hv
photonics
Review
Progress in Rapidly-Tunable External
Cavity Quantum Cascade Lasers with a
Frequency-Shifted Feedback
Arkadiy Lyakh 1,2,3, *, Rodolfo Barron-Jimenez 1 , Ilya Dunayevskiy 1 , Rowel Go 1,2,3 ,
Eugene Tsvid 1 and C. Kumar N. Patel 1, *
1
2
3
*
Pranalytica, Inc., 1101 Colorado Ave., Santa Monica, CA 90401, USA; rbarron@pranalytica.com (R.B.-J.);
idunay@pranalytica.com (I.D.); rowel.go@knights.ucf.edu (R.G.); etsvid@pranalytica.com (E.T.)
NanoScience Technology Center, University of Central Florida, 12424 Research Pkwy, Orlando,
FL 32826, USA
College of Optics and Photonics, University of Central Florida, Scorpius St 304, Orlando, FL 32826, USA
Correspondence: arkadiy.lyakh@ucf.edu (A.L.); patel@pranalytica.com (C.K.N.P.);
Tel.: +1-407-823-0699 (A.L.); +1-310-458-4493 (C.K.N.P.)
Received: 21 March 2016; Accepted: 13 April 2016; Published: 18 April 2016
Abstract: The recent demonstration of external cavity quantum cascade lasers with optical feedback,
controlled by an acousto-optic modulator, paves the way to ruggedized infrared laser systems with
the capability of tuning the emission wavelength on a microsecond scale. Such systems are of
great importance for various critical applications requiring ultra-rapid wavelength tuning, including
combustion and explosion diagnostics and standoff detection. In this paper, recent research results
on these devices are summarized and the advantages of the new configuration are analyzed in the
context of practical applications.
Keywords: quantum cascade laser; tunable laser; spectroscopy; combustion and explosion
diagnostics; standoff detection
1. Introduction
The development of quantum cascade laser (QCL) configurations with a broad spectral coverage,
and compatible with ultra-fast spectroscopic measurements, is an area of active research, primarily
driven by QCL applications in infrared standoff detection and combustion/explosion diagnostics.
One of the most promising developments in the field has been the demonstration of QCL-based
frequency combs [1] and ultra-fast dual-comb QCL sensors [2]. The dual-comb QCL approach promises
lab-on-a-chip sensors with a short measurement time. These sensors can potentially be integrated
in various compact infrared platforms. The biggest drawback of this approach is a relatively small
spectral coverage, on the order of tens of wavenumbers. This coverage may be insufficient to uniquely
identify compounds with broad absorption spectra.
The widest wavelength tunability, covering a substantial fraction of the QCL gain bandwidth,
has been demonstrated for lasers in the external cavity configuration with a mechanical grating
controlling a wavelength-dependent feedback [3]. This configuration has been in wide use with
all types of gain media including solids, liquids and gases for over forty years. Such lasers are
easy to construct and utilize using proven optical components. Such systems, however, suffer
from two important disadvantages arising from the need to physically rotate a macroscopic spectral
dispersive component—the diffraction grating. These disadvantages are (1) slow wavelength tuning
and (2) sensitivity to mechanical vibrations and shocks. The highest tuning rate reported for traditional
external cavity (EC) QCLs with a mechanical grating is 5 kHz, achieved by tilting a small intracavity
mirror [4]. However, the tuning range is limited. Using a micro opto-electro-mechanical systems
Photonics 2016, 3, 19; doi:10.3390/photonics3020019
www.mdpi.com/journal/photonics
Photonics 2016, 3, 19
Photonics 2016, 3, 19
2 of 12
2 of 12
mirror [4]. However, the tuning range is limited. Using a micro opto-electro-mechanical systems
(MOEMS)scanning
scanninggrating,
grating,Grahmann
Grahmannetetal.al.achieved
achieveda scanning
a scanning
rate
1 kHz
The
improvement
(MOEMS)
rate
ofof
1 kHz
[5].[5].
The
improvement
in
in tuning
speed
demonstrated
for devices,
these devices,
while impressive,
is still insufficient
for many
tuning
speed
demonstrated
for these
while impressive,
is still insufficient
for many important
important applications,
laboratory applications,
including combustion/explosion
a full
spectral
laboratory
including combustion/explosion
diagnosticsdiagnostics
where a fullwhere
spectral
scan
needs
scan
to be performed
onof
the
order
of tens of microseconds
(see, for example,
Inpractical
practical
to
be needs
performed
on the order
tens
of microseconds
(see, for example,
ReferenceRef.6).
[6]). In
situations,for
forapplications
applications
involving
optical
detection
of improvised
explosive
devicesdemand
(IEDs)
situations,
involving
optical
detection
of improvised
explosive
devices (IEDs)
demand
sub-millisecond
spectral
scan
of
a
remote
object.
Sensitivity
to
vibrations
and
shocks,
sub-millisecond spectral scan of a remote object. Sensitivity to vibrations and shocks, especially during
especially
the tunable
laser operation
precludes the
theQCLs
traditional
EC QCLs
the
tunableduring
laser operation
precludes
the consideration
of consideration
the traditionalofEC
in applications
in
applications
where
the
tunable
laser
is
deployed
on
moving
platforms,
such
as
a
vehicle
travelling
where the tunable laser is deployed on moving platforms, such as a vehicle travelling on unpaved
on unpaved
typical ofenvironment
a battlefieldwhere
environment
where
thecan
tunable
lasers can
be to
significant
roads,
typicalroads,
of a battlefield
the tunable
lasers
be significant
help
soldiers
help
to soldiers
in from
protecting
them from
the dangers
of IEDs.
Deployment
on aircraft
including
in
protecting
them
the dangers
of IEDs.
Deployment
on aircraft
including
helicopters
and
helicopters
and
unmanned
aerial
vehicles
(UAVs)
also
requires
ruggedness
derived
from
immunity
unmanned aerial vehicles (UAVs) also requires ruggedness derived from immunity to vibrations
to vibrations
and
shocks. and shocks.
Both of
of the
the above
above needs
needs require
require aa novel
novel tuning
tuning mechanism,
mechanism, which
which does
does not
not rely
rely on
on physical
physical
Both
rotation
of
a
macroscopic
optical
element.
An
ideal
tunable
laser
system
would
be
one
which
is
rotation of a macroscopic optical element. An ideal tunable laser system would be one which is
electronically tuned
tuned with
with no
no moving
moving parts.
parts. Moreover,
Moreover, to
to fully
fully exploit
exploit the
the QCL
QCL gain
gain bandwidth,
bandwidth, the
the
electronically
new
approach
to
a
rapid
QCL
wavelength
control
should
be
compatible
with
the
EC
configuration.
new approach to a rapid QCL wavelength control should be compatible with the EC configuration. Use
Use
of electronic
key tuning
to fastand
tuning
and of
absence
ofparts
moving
parts can
assure
of
electronic
tuningtuning
could could
be the be
keythe
to fast
absence
moving
can assure
immunity
immunity
to and
vibrations
to
vibrations
shocks.and shocks.
2. Experimental
2.
ExperimentalSection
Section
We have
have recently
recently demonstrated
demonstrated aa significantly
significantly increased
increased scanning
scanning rate
rate using
using an
an all-electronic
all-electronic
We
grating as
as aa wavelength
wavelength selective element. The configuration
grating
configuration explored
explored in
in [7,8] is shown in Figure 1.
As with the
ofof
a single
Fabry-Perot
(FP)(FP)
QCL
chipchip
withwith
an ARAs
the EC
EC approach,
approach,emission
emissionwavelength
wavelength
a single
Fabry-Perot
QCL
an
coated back
facet
is controlled
using
a dispersive
AR-coated
back
facet
is controlled
using
a dispersiveoptical
opticalelement
elementlocated
locatedoutside
outsidethe
the laser
laser chip.
However, in
traditional
ECEC
QCLs
withwith
a moving
grating,
the rapid
tuning
However,
in contrast
contrasttotothe
the
traditional
QCLs
a moving
grating,
the wavelength
rapid wavelength
in the demonstrated
externalexternal
cavity configuration
is achieved
using anusing
electrically
controlled
acoustotuning
in the demonstrated
cavity configuration
is achieved
an electrically
controlled
optic modulator
(AOM). In
the AOM,
anAOM,
acoustic
at awave
radioatfrequency
is generated
by applying
acousto-optic
modulator
(AOM).
In the
an wave
acoustic
a radio frequency
is generated
by
the RF electronic
signal to
a piezoelectric
transducer
attached
to anto optical
crystal,
which
is
applying
the RF electronic
signal
to a piezoelectric
transducer
attached
an optical
crystal,
which
transparent
atatthe
is
transparent
thewavelength
wavelengthatatwhich
whichthe
thelaser
laserneeds
needstotooperate
operateand
andalso
alsohas
has low
low acoustic
acoustic losses.
The acoustic
acoustic wave
wave (alternate
(alternate compressions
compressions and rarefactions
rarefactions in the crystal) represents a phase grating,
The
from which
which the
the optical
optical beam
beam traversing
traversing the
the AOM
AOM crystal
crystal can
canbe
bediffracted
diffracted(see
(seeFigure
Figure1).
1).
from
Figure 1.
1. Schematic
Schematic of
of travelling
travelling wave
wave acousto-optic
acousto-optic modulator
modulator tuned
tuned quantum
quantum cascade
cascade laser.
laser.
Figure
The angle of diffraction, θB, is given by:
The angle of diffraction, θ B , is given by:
𝑚𝑛𝜆0
mnλ0
sin⁡𝜃
=
𝐵
sin θ B “
ΛΛ
(1)
(1)
where m is the order of diffraction, n is the optical refractive index of the AOM material, λ0 is the
where m is the order of diffraction, n is the optical refractive index of the AOM material, λ0 is the
optical wavelength and Λ is the acoustic wavelength. For an optical wavelength, λ0 = 10 μm, refractive
optical wavelength and Λ is the acoustic wavelength. For an optical wavelength, λ0 = 10 µm, refractive
Photonics 2016, 3, 19
3 of 12
index of ~4.0 for the germanium AOM crystal, and an acoustic wavelength of ~135 µm (assume an
acoustic driving frequency of 35 MHz), we get a diffraction angle of θ B = ~7.5˝ .
Typical diffraction efficiency of 85% is possible for an acoustic power level of ~30 W into
a germanium AOM crystal, which is comparable to the diffraction efficiency of a mechanical
diffraction grating.
AOMs have been successfully used for rapid and random access discreet tuning of CO2 lasers on
their 9.6 µm and 10.6 µm laser bands [9] and continuous tuning of Ti-sapphire lasers [10]. Reference [10]
also provides earlier data on tuning of dye lasers using acousto-optic modulators.
For practical applications, the demonstrated approach has several key advantages, over either the
QCL arrays or the traditional EC QCLs with a mechanical grating.
‚
‚
‚
‚
‚
Tuning speed. Typical modulation bandwidth for commercially available AO modulators is in
the range of tens of MHz. Hence, the required switching time, between arbitrary wavelengths,
of under 1 µs can be achieved. This switching time is approximately three orders of magnitude
faster than that for EC QCLs with a moving grating. The scanning rate is also much faster than
characteristic mechanical vibrational frequencies of the system. As a consequence, the mirror
vibration does not influence single sweep measurement results.
Ruggedness. In contrast to the EC QCLs with a moving grating, the wavelength tuning
mechanism does not require any mechanical motion. The module can therefore be ruggedized to
meet the most stringent requirements for field applications.
Yield (cost). The module comprises a single QCL chip. The standard single-QCL fabrication
process has a relatively high yield, exceeding 50% for watt-level long-wave infrared devices
(LWIR) devices. In the case of QCL DFB arrays, an alternative approach to fast wavelength tuning
with a broad spectral coverage, however, the yield quickly drops with an increase in number of
elements. For comparison, even a 10-element array processed from the same material would have
a yield of less than 10%. As a consequence, DFB QCL arrays with a broad spectral coverage are
projected to have low yield, making this technical approach impractical for some applications.
Reliability and manufacturability. All the components in the proposed module have been
commercially used for an extended period of time and have proven their long-term reliability.
The high power FP QCLs have been in commercial production for nearly a decade and theAOMs
for LWIR region have been commercially deployed for several decades. Thus, the two critical
components in the proposed configuration, namely FP QCLs and AOMs, represent mature
technologies compatible with a large throughput production.
Practical Implications: Since dispersion is provided by the acoustic grating, the same mechanical
construction can be used for QCLs with different center wavelengths (and different tuning ranges)
unlike the mechanical diffraction grating where a different grating is needed for a different
wavelength range.
3. Results
3.1. Response Time
The response time of the change in the optical wavelength with a change in the AOM frequency is
the main advantage of the new configuration. The response time has two components: (1) propagation
time of the acoustic wave from the piezoelectric transducer to the edge of the optical beam going
through the AOM (this propagation distance is designated here as L); and (2) the propagation time for
the acoustic wave across the optical beam (diameter of the beam is designated as Db ). The first “delay”
is the latency time and does not represent the response time of the change of the optical wavelength
when AOM frequency is changed. The latency time can be shortened, almost arbitrarily, by reducing
the distance between the acoustic transducer and the position of the optical beam. The actual response
time, therefore, is determined by the acoustic wave transit time across the optical beam. For the
Photonics 2016, 3, 19
4 of 12
present case, the time it takes an acoustic wave to propagate from the piezoelectric transducer across
the germanium material to the edge of the optical beam is t1 and time for the acoustic wave to cross
the optical beam is t2 . The time t2 is the actual response time of the AOM for changing the optical
wavelength. As mentioned above, t1 can be shortened to almost zero. However, there are limitations
on how short t2 can be. If t2 is made too short by making the optical beam diameter small, the optical
wave will interact with a fewer number of acoustic waves and therefore the linewidth of the output
will increase. The linewidth of the optical output can be reduced by making the optical beam diameter
larger, but that
occurs
at the expense of the response time. The optimal linewidth/response
time
Photonics
2016, 3, 19
4 of 12
balance is application driven.
the acoustic wave to cross the optical beam is t2. The time t2 is the actual response time of the AOM
Figure 2forshows
measured response time of the system as AOM frequency switches from 35 MHz
for changing the optical wavelength. As mentioned above, t1 can be shortened to almost zero.
(outside of the
tuning
curve)
to 45 on
MHz
ofbe.the
curve).
The
blue
However,
there
are limitations
how (center
short t2 can
If t2tuning
is made too
short by
making
thecurve
optical in Figure 2
beam diameter
small,
the optical
waveand
will the
interact
a fewer
number
acoustic signal
waves and
shows the AOM
frequency
control
signal
redwith
curve
shows
theofoptical
from the laser
therefore the linewidth of the output will increase. The linewidth of the optical output can be reduced
measured with
a room temperature mercury-cadmium-telluride (MCT) detector, which consists of
by making the optical beam diameter larger, but that occurs at the expense of the response time. The
pulses of duration
ns and repetition
period
of 200 ns.
AOM frequency is initially equal to 35 MHz,
optimal 100
linewidth/response
time balance
is application
driven.
Figure 2curve
shows measured
response
timeoptical
of the system
as AOM
switches
35 MHz
outside of the tuning
of the laser.
The
signal
levelfrequency
from the
laserfrom
is at
zero level since
(outside of the tuning curve) to 45 MHz (center of the tuning curve). The blue curve in Figure 2 shows
the Bragg condition is not satisfied anywhere in the laser gain spectrum. AOM frequency is abruptly
the AOM frequency control signal and the red curve shows the optical signal from the laser measured
changed to 45
MHz
at time
t = 0. AOM
frequency of 45 MHz
the peak
of the
with
a room
temperature
mercury-cadmium-telluride
(MCT) corresponds
detector, which to
consists
of pulses
of gain curve.
duration 100 ns
repetition
period
of
200
ns.
AOM
frequency
is
initially
equal
to
35
MHz,
outside
It takes approximately
t1and
= L/v
=
1.25
µs
for
the
acoustic
signal
with
the
new
frequency
to reach the
s
of
the
tuning
curve
of
the
laser.
The
optical
signal
level
from
the
laser
is
at
zero
level
since
the
Bragg
area where the optical beam is incident on the crystal (vs is sound velocity). When the sound signal
condition is not satisfied anywhere in the laser gain spectrum. AOM frequency is abruptly changed
with the newto frequency
reaches
that
area, the
signal starts
growing.
Itcurve.
reaches
its maximum
45 MHz at time
t = 0. AOM
frequency
of 45 optical
MHz corresponds
to the peak
of the gain
It takes
value at t2 = approximately
550 ns + t1 , tin
a good
agreement
with
thewith
calculated
travelling
time
the beam area
1 = L/v
s = 1.25 μs
for the acoustic
signal
the new frequency
to reach
theacross
area where
beam ismm/µs)
incident on~the
crystal
(vs is sound
velocity).
When frequency
the sound signal
with the back
new to 35 MHz
t2 = DB /vs =the
3.4optical
mm/(5.5
600
ns (Figure
2a).
The AOM
switches
frequency reaches that area, the optical signal starts growing. It reaches its maximum value at t2 = 550
at t = 3 µs. The
process is repeated: It takes about t1 + t2 =1.8
µs for the laser to completely shut down
ns + t1, in a good agreement with the calculated travelling
time across the beam area t2 = DB/vs = 3.4
(Figure 2b). mm/(5.5 mm/μs) ~ 600 ns (Figure 2a). The AOM frequency switches back to 35 MHz at t = 3 μs. The
process is repeated: It takes about t1 + t2 =1.8 μs for the laser to completely shut down (Figure 2b).
Figure 2. Switching time data for the AOM-controlled EC QCL setup. (a) AOM frequency switches
Figure 2. Switching
time
datathe
for
the AOM-controlled
EC QCLtosetup.
(a)ofAOM
frequency
switches
from 35 MHz
outside
tuning
curve to 45 MHz corresponding
the center
the tuning
curve.
from 35 MHz
the the
tuning
curvewith
to a451.8MHz
corresponding
thewave
center
of the tuning
curve.
Theoutside
laser follows
AOM signal
μs delay,
resulting from a to
sound
propagation
to
the edge
of the
lasersignal
beam (1.25
μs) a
and
the laserresulting
beam (550 ns);
(b) a
when
AOMwave
signal propagation
switches
The laser follows
the
AOM
with
1.8across
µs delay,
from
sound
to the
back to 35 MHz the process is repeated and it takes 1.8 μs for the laser to shut down.
edge of the laser beam (1.25 µs) and across the laser beam (550 ns); (b) when AOM signal switches back
to 35 MHz the process is repeated and it takes 1.8 µs for the laser to shut down.
Photonics 2016, 3, 19
5 of 12
As mentioned above, the latency time t1 is proportional to L and can be almost arbitrarily
minimized by designing the AOM so that the piezoelectric transducer generating the sound wave is
positioned closer to the beam area. Time t2 , which is the actual wavelength switching time, on the
other hand, is determined by the AOM material and beam size. It is, however, already well below 1 µs.
Thus, if we change the AOM frequency between two values, say f1 and f2 , the laser wavelength will
switch between two corresponding values in a time less than 1 µs, the latency time notwithstanding.
3.2. Continuous Wave Operation
Two variations of the AOM approach are possible. The first one is when the acoustic wave
reflects from the side opposite to the sound wave transducer, which creates a standing sound wave
in the crystal. The standing acoustic waves occur at discrete acoustic frequencies defined by crystal
dimensions and sound wave speed. Therefore, corresponding laser wavelengths are also discrete.
A more interesting approach, explored in References [7,8], is when a sound wave absorber is positioned
opposite end of the AOM crystal from the transducer. This is the travelling sound wave case. Since the
standing wave resonance does not need to be satisfied, continuous wavelength tuning is available for
this configuration.
A corollary property of the travelling wave approach is a frequency-shifted feedback. As the
laser beam propagates through the crystal, it experiences the Doppler shift equal to the sound wave
frequency. The frequency shift is doubled when the beam reflected by the high reflectivity back mirror
propagates through the crystal again before coupling back into the laser waveguide. The frequency
shift prevents constructive interference and formation of axial optical modes in the laser (corresponding
to the round trip time for the overall laser cavity) when zero-reflectivity anti-reflection coating is used
on the laser facet facing the AOM. As a consequence, emission spectrum for the AOM controlled lasers
in the travelling wave configuration is expected to be different from that for traditional Fabry-Perot
and external cavity devices.
Emission spectrum of lasers, including AOM-controlled QCLs, is seeded by spontaneous emission
events in the gain medium. Therefore, phase consideration between these random events is not
important. Each of the spontaneous emission events results into the formation of a broad spectral line
as the beam propagates back and forth in the cavity. The resultant (integrated) spectrum is defined by
the gain/loss spectrum, the AOM frequency response, and the frequency shift.
It was shown in Reference [11] that the integrated emission spectrum for an AOM-controlled laser,
seeded by a broad spontaneous emission, has a nearly Gaussian shape with an estimated spectral full
width at half maximum δϑ given by
´
¯1{3
δϑ « ∆¨ Γ2f
(2)
where ∆ is the roundtrip frequency shift and Γ f is the spectral width of the AOM frequency response.
The expression in Equation (2) was derived assuming that material gain/loss spectral width was much
larger than Γ f . As a consequence, Equation (2) is independent of gain spectrum width. In addition, it
was also assumed in Equation (2) that Γ f was much larger than both ∆ and free spectral range of the
external cavity setup. These assumptions are generally applicable to AOM-controlled QCLs. For the
external cavity setup described in the present work, ∆ = 2 ˆ 45 MHz = 90 MHz and Γ f « 25 cm´1 .
Both parameters are calculated assuming that the beam propagated twice through the modulator
during the roundtrip. For these values of ∆ and Γ f , calculated value for δϑ is approximately 1.3 cm´1 .
Figure 3 shows that measured continuous-wave (CW) linewidth for a 3 mm ˆ 10 µm
AOM-controlled QCL epi-down mounted on an AlN submount at a high power level (output power
~100 mW; driving conditions 1.17 A, 9.3 V, 300 K; AOM frequency 45 MHz) was approximately equal to
1.7 cm´1 , in a good agreement with the model results. The spectra were measured using a Bruker FTIR
spectrometer with approximately 30 s scanning time. The inset for Figure 3 also shows that linewidth
narrowed to ~1 cm´1 in the vicinity of laser threshold as the spectral region where laser gain equals
mirror losses becomes comparable to or less than the spectral width of the AOM response. Linewidth
Photonics 2016, 3, 19
Photonics 2016, 3, 19
6 of 12
6 of 12
did not
change
appreciably
with power level when the laser operated significantly above threshold.
did
not
change
Photonics
2016,
3, 19 appreciably with power level when the laser operated significantly above threshold.
6 of 12
The emission
spectrum
is
slightly
asymmetric,
asaspredicted
[12],and
and
modulated
at free
The emission spectrum is slightly
asymmetric,
predicted in
in Reference
Reference [12],
modulated
at free
didspectral
not change
appreciably
level
when
the laser operated
significantlycoated
above
threshold.
range
of
the laser
chip
due
a residual
reflectivity
of the
the anti-reflection
facet.
For aFor a
spectral
range
of
the
laser
chipwith
duepower
to atoresidual
reflectivity
of
anti-reflection
coated
facet.
Thezero
emission
spectrum
is slightly
asymmetric,
asline
predicted
in Reference [12], and modulated at free
reflectivity
coating,
a broad
single
emission
line
expected.
zero reflectivity
coating,
a broad
single
emission
isisexpected.
spectral range of the laser chip due to a residual reflectivity of the anti-reflection coated facet. For a
zero reflectivity coating, a broad single emission line is expected.
Figure 3. Emission line of the CW AOM-controlled QCL. Inset shows linewidth variation with change
Figure 3. Emission line of the CW AOM-controlled QCL. Inset shows linewidth variation with change
in optical power.
in optical power.
Figure 3. Emission line of the CW AOM-controlled QCL. Inset shows linewidth variation with change
The relatively large linewidth could be a limitation for the usefulness of AOM-controlled EC
in optical power.
with alarge
frequency
shifted
feedback
for some applications.
This linewidth
is, however, EC
TheQCLs
relatively
linewidth
could
be a limitation
for the usefulness
of AOM-controlled
sufficiently narrow for numerous other applications, including improvised explosive device (IED)
QCLs with
a frequency
shifted
feedback
for be
some
applications.
This
linewidth
is, however, sufficiently
The
relatively large
linewidth
could
a limitation
for the
usefulness
of AOM-controlled
EC
detection and combustion/explosion diagnostics, where absorption features are typically in the range
narrow
for numerous
other applications,
including
improvised
explosive
(IED)
QCLs
with a frequency
shifted feedback
for some
applications.
This device
linewidth
is, detection
however, and
of tens of wavenumbers. If necessary, the linewidth can be significantly reduced employing the
sufficiently narrow for
numerous other
applications,
including
improvised
explosive
combustion/explosion
diagnostics,
where
absorption
featurestime.
are typically
in the device
range (IED)
of tens of
standing wave configuration,
without
sacrificing
the response
detection
and
combustion/explosion
diagnostics,
where
absorption
features
are
typically
instanding
the rangewave
wavenumbers.
If
necessary,
the
linewidth
can
be
significantly
reduced
employing
the
The CW tuning curve for the laser at a 100-mW power level is shown in Figure 4. The data were
of tens of without
wavenumbers.
If necessary,
the linewidth can be significantly reduced employing the
configuration,
sacrificing
the response
collected under the
same conditions
as the time.
data in Figure 3. The laser tuning range under these
standing
wave
configuration,
without
sacrificing
the1070
response
time.
parameters
covered
spectral
from
cm−1level
to
1135
cm−1 when
AOM frequency
waswere
Thedriving
CW tuning
curve
for thea laser
at range
a 100-mW
power
is shown
in Figure
4. The data
The
CW
tuning
curve
for
the
laser
at
a
100-mW
power
level
is
shown
in Figure 4. The data were
43.5 MHz
to 46.5 as
MHz.
collectedchanged
under from
the same
conditions
the data in Figure 3. The laser tuning range under these driving
collected under the same conditions as the data
in Figure 3. ´
The
laser tuning range under these
1
parameters
a spectral
fromrange
1070 from
cm´11070
to 1135
AOMAOM
frequency
was was
changed
drivingcovered
parameters
coveredrange
a spectral
cm−1 cm
to 1135when
cm−1 when
frequency
from changed
43.5 MHz
to
46.5
MHz.
from 43.5 MHz to 46.5 MHz.
Figure 4. Tuning curve of the CW AOM-controlled laser under the same laser driving conditions as
the ones listed in Figure 3.
Figure 4. Tuning curve of the CW AOM-controlled laser under the same laser driving conditions as
Figure 4. Tuning curve of the CW AOM-controlled laser under the same laser driving conditions as the
the ones listed in Figure 3.
ones listed in Figure 3.
Figure 5 compares transmission spectra for refrigerant R134a (CAS no. 811-97-2) collected with
a slow (1 ms) and fast (16 μs) sweep time. It is evident from the figure that the measured spectrum
gets smoother as sweep time reduces. The effect can be explained noting that the effective linewidth
is expected to increase for short sweep times. The increase in the linewidth required to match the
Photonics 2016, 3, 19
7 of 12
simulated and experimental results is due to the chirp in the sound wave frequency across the area
on the crystal illuminated by the beam. The chirp is caused by the continuous frequency sweep from
oneFigure
edge of
the tuningtransmission
curve to thespectra
other one
a short
period
time.
This leads
to a spatial
5 compares
for within
refrigerant
R134a
(CASofno.
811-97-2)
collected
with
in and
the fast
phase
grating
period
created
by from
the sound
wave.
frequency spectrum
spread is
a variation
slow (1 ms)
(16 µs)
sweep
time. It
is evident
the figure
thatThe
the measured
approximately
product
of scanning
rate effect
and time
for sound
to cross
the illuminated
gets
smoother asasweep
time
reduces. The
can required
be explained
notingwave
that the
effective
linewidth
In terms
of the parameters
controlled
in the
broadening
𝛿𝜗the
can
isarea.
expected
to increase
for short sweep
times.
Theexperiment,
increase in the
theadditional
linewidth line
required
to match
be estimated
simulated
and as:
experimental results is due to the chirp in the sound wave frequency across the area on
the crystal illuminated by the beam. The chirp is caused by𝑡 the continuous frequency sweep from one
𝑡𝑟
𝛿𝜗 ≈ a⁡𝛥𝜗
(3)
edge of the tuning curve to the other one within
short
𝑇𝑅 ⁡period of time. This leads to a spatial variation
𝑇𝑠𝑤𝑒𝑒𝑝
in the phase grating period created by the sound wave. The frequency spread is approximately a
−1), T
product
andrange
time required
for sound
wave
tosweep
cross
thetime
illuminated
In terms
of
where of
𝛥𝜗scanning
tuning
for the laser
(~70 cm
is the
taken toarea.
sweep
emission
𝑇𝑅 is the rate
the
parameters
controlled
in tuning
the experiment,
thettradditional
linewave
broadening
δϑ can
be estimated
frequency
across
the entire
range, and
is the sound
transit time
across
the beam.as:
Comparison between the experimental and simulation data (Figure 5) shows that Equation (3)
ttr
well describes the additional line broadening
reduction in sweep time: the additional (3)
line
δϑ « ∆ϑwith
TR
Tsweep
−1
broadening for 16 μs sweep time compared to the 1-ms case is estimated to be 2.5 cm , in a good
agreement with 2.1 cm−1 deduced from Figure 5. Therefore, the effective line broadening effect can be
where ∆ϑTR is the tuning range for the laser (~70 cm´1 ), Tsweep is the time taken to sweep emission
taken into account using Equation (3) when ultra-short scanning rates are used.
frequency across the entire tuning range, and ttr is the sound wave transit time across the beam.
Figure
Figure5.5.Comparison
Comparisonbetween
betweenR134a
R134aabsorption
absorptionspectra
spectrameasured
measuredfor
forfast
fast(16
(16µs;
μs;top)
top)slow
slow(1(1ms;
ms;
bottom)
sweep
times.
bottom) sweep times.
3.3. Comparison
Quasi-Continuous
Wave
between
theOperation
experimental and simulation data (Figure 5) shows that Equation (3) well
describes
lineactive
broadening
reductionrise
in sweep
thethe
additional
line broadening
Duethe
to additional
a significant
regionwith
temperature
(up totime:
100 K),
tuning range
of CW EC
´1 , in a good agreement
for
16
µs
sweep
time
compared
to
the
1-ms
case
is
estimated
to
be
2.5
cm
QCLs is limited. It can be significantly extended by operating the QCL in a quasi-cw (QCW), as
´1 deduced from Figure 5. Therefore, the effective line broadening effect can be taken into
with
2.1 cmto
opposed
CW, mode of operation with a MHz-scale repetition rate and a duty cycle in the range of
account
using
Equation
(3) when allows
ultra-short
scanning
rates
are used.
20%–60%. The
QCW operation
the laser
active
region
to cool down during the “off-cycles”,
which in turn leads to a reduced average active region temperature and a higher laser performance
3.3. Quasi-Continuous Wave Operation
[13]. The change of the QCL operation from the CW regime to the QCW regime is expected to lead to
Due to in
a significant
(up to 100
K), the
tuning
range ofpower
CW ECcycling.
QCLs
a change
linewidthactive
due region
to the temperature
wavelengthrise
(thermal)
chirp
during
electrical
isWavelength
limited. It can
be significantly
extended
operating
the QCL
in a quasi-cw
(QCW),
as opposed to
tuning
and linewidth
for theby
QCW
operation
are separately
discussed
below.
CW, mode of operation with a MHz-scale repetition rate and a duty cycle in the range of 20%–60%.
The QCW operation allows the laser active region to cool down during the “off-cycles”, which in turn
leads to a reduced average active region temperature and a higher laser performance [13]. The change
of the QCL operation from the CW regime to the QCW regime is expected to lead to a change in
Photonics 2016, 3, 19
8 of 12
linewidth due to the wavelength (thermal) chirp during electrical power cycling. Wavelength tuning
and linewidth for the QCW operation are separately discussed below.
Photonics 2016, 3, 19
8 of 12
Figure 6 shows a measured tuning curve for the same laser in QCW mode (350 ns pulses and 50%
duty cycle) as
AOM
frequency
was changed
from
MHz
to 48.5
MHz
(approximately
1/3 of the
Figure
6 shows
a measured
tuning curve
for 41.7
the same
laser
in QCW
mode
(350 ns pulses and
duty
cycle)
as
AOM
frequency
was
changed
from
41.7
MHz
to
48.5
MHz
(approximately
1/3
available50%
frequency
range).
AOM
input
power
was
fixed
at
35
W,
which
corresponded
to
approximately
Photonics 2016, 3, 19
8 ofof
12
the
available
frequency
range).
AOM
input
power
was
fixed
at
35
W,
which
corresponded
to
90% diffraction efficiency and estimated 70% effective coupling efficiency of the external cavity.
The
approximately
90% diffraction
efficiency
and for
estimated
70%
effective
coupling
efficiency
of and
the
´
1 (9.8
´1pulses
Figure 6 shows
a measured
tuning
curve
the same
laser
in µm)
QCW
mode
emission
wavelength
for the
EC laser
tuned
from
1020
cm
to
1170(350
cmns
(8.5
µm),
twice
−1
external
cavity.
emission
wavelength
for the EC
laser
tuned
from
1020MHz
cm−1 (approximately
(9.8 μm) to 11701/3
cmof
50% duty
cycle)The
as AOM
frequency
was changed
from
41.7
MHz
to 48.5
that of CW operation for the same laser. These results clearly demonstrate that a much wider tuning
(8.5
twicefrequency
that of CWrange).
operation
for the
same
laser. was
Thesefixed
results
that a much
the μm),
available
AOM
input
power
at clearly
35 W, demonstrate
which corresponded
to
range can
be achieved
in can
QCW
mode
without
sacrificing
average
optical
power.
wider
tuning range
be achieved
in QCWand
mode
without
sacrificing
average
optical
power. of the
approximately
90%
diffraction
efficiency
estimated
70%
effective
coupling
efficiency
external cavity. The emission wavelength for the EC laser tuned from 1020 cm−1 (9.8 μm) to 1170 cm−1
(8.5 μm), twice that of CW operation for the same laser. These results clearly demonstrate that a much
wider tuning range can be achieved in QCW mode without sacrificing average optical power.
Figure 6. Tuning curve of the QCW AOM-controlled laser.
Figure 6. Tuning curve of the QCW AOM-controlled laser.
Measured QCL emission spectrum linewidth for AOM central frequency of 45 MHz was 4.7 cm−1.
The
increase
linewidth
compared
to curve
that for
CW
mode
of
operation
is attributed
the thermal
Measured QCLinemission
spectrum
linewidth
forQCW
AOM
central
frequency
of 45toMHz
was 4.7 cm´1 .
Figure
6. Tuning
of the
AOM-controlled
laser.
wavelength
chirp.
The
linewidth
can
be
changed
by
aligning
the
system
so
that
it
operates
at athermal
The increase in linewidth compared to that for CW mode of operation is attributed to the
different
AOMQCL
central
frequency
(different
Bragg
condition).
a higher
Measured
emission
spectrum
linewidth
for AOM
central Employment
frequency of 45ofMHz
was 4.7AOM
cm−1.
wavelength chirp. The linewidth can be changed by aligning the system so that it operates at a different
frequency
willinresult
into acompared
larger number
of for
illuminated
acoustic
gratingisperiods
and,totherefore,
to
The increase
linewidth
to that
CW mode
of operation
attributed
the thermal
AOM central
frequency
(different
Bragg
condition).
Employment
of a higher
AOM
will
awavelength
proportionately
linewidth.
effect we
for frequency
different
chirp.narrower
The linewidth
canTo
bedemonstrate
changed bythis
aligning
themeasured
system solinewidth
that it operates
at a
result into
a
larger
number
of
illuminated
acoustic
grating
periods
and,
therefore,
to
a
proportionately
AOM
central
frequencies.
The increase
in the central
frequency
from 35 MHz
MHz led
to
different
AOM
central frequency
(different
BraggAOM
condition).
Employment
of toa 55
higher
AOM
afrequency
linear
reduction
emission
linewidth
cm−1measured
to 3.9acoustic
cm−1, as
shown in
Figures
7 and
8. During
narrower
linewidth.
Toindemonstrate
thisfrom
effect
we
linewidth
for different
AOMto central
will result
into a larger
number
of5.7
illuminated
grating
periods
and,
therefore,
these
measurements,
the
position
of the
high frequency
reflectivity
mirror
adjusted
to keep
the
lasing
a proportionately
narrower
linewidth.
To
demonstrate
this effect
wewas
measured
linewidth
for
different
frequencies.
The increase
in the
central
AOM
from
35
MHz
to
55
MHz
led
to a linear
wavelength
unchanged.
´
1
´
1
AOM
central
frequencies.
The
increase
in
the
central
AOM
frequency
from
35
MHz
to
55
MHz
led
reduction in emission linewidth from 5.7 cm to 3.9 cm , as shown in Figures 7 and 8.toDuring
a linear reduction in emission linewidth from 5.7 cm−1 to 3.9 cm−1, as shown in Figures 7 and 8. During
these measurements, the position of the high reflectivity mirror was adjusted to keep the lasing
these measurements, the position of the high reflectivity mirror was adjusted to keep the lasing
wavelength
unchanged.
wavelength
unchanged.
Figure 7. Spectral measurements at three AOM drive frequencies, keeping laser wavelength constant
by adjusting the back reflecting mirror tilt.
Figure 7. Spectral measurements at three AOM drive frequencies, keeping laser wavelength constant
Figure 7.bySpectral
at mirror
three AOM
drive frequencies, keeping laser wavelength constant
adjustingmeasurements
the back reflecting
tilt.
by adjusting the back reflecting mirror tilt.
Photonics 2016, 3, 19
Photonics 2016, 3, 19
Photonics 2016, 3, 19
-1
-1
Linewidth,
[cm
Linewidth,
[cm] ]
9 of 12
9 of 12
9 of 12
5.5
5.5
5.0
5.0
4.5
4.5
4.0
4.0
35
35
40
45
50
40
45
50
AOM frequency, [MHz]
AOM frequency, [MHz]
55
55
Figure 8.
8. Linewidth
Linewidth dependence
dependence on
on AOM
AOM frequency.
frequency.
Figure
Figure
8. Linewidth
dependence on
AOM frequency.
3.4. Preliminary Stand-Off Detection Studies
3.4. Preliminary
Preliminary Stand-Off
Stand-Off Detection
Detection Studies
Studies
3.4.
One of the potential applications for the AOM tuned QCLs is stand-off detection. We have
One of
forfor
thethe
AOM
tuned
QCLs
is stand-off
detection.
We have
One
of the
thepotential
potentialapplications
applications
AOM
tuned
QCLs
is stand-off
detection.
Wecarried
have
carried out preliminary studies of short-range stand-off detection utilizing the experimental
out
preliminary
studies
of
short-range
stand-off
detection
utilizing
the
experimental
geometry
shown
carried out preliminary studies of short-range stand-off detection utilizing the experimental
geometry shown in Figure 9. The target was placed at a distance of 50 cm from the laser source. The
in Figure 9.
The target
was 9.
placed
at a distance
of 50atcm
from theoflaser
source.
laser source.
output was
geometry
shown
in Figure
The target
was placed
a distance
50 cm
from The
the laser
The
laser output was attenuated using an appropriate neutral density filter to assure that the detector was
attenuated
using
an
appropriate
neutral
density
filter
to
assure
that
the
detector
was
not
saturated.
laser output was attenuated using an appropriate neutral density filter to assure that the detector was
not saturated. A 1 cm diameter, 50 mm focal length lens, located 12 cm from the target was used to
A 1 saturated.
cm diameter,
mm
focal length
lens,
located
12lens,
cm from
the 12
target
was used
to collect
laser
not
A 150cm
diameter,
50 mm
focal
length
located
cm from
the target
wasthe
used
to
collect the laser light scattered from the target and focus light on a cooled MCT detector. The total
light scattered
target and
focus
light on
a cooled
MCTon
detector.
The
totaldetector.
distanceThe
fromtotal
the
collect
the laserfrom
lightthe
scattered
from
the target
and
focus light
a cooled
MCT
distance from the target to the detector was 25 cm.
target to from
the detector
wasto25
cm.
distance
the target
the
detector was 25 cm.
Figure 9. Schematic of an AOM-controlled QCL setup for short-range standoff detection.
Figure 9.
9. Schematic
Schematic of
of an
an AOM-controlled
AOM-controlled QCL setup for short-range standoff detection.
Figure
Figure 10 shows the data collected for a dextrose tablet (top graph), dissolved in water dextrose
Figure
10
shows
the
collected
for
tablet
(top
dissolved
in
water
Figure
10an
shows
the data
data
collected
for aa dextrose
dextrose
tablet
(top graph),
graph),
dissolved
indeposited
water dextrose
dextrose
deposited on
aluminum
substrate
(middle
graph), and
dissolved
in water
aspirin
on an
deposited
on
an
aluminum
substrate
(middle
graph),
and
dissolved
in
water
aspirin
deposited
on
deposited
on
an
aluminum
substrate
(middle
graph),
and
dissolved
in
water
aspirin
deposited
on an
an
aluminum substrate (bottom graph). For the latter two samples, after they were dissolved in water,
aluminum
substrate
(bottom
graph).
For
the
latter
two
samples,
after
they
were
dissolved
in
water,
aluminum
(bottom
graph).
the latter
twoComposition
samples, afterof
they
dissolved
in water,
they weresubstrate
subsequently
dried
out For
before
testing.
thewere
tested
substances
wasthey
not
they
were subsequently
dried
outtesting.
before Composition
testing. Composition
of substances
the tested was
substances
was not
were
subsequently
dried
out
before
of
the
tested
not
controlled
in
controlled in this prove-of-concept experiment. The measured spectra for the dextrose tablet and
controlled
in
this
prove-of-concept
experiment.
The
measured
spectra
for
the
dextrose
tablet
and
this
prove-of-concept
experiment.
measured substrate
spectra forare
thedifferent
dextrose because
tablet and
dextrose
dissolved
dextrose deposited
on The
an aluminum
of dissolved
binders and
other
dissolved
dextrose
deposited
on an aluminum
substrate
are binders
differentand
because of
binders impurities
and other
deposited
an aluminum
different
because
unknown
unknown on
impurities
in the substrate
tablet. Theare
scan
time for
each ofofthe
traces wasother
approximately
700 μs. The
unknown
impurities
in time
the tablet.
Theof
scan
time
forwas
each of the traces was
approximately
700 μs.data,
The
in
the tablet.
The scan
for each
thefilter,
traces
700
µs. The
preliminary
preliminary
data,
obtained
using
an ND2
clearly approximately
demonstrate the
“single
shot”
data gathering
preliminary
data, obtained
using
an ND2
filter, clearly
demonstrate
the
shot” data gathering
obtained
ND2 filter,
clearly
demonstrate
the “single
data“single
gathering
made
capabilityusing
madeanpossible
using
the AOM
tuned QCL
system.shot”
The data
also
show capability
that even at
this
capability
made
possible
using
the
AOM
tuned
QCL
system.
The
data
also
show
that of
even
atearly
this
possible
using
the
AOM
tuned
QCL
system.
The
data
also
show
that
even
at
this
stage
very
stage of very early demonstration of the capability of AOM-controlled QCLs (VeloXscan), it is very
stage
of very early
demonstration
the capability of
AOM-controlled
QCLs
(VeloXscan),
it is very
demonstration
of the
capability
of of
AOM-controlled
QCLs
(VeloXscan),
it is very
easy to distinguish
easy to distinguish
between
various
compounds in Figure
10.
easy
to distinguish
between various
compounds
in Figure 10.
between
various compounds
in Figure
10.
Photonics 2016, 3, 19
Photonics 2016, 3, 19
10 of 12
10 of 12
Figure10.10.
Measured
short-range
stand-off
detection
signal
for dextrose
tablet
(top),dissolved
dextrose
Figure
Measured
short-range
stand-off
detection
signal for
dextrose
tablet (top),
dextrose
dissolved
in
water
(middle),
and
aspirin
dissolved
in
water
(bottom).
in water (middle), and aspirin dissolved in water (bottom).
Using the present setup we estimate that data of comparable quality can be easily obtained for
Using the present setup we estimate that data of comparable quality can be easily obtained for a
a stand-off distance of 2 meters by simply removing the ND2 filter and using a 12′ diameter scattered
stand-off distance of 2 meters by simply removing the ND2 filter and using a 2 diameter scattered
light-collecting lens.
light-collecting lens.
3.5. Future Directions for AOM Tuned QCLs
3.5. Future Directions for AOM Tuned QCLs
For the configuration shown in Figure 1, a substantial fraction of the optical power is lost with
For the configuration shown in Figure 1, a substantial fraction of the optical power is lost with
the undiffracted beam (diffraction efficiency ~80%). The undiffracted beam can be used as the output
the undiffracted beam (diffraction efficiency ~80%). The undiffracted beam can be used as the output
beam, as shown in Figure 11. A very interesting aspect of the new configuration is that AOM RF
beam, as shown in Figure 11. A very interesting aspect of the new configuration is that AOM RF power
power can be used to balance the laser feedback and output power. For example, AOM RF power
can be used to balance the laser feedback and output power. For example, AOM RF power can be
can be (dynamically) tuned to reach the optimal efficiency for each emission wavelength under the
(dynamically) tuned to reach the optimal efficiency for each emission wavelength under the tuning
tuning curve. This is analogous to controlling front facet reflectivity for the traditional Fabry-Perot
curve. This is analogous to controlling front facet reflectivity for the traditional Fabry-Perot chips.
chips. However, instead of being fixed by coating layers thicknesses, in this case, the effective
However, instead of being fixed by coating layers thicknesses, in this case, the effective reflectivity can
reflectivity can be easily electronically adjusted on a microsecond scale.
be easily electronically adjusted on a microsecond scale.
In contrast to the mechanical grating configuration, a single AOM can control several lasers
emitting at different wavelengths at the same time. As shown in Figure 12, two or more emitters with
complementary spectral coverages can be combined in a single system without increasing system
size and weight. Each of these emitters can be designed to have a broad gain and, therefore, broad
tuning range capability [14,15]. Their output optical beams can be combined in a single beam using
either the polarization beam combining approach or the dichroic beam combining approach. The
multiple-emitter configuration is projected to result into a tunable laser source with an ultra-broad
coverage (4–12 μm) and ultra-rapid wavelength tuning.
Additionally, unlike mechanical grating tuned QCLs, the AOM-tuned QCLs can be made to
operate at more than one wavelength at a time by driving the AOM crystal with a combination of
frequencies. Each one of the RF drive frequency sets up an independent grating period, which forces
the QCL to operate at different wavelengths simultaneously. As a consequence, several absorption
features2016,
can3,be19tracked at the same time. Again, the wavelengths can be changed, independently,
Photonics
11 ofon
12
a microsecond time scale.
Photonics 2016, 3, 19
11 of 12
Additionally, unlike mechanical grating tuned QCLs, the AOM-tuned QCLs can be made to
operate at more than one wavelength at a time by driving the AOM crystal with a combination of
frequencies. Each one of the RF drive frequency sets up an independent grating period, which forces
the QCL to operate at different wavelengths simultaneously. As a consequence, several absorption
features can be tracked at the same time. Again, the wavelengths can be changed, independently, on
a microsecond time scale.
Figure 11.
11. Configuration
for extracting
extracting higher
higher power
power output
output form
form the
the AOM-tuned
AOM-tuned QCL.
QCL.
Figure
Configuration for
In contrast to the mechanical grating configuration, a single AOM can control several lasers
emitting at different wavelengths at the same time. As shown in Figure 12, two or more emitters with
complementary spectral coverages can be combined in a single system without increasing system
size and weight. Each of these emitters can be designed to have a broad gain and, therefore, broad
tuning range capability [14,15]. Their output optical beams can be combined in a single beam using
either the polarization beam combining approach or the dichroic beam combining approach. The
multiple-emitter configuration is projected to result into a tunable laser source with an ultra-broad
coverage (4–12
and ultra-rapid
tuning.
Figureµm)
11. Configuration
for wavelength
extracting higher
power output form the AOM-tuned QCL.
Figure 12. Dual-laser AOM-tuned QCL setup.
The AOM that was used in References [7,8] required water-cooling. However, our experimental
results showed that input RF power of only 20–25 W was sufficient for efficient module operation.
That level of power dissipated in the crystal can be removed using thermo-electrical cooling
(TEC)/air-cooling, a much more compact and power efficient approach. The air-cooled AOMcontrolled QCL can be packaged in a “shoebox-size” platform designed for handheld applications.
4. Conclusions
Figure 12. Dual-laser AOM-tuned QCL setup.
The control of an all electronically tuned external cavity QCL using an AOM opens up unique
Figure 12. Dual-laser AOM-tuned QCL setup.
opportunities
in the
infrared
spectroscopy
field.
A compact
(handheld),
reliable,
ruggedized
Additionally,
unlike
mechanical
grating
tuned
QCLs, the
AOM-tuned
QCLs and
can be
made to
tunable
system
can
be
designed
to
cover
the
entire
MWIR
and
LWIR
spectral
regions
with
an ultraTheat
AOM
was
used
in References
water-cooling.
However,
experimental
operate
morethat
than
one
wavelength
at a [7,8]
time required
by driving
the AOM crystal
with our
a combination
of
fast tuning
time
and
a of
very
flexible
(multi)
wavelength
control.
All the
blocks
of
the
system
results
showed
that
input
RF
power
of
only
20–25
sufficient
forbuilding
efficientperiod,
module
operation.
frequencies.
Each
one
the
RF
drive
frequency
setsWupwas
an independent
grating
which
forces
represent
technologies
with
highAs
yield,
large thermo-electrical
throughput
production.
That
leveltomature
of
power
inare
thecompatible
crystal
can
bea removed
using
cooling
the
QCL
operate
at dissipated
differentthat
wavelengths
simultaneously.
a consequence,
several
absorption
(TEC)/air-cooling,
a much
and
power
efficient
approach.
Thecontract
air-cooled
AOMfeatures
can be tracked
atreported
themore
same
time.
Again,
wavelengths
can
changed,
independently,
on a
Acknowledgments:
Work
incompact
this paper
wasthe
supported,
in part,
bybe
Army
STTR
#W911SR-14C-0067
under
the
oversight
of
Alan
Samuels
(Edgewood
Chemical
Biological
Center).
controlled
QCL
can
be
packaged
in
a
“shoebox-size”
platform
designed
for
handheld
applications.
microsecond time scale.
The
AOM that was
used in
References
required
water-cooling.
our experimental
Author
Contributions:
A. Lyakh
and
C. Kumar [7,8]
N. Patel
conceived
and designedHowever,
the experiments,
A. Lyakh, R.
4.
Conclusions
results
showed
that
input
RF
power
of
only
20–25
W
was
sufficient
for
efficient
module
Barron-Jimenez, I. Dunayevskiy, R. Go, and E. Tsvid performed the experiment and analyzed data. operation.
ThatThe
level of power
dissipated
in the crystal
can be
removed
using an
thermo-electrical
of an
electronically
external
cavity
QCL using
AOM opens up cooling
unique
Conflicts control
of Interest:
Theall
authors
declare notuned
conflict
of interest.
(TEC)/air-cooling,
a
much
more
compact
and
power
efficient
approach.
The air-cooled
AOMopportunities in the infrared spectroscopy field. A compact (handheld), reliable,
and ruggedized
controlled
QCL
can
be
packaged
in
a
“shoebox-size”
platform
designed
for
handheld
applications.
tunable system can be designed to cover the entire MWIR and LWIR spectral regions with an ultrafast tuning time and a very flexible (multi) wavelength control. All the building blocks of the system
represent mature technologies that are compatible with a high yield, large throughput production.
Acknowledgments: Work reported in this paper was supported, in part, by Army STTR contract #W911SR-14C-0067 under the oversight of Alan Samuels (Edgewood Chemical Biological Center).
Author Contributions: A. Lyakh and C. Kumar N. Patel conceived and designed the experiments, A. Lyakh, R.
Barron-Jimenez, I. Dunayevskiy, R. Go, and E. Tsvid performed the experiment and analyzed data.
Conflicts of Interest: The authors declare no conflict of interest.
Photonics 2016, 3, 19
12 of 12
4. Conclusions
The control of an all electronically tuned external cavity QCL using an AOM opens up unique
opportunities in the infrared spectroscopy field. A compact (handheld), reliable, and ruggedized
tunable system can be designed to cover the entire MWIR and LWIR spectral regions with an ultra-fast
tuning time and a very flexible (multi) wavelength control. All the building blocks of the system
represent mature technologies that are compatible with a high yield, large throughput production.
Acknowledgments: Work reported in this paper was supported, in part, by Army STTR contract
#W911SR-14-C-0067 under the oversight of Alan Samuels (Edgewood Chemical Biological Center).
Author Contributions: A. Lyakh and C. Kumar N. Patel conceived and designed the experiments, A. Lyakh, R.
Barron-Jimenez, I. Dunayevskiy, R. Go, and E. Tsvid performed the experiment and analyzed data.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Andress, H.; Gustavo, V.; Stephane, B.; Liu, H.C. Mid-infrared frequency comb based on a quantum cascade
laser. Nature 2012, 492, 229–233.
Villares, G.; Wolf, J.; Kazakov, D.; Suess, M.; Hugi, A.; Beck, M.; Faist, J. On-chip dual-comb based on
quantum cascade laser frequency combs. Appl. Phys. Lett. 2015, 107. [CrossRef]
Luo, G.P.; Peng, C.; Le, H.Q.; Pei, S.S.; Hwang, W.-Y.; Ishaug, B.; Um, J.; Baillargeon, J.N.; Lin, C.-H.
Grating-tuned external-cavity quantum-cascade semiconductor lasers. Appl. Phys. Lett. 2001, 78, 2834–2836.
[CrossRef]
Tsai, T.; Wysocki, G. External-cavity quantum cascade lasers with fast wavelength scanning. Appl. Phys.
2010, 100, 243–251. [CrossRef]
Grahmann, J.; Merten, A.; Herrmann, A.; Ostendorf, R.; Bleh, D.; Drabe, C.; Kamenz, J. Large MOEMS
Diffraction Grating Results Providing an EC-QCL Wavelength Scan of 20%. Proc. SPIE 2015, 9375. [CrossRef]
Koroglu, B.; Pryor, O.; Lopez, J.; Nash, L.; Vasu, S. Shock tube ignition delay times and methane time-histories
measurements during excess CO2 diluted oxy-methane combustion. Combust. Flame 2016, 164, 152–163.
[CrossRef]
Lyakh, A.; Barron-Jimenez, R.; Dunayevskiy, I.; Go, R.; Patel, C.; Kumar, N. Continuous wave operation of
quantum cascade lasers with frequency shifted feedback. AIP Adv. 2016, 6, 29715. [CrossRef]
Lyakh, A.; Barron-Jimenez, R.; Dunayevskiy, I.; Go, R.; Patel, C.; Kumar, N. External cavity quantum cascade
lasers with ultra rapid acousto-optic tuning. Appl. Phys. Lett. 2015, 106, 2834. [CrossRef]
Thompson, D.C.; Bush, G.E.; Hewitt, C.G.; Remelius, D.K.; Shimada, T.; Strauss, C.E.; Wilson, C.W.; Zaugg, T.J.
High-speed random access laser tuning. Appl. Opt. 1999, 38, 2545–2553. [CrossRef]
Yumoto, M.; Maeda, Y.; Saito, N.; Norihito, S.; Takyo, O.; Masafumi, Y.; Satoshi, W. Electronic Wavelength
Tuning of Tunable Laser with Acousto-Optic Tunable Filter. Jpn. J. Appl. Phys. 2008, 11, 8411–8415. [CrossRef]
Yatsenko, L.P.; Shore, B.W.; Bergmann, K. Theory of a frequency-shifted feedback laser. Opt. Commun. 2004,
236, 183–202. [CrossRef]
Streifer, W.; Whinnery, J.R. Analysis of a dye laser tuned by acousto-optic filter. Appl. Phys. Lett. 1970, 17,
335–337. [CrossRef]
Maulini, R.; Lyakh, A.; Tsekoun, A.; Go, R.; Patel, C.; Kumar, N. High average power uncooled mid-wave
infrared quantum cascade lasers. Electron. Lett. 2011, 47, 395–397. [CrossRef]
Maulini, R.; Beck, M.; Faist, J.; Gini, E. Broadband tuning of external cavity bound-to-continuum quantum
cascade lasers. Appl. Phys. Lett. 2004, 84, 1659–1661. [CrossRef]
Maulini, R.; Yarekha, D.A.; Bulliard, J.-M.; Giovannini, M.; Faist, J. Continuous-wave operation of a broadly
tunable thermoelectrically cooled external cavity quantum cascade laser. Opt. Lett. 2005, 30, 2584–2586.
[CrossRef] [PubMed]
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).