling time only depends on the edge time of pulses (5 ns in this
section), while rising time is determined by the cavity length and
the length of the fiber between the two SOAs. In our setup, each
ring has a cavity length of 20 m corresponding to a round-trip
time of about 100 ns. Experimental measurements show [see
Fig. 4(a)] that the building-up process of one state takes place
step by step and each step corresponds to a cavity round-trip
time equal to 100 ns. The total rising edge behavior lasts several
hundreds of nanoseconds. The experimental falling edge behavior is shown in Fig. 4(b), with a transition time of 5 ns, equal to
the input pulse edge.
The dynamic behavior of the two SOA-based coupled lasing
cavities has been analyzed through simulations as well. Steadystates behavior of the two coupled lasing cavities using two coupled sets of rate equations for electron and photon density has
been analyzed in [14]. We supposed identical SOAs in both the
cavities and symmetric coupled rings, 1500 µm-long commercial SOAs segmented into ten sections [15], [16] (ten sections
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.
810
IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS, VOL. 14, NO. 3, MAY/JUNE 2008
Fig. 2. Experimental results: optical spectra of the two states (top) and output power of both the lasers versus input power injected into cavity 1 (bottom, left)
and into cavity 2 (bottom, right) of the all-optical flip-flop.
Fig. 3. Experimental results of the all-optical flip-flop output with an input
pulses repetition rate of 50 KHz; set: set pulse; reset: reset pulse; ring 1: output
of ring laser 1; ring 2: output of ring laser 2.
provide optimum performance in terms of speed and accuracy)
and each SOA section acting as a punctiform amplifier in the
cavity. Assuming the same parameters of the experimental setup
(cavity length and cavity loss, injected pulses edge time, and average power), as can be observed in Fig. 4(c) and (d), simulation
results for rising and falling edges are in rather good agreement
with experimental measurements, confirming the step behavior
of the rising edge, and at the same time, a falling edge as fast as
the input pulse edge.
In order to simulate an integrated version of the flip-flop, a
12 ps edge pulse with different average powers is supposed to
be injected as set and reset signals and SOA gain recovery time
is studied. The cavity length and SOA length are assumed to be
2 and 0.5 mm, respectively. A 12 ps falling time [Fig. 5 (left),
from A to B] and ∼ 40 ps rising time [Fig. 5 (left), from C to D]
for the gain recovery can be observed for an input pulse power of
−8 dBm. Increasing the injected input pulse power up to
−4 dBm, no gain recovery time reduction has been obtained,
whereas decreasing the input power to −12 dBm, gain recovery
time slows down to 72 ps [Fig. 5 (left), from C to E] because of
a partial depleting of SOA carrier density. We deduce a critical
injected pulse power below which rising time increases, and a
rising time with the lowest value of about 40 ps as well. Furthermore, in both discrete-device-based and integrated implementations, the injected input pulses have to maintain the SOA saturation condition of a cavity until the other cavity reaches a steady
lasing condition. Therefore, the shortest injected pulsewidth is
∼40 ps. Fig. 5 (right) reports the fastest achievable rising and
falling edges for the integrated solution of the flip-flop exploiting SOA-based coupled ring cavities. Switching energy being
proportional to the cavity length, the value of injected input
pulse energy (15.6 fJ in our simulation) is comparable with one
of the latest optical flip-flop integrated version one [7].
In the next section, we propose a solution for the slow rising
time issue and demonstrate an ultrafast optical flip-flop with fast
dynamics for both rising and falling transitions.
III. ULTRAFAST SOA-BASED ALL-OPTICAL FLIP-FLOP
In Section II, it is pointed out that an all-optical flip-flop based
on two coupled ring lasers presents a fast falling edge (as fast
as the input pulse rising edge), but a slow rising edge (several
round-trip times), which mainly limits the flip-flop operating
speed for optical packet switching (OPS). In this section, using
two SOA-based optical NOT logic gates and two identical slow
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.
MALACARNE et al.: 20 PS TRANSITION TIME ALL-OPTICAL SOA-BASED FLIP-FLOP USED
Fig. 4.
811
Edges behavior of the flip-flop output. (a) Measured and rising. (b) Measured and falling. (c) Simulated and rising. (d) Simulated and falling.
Fig. 5. Normalized gain recovery of the SOA in a 2000-µm-long cavity for various external input power (left) and output rising and falling edge behavior for
flip-flop SOA-based integrated version (right).
flip-flops, we obtain an optical flip-flop with ultrafast transition
times for both rising and falling edges. The experimental setup
is shown in Fig. 6, while the operating principle is described
in Fig. 7. Flip-flop 1 is controlled by reset and assistant pulses
whereas flip-flop 2 is controlled by assistant and set pulses.
Exploiting a 10 GHz pattern generator, we produce a 16 ps edge
pulsed sequence with a pulsewidth of 1 µs and a repetition rate of
50 KHz. Such a wide pulse has been set in order to maintain the
gain saturation level into the ring laser to be quenched for several
round-trip time, allowing to reach a lasing steady condition. The
reset pulse is delayed by 10 µs (Td1 ) with respect to the set pulse
whereas the assistant pulse is delayed by 15 µs (Td1 + Td2 )
with respect to the set pulse. As shown in Fig. 7, a set pulse
is firstly injected into ring 3 switching OFF signal B. Secondly,
a reset pulse is injected into ring 1 switching off signal A.
Then, two assistant pulses are injected into ring 2 and ring 4
simultaneously. They switch OFF ring 2 and ring 4, switching
ON ring 1 and ring 3, respectively. Consequently, signals A
and B are switched ON at the same time. As we pointed out
in Section II, both signals A and B have a fast falling edge,
but a slow rising edge. Exploiting the optical NOT logic gate 1,
signal A is inverted in order to obtain signal C, which, therefore,
presents a fast rising edge and a slow falling edge. Since signals
A and B are switched ON by two assistant pulses simultaneously,
the slow falling edge of signal C is almost synchronized with the
slow rising edge of signal B, and when they are added together,
the slow edges compensate each other in terms of intensity
profile. This way, signal D (the sum of signals B and C) has a
fast rising edge due to signal C and a fast falling edge coming
from signal B. The wavelengths of signals A, B, and C are
1550, 1558.2, and 1557.4 nm, respectively; thus, signal D is
made by two different wavelengths, as highlighted in Fig. 6,
and a tunable filter with −3 dB bandwidth of 4.5 nm is used
to filter and equalize these two wavelength components. Using
NOT logic gate 2, we invert signal D and thus obtain signal
E, at the same time convert it to one single wavelength λE =
1560 nm. Signal E is switched ON and OFF by the set and reset
pulses, respectively, showing fast rising and falling edges.
The optical NOT logic gates are implemented exploiting XGM
in SOAs. Concerning NOT logic gate 1, in SOA 5, a CW probe
light counter propagates with respect to signal A. The gain of
SOA 5 is modulated by the intensity profile of signal A through
XGM. In particular, when signal A has a low input power,
the gain provided by SOA 5 for the CW probe will be high,
whereas when signal A has a high power, the CW probe will
experience a lower gain. Ultimately, the CW probe undergoes
the gain variations obtaining the inversion of signal A, i.e.,
signal C. At the gate output (port 3 of circulator), selecting the
right polarization state through a polarization state controller
(PC) and a polarization beam splitter (PBS), it is possible to
increase the extinction ratio through XPM-based polarization
rotation [17], [18] and to limit noise induced by reflections on
the fiber–SOA interface. The principle of this all-optical NOT
logic gate has been described in [19].
Signals from A to E are shown in Fig 8. Since the slow edges
of signals B and C do not have a linear behavior, their sum gives
rise to a residual peak during the high level of signal D. After
NOT logic gate 2, this dynamic is suppressed because of the gain
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.
812
IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS, VOL. 14, NO. 3, MAY/JUNE 2008
Fig. 6. Experimental setup of the ultrafast all-optical flip-flop. SOA: semiconductor optical amplifier; BPF: bandpass filter; PC: polarization state controller;
PBS: polarization beam splitter; DSF: dispersion shifted fiber; EDFA: erbiumdoped fiber amplifier. Signal A is inverted by NOT logic gate 1 obtaining signal
C and added with signal B. Signal D (B+C) is inverted by NOT logic gate 2
obtaining signal E.
Fig. 7. Working principle of the ultrafast all-optical flip-flop. Signals A, B, C,
D, and E. T d 1 : delay between set and reset pulses; T d 2 : delay between reset and
assistant pulses; T O N : rising time; T O F F : falling time; ∆T = T d 1 + T O F F .
saturation level of SOA 6. The CW probe power injected into
SOA 6 has been set in order to optimize its saturation level (as
CW probe injected into SOA 5). Exploiting input set and reset
pulsewidths of 1 µs with edge time of 16 ps, signal E presents rising and falling times of 18.8 and 21.9 ps, respectively, as shown
in Fig. 9 (measured with a total bandwidth of 53 GHz), preserving a contrast ratio of 17.5 dB. It is possible to obtain a higher
contrast ratio just decreasing the CW probe signal powers in
SOA 5 and SOA 6, reducing their gain saturation level, with the
drawback of slower switching times [19]. Moreover, integrated
coupled ring lasers would experience a round-trip time in the
picosecond range (instead of 100 ns as in our experiment), allowing to use an injected pulsewidth in the picosecond range too.
IV. 10 Gb/s SWITCHING OPERATION WITH NO BIT LOSS
EXPLOITING ULTRAFAST ALL-OPTICAL FLIP-FLOP
Fast dynamics (rising and falling times of 20 ps) and quite
high extinction ratio (17.5 dB) make the ultrafast all-optical flipflop suitable to be exploited to control a 2×2 SOA-based alloptical switch [20]. In this section, we demonstrate an error-free
Fig. 8. Signals A, B, C, D, and E. Signal C = NOT (signal A); signal D =
signal B + signal C; signal E = NOT (signal D).
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.
MALACARNE et al.: 20 PS TRANSITION TIME ALL-OPTICAL SOA-BASED FLIP-FLOP USED
813
Fig. 10. All-optical switching operation experimental setup using a 2×2 SOAbased optical switch controlled by the ultrafast all-optical flip-flop. PC: polarization controller; BPF: bandpass filter; VOA: variable optical attenuator; EDFA:
erbium-doped fiber amplifier; ODL: optical delay line.
Fig. 11. Output 1 of the 2×2 all-optical switch, when just input 1 is connected
(input 2 is disconnected). Insets shows the fast (a) switching ON and (b) switching
OFF transitions.
Fig. 9. (a) Signal E rising and (b) falling edges. The rising time is 18.8 ps and
the falling time is 21.9 ps.
10 Gb/s optical switching operation using the aforementioned
2×2 optical switch driven by the ultrafast optical flip-flop. Our
experimental setup is shown in Fig. 10.
The switching operation is based on the XGM effect in two
different SOAs. Depending on the high- or low-intensity level
of the control signal (pump), in an SOA, the gain is strongly
reduced while the other SOA is not saturated.
The two input signals are generated by splitting a single 10Gb/s non-return-to-zero (NRZ) continuous data stream.
The stream is generated by modulating a CW laser at λIN =
1550 nm by means of a Mach–Zehnder (MZ) modulator driven
by a 10Gb/s pattern generator running in (231 − 1) long pseudo
random bit sequences (PRBS) mode. At the same time, the ultrafast flip-flop output is used as a pump signal of the optical
switch and controls the switch state (BAR or CROSS). The inverted pump signal (needful for switching operation [20]) is
obtained within the optical switch block (Fig. 10) through signal inversion by means of XGM in an SOA. The data streams
average power at the switch inputs is set to −7 dBm, while the
high pump level is 11.5 dBm.
We have chosen continuous data streams instead of packet
traffic to demonstrate and point out that it is possible to obtain a switching operation without any bit loss, exploiting the
20 ps fast dynamics of the flip-flop. Indeed, as can be observed in
Fig. 11, we can confirm a fast switching operation (faster than the
10 Gb/s single bit edge), connecting only input 1 (disconnecting input 2) of the switch and visualizing output 1 on a sampling oscilloscope, switching the output data signal ON and OFF
within one bit time. The contrast ratio between switched ON
and switched OFF signal is about 14 dB. This way we avoid any
distorted transition bit between switched ON and switched OFF
output signals, and vice versa.
Connecting both inputs 1 and 2 of the switch, high- or lowintensity level of the input pump signal sets the switch in BAR or
CROSS state. During BAR state, input 1 of the switch is routed to
output 1 (and input 2 is routed to output 2), while during CROSS
state input 2 is routed to output 1 (and input 1 is routed to output
2). Fig. 12 shows both input data eye diagrams and output 1
eye diagrams in BAR and CROSS configurations, measured by
a wideband photodiode and a sampling oscilloscope. As it can
be noticed, the output signal is not affected by pattern effects,
showing clearly open eye diagrams, confirming the effectiveness
of the scheme.
Fig. 13 shows the BER measurements at output 1 of the
switch, in both BAR and CROSS configurations. The used
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.
814
IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS, VOL. 14, NO. 3, MAY/JUNE 2008
Fig. 12. Eye diagrams of data frames. (a) Input 1. (b) Input 2. A 2 × 2 alloptical switch. (c) Output 1 in BAR configuration. (d) Output 1 in CROSS
configuration.
an extinction ratio of 17.5 dB. In particular, exploiting two flipflops, each of them based on two coupled SOA-based ring lasers
and two SOA-based NOT logic gates, both rising and falling
edges arise from the fast falling edge of the original flip-flop
outputs, overcoming the intrinsic slow rising time that limits the
flip-flop configuration based on the coupled ring lasers. Each
NOT logic gate exploits XGM and cross polarization rotation in
a single SOA.
Finally, exploiting a 2×2 SOA-based all-optical spatial
switch, we experimentally demonstrated a 10 Gb/s optical
switching operation, showing a switching time faster than the
10Gb/s single bit edge, without any bit loss or distorted transition bit. Clearly, open eye diagrams and error-free BER measurements for both CROSS and BAR switch configurations are
reported, showing a power penalty of about 1 dB, making the
switch driven by the ultrafast all-optical flip-flop suitable for cascaded schemes. The whole experimental setup is SOA-based,
making the scheme suitable for integrated solutions. Integrated
schemes should also improve the stability and reduce detrimental reflections, requiring low switching energies.
REFERENCES
Fig. 13. BER curves in the back-to-back (B to B) case and at switch output 1
in BAR and CROSS configurations.
receiver is composed by an optical preamplifier with 5 dB noise
figure, followed by a variable optical attenuator (VOA), a BPF,
and a photoreceiver, whose input power is kept constant (by
means of the VOA) at −16.7 dBm in order to avoid thermal
noise. As shown in Fig. 13, making a comparison with the backto-back case, the maximum penalty at BER = 10−9 is about
1 dB, making the switch driven by the ultrafast all-optical flipflop suitable for cascaded schemes.
V. CONCLUSION
A novel ultrafast all-optical flip-flop and its application for
photonic 10 Gb/s switching operation with no bit loss are presented and demonstrated. Starting from a well-known flip-flop
configuration exploiting SOA-based coupled ring lasers, with
slow rising transition time, using proper time delays and SOAbased optical NOT logic gates, we obtain an optical flip-flop with
20 ps fast transition times for both rising and falling edges and
[1] H. J. S. Dorren et al., “Optical packet switching and buffering by using
all-optical signal processing methods,” J. Lightw. Technol., vol. 21, no. 1,
pp. 2–12, Jan. 2003.
[2] F. Ramos et al., “IST-LASAGNE: Towards all-optical label swapping
employing optical logic gates and optical flip-flops,” J. Lightw. Technol.,
vol. 23, no. 10, pp. 2993–3011, Oct. 2005.
[3] A. Bogoni et al., “A synchronous all-optical 160 Gb/s photonic interconnection network,” in Opt. Fiber Commun. Conf. Tech. Dig., Anaheim, CA,
pp. 1–3.
[4] Y. Liu, et al., “All-optical switching of 80 Gb/s data packets using
a wavelength converter controlled by a monolithically integrated optical flip-flop,” in Proc. ECOC 2005, Sep. 25–29, vol. 6, pp. 27–
28.
[5] J. Herrera et al., “160-Gb/s all-optical packet-switching with in-band
filter-based label extraction and a hybrid-integrated optical flip-flop,”
IEEE Photon. Technol. Lett., vol. 19, no. 13, pp. 990–992, Jul.
2007.
[6] A. Malacarne et al., “Erbium-ytterbium doped fiber-based optical flipflop,” IEEE Photon. Technol. Lett., vol. 19, no. 12, pp. 904–906, Jun.
2007.
[7] M. T. Hill et al., “A fast low-power optical memory based on coupled
micro-ring lasers,” Nature, vol. 432, pp. 206–208, Nov. 2004.
[8] Y. Liu et al., “Packaged and hybrid integrated all-optical flip-flop memory,” Electron. Lett., vol. 42, no. 24, pp. 1399–1400, Nov. 2006.
[9] M. T. Hill et al., “Integrated two-state AWG-based multiwavelength
laser,” IEEE Photon. Technol. Lett., vol. 17, no. 5, pp. 956–958, May
2005.
[10] V. Van et al., “Optical signal processing using nonlinear semiconductor
microring resonators,” IEEE J. Sel. Topics Quantum Electron., vol. 8,
no. 3, pp. 705–713, May/Jun. 2002.
[11] R. W. Keyes, “Fundamental limits of silicon technology,” Proc. IEEE,
vol. 89, no. 3, pp. 227–239, Mar. 2001.
[12] C. W. Lowry et al., “Picosecond on–off switching using a pair of
picosecond-on/nanosecond-off GaAs etalons,” Appl. Phys. Lett., vol. 57,
no. 26, pp. 2759–2761, Dec. 1990.
[13] Y. Liu et al., “Three-state all-optical memory based on coupled ring
lasers,” IEEE Photon. Technol. Lett., vol. 15, no. 10, pp. 1461–1463,
Oct. 2003.
[14] M. T. Hill et al., “All optical flip-flop based on coupled laser diodes,”
IEEE J. Quantum Electron., vol. 37, no. 3, pp. 405–413, Mar. 2001.
[15] A. Das Barman et al., “Modelling and implementation of photonic digital
subsystem for bit comparison,” in Proc. Photon. Switching, 2007, Paper
TuB2.4, pp. 61–62.
[16] W. Mathlouthi et al., “Fast and efficient dynamic WDM semiconductor
optical amplifier model,” J. Lightw. Technol., vol. 24, no. 11, pp. 4353–
4365, Nov. 2006.
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.
MALACARNE et al.: 20 PS TRANSITION TIME ALL-OPTICAL SOA-BASED FLIP-FLOP USED
[17] H. J. S. Dorren et al., “Nonlinear polarization rotation in semiconductor
optical amplifiers: Theory and application to all-optical flip-flop memories,” IEEE J. Quantum Electron., vol. 39, no. 1, pp. 141–148, Jan.
2003.
[18] H. Soto et al., “Cross-polarization modulation in semiconductor optical
amplifiers,” IEEE Photon. Technol. Lett., vol. 11, no. 8, pp. 970–972,
Aug. 1999.
[19] G. Berrettini et al., “Ultrafast integrable and reconfigurable XNOR, AND,
NOR, and NOT photonic logic gate,” IEEE Photon. Technol. Lett., vol. 18,
no. 8, pp. 917–919, Apr. 2006.
[20] G. Berrettini et al., “Ultra-fast integrable 2’2 all-optical switch,” presented
at the ECOC 2006, Cannes, France, Sep. 24–28.
Antonio Malacarne was born in Livorno, Italy, in
1978. He received the M.S. degree in telecommunications engineering from the University of Pisa,
Pisa, Italy, in 2004. He is currently working toward
the Ph.D. degree at the Scuola Superiore Sant’Anna,
Ultra-Fast Optical Subsystems Group, Centre of Excellence for Information and Communication Engineering (CEIIC), Pisa.
During 2006, he was a visiting student at the
Photonics Research Center, Systems Technology Division, Korea Institute of Science and Technology
(KIST), Seoul, Korea. His current research interests include ultrafast optical
time-division multiplexing (OTDM) systems, ultrashort optical pulses generation, all-optical sampling, optical pulse shaping, erbium-doped fibers properties, all-optical multistable devices, all-optical technologies, and optical packet
switching (OPS).
Jing Wang was born in Shanghai, China, in 1983. He
received the Bachelor’s degree in electronic engineering in 2006 from Tsinghua University, Beijing, China,
where he is currently working toward the Ph.D. degree at the Department of Electronic Engineering.
Since July 2007, he has been a visiting student at
the Photonics Networks National Laboratory, Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), Pisa, Italy. His current research
interests include all-optical flip-flop, optical packet
switching networks applications, and all-optical
processing area.
Yuancheng Zhang was born in Inner Mongolia,
China, in 1984. He received the Bachelor’s degree in
electronic engineering in 2006 from Tsinghua University, Beijing, China, where he is currently working
toward the Ph.D. degree at the Department of Electronic Engineering.
Since July 2007, he has been a visiting student at
the Photonics Networks National Laboratory, Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), Pisa, Italy. His current research
interests include all-optical bistability, optical signal processing technologies, and ultrafast optical time-division multiplexing
(OTDM) systems.
Abhirup Das Barman received the M.Tech. degree in electrical engineering from Indian Institute
of Technology (IIT), Kanpur, India, in 1995.
In 1996, he joined the Indian Engineering Service
(IES), Government of India, and worked in the area
of satellite communications, digital terrestrial transmission, etc. In 2004, he became a faculty member at
the Institute of Radio Physics and Electronics, University of Calcutta. Since 2007, he has been a Visiting
Research Scholar at the Photonic Networks National
Laboratory, Consorzio Nazionale Interuniversitario
per le Telecomunicazioni (CNIT), Pisa, Italy. He is also with the Scuola Superiore Sant’Anna, Centre of Excellence for Information and Communication
Engineering (CEIIC), Pisa. His current research interests include the area of
modeling of wavelength division multiplexed fiber optic links and networks and
some studies on related photonic devices.
815
Gianluca Berrettini was born in Pescia (PT), Italy,
in 1979. He received the M.S. degree in telecommunications engineering from the University of Pisa,
Pisa, Italy, in 2004. He is currently working toward
the Ph.D. degree at the Scuola Superiore Sant’Anna,
Centre of Excellence for Information and Communication Engineering (CEIIC), Pisa.
He is the author or coauthor of more than 20
papers published in international conference proceedings and journals. His current research interests
include the area of fiber optic transmission with particular interest in generation of ultrashort pulse sources, ultrafast optical timedivision multiplexing (OTDM) systems, optical packet switching networks
applications and all-optical technologies.
Luca Potı̀ (M’98) was born in Parma, Italy, in 1971.
He received the M.S. degree in electronics engineering from the University of Parma, Parma, in 1997.
He is also an external collaborator of the Scuola
Superiore Sant’Anna, Centre of Excellence for Information and Communication Engineering (CEIIC),
Pisa. During 1997, he was a Visiting Researcher at
the Centre d’Optique, Photonique et Laser (COPL),
Quebéc, QC, Canada, where he is engaged in recirculating loop experiment for ultralong haul transmission. From 1999 to 2000, he was with the Optical Communications Laboratory, University of Parma. Since 2001, he has
been a Senior Researcher with the Photonic Networks National Laboratory,
Consorzio Nazionale Interuniversitario per le Telecomunicazioni (CNIT), Pisa,
Italy, where he is currently the Head of Research. During 2002, he was a Visiting Researcher at the National Institute of Information and Communications
Technology (NICT), Tokyo, Japan, working on ultrafast phase comparison for
160 GHz signals. In 2005, he started, together with two researchers, PhoTrix
S.r.l., an Italian company producing pulsed fiber lasers and ultrafast subsystems.
He is the author or coauthor of more than 100 published international journal
papers, conference papers, and patents. He was a Scientific Coordinator for the
European Union Project “Large optical bandwidth by amplifier systems based
on tellurite fibers doped with eare earths” (LOBSTER). He was also involved
in several projects supported by the Italian Ministry of University and Research
(MIUR) and the Ministry of Foreign Affairs (MAE). He is currently managing
the Fondo per gli Investmenti della Ricerca di Base (FIRB) project “Photonic enabling devices for regeneration and optical switching” (PEDROS) supported by
the MIUR. His research interests were focused on ultrafast communication systems. In 2001, his group demonstrated first Italian transmission system working
at 160 Gb/s. His current research interests include all-optical processing subsystems for ultrafast communications and optical packet switching networks.
Mr. Potı̀ received a Marconi Communication grant in 1998 on the topic
“WDM communication systems on optical fiber” at the Marconi Laboratory,
Parma University, working on nonlinear effects due to fibers and erbium-doped
fiber amplifiers (EDFAs) in wavelength-division multiplexing (WDM) systems.
Antonella Bogoni was born in Mantova, Italy, in
1972. She received the M.S. degree in electronics engineering and the Ph.D. degree in information technologies from the University of Parma, Parma, Italy,
in 1997 and 2004, respectively.
From 1998 to 1999, she was a grantee of Marconi
S.p.a. at the University of Parma. From 2000 to 2006,
she was a Researcher with the Consorzio Nazionale
Interuniversitario per le Telecomunicazioni (CNIT),
Pisa, Italy, where she was a Researcher at the University of Parma up to 2001, and then, at the Photonic
Networks National Laboratory. She is currently the Head of research of the
CNIT at the Integrated Research Center for Photonic Networks and Technologies. She was a Scientific Coordinator for the National Project “Realization of
ultrashort pulsed source prototypes” and has been involved in several European
Union and national projects. She is the author or coauthor of more than 30
papers published in international journals, 100 contributions for international
conferences, and 20 international patents. Her current research interests include
the area of fiber optical transmissions, especially in ultrafast all-optical signal
processing and pulsed source generation.
Dr. Bogoni has been a member of revision committees of international conferences and is also a reviewer for international journals and for the European
Commission within FP7.
Authorized licensed use limited to: UNIVERSITA PISA S ANNA. Downloaded on November 9, 2009 at 10:29 from IEEE Xplore. Restrictions apply.