Henini M. Handbook of Self Assembled Semiconductor Nanostructures for Novel devices in Photonics and Electronics
Подождите немного. Документ загружается.
Quantum Dot-based Mode-locked Lasers and Applications 609
20.2.4 Low timing jitter of quantum dot mode-locked lasers
For many applications such as electro-optic sampling, high bit rate Optical Time Division
Multiplexing (OTDM), and optical interconnect, a low timing jitter of ultra-short pulses is
required because the fl uctuation in the time arrival degrades the bit error rate (BER) and time
resolution [33] . This parameter represents the deviation from a perfect train of identical pulses
[12] . Here we rapidly stress the origin of low timing jitter of QD MLL and present the main
results.
Theoretical investigations showed that QD amplifi ers exhibit improved performance in terms of
noise [10, 37] . Noise fi gures better than 4.5 dB are expected owing to three characteristics [10] :
low modal gain due to a lower confi nement factor Γ (i.e. 1%), a lower differential gain owing to
a reduced Γ , and a large energy separation between GS, ES and WL [10] . The energy diagram of
this material system is thus equivalent to a three-level system allowing a high degree of popula-
tion inversion of the GS which implies a population inversion factor n
sp
close to 1. These features
are also present in QD MLLs.
The low confi nement factor leads to a reduced coupling of the amplifi ed spontaneous emission
to optical mode, thereby reducing the noise level of QD lasers compared to QW lasers. The low
differential gain implies a large saturation energy of the gain section, which is useful as the avail-
able high optical power will reduce the timing jitter [33] . Timing jitter as well as intensity noise
in MLLs arises from spontaneous emission, random fl uctuations of driving current and external
circuitry [12, 38] . While the absolute timing jitter refers to all sources of noise, residual timing
jitter represents the noise contribution of the laser alone [12] . Here we will focus on the residual
timing jitter to underline the potential of QD MLLs over QW MLLs.
Gain or photon density variation can also impact on the timing jitter through the phase-ampli-
tude coupling factor or so-called linewidth enhancement factor (LEF). This parameter is expected
to be lower than in QW-based counterparts as discussed in section 20.2.2.2. Experimentally, we
have already shown that the Henry factor can be similar to that of the best QW lasers at high
currents (i.e. 3) using high modal gain structures to prevent GS gain saturation [24] . The dem-
onstrated reduced timing jitter of QD MLLs cannot hence solely be attributed to a low LEF.
Early measurements of timing jitter in passive QD MLLs emitting at 1 . 3 μ m have yielded
values lower than 1 ps [33, 34, 39] and lower than 2 ps at a repetition rate of 35 GHz [16] .
This is smaller than that of passive QW MLLs, integrated over a frequency range from 150 kHz
to 50 MHz [40] . Thompson et al. measured the timing jitter of 18 GHz monolithic passive QD
mode-locked lasers emitting at 1.3 μ m [39] . They estimated an upper limit of 600 fs for the RMS
timing jitter from the single-sideband phase noise (SSBPN) spectrum over a frequency range of
2.5–50 MHz. In this regime, near transform-limited optical pulses of 10 ps (time bandwidth
product of 0.315) were demonstrated in this passive monolithic QD ML laser with ten QD
layers. The 24 μ m wide waveguide supported several tranverse modes and limited the peak power
to 3 m W.
Timing and amplitude jitter of pulses from narrow ridge devices were extracted by comparison
of autocorrelation and cross-correlation traces [16] . The uncorrelated jitter was found to be less
than 1 ps in active ML at 20 GHz and less than 2 ps in passive ML at 35 GHz. The authors did not
indicate the integration range of the SSB PN spectrum.
The fi rst unambiguous demonstration of sub-picosecond timing jitter in passive QD MLLs was
reported in 2005 by Zhang et al. at a 5 GHz repetition rate [33] . The integration of the SSBPN
spectral density L ( f ) (fourth harmonic) over the offset frequency range of 30 kHz to 30 MHz led to
a root mean squared timing jitter of 0.91 ps. Compared to the standard timing jitter of QW pas-
sively MLLs (12.5 ps over [150 kHz; 50 MHz]), these QD ML lasers exhibit more than one order
of magnitude improvement [33] . The results are explained by the Dirac function-like density of
states in QDs as well as low transparency current implying that only a reduced part of carriers is
involved in non-stimulated emission [33] . These arguments confi rm the origin of the low noise
fi gure evoked in QD SOAs by Berg et al. [10] .
More recently, very impressive results were reported with two-section lasers at a 40 GHz
repetition frequency by Thompson et al. [26] . Pulsewidths of 2.5 ps are measured with a time
bandwidth product of 0.5 for passive and hybrid QD MLLs. Timing jitter as low as 219 fs was
CH020-I046325.indd 609CH020-I046325.indd 609 6/25/2008 10:03:37 AM6/25/2008 10:03:37 AM
610 Handbook of Self Assembled Semiconductor Nanostructures for Novel Devices in Photonics and Electronics
demonstrated in the [16 MHz; 320 MHz] range. Its value amounts to 843 fs in the [150 kHz; 320
MHz] range (integration range specifi ed by ITU-T for a passive MLL) [26] . This is signifi cantly
lower than any other reported result in passive QW MLLs at 40 GHz where timing jitter is in
the range of 12 ps. The timing jitter for hybrid ML at 40 GHz (124 fs in [20 kHz; 320 MHz] range)
was not as good as that of hybrid QW MLLs (73 fs). Further optimization of the fabrication
of the absorber should lead to higher absorber modulation and improved timing jitter in the
hybrid regime.
An interesting study performed by Thompson et al. [34] showed that the timing jitter of passive
QD MLLs can be optimized through a careful adjustment of the absorber length at the expense of
the pulsewidth. Indeed, a trade-off exists between the pulse duration and the timing jitter: while
the pulsewidth decreases with an increased absorber length for a given reverse voltage, the tim-
ing jitter is found to decrease when the absorber length is reduced. This implies that the lowest
timing jitter is obtained for longer pulsewidths. Typical timing jitter integrated over [20 kHz;
50 MHz] of 100 ps are extracted for 800 fs pulses generated with a 520 μ m long absorber, while
timing jitter below 400 fs is obtained for 1.8 ps pulses for a 260 μ m long absorber [34] . This trend
is attributed to the decrease in overall cavity loss for shorter absorber devices.
Finally, one must mention extraction of the residual timing jitter of an external cavity
12.8 GHz actively mode-locked laser performed by Delfyett and co-workers [41] : it amounts to
7.5 fs in the 1 Hz–10 MHz frequency range. All these results show that QD MLLs are ideal sources
for both short pulses and low noise optical pulse trains.
As discussed above, the evaluation of the timing jitter using the spectral technique is par-
ticulary suited for repetition rates lower than the typical bandwidth of high-speed photodiodes
(50 GHz). Based on a cross-correlation technique, Tourrenc et al. measured the timing jitter of
passive ML QD lasers [42] .
Interestingly, the two techniques lead to similar conclusions regarding the evolution of the tim-
ing jitter versus the gain current for a given absorber bias [33, 42] . The jitter appears to improve
with an increase of the gain leading to typical values as low as 1 ps for a 5 GHz repetition rate
[33] and as low as 20 fs/pulse cycle at an 18 GHz repetition rate [42] .
In summary, three main properties arise from the application of the reverse voltage to the QD
saturable absorber. First, the low differential gain results in high saturation energy of the gain sec-
tion. Moreover, the absorber recovery time decreases exponentially with the reverse voltage, which
implies an exponential decrease of the pulsewidth with voltage. Finally, reduction of the optical
confi nement factor implies a low phase noise with direct consequence on the timing jitter. All these
trends prove that the carrier dynamics of QD saturable absorbers is a key element for the generation
of both sub-ps pulses and sub-ps timing jitter, provided high GS modal gain is again available.
20.2.5 Summary
Table 20.1 summarizes the performances discussed above for the sake of comparison. To match
the requirement of desired applications (optically interconnected systems, optical time division
multiplexing, etc.), the active layer structure, the regime of mode locking, the pulsewidth, the
repetition rate, the time bandwidth product and timing jitter are reported.
Different applications require a combination of pulse characteristics, and some applications
may require low timing jitter, high peak power or ultra short-pulse duration, and sometimes a
combination of all these three parameters [34] .
20.3 InAs/InP quantum dash mode-locked lasers emitting at 1.55 μ m
20.3.1 Sub-picosecond pulse generation at very high repetition rate
Growth of InAs quantum dots on InP (100) substrates generally leads to the formation of quan-
tum “ dash ” nanostructures (elongated dots) and the litterature has been less fertile in device
results related to 3D confi nement of the carriers [43–47] .
These nanostructures however, exhibit very interesting properties pertaining to short pulse
generation using mode-locked lasers.
CH020-I046325.indd 610CH020-I046325.indd 610 6/25/2008 10:03:38 AM6/25/2008 10:03:38 AM
Quantum Dot-based Mode-locked Lasers and Applications 611
Table 20.1 Summary of the performances of the QD MLLs emitting at 1.3 μ m.
Reference Active
layer
Regime of
mode
locking
Pulsewidth Repetition
rate
Peak
power
Time bandwidth
product
Timing
jitter
[11] Two
DWELL
layers
Passive 17 ps 7.4 GHz 20 mW 3.1
(1 nm 17 ps)
n/a
[16] Five QD
layers
Passive 8–18 ps 35 GHz 6 mW 0.44
(180 pm 13 ps)
2ps
[16] Hybrid 7 ps 20 GHz 1ps
[6] Five QD
layers
Passive 390 fs 21 GHz 3 W 1 (14 nm 2 ps) n/a
[33] Six
DWELL
layers
Passive 5.7 ps 5.1 GHz 290 mW 0.91 ps [30 kHz;
30 MHz]
[33] Passive 4 ps 5 GHz 3 ps
[28] Five QD
layers
Passive 800 fs 24 GHz 500 mW 0.5 n/a
[34] Five QD
layers
Passive 5.5 ps 20.5 GHz 45 mW 390 fs [20 kHz;
50 MHz]
[26] Ten QD
layers
Passive 2.5 ps 40 GHz 0.5–0.9 219 fs [16 MHz;
320 MHz ]
843 fs [150 kHz;
320 MHz]
[26] Ten QD
layers
Hybrid 124 fs [20 kHz;
320 MHz]
[26] Ten QD
layers
Passive 2.2 ps 80 GHz 11 mW 1.5 n/a
The “ elongated dots ” provide a higher optical confi nement factor compared to that of QDs
grown on GaAs susbstrates. This results in high modal gain (48 cm
1
for a nine dash-in-a-well
layer structure), leading to static performances (threshold current and quantum effi ciency) simi-
lar to those of QW-based lasers [48] . An optimized laser structure consisting of a stack of six
QDash layers has in particular allowed laser emission for very short cavity lengths of 200 μ m
owing to the available high modal gain (typically 36 cm
1
) [48] . This is an asset for very high
achievable repetition rates of MLLs, in excess of 80 GHz, compared to the lower modal gain
obtained for QDs grown on GaAs.
QDash lasers further exhibit an inherently wider inhomogeneous broadening due to overlap-
ping states in the high energy portion of the spectrum [49] . The expected wider gain spectrum
of the QDash lasers should be an advantage compared to QD lasers for the generation of a much
shorter pulsewidth.
The fi rst demonstration of passive QDash-based MLLs at 1.56 μ m was reported by Martinez et al. in
a two-section confi guration based on gain and saturable absorber sections [7] . Narrow ridge single-
mode waveguides yielded sub-picosecond pulses of 500 fs at a 53 GHz repetition rate without any
compression scheme. The reported full width at half maximum of the emission spectrum is 1 1 n m
for this pulse duration highlighting the potential of QDash material for short pulse generation.
At the same time, based on a similar material system, Renaudier et al. [50] showed self-pulsa-
tion at 45 GHz in a single section Fabry–Perot laser, i.e. without resorting to an absorber section.
In this latter confi guration, enhanced four-wave mixing (FWM) is invoked as the driving mecha-
nism of the passive mode locking which induces a strong correlation between the longitudinal
modes (section 20.3.3).
CH020-I046325.indd 611CH020-I046325.indd 611 6/25/2008 10:03:38 AM6/25/2008 10:03:38 AM
612 Handbook of Self Assembled Semiconductor Nanostructures for Novel Devices in Photonics and Electronics
Indeed, InAs/InP quantum dash-based lasers have recently demonstrated sub-picosecond
pulse generation (800 fs) at a very high repetition rate of 134 GHz with an extinction ratio better
than 20 dB [51, 52] . This very short pulse duration was achieved using a one-section Fabry–Perot
laser without resorting to any compression scheme. The time bandwidth product of 0.46 indi-
cates that the pulses are close to being Fourier transform limited. These high performances were
not obtained at the expense of output power which amounts to about 10 mW (implying a peak
power of 0.1 W).
In this material system, all that remains is to conduct systematic investigations in two-section
devices on how the bias current and the absorber reverse voltage modify the temporal properties
of the pulses, as carrier dynamics of the absorber will play an important role.
20.3.2 Extremely narrow radio frequency spectrum
Analysis of the radio frequency photocurrent spectrum has shown a record low linewidth of
47 kHz for a Fabry–Perot laser emitting pulses at 43 GHz repetition rate [7, 52] ( Fig. 20.8 ). This
was the lowest linewidth ever reported for any passively mode-locked laser. Effi cient phase corre-
lation between longitudinal modes involved in the mode-locking process is found responsible for
such performances, as it will be discussed in the next part. With such a low linewidth, the high-
frequency timing jitter can be estimated to be less than 200 fs [38] .
The investigation of mode locking in a two-section laser including an unbiased short absorber
reveals similar performances with a spectral linewidth of 47 kHz for a 43.6 GHz pulsation fre-
quency [52] . Passive mode locking was confi rmed through the measurement of the linewidth
of the longitudinal mode which amounts to 17 MHz, approximately two orders of magnitude
higher than the FWHM of the RF photocurrent spectrum.
55
60
65
70
75
80
85
400 200 0
Relative frequency (kHz)
200 400
Spectral power (dBm)
f
SP
42.20 GHz
v 50 kHz
Figure 20.8 Photocurrent spectrum of passively mode-locked laser around the fundamental frequency (resolution
bandwidth 10 kHz) [52] (Reused with permission from C. Gosset et al. , Appl. Phys. Lett. 88 , 241105 (2006).
Copyright 2006, American Institute of Physics.)
Similar experiments performed by Renaudier et al. [50] demonstrated linewidths of the mode-
beating spectrum as low as 70 kHz using a single-section laser generating pulses at a 45 GHz
frequency. Even lower linewidths of 15 kHz have been evidenced in lower confi nement QDash
laser structures [47] . This is attributed to a reduced coupling of the spontaneous emission to the
lasing mode owing to the low optical confi nement factor of QDash-based MLLs.
As a conclusion, at this stage of the investigation of QDash MLLs, it appears that the absorber
section does not play a major role in the reported mode locking with narrow FWHM of the RF spec-
trum. The dominant mechanism arises from the gain section. Indeed, reduction of the dimension of
active medium from two dimensions to one dimension should be responsible for the enhancement
of four-wave mixing wavelength conversion effi ciency in discrete level systems [52, ref. 16 therein].
Besides passive mode locking, active mode locking performed in a six QDash layer FP laser [50]
indicates a high-frequency timing jitter as low as 82 fs around the fundamental frequency of 45 GHz,
thus demonstrating the reduced intrinsic noise owing to the narrow linewidth of the RF spectrum.
CH020-I046325.indd 612CH020-I046325.indd 612 6/25/2008 10:03:38 AM6/25/2008 10:03:38 AM
Quantum Dot-based Mode-locked Lasers and Applications 613
The record low linewidth of the RF photocurrent highlights the potential of the QDash mate-
rial system for extremely low jitter applications. In particular, we will present in the next section
how jitter fi ltering effects can be useful to realize all optical clock recovery at 40 Gb/s.
20.3.3 Phase amplitude characterization
Enhanced four-wave mixing which favours passive phase-locking between longitudinal modes is
invoked to be the main mechanism responsible for QDash mode locking. Recent experiments dem-
onstrated FWM effi ciency in QDash semiconductor optical amplifi ers (QDash SOA) similar to that
of bulk counterpart, although the optical confi nement factor is 30 times lower [47] .
Evidence of phase correlation in QDash MLLs was further brought by the measurement of the
optical spectral phase of longitudinal modes. This relies on the successive analysis of the auto-
correlation intensity signal of groups of three adjacent modes [53] . The investigated laser was
a one-section passive MLL at 42 GHz pulsation frequency. The evolution of the spectral phase
shows a linear dependence versus wavelength in the central part of the spectrum, leading to a
transform limited pulse reconstructed through modelling. A non-linear dependence of the spec-
tral phase is evidenced in the lower and higher portion of the spectrum. Simulations also con-
fi rm that these uncorrelated modes give rise to a pedestal in the intensity autocorrelation trace
and result in non-transform-limited pulses. The phase variation is attributed to inhomogeneous
broadening, typical of the QDash active medium [53] .
It can be argued that fi ltering out the uncorrelated longitudinal modes should improve the
quality of the pulses as well as the performances of the QDash MLLs. Indeed, very recent experi-
ments demonstrated greatly improved extinction ratios as well as time bandwidth product by an
adequate spectrum fi ltering of undesirable groups of longitudinal modes [54] . The tested device
was a 1 mm long Fabry-Perot laser processed from a six DWELL layer structure. Careful selection
of the centre frequency and spectral width of the fi lter results in an improved extinction ratio of
the intensity autocorrelation signal from 2 dB to 17 dB, with a time bandwidth product of 0.44.
Spectrum fi ltering also warrants intensity autocorrelation traces without pedestal, thus confi rm-
ing the above arguments, and leads to an RF photocurrent linewidth as narrow as 30 kHz.
In summary, QDash-based MLLs have demonstrated very promising performances such as sub-
picosecond widths of transform limited pulses, very low RF linewidths implying extremely low
high-frequency timing jitter and over 100 GHz repetition rates. Systematic investigation of the mode-
locked mechanisms is still under way to fully exploit all the potential of this novel material system.
20.4 Applications
Practical applications of mode-locked lasers rely on short pulse duration and low timing jitter in
such diverse areas as high bit rate optical communications (optical time division multiplexing,
optical sources, clock recovery), microwave signal generation, optical sampling, optical intercon-
nects, etc.
Two specifi c applications will be highlighted, namely pulse generation for optical interconnect
and clock distribution in integrated circuits (ICs) [55] and clock recovery at 40 Gbit/s.
In(Ga)As/GaAs quantum dot mode-locked lasers operating up to 80°C have recently been
demonstrated, using an undoped active layer [56] . Further improvement in the temperature
behaviour is, however, still needed. A main requirement for this former application is the capabil-
ity to operate the devices above about 100°C.
The second part of this section will present recent exploitation of jitter fi ltering effect of QDash
MLLs for clock recovery at 40 Gbit/s in fi bre optic communications.
20.4.1 Optical interconnects applications
In the context of optical interconnects [55] an important requirement for integrated circuits
manufacturers regarding clock generation and distribution is the ability to operate at least at
CH020-I046325.indd 613CH020-I046325.indd 613 6/25/2008 10:03:38 AM6/25/2008 10:03:38 AM
614 Handbook of Self Assembled Semiconductor Nanostructures for Novel Devices in Photonics and Electronics
100°C. Another important issue is related to the value of the IC ’ frequency, 5 GHz for the next
generation processors. This implies cavity lengths of the order of 8 mm which requires specifi c
processing and mounting of the devices. Indeed, MLLs are ideal candidates for interconnects
owing to their compactness, low power consumption and DC-biased operation [33] . Moreover,
monolithic MLLs based on a two-section confi guration offer competitive advantage in terms of
reduced cost for volume production.
As CMOS transistor technology improves, problems such as frequency and distance-dependent
loss and distortion appear to be a major problem for operations at frequencies in the GHz range
[55] . Optical interconnected systems offer potential benefi ts because optical signal quality is pre-
served with distance. QD MLLs are very attractive for such applications since they can provide
sub-picosecond pulsewidth at repetition rates from 5 GHz up to 80 GHz, and peak power up to a
few watts. The ultra-short generated pulses offer advantages over QW MLLs for “ optical signal-
ling ” : receiver sensitivity, signal retiming and time division multiplexing are important benefi ts.
Moreover, the sub-picosecond timing jitter of these lasers is interesting for optical clock distribu-
tion and precise non-contact chip testing.
The fi rst report of MLLs for optical interconnects was made in 2005 [33] , where sub-picosec-
ond timing jitter and pulse duration shorter than 10 ps at 5 GHz were demonstrated. Growth
conditions of the InAs/GaAs six DWELL layers were adjusted to target two different applications:
intra-chip/inter-chip clocking and signalling applications at 1250 nm (transparent Si waveguide
and SiGe photodetector) and signal generation and processing for data- and telecommunication.
As discussed in the introduction to this chapter, owing to the large energy separation between
the ground state (GS) and the excited state (ES), quantum dot directly modulated lasers and QD
mode-locked lasers are expected to demonstrate reduced temperature sensitivity compared to
their QW counterparts. This potentially opens the way to fabrication of these devices without
a thermoelectric cooler, thereby reducing the complexity and the cost of the next generation of
optical sources.
However, until recently QD DMLs exhibited a standard characteristic temperature of 5 0 K
[56] . Deppe et al. showed theoretically that the temperature sensitivity of InAs/GaAs QD lasers
was largely determined by the closely spaced energy levels of holes in the valence band [57] .
Hence, thermal hole spill-out is the main limiting physical mechanism which reduces the gain
and the differential gain of InAs/GaAs QD lasers when the temperature increases [57] . Moreover,
as GS gain saturation was evidenced to be responsible for pulse broadening in MLLs, one may
expect degradation of the performances of the QD MLLs with temperature. Improvement of the
temperature stability of QD lasers may be achieved by p -type doping of the active layer [3, 4,
57–60] .
However, the demonstration of greater resilience of QD MLLs to temperature up to 80°C and
investigation of the temperature dependence of pulse duration was recently reported in undoped
layers.
Mode-locking operation using a two-section QD laser in the temperature range 20°C to 80°C
was demonstrated [56] . The active region is based on an undoped fi ve QD layer structure and the
two-section confi guration is similar to that in previous QD MLLs which showed high power sub-
ps pulse generation at 21 GHz [6] . The characteristic temperature T
0
amounts to 41 K and does
not change signifi cantly with the reverse bias in the absorber section. This work led to the obser-
vation of stable mode locking from 20°C to 70°C with signal to noise ratios greater than 20 dB
and FWHM of the RF spectra smaller than 80 kHz. At 80°C, the mode locking became less sta-
ble with an SNR of 15 dB and an FWHM of the beating spectrum of 700 kHz [56] . To maintain
mode locking with temperature above 70°C for a given constant current, the reverse bias voltage
has to be reduced to avoid excess cavity losses which prevents mode locking.
More recently, experimental characterizations of the temporal width of the preceding pulses
demonstrated that the pulse duration in this QD MLL decreases with temperature [61] . The current
is kept at a constant value of 190 mA while the reverse voltage is adjusted to its optimum value in
terms of pulse duration for each temperature. The minimum pulse width decreases from 10 ps
to 6.5 ps from 25°C to 70°C, resulting in a seven-fold decrease of the time bandwidth product
from 20 to 3.5. The spectra were found to become narrower with the temperature (Fig. 20.9).
Hence, the reduction in the pulse duration may be attributed to an increase in the homogeneous
CH020-I046325.indd 614CH020-I046325.indd 614 6/25/2008 10:03:38 AM6/25/2008 10:03:38 AM
Quantum Dot-based Mode-locked Lasers and Applications 615
linewidth, or to thermal coupling of carriers with the wetting layer which impacts on the number
of modes experiencing gain at threshold [61] .
The p -type doping approach was proven to be successful to improve the temperature stabil-
ity of QD lasers with the demonstration of very high characteristic temperature [3, 4] as well as
temperature-independent 10 Gb/s direct modulation up to 85°C [58, 59] . While this technique
has been applied to DML lasers, there are no reports to date of QD MLLs based on p -type doping
of the active region. One has to point out that this technique is accompanied by substantial inter-
nal losses [3, 4] which may prevent mode locking. Thus a trade-off must be found in the future
between improved characteristic temperature and reasonable losses to achieve MLL using p -type-
doped QD active medium.
20.4.2 40 Gb/s all-optical clock recovery at 1.55 μ m
Of particular signifi cance is the recent demonstration of 40 Gbit/s all-optical clock recovery [62,
63] , compliant with ITU-T recommendation. This used a one-section self-pulsating QDash laser
which realizes high-frequency jitter suppression owing to its inherently narrow free-running
spectral linewidth. This result paves the way to all-optical clock recovery up to 160 Gbit/s.
The potential of jitter suppression for clock recovery at 40 Gb/s was demonstrated through the
phase noise measurement of a six QDash layer self-pulsating Fabry–Perot laser. The lasing wave-
length is centred at 1.5 μ m and the laser exhibits a free-running spectral linewidth of 20 kHz
for a fundamental frequency of 40 GHz [62] . The experiment relies on the generation of a refer-
ence clock at 40 GHz whose jitter can be monitored on the incoming data and then measured on
the recovered clock using a jitter analyser in the range [100 Hz; 500 MHz].
Preliminary calibrations require the evaluation of the recovered clock through the phase noise
measurement of the unjittered reference clock and of the self-pulsating QDash laser. The phase
noise curves of the MLL indicate three different regimes [62] as evidenced in an earlier paper
[63] . In the low-frequency region, i.e. below 60 kHz, the phase noise is dominated by the noise
of the incoming signal. The high-frequency region ( 4 MHz), where the noise of the laser itself
dominates, shows a jitter suppression defi ned by a 1/ f
2
slope, attributed to the narrow linewidth
of the RF spectrum. A previous work indeed measured a cut-off frequency of 9 MHz for this laser,
much better than the 100 MHz typical cut-off frequency of bulk DBR lasers for the same repeti-
tion rate [63] . In the intermediate region, the noise originates from both contributions of clock
and self-pulsating laser. The RMS jitter amounts to 0.16 ps and is independent of PRBS length
of incoming OTDM signal at 40 Gbit/s.
Experiments performed using jittered incoming signal evidenced a drastic decrease of the
timing jitter attributed to the fi ltering effect of the QDash laser. The RMS jitter is reduced from
1.37 ps for the incoming data to 0.31 ps for the recovered clock from the analysis of phase noise
curves. Furthermore, the authors report for the fi rst time the jitter transfer function which com-
plies with the ITU standard for clock recovery at 40 Gb/s.
These results highlight the real potential of QDash-based MLLs for all-optical clock recovery at
40 GHz based on the narrow linewidth of the RF spectrum.
30
6
7
8
9
10
40 50 60 70
4
8
12
16
20
Temperature (°C)
Pulsewidth (ps)
Current 190 mA
Time–bandwidth produc
t
Figure 20.9 Pulse duration and time bandwidth product versus temperature [61]. ( Reused with permission from
M.A. Cataluna et al. , Appl. Phys. Lett. 90 , 101102 (2007). Copyright 2007, American Institute of Physics.)
CH020-I046325.indd 615CH020-I046325.indd 615 6/25/2008 10:03:38 AM6/25/2008 10:03:38 AM
616 Handbook of Self Assembled Semiconductor Nanostructures for Novel Devices in Photonics and Electronics
20.5 Conclusion and perspectives
Quantum dot-based monolithic mode-locked lasers have recently been the subject of intense
investigations owing to the unique properties expected from 3D confi nement of carriers.
Signifi cant advances have readily been achieved, opening the way to a large range of photonic
applications.
Both the InAs/GaAs and InAs/InP QD material systems have been investigated with the dem-
onstration of sub-picosecond pulse generation, multi-GHz repetition rates up to 134 GHz for fun-
damental mode locking, and more interestingly record timing jitter which is of prime importance
for many applications. Specifi c confi gurations have also led to improved output powers [6] .
Based on these achievements, the fi rst applications such as clock recovery at 40 Gbit/s in fi bre
optic communication systems, and very short pulse emitters for optical interconnects, have
already been reported.
Further improvements are still expected in device characteristics such as temperature stability,
output power, control of pulse characteristics, etc. to take full advantage of this new generation
of compact, ultra-short pulse sources based on QD material.
Acknowledgements
The authors acknowledge the EC SANDiE Network of Excellence for partial support of this work.
References
1. D. Bimberg , N. Kirstaedter , N.N. Ledentsov , Zh.I. Alferov , P.S. Kop’ev , and V.M. Ustinov , InGaAs-
GaAs quantum-dot lasers , IEEE J. Sel. Topics in Quantum Electron. 3 ( 2 ) , 196 – 205 ( 1997 ) .
2. G.T. Liu , A. Stintz , H. Li , K.J. Malloy , and L.F. Lester , Extremely low room-temperature threshold
current density diode lasers using InAs dots in In
0.15
Ga
0.85
As quantum well, IEE Electron. Lett.
35 ( 14 ) , 1163 – 1165 ( 1999 ) .
3. O.B. Shchekin and D.G. Deppe , Low-threshold high-T
0
1.3- μ m InAs quantum-dot lasers due to
p-type modulation doping of the active region, IEEE Photon. Technol. Lett. 14 ( 19 ) , 1231 – 1233
( 2002 ) .
4. I.I. Novikov , N.Y.u. Gordeev , L.Ya. Karachinski , M.V. Maksimov , Yu.M. Shernyakov , A.R. Kovsh ,
I.l. Krestnikov , A.V. Kozhukhov , S.S. Mikhrin , and N.N. Ledentsov , Effect of p-type doping of the
active region on the temperature stability of InAs/GaAs QD lasers ” , Semiconductors 39 , 477 – 480
( 2005 ) .
5. M. Ishida , N. Hatori , K. Otsubo , T. Yamamoto , Y. Nakata , H. Ebe , M. Sugawara , and Y. Arakawa ,
Low-driving-current temperature-stable 10 Gbit/s operation of p-doped 1.3 μ m quantum dot
lasers between 20 and 90°C, IEE Electron. Lett. 43 ( 4 ) , 219 – 221 ( 2007 ) .
6. E.U. Rafailov , M.A. Cataluna , W. Sibbett , N.D. Il’inskaya , Yu.M. Zadiranov , A.E. Zhukov ,
V.M. Ustinov , D.A. Livshits , A.R. Kovsh , and N.N. Ledentsov , High-power picosecond and femto-
second pulse generation from a two-section mode-locked quantum-dot laser , Appl. Phys. Lett. 87 ,
081107 ( 2005 ) .
7. A. Martinez, C. Gosset, K. Merghem, F. Lelarge, J. Landreau, G. Aubin, and A. Ramdane, Sub-
picosecond pulse generation at 1.56 μ m using a mode-locked quantum dot laser, Techni. Digest of
European Conference on Optical Communications 2005 , paper Tu1.5.4 (2005).
8. T. Akiyama , H. Kuwatsuka , T. Simoyama , Y. Nakata , K. Mukai , M. Sugawara , O. Wada , and
H. Ishikawa , Linear and non linear properties of QD material , IEEE J. Quantum. Electron. 37 ( 8 ) ,
1059 – 1065 ( 2001 ) .
9. D.B. Malins , A. Gomez-Iglesias , S.J. White , W. Sibbett , A. Miller , and E.U. Rafailov , Ultrafast electro-
absorption dynamics in an InAs quantum dot saturable absorber at 1.3 μ m , Appl. Phys. Lett. 89 ,
171111 ( 2006 ) .
10. T.W. Berg and J. Mørk , Quantum dot amplifi ers with high output power and low noise , Appl. Phys.
Lett. 82 ( 18 ) , 3083 – 3085 ( 2003 ) .
CH020-I046325.indd 616CH020-I046325.indd 616 6/25/2008 10:03:39 AM6/25/2008 10:03:39 AM
Quantum Dot-based Mode-locked Lasers and Applications 617
11. X. Huang , A. Stintz , H. Li , L.F. Lester , J. Cheng , and K.J. Malloy , Passive mode-locking in 1.3 μ m
two-section InAs quantum dot lasers, Appl. Phys. Lett. 78 ( 19 ) , 2825 – 2827 ( 2001 ) .
12. P. Vasil’ev . Ultrafast Diode Lasers: Fundamentals and Applications , ( Artech House Publishers ,
Chapters 3 and 4 , 1995).
13. A.E. Siegman and D.J. Kuizenga , Simple analytical expression for AM and FM mode-locked pulses
in homogenous lasers , Appl. Phys. Lett. 14 ( 6 ) , 181 – 182 ( 1969 ) .
14. D.J. Kuizenga and A.E. Siegman , FM and AM mode-locking of the homogeneous laser – Part I:
Theory , IEEE J. Quantum Electron. 6 , 694 – 708 ( 1970 ) .
15. H.A. Haus , A theory of forced mode-locking , IEEE J. Quantum Electron. 11 , 323 – 330 ( 1975 ) .
16. M. Kuntz , G. Fiol , M. Lämmlin , D. Bimberg , M.G. Thompson , K.T. Tan , C. Marinelli , R.V. Penty ,
I.H. White , V.M. Ustinov , A.E. Zhukov , Y.M. Shernyakov , and A.R. Kovsh , 35 GHz mode-locking of
1.3 μ m quantum dot lasers, Appl. Phys. Lett. 85 ( 5 ) , 843 – 845 ( 2004 ) .
17. K.A. Williams , M.G. Thompson , and I.H. White , Long wavelength monolithic mode-locked diode
lasers , New Journal Physics , 6 ( 70 ) , 1 – 30 ( 2004 ) .
18. A. Markus , J.X. Chen , O. Gauthier-Lafaye , J.G. Provost , C. Paranthoën , and A. Fiore , Impact of
intraband relaxation on the performance of a quantum-dot laser , IEEE J. Sel. Topics in Quantum
Electron. 9 ( 5 ) , 1308 – 1314 ( 2003 ) .
19. A. Martinez , A. Lemaître , K. Merghem , L. Ferlazzo , C. Dupuis , A. Ramdane , J.-G. Provost ,
B. Dagens , O. Le Gouezigou , and O. Gauthier-Lafaye , Static and dynamic measurements of the
α -factor of 5-quantum-dot-layer single-mode lasers emitting at 1.3 μ m on GaAs, Appl. Phys. Lett.
86 , 211115 ( 2005 ) .
20. B. Dagens , A. Markus , J.X. Chen , J.-G. Provost , D. Make , O. Le Gouezigou , J. Landreau , A. Fiore ,
and B. Thedrez , Giant linewidth enhancement factor and purely frequency modulated emission
from quantum dot laser , IEE Electron. Lett. 41 ( 6 ) , 323 – 324 ( 2005 ) .
21. D.-Y. Cong , A. Martinez , K. Merghem , G. Moreau , A. Lemaître , J.-G. Provost , O. Le Gouezigou ,
M. Fischer , I. Krestnikov , A.R. Kovsh , and A. Ramdane , Optimisation of a- factor for quantum dot
InAs/GaAs Fabry–Pérot lasers emitting at 1.3 μ m , IEE Electron. Lett. 43 ( 4 ) , 222 – 224 ( 2007 ) .
22. M.-T. Choi , W. Lee , J.-M. Kim , and P.J. Delfyett , Ultrashort, high-power pulse generation from a
master oscillator power amplifi er based on external cavity mode locking of a quantum-dot two-
section diode laser , Appl. Phys. Lett. 87 , 221107 ( 2005 ) .
23. M.A. Cataluna , W. Sibbett , D.A. Livshits , J. Weimert , A.R. Kovsh , and E.U. Rafailov , Stable mode
locking via ground or excited-state transitions in a two-section QD laser , Appl. Phys. Lett.
89 ,
081124 ( 2006 ) .
24. M.A. Cataluna, A.D. McRobbie, W. Sibbett, D.A. Livshits, A.R. Kovsh, and E.U. Rafailov, New mode
locking regime in quantum-dot laser: enhancement by simultaneous cw excited-state emission,
Techni. Digest Conference on Laser and Electro-Optic 2006 , paper CThH3 (2006).
25. M. Kuntz , G. Fiol , M. Lämmlin , D. Bimberg , M.G. Thompson , K.T. Tan , C. Marinelli , A. Wonfor ,
R. Sellin , R.V. Penty , I.H. White , V.M. Ustinov , A.E. Zhukov , Yu.M. Shernyakov , A.R. Kovsh ,
N.N. Ledentsov , C. Shubert , and V. Marembert , Direct modulation and mode locking of 1.3 μ m
quantum dot lasers, New J. Phys. 6 ( 181 ) , 1 – 11 ( 2004 ) .
26. M.G. Thompson, D. Larson, A.R. Rae, K. Yvind, R.V. Penty, I.H. White, J. Hvam, A.R. Kovsh,
S.S. Mikhrin, D.A. Livshits, and I.L. Krestnikov, Monolithic hybrid and passive mode-locked 40 GHz,
Techni. Digest of European Conference on Optical Communications 2006 , paper We4.6.3 (2006).
27. M. Laemmlin , G. Fiol , C. Meuer , M. Kuntz , F. Hopfer , A.R. Kovsh , N.N. Ledentsov , and D. Bimberg ,
Distortion-free optical amplifi cation of 20–80 GHz mode-locked laser pulses at 1.3 μ m using quan-
tum dots, IEE Electron. Lett. 42 ( 12 ) , 697 – 699 ( 2006 ) .
28. M.G. Thompson , A. Rae , R.L. Sellin , C. Marinelli , R.V. Penty , I.H. White , A.R. Kovsh , S.S. Mikhrin ,
D.A. Livshits , and I.L. Krestnikov , Subpicosecond high-power mode locking using fl ared waveguide
monolithic quantum-dot lasers , Appl. Phys. Lett. 88 , 133119 ( 2006 ) .
29. A.A. Lagatsky , E.U. Rafailov , W. Sibbett , D.A. Livshits , A.E. Zhukov , and V.M. Ustinov , Quantum-
dot-based saturable absorber with p-n junction for mode-locking of solid-state lasers , IEEE Photon.
Technol. Lett. 17 ( 2 ) , 294 – 296 ( 2005 ) .
30. Y.-C. Xin, L.F. Lester, A.L. Gray, and L. Zhang, Characterization of the static and dynamic parame-
ters in a 1.3 μ m quantum dot mode-locked laser, Techni. Digest Conference on Laser and Electro-Optic
2007 , paper CMM3 (2007).
CH020-I046325.indd 617CH020-I046325.indd 617 6/25/2008 10:03:39 AM6/25/2008 10:03:39 AM
618 Handbook of Self Assembled Semiconductor Nanostructures for Novel Devices in Photonics and Electronics
31. N. El Dahdah , G. Aubin , J.-C. Harmand , A. Ramdane , A. Shen , F. Devaux , A. Garreau , and
B.-E. Benkelfat , Ultrafast InGaAs/InGaAlAs multiple-quantum-well electro-absorption modulator
for wavelength conversion at high bit rates , Appl. Phys. Lett. 84 ( 21 ) , 4268 – 4270 ( 2004 ) .
32. M.T. Todaro , J.-P. Tourrenc , S.P. Hegarty , C. Kelleher , B. Corbett , G. Huyet , and J.G. McInerney ,
Simultaneous achievement of narrow pulse width and low pulse-to-pulse timing jitter in 1.3 μ m
passively mode-locked quantum-dot lasers, OSA Optics Lett , 31 ( 21 ) , 3107 – 3109 ( 2006 ) .
33. L. Zhang, L. Cheng, A.L. Gray, S. Luong, J. Nagyvary, F. Nabulsi, L. Olona, K. Sun, T. Tumolillo,
R. Wang, C. Wiggins, J. Zilko, Z. ZOu, and P.M. Varangis, Low timing jitter, 5 GHz optical pulses from
monolithic two-section passively mode-locked 1250/1310 nm quantum dot lasers for high speed
optical interconnects, Techn. Digest of Optical Fiber Communications 2005 , paper OWM4 (2005).
34. M.G. Thompson, A. Rae, R.V. Penty, I.H. White, A.R. Kovsh, S.S. Mikhrin, D.A. Livshits, and I.L.
Krestnikov, Absorber length optimisation for sub-picosecond pulse generation and ultra-low jitter
performance in passively mode-locked 1.3 μ m quantum-dot laser diodes, Techn. Digest of Optical
Fiber Communications 2006 , paper OThG3, (2006).
35. M.-T. Choi , W. Lee , J.-M. Kim , and P.J. Delfyett , Ultrashort, high-power pulse generation from a
master oscillator power amplifi er based on external cavity mode locking of a quantum-dot two-
section dioed laser , Appl. Phys. Lett. 87 , 221107 ( 2005 ) .
36. J. Kim , M.-T. Choi , and P.J. Delfyett , Pulse generation and compression via ground and excited
states from a grating coupled passively mode-locked quantum dot two-section diode laser , Appl.
Phys. Lett. 89 , 261106 ( 2006 ) .
37. T. Akiyama , M. Ekawa , M. Sugawara , K. Kawaguchi , H. Sudo , A. Kuramata , H. Ebe , and
Y. Arakawa , An ultrawide-band semiconductor optical amplifi er having an extremely high pen-
alty-free output power of 23 dBm achieved with quantum dots , IEEE Photon. Technol. Lett. 17 ( 8 ) ,
1614 – 1616 ( 2005 ) .
38. D. Von der Linde , Characterization of the noise in continuously operating mode-locked lasers ,
Appl. Phys. B , 39 , 201 – 217 ( 1986 ) .
39. M.G. Thompson , K.T. Tan , C. Marinelli , K.A. Williams , R.V. Penty , I.H. White , M. Kuntz , D. Ouyang ,
D. Bimberg , V.M. Ustinov , A.E. Zhukov , A.R. Kovsh , N.N. Ledentsov , D.-J. Kang , and M.G. Blamire ,
Transform-limited optical pulses from 18 GHz monolithic modelocked quantum dot lasers operat-
ing at 1.3 μ m , IEE Electron. Lett. 40 ( 5 ) , 346 – 347 ( 2004 ) .
40. D.J. Derickson , P.A. Morton , J.E. Bowers , and L.R. Thornton , Comparison of timing jitter in
external and monolithic cavity mode-locked semiconductor-lasers , Appl. Phys. Lett. 59 ( 2 6 ) ,
3372 – 3374 ( 1991 ) .
41. M.-T. Choi , J.-M. Kim , W. Lee , and P.J. Delfyett , Ultralow noise optical pulse generation in an
actively mode-locked quantum-dot semiconductor laser , Appl. Phys. Lett. 88 , 131106 ( 2006 ) .
42. J.-P. Tourrenc , S. O’Donoghue , M.T. Todaro , S.P. Hegarty , M.B. Flynn , G. Huyet , J.G. McInerney ,
L. O’Faolain , and T.F. Krauss , Cross-correlation timing jitter measurement of high power passively
mode-locked two-section quantum-dot lasers , IEEE Photon. Technol. Lett. 18 ( 31 ) , 2317 – 2319
( 2006 ) .
43. R.H. Wang , A. Stintz , P.M. Varangis , T.C. Newell, H. Li , K.J. Malloy , and L.F. Lester ,
IEEE Photon.
Technol. Lett. 13 ( 8 ) , 767 – 769 ( 2001 ) .
44. S. Deubert , A. Somers , W. Kaiser , R. Schwertberger , J.P. Reithmaier , and A. Forchel , J. Cryst. Growth
278 , 346 ( 2005 ) .
45. F. Lelarge , B. Rousseau , B. Dagens , F. Poingt , F. Pommereau , and A. Accard , Room temperature
continuous-wave operation of buried ridge stripe lasers using InAs-InP (100) quantum dots as
active core , IEEE Photon. Technol. Lett. 17 ( 7 ) , 1369 – 1371 ( 2005 ) .
46. J.P. Reithmaier , A. Somers , S. Deubert , R. Schwertberger , W. Kaiser , A. Forchel , M. Calligaro ,
P. Resneau , O. Parillaud , S. Bansropun , M. Krakowski , R. Alizon , D. Hadass , A. Bilenca , H. Dery ,
V. Mikhelashvili , G. Eisenstein , M. Gioannini , I. Montrosset , T.W. Berg , M. van der Poel , J. Mørk , and
B. Tromborg , InP based lasers and optical amplifi ers with wire-/dot-like active regions , J. Phys. D:
Appl. Phys. 38 , 2088 – 2102 ( 2005 ) .
47. F. Lelarge , B. Dagens , J. Renaudier , R. Brenot , A. Accard , F. van Dijk , D. Make , O. Le Gouezigou ,
J.-G. Provost , F. Poingt , J. Landreau , O. Drisse , E. Derouin , B. Rousseau , F. Pommereau , and
G.-H. Duan , Recent advances on InAs/InP quantum dash based semiconductor lasers and optical
amplifi ers operating at 1.55 μ m , IEEE J. Sel. Topics in Quantum Electron. 13 ( 1 ) , 111 – 124 ( 2007 ) .
CH020-I046325.indd 618CH020-I046325.indd 618 6/25/2008 10:03:39 AM6/25/2008 10:03:39 AM