Light induced tuning of quantum cascade lasers

Article abstract of DOI:10.1063/1.3473780

A method for tuning the emission of a midinfrared quantum cascade laser using ultraviolet light is presented. The method uses a quantum cascade laser where a photochromic material is deposited as the top waveguide cladding. Changing the state of the cladding causes variations in the absorbance of the cladding. Wavelength dependent changes in the waveguide losses occur, forcing the laser to emit at a wavelength correlated with an absorbance minimum of the cladding. A blueshift by about 6 cm^-1 was obtained at room temperature under UV irradiation which was fully reversed by exposure to visible light or room temperature relaxation.

Full text of DOI:10.1063/1.3473780

Light induced tuning of quantum cascade lasers
B. Basnar, E. Mujagic, A. M. Andrews, T. Roch, W. Schrenk, and G. Strasser
Citation: Applied Physics Letters 97, 051106 (2010); doi: 10.1063/1.3473780
View online: http://dx.doi.org/10.1063/1.3473780
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/97/5?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Electrical laser frequency tuning by three terminal terahertz quantum cascade lasers
Appl. Phys. Lett. 104, 011107 (2014); 10.1063/1.4861122
Extended electrical tuning of quantum cascade lasers with digital concatenated gratings
Appl. Phys. Lett. 103, 231110 (2013); 10.1063/1.4841635
Wavelength tuning and thermal dynamics of continuous-wave mid-infrared distributed feedback quantum
cascade lasers
Appl. Phys. Lett. 103, 031111 (2013); 10.1063/1.4813753
Widely tuned room temperature terahertz quantum cascade laser sources based on difference-frequency
generation
Appl. Phys. Lett. 101, 251121 (2012); 10.1063/1.4773189
Mode tuning of quantum cascade lasers through optical processing of chalcogenide glass claddings
Appl. Phys. Lett. 89, 041115 (2006); 10.1063/1.2236296
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.143.199.160 On: Mon, 15 Dec 2014 10:59:26

APPLIED PHYSICS LETTERS 97, 051106 ͑2010͒
Light induced tuning of quantum cascade lasers
B. Basnar,1,a͒ E. Mujagic,1 A. M. Andrews,1 T. Roch,2 W. Schrenk,1 and G. Strasser1,3
1Center for Micro- and Nanostructures, Vienna University of Technology, A-1040 Vienna, Austria
2Department of Experimental Physics, Comenius University, 84248 Bratislava, Slovakia
3Departments of Electrical Engineering and Physics, The State University at Buffalo, Buffalo,
New York 14260-1920, USA
͑Received 25 June 2010; accepted 8 July 2010; published online 3 August 2010͒
A method for tuning the emission of a midinfrared quantum cascade laser using ultraviolet light is
presented. The method uses a quantum cascade laser where a photochromic material is deposited as
the top waveguide cladding. Changing the state of the cladding causes variations in the absorbance
of the cladding. Wavelength dependent changes in the waveguide losses occur, forcing the laser to
emit at a wavelength correlated with an absorbance minimum of the cladding. A blueshift by about
6 cm−1 was obtained at room temperature under UV irradiation which was fully reversed by
exposure to visible light or room temperature relaxation. © 2010 American Institute of Physics.
͓doi:10.1063/1.3473780͔
Quantum cascade lasers ͑QCLs͒ are versatile light
sources for the midinfrared to far-infrared wavelength re-
gime. Since their invention in 1994,1 tremendous progress
has been made in regards to achievable wavelengths, operat-
ing temperatures, optical power, and wall-plug efficiency.
Additionally for spectroscopy, wavelength selection and tun-
ing are crucial parameters.
The QCL wavelength can be designed by varying the
thickness or composition of the individual quantum wells
and barriers. Out of the broad gain spectrum, a single wave-
length can then be chosen using different techniques. The
most common methods involve the fabrication of an
index-coupled2 or loss-coupled3 distributed feedback grating.
In these cases, the wavelength is determined by the grating
period and by the effective refractive index of the optical
mode.
bance within the gain region. A chromic compound can un-
dergo reversible changes in the absorbance which, in turn,
causes reversible changes in the emitted wavelength.
Photochromism is the reversible change in the absor-
bance as a function of illumination. Such systems are under
investigation for optical data storage.9 Spiropyrans have
been intensively researched in the visible light regime10
but they also exhibit significant spectral changes in
midinfrared. However, the photochromic behavior is not
present in the crystalline state, only a solution is capable
of the photoinduced transformation. Thus, the spiropyran
was dissolved in a mixture of polystyrene ͑PS͒ and toluene.
PS is, due to the aromatic ring, a good matrix for the spiro-
pyran, which is also of aromatic nature. Additionally, it
does not exhibit absorbance in the wavelength range
of interest ͑1250–1350 cm−1͒. The coating solution was
prepared by dissolving 25 mg 1,3,3-trimethylindolino-6 -
Ј
In order to tune the wavelength, the effective refractive
index can be modified. This is usually done by utilizing the
temperature dependence of the refractive index. Tuning rates
of about 0.5 nm/K have been achieved.2 Similarly, lasers
with leaky waveguides can be tuned by changing the refrac-
tive index of the surrounding media, thus altering the effec-
tive index for the laser mode. Liquids with differing refrac-
tive indices have enabled a tuning by about 1.15 cm−1.4
Deposition of a chalcogenide ͑As2S3͒ cladding5 led to a
3.8 cm−1 shift in emission and a subsequent illumination
with UV light allowed for an additional shift of 1.31 cm−1.
A more sophisticated approach to wavelength selection
and tuning utilizes an external cavity with a grating. External
cavity QCLs with an especially broad gain spectrum showed
continuous single-mode tuning over a range of more than
250 cm−1.6 However, they require high quality antireflection
coatings and complex mechanical set-ups.
nitrobenzopyrylospiropyran ͑obtained from TCI Europe͒ and
35 mg PS in 1 ml toluene. Evaporation of the toluene during
the spin-coating procedure led to a transparent film where the
spiropyran is dissolved in the PS matrix, thus enabling the
photochromic transition of the spiropyran.
The spiro state of the compound ͑i.e., two organic rings
sharing a single carbon atom͒ is colorless with several dis-
tinct absorbance peaks between 1000 and 1400 cm−1. Upon
illumination with UV light ͑mercury lamp at 365 nm wave-
length with optical power density of 0.7 mW/cm2͒ a ring-
opening reaction occurs, forming the blue merocyanine state.
Figure 1 shows the spectra of the two states. The inset de-
picts the shift over time of the absorbance minimum around
1320 cm−1.
The laser used is an InGaAs/InAlAs QCL lattice
matched to InP with a gain maximum around 1300 cm−1 at
room temperature. Laser ridges were processed with an air
waveguide on top. They had a length of 2.55 mm and widths
between 30 and 60 ␮m. After indium bonding to a copper
submount and wire bonding of the ridges, the laser bars were
spin-coated with the spiropyran/PS cladding to obtain a 500
nm thick film.
We have recently developed a method for the wave-
length selection and tuning based on chromic claddings.7,8
This method utilizes the interaction of the evanescent field of
a laser with the chromic material. The absorbance of the
cladding influences the light amplification within the laser,
focusing the emission wavelength to an area of low absor-
The laser was operated in pulsed mode ͑100 ns pulses, 5
kHz repetition rate͒. The uncoated laser shows a Fabry–Perot
type spectrum between 1270 and 1320 cm−1 ͑see Fig. 2͒.
a͒
Electronic mail: basnar@tuwien.ac.at.
0003-6951/2010/97͑5͒/051106/3/$30.00
97, 051106-1
© 2010 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.143.199.160 On: Mon, 15 Dec 2014 10:59:26

051106-2
Basnar et al.
Appl. Phys. Lett. 97, 051106 ͑2010͒
FIG. 3. ͑Color online͒ Shift in the emission of a 60 ␮m ridge QCL with
spiropyran coating as a function of time of irradiation as follows: ͑a͒ shows
the spectra recorded at room temperature under irradiation with a 365 nm
UV lamp, ͑b͒ shows the shift in the center wavelength, and ͑c͒ shows the
emitted power ͓area underneath the laser peak in part ͑a͔͒ with time of
irradiation.
FIG. 1. ͑Color online͒ Absorbance spectra of a 230 nm thick film of spiro-
pyran and merocyanine in PS. The inset depicts the shift in the absorbance
minima around 1320 cm−1 as a function of time of irradiation with 365 nm
UV light.
After coating, the laser emission maximum is blueshifted to
1314 cm−1 with a bandwidth of approximately 4 cm−1. Ex-
posure to UV light causes an additional blueshift by 6 cm−1
to 1320 cm−1.
Looking at the evolution of the emission during the tran-
sition between spiro and mero form ͑Fig. 3͒, one can see
several features. First, the emission spectrum still consists of
2–4 Fabry–Perot modes. Second, in the beginning of the
transition, a small extra peak at 1290 cm−1 appears, which
correlates to another minimum in the absorbance spectrum.
Although it experiences higher losses, the closer proximity to
the gain maximum favors emission. This peak disappears
over UV irradiation time as the losses in this area increase
further.
Plotting the power ͑calculated as the integral of the IR
peaks͒ and the peak center position over time, it is evident
that the shift in peak position comes at the expense of emit-
ted power. This is easily understood, keeping in mind that
both the losses increase and the peak position moves away
from the gain maximum as the compound is converted from
the spiro to the mero state.
The time for the transformation from spiro to mero is
about 20 min, which is the same as for the absorbance
changes in the cladding. The reverse direction, however, re-
quires an order of magnitude more time. This is due to the
low power of the halogen lamp used for this step. Neverthe-
less, after 300 min the laser has fully reverted to its original
state. The transition from mero to spiro can also occur via
thermal relaxation. At room temperature, this process re-
quires over 10 h, but at 80 °C less than 60 s are needed.
These times are not intrinsic to the spiropyran but are
rather caused by the low power light sources and the mount-
ing in a cryostat for subsequent measurements at lower tem-
peratures. Switching times in the range of 3 min ͑UV͒ and 9
min ͑visible͒ were achieved without the cryostat due to
closer proximity and, thus, higher power density of the
lamps. Using high power UV lamps ͑12 mW/cm2 at 365
nm͒ switching times below 1 min were achieved.
The emitted wavelength follows the shift in the absor-
bance minimum qualitatively. In absolute numbers, however,
the laser wavelength has an offset of several cm−1. For the
spiro state of the cladding, the laser emits at 1314 cm−1,
which is still inside of the minimum but not at the wave-
length of lowest loss. However, the gain at this wavelength is
higher than at 1316 cm−1, thus compensating the slightly
higher losses.
In the mero-state, the situation is different. As the mero
compound exhibits a higher UV absorbance than the spiro
state, the top layers of the cladding effectively shield the
underlying material from the UV light, thus blocking the
photochromic transition. Layers up to 200 nm can be con-
verted fully. The cladding on our lasers ͑approx 500 nm͒,
however, is only partially transformed, leading to a superpo-
sition of the spectra of the two states. Using the actual wave-
length of the laser at maximum switching, one can estimate
that the average composition the laser experiences is about
70% spiro and 30% mero, leading to a shift in the minimum
from 1316.5 to 1322 cm−1.
Looking at the room temperature light-current-voltage
͑LIV͒ curves ͑Fig. 4͒ a strong increase in the threshold cur-
rent density and drop in the maximum output power is seen
when coating the laser with the cladding material. This is
caused by the fact that the gain maximum coincides with the
FIG. 2. ͑Color online͒ Emission spectra for QCLs at room temperature with
͑a͒ 60 ␮m, ͑b͒ 50 ␮m, and ͑c͒ 40 ␮m ridge widths. The top spectrum
shows the emission of the uncoated laser, the middle spectrum stems from
the same lasers after spiropyran coating. The bottom spectrum is obtained
FIG. 4. ͑Color online͒ LIV characteristics at room temperature for an ͑a͒ 60
and ͑b͒ 40 ␮m wide ridge laser without cladding ͑native͒, with native clad-
after irradiation with UV light, converting the cladding to its mero state.
ding ͑spiro͒, and with cladding after UV irradiation ͑mero͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.143.199.160 On: Mon, 15 Dec 2014 10:59:26

051106-3
Basnar et al.
Appl. Phys. Lett. 97, 051106 ͑2010͒
the local minima of the photochromic cladding was mea-
sured in transmission IR spectroscopy ͑Fig. 1͒ and values of
410 cm−1 ͑100% spiro͒ and 728 cm−1 ͑30% mero and 70%
spiro͒ were obtained. These lead to refractive indices of
1.59+0.025i and 1.59+0.041i, respectively. With these val-
ues waveguide losses of 8.8 cm−1 ͑native laser͒, 13.3 cm−1
͑spiro͒, and 16.6 cm−1 ͑30% mero͒ were obtained from
simulation ͑modal overlap with the cladding about 0.3%͒,
yielding threshold current densities of 3.0 kA/cm2,
4.0 kA/cm2, and 4.7 kA/cm2, respectively. This is in good
agreement with the data obtained with the 60 ␮m ridge.
These results point toward ways of improving the per-
formance of the laser. Thinner claddings with a lower
amount of spiropyran lead to a more complete conversion
͑i.e., larger tuning͒ accompanied by lower overall losses.
This will significantly reduce the drop in the laser output
power. Using UV-light-emitting diodes for illumination will
lead to faster switching due to higher optical power ͑approx
1 order of magnitude larger than for the UV lamp used in this
experiment͒. Ramping the heat sink temperature to higher
temperatures will allow faster resetting of the cladding state
from mero to spiro.
FIG. 5. ͑Color online͒ Shift in the emission of a 60 ␮m ridge QCL with
spiropyran coating as a function of UV irradiation time at 180 K as follows:
͑a͒ shows the spectra recorded at room temperature under irradiation with a
365 nm UV lamp, ͑b͒ shows the shift in the center wavelength and the
emitted power ͓area underneath the laser peak in part ͑a͔͒ with the irradia-
tion time, and ͑c͒ shows the LIV characteristics in the spiro and mero state.
position of an absorbance peak ͑1300 cm−1͒. When switch-
ing the cladding, the increase in absorbance and further shift
in the emission wavelength away from the gain maximum
lead to an additional increase in the threshold and corre-
sponding decrease in the power.
These changes in the threshold and power are more pro-
nounced for broader ridges than for narrower ones. This is
due to the larger opening in the top metal contact for wider
ridges, which leads to a larger evanescent field and, thus, to
higher losses.
In summary, a method was developed which allows the
tuning of QCLs by irradiation with UV and visible light.
Reversible tuning by about 6 cm−1 was achieved. The emit-
ted wavelength is nearly constant between 180 K and room
temperature with a shift of only 2.5 cm−1 over this tempera-
ture range.
Operating at lower temperatures shifts the gain maxi-
mum closer to the absorbance minimum of the cladding. Fig-
ure 5 shows the results for a 60 ␮m ridge operated at 180 K.
The laser wavelength for the spiro state is at the position of
minimum absorbance. When illuminating the cladding with
UV light, the wavelength shifts by about 6 cm−1 to reach a
final value of 1322 cm−1. These values coincide with the
values for the absorbance minimum and they are shifted by
about 2.5 cm−1 as compared to room-temperature operation.
The reverse direction ͑mero to spiro͒ is nearly completely
suppressed at 180 K as the thermal energy, necessary to en-
able the ring closing reaction, is reduced.
The drop in the optical power with UV illumination time
is about 30%, which is significantly smaller than what was
observed at room temperature ͑about 65% for the 60 ␮m
ridge͒. The reason is that at room temperature the drop in the
output power is caused by both an increase in absorbance
and the shift to an area with reduced gain. At 180 K the gain
maximum is close to the absorbance minimum. Thus, only
the change in absorbance affects the power, while the gain
itself remains basically constant. This is also reflected in a
smaller increase in the threshold current density at 180 K as
compared to room temperature.
This work was supported by the Federal Ministry for
Transport, Innovation, and Technology ͑BMVIT͒ and the
FFG through the Nanoinitiative project PLATON ͑Project
No. 819654͒.
1J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, and A. Y.
Cho, Science 264, 553 ͑1994͒.
2J. Faist, C. Gmachl, F. Capasso, C. Sirtori, D. L. Sivco, J. N. Baillargeon,
and A. Y. Cho, Appl. Phys. Lett. 70, 2670 ͑1997͒.
3L. Ajili, J. Faist, H. Beere, D. Ritchie, G. Davies, and E. Linfield, Elec-
tron. Lett. 41, 419 ͑2005͒.
4L. Diehl, B. Lee, P. Behroozi, M. Lončar, M. Belkin, F. Capasso, T.
Aellen, D. Hofstetter, M. Beck, and J. Faist, Opt. Express 14, 11660
͑2006͒.
5S. Song, S. S. Howard, Z. Liu, A. O. Dirisu, C. F. Gmachl, and C. B.
Arnold, Appl. Phys. Lett. 89, 041115 ͑2006͒.
6R. Maulini, A. Mohan, M. Giovannini, J. Faist, and E. Gini, Appl. Phys.
Lett. 88, 201113 ͑2006͒.
7B. Basnar, S. Schartner, M. Austerer, A. M. Andrews, T. Roch, W.
Schrenk, and G. Strasser, Opt. Express 16, 8557 ͑2008͒.
8B. Basnar, E. Mujagic, A. M. Andrews, P. Klang, W. Schrenk, and G.
Strasser, Conference on Lasers and Electrooptics/Quantum Electronics
Conference, May 31, 2009, Baltimore, MD, OSA Technical Digest ͑CD͒
͑Optical Society of America, 2009͒, paper CThC7.
9G. Jiang, Y. Song, X. Guo, D. Zhang, and D. Zhu, Adv. Mater. ͑Weinheim,
Ger.͒ 20, 2888 ͑2008͒.
10B. S. Lukyanov and M. B. Lukyanov, Chem. Heterocycl. Compd. 41, 281
͑2005͒.
Slab waveguide calculations allow estimating the wave-
guide losses. Toward this end, the absorption coefficient for
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.143.199.160 On: Mon, 15 Dec 2014 10:59:26