◆ The Quantum Cascade Laser: A Versatile

High-Power Semiconductor Laser

for Mid-Infrared Applications
Oana Malis, Claire Gmachl, Deborah L. Sivco, Loren N. Pfeiffer,
A. Michael Sergent, and Kenneth W. West

Since its invention at Bell Labs in the mid-nineties, the quantum cascade laser
(QCL) has evolved rapidly to become a viable commercial alternative to other
solid-state mid-infrared light sources. QCLs are compact, high-power,
wavelength-agile laser devices that are ideally suited for mid-infrared
applications such as chemical sensing and free-space telecommunications.
This paper summarizes our recent progress in QCL technology, with a
particular emphasis on device design and fabrication. The current state-of-
the-art in performance of InP lattice-matched QCL is reviewed. In order to
expand the functionality of QCLs, new materials and device concepts are
being explored. In particular, nonlinear light generation in QCLs is discussed
in detail. The experimental results confirm the potential of nonlinear QCLs
to extend the wavelength range of InP-based devices. Ongoing efforts to
improve the fabrication of InP-based QCLs and to develop new QCL
materials are presented. © 2005 Lucent Technologies Inc.

Introduction
Quantum cascade lasers (QCLs) are compact, currently offered commercially by several Lucent
rugged, high-power, semiconductor light sources op- licensees. This rapid progress was made possible not
erating in the mid- to far-infrared range of the spec- only by the extremely active worldwide research on
trum. Unlike quantum-well semiconductor lasers, band-engineering, but also by the continuous devel-
QCLs can be designed to operate at any wavelength opment of molecular beam epitaxy (MBE), the semi-
within a broad spectral range. While QCL emission conductor growth technique that enables accurate
has been demonstrated down to the terahertz range, nanometer control and the reproducibility of complex
this paper will focus on mid-infrared Type I QCLs. In semiconductor structures. Today, QCL technology
the approximately ten years since its invention at Bell promises to have a major impact in fields ranging from
Labs [11], QCL technology has evolved rapidly and chemical sensing to free-space telecommunications.
has been extremely helpful in extending our knowl- This paper presents an overview of QCL technol-
edge of semiconductor physics. Moreover, the swift ogy, with a particular focus on current research ac-
development of QCLs has led to a technology transfer tivities at Bell Labs. First, the principles of QCL
from research to production; mid-infrared QCLs are operation are introduced. Then, material growth,

Bell Labs Technical Journal 10(3), 199–214 (2005) © 2005 Lucent Technologies Inc. Published by Wiley Periodicals,
Inc. Published online in Wiley InterScience (www.interscience.wiley.com). • DOI: 10.1002/bltj.20114
device processing, and device characterization are
discussed in detail. Next, current state-of-the-art de- Panel 1. Abbreviations, Acronyms, and Terms
vice performance is presented and the factors limiting CVD—Chemical-vapor deposition
device performance are investigated. It turns out that cw—Continuous wave
QCL light characteristics and operating range are lim- DFB—Distributed feedback
DIAL—Differential-absorption LIDAR
ited by some fundamental properties of the materials
epi—Epitaxial
of choice. A separate section is devoted to ongoing HR—High reflectance
research efforts at Bell Labs to push the limits of QCL L-I—Output versus current
operation and to explore new materials and device LIDAR—Laser radar
concepts. The following section briefly discusses LO—Longitudinal optical
applications of QCLs. The final section presents some MBE—Molecular beam epitaxy
MOCVD—Metal-organic chemical-vapor
concluding remarks and offers a perspective on future
deposition
research and development. ppb—parts per billion
QCL Principles of Operation QCL—Quantum cascade laser
QW—Quantum wells
QCLs are unipolar semiconductor devices con- RADAR—Radio detection and ranging
sisting of complex layered structures of two or more SH—Second harmonic
semiconductor alloys [11, 15]. Figure 1 shows a laser TEM—Transmission electron microscopy
bar containing several QCLs mounted on a copper TILDAS—Tunable infrared laser diode
heat-sink (a), an scanning-electron microscopy (SEM) absorption spectroscopy
TM—Transverse magnetic
image of an individual laser device (b), and a trans-
VCSEL—Vertical cavity surface-emitting lasers
mission electron microscopy (TEM) image of the
semiconductor layered structure inside the active
waveguide core of the device (c). The light is gener-
ated in the active region by an intersubband transition a QCL biased at operating voltage. In this basic design,
of the single charge carriers (i.e., the electrons) be- the active region of the device consists of three strongly
tween two quantized levels in the conduction band. coupled QWs. The light is emitted by a transition be-
Because the energy difference between the two quan- tween levels 3 and 2. In order to achieve population
tized levels is determined by the specific structure de- inversion for lasing, the electrons must be injected rap-
sign (i.e., the quantum well [QW] and barrier widths), idly into the upper level 3 and then rapidly extracted
the wavelength emitted can be tailored by band- from level 2. Fast depopulation of level 2 is achieved
engineering to be any value within a broad spectral by means of longitudinal-optical (LO) phonon scat-
range. This range is limited only by the conduction- tering from level 2 into level 1, which is located ap-
band offset available between the two materials of proximately 35 meV below level 2. The injector region,
choice and the Reststrahlen region. here consisting of 5 QWs, has the role of facilitating the
This section will be using InP lattice-matched al- transport of the electrons through the device by in-
loys to exemplify the QCL principles of operation, but suring fast extraction from both level 1 and level 2 and
the general concepts are the same for any other mate- fast re-injection into level 3 of the down-stream ac-
rial system. In InP-based QCLs, the active region con- tive region. To maximize the gain, 30 to 100 active
sists of In0.53Ga0.47As QWs and In0.52Al0.48As barriers regions are typically chained (i.e., “cascaded”) together.
(see Figure 1c). The conduction-band offset for these The electrons are recycled from one active region to
two materials is 0.52 eV at low temperatures and cor- the next, emitting more than one photon per pass
responds to an operation range between 5 mm and through the device. This process is the reason for the
20 mm. It is noteworthy that QCL operation has also high emission power of these devices.
been achieved in strained InGaAs/InAlAs devices on The basic QCL structure exemplified in Figure 2
InP as well as in GaAs/AlGaAs and Sb-based material has evolved over the years into many more sophisti-
systems. Figure 2 shows the detailed band structure of cated designs. The three-level active region can be

200 Bell Labs Technical Journal

 (a)

(c)

15 m

(b)

DFB-QCL—Distributed feedback-quantum cascade laser

QCL—Quantum cascade laser
SEM—Scanning electron microscopy
TEM—Transmission electron microscopy

(a) QCL bar containing several devices mounted on a copper heat-sink and wire-bonded to
Au-plated alumina contact pads
(b) SEM image of the front facet portion of a deep-etched ridge-waveguide DFB-QCL
(c) TEM image of a portion of the active waveguide core of a QCL

Figure 1.
The quantum cascade laser is a compact, rugged, high-power, semiconductor light source.

replaced by a superlattice to allow large current- the matrix dipole moment for the 3-2 transition, nef the
carrying capabilities [40]. In this case, the lasing tran- effective refractive index at wavelength l, Lp the thick-
sition occurs between two extended states of the ness of one active region and injector, e0 the vacuum
superlattice minibands. If the transition is between a dielectric constant, e the elementary charge, and 2g32
localized state and an extended superlattice state, the the full width at half maximum of the luminescence
design is referred to as a bound-to-continuum transi- spectrum. The lifetimes are determined by LO-phonon
tion [9]. Moreover, the transition can be either verti- interband scattering and are of the order of ps.
cal, as in the above examples, or diagonal, a design Laser action requires that the material gain over-
that increases the upper-laser-state lifetime [10]. come the device losses. The losses come from two main
Using the usual rate-equation formalism, the sources: the nonresonant free-carrier losses in the wave-
material gain coefficient can be calculated as [15] guide, aw, typically calculated with a Drude model, and
the mirror losses, am, given by am
(1
L)ln(R), where
t2 4pez232 1 R is the mirror reflectivity R
((n  1)
(n  1))2. The
g
t3 a1  b ,
t32 le0nef Lp 2g32 threshold current density Jth is then given by

where ␶3 is the total upper-state lifetime, ␶2 the total life- am  aw

Jth
,
time of level 2, ␶32 the lifetime of the 3-2 transition, Z32 g

Bell Labs Technical Journal 201

 3

Itop

2
e 1
3 Ibott

Itop Active
region
2
e 1 Injector

Ibott
Active
region

Injector

The laser transition is indicated by the arrow between levels 3 and 2. The layer thicknesses
in nanometers of one period of active material and injector are, from right to left starting
from the injection barrier: 3.8/2.1/1.2/6.5/1.2/5.3/2.3/4.0/1.1/3.6/1.2/3.2/1.2/3.0/1.6/3.0. The
underlined layers are n-type doped. The barriers are indicated in bold symbols; italic style
shows the injector region. The applied electric field is 65 kV cm1.

Figure 2.
Conduction band diagram of the active and injector regions of an 8 ␮m QCL and the moduli squared of the most
significant wavefunctions.

where  is the confinement factor, typically of the of the various waveguide layers depend on the specific
order of 0.5. For the structure in Figure 2, the losses application. The waveguide in Figure 3 is optimized
can be estimated as am
5.1 cm1 (for a 2.5 mm-long for low loss and maximum overlap of the TM00 mode
cavity) and aw
19 cm1, resulting in a threshold with the active region.
current density of 1.6 kA/cm2. This typical value for a The dielectric waveguide shown in Figure 3 is a
three-well design emphasizes the main limiting as- Fabry-Perot-type resonator that emits a multimode
pects of QCL devices, i.e., their high threshold and laser spectrum. For certain applications, however, it is
operating currents. desirable to have single-mode operation. This is
In order to minimize device losses and confine achieved with a distributed feedback (DFB) resonator.
the mid-infrared light to the gain material, the active A DFB resonator consists of a grating of period
region is inserted into a waveguide. Figure 3 shows 
lB
(2  neff) etched in the top cladding of the
the structure of a typical QCL waveguide, together device. For practical processing reasons, DFB top
with the profiles of the refractive index and of the claddings are typically significantly thinner than that
lowest-order transverse-magnetic mode (TM00). The shown in Figure 3. The grating induces a modulation
waveguide core consists of the active section stacked in the refractive index and losses along the laser ridge
between two InGaAs layers. The claddings are given that selects a unique emission wavelength for each
on one side by the InP substrate and on the other by lateral and transverse mode of the cavity. Because the
either an InAlAs or another InP layer. A highly-doped refractive index depends on temperature, the QCL
plasmon-enhanced InGaAs layer also contributes to wavelength also varies with temperature. QCL tem-
the confinement. The thicknesses and doping profiles perature tunability, which is particularly important

202 Bell Labs Technical Journal

 n InP, 1020 cm3, 10 nm
Ti/Au top contact

n InP, 5 1018 cm3, 1 m

n InP, 1017 cm3, 1.6 m

Thickness (m)
n InGaAs, 5 1016 cm3, 0.35 m

Active regions and injectors

30 stages 1.329 m

n InGaAs, 5 1016 cm3, 0.35 m

n InP, 1-2 1017 cm3, substrate

0 3
(a) (b) Refractive index and
mode profile
QCL—Quantum cascade laser
TM—Transverse magnetic

(a) Layer structure of a low-loss waveguide for the 8 m QCL shown in Figure 2
(b) The corresponding profiles of the refractive index and magnetic field of
the TM00 mode

Figure 3.
Waveguide structure of a QCL.

for spectroscopy applications, is achieved by changing x-ray scattering for the alloy composition and Hall
the active region temperature by adding a DC ramp to measurements for doping calibration. It is worth not-
the laser current drive. ing that QCLs have also been successfully grown by
metal-organic chemical-vapor-deposition (MOCVD)
QCL Material Growth [35, 20].
The complex QCL structures are grown by solid- For InP-containing structures, such as InP top-
source MBE on n-type InP (100) substrates doped at cladding devices, we have been using a two-step MBE
the level of 1–2  1017/cm3. MBE has so far proven to growth process. First, the P-free core of the device is
be the growth technique of choice, because it allows grown in an arsenide machine and the material is
accurate control of sub-nanometer size semiconductor capped with a low-temperature, pure As layer. Then,
layers. Moreover, MBE has high reproducibility over the sample is transferred into a second MBE machine
the entire structure, which may be 5 mm to 10 mm equipped with a solid P-source and the top InP-
thick and contain more than 500 layers. Figure 1c containing layers are grown. InP has the advantage of
shows the TEM image of several active regions of a having lower mid-infrared losses and a lower refrac-
QCL structure. The growth of InP lattice-matched ma- tive index than InAlAs as well as higher thermal
terials is extremely demanding and requires precise conductivity. This results in a better optical confine-
control of the InGaAs and InAlAs alloy composition. ment of the wave and ultimately in improved laser
The MBE-grown structures are characterized using performance. The two-step technique has been most

Bell Labs Technical Journal 203

successful in growing buried-grating DFB lasers. To Spectrometer fitted with cooled HgCdTe and InSb
improve the thermal properties of the devices, we detectors for the long- and short-wavelength ranges,
have also used InP overgrowth by MOCVD; this respectively. The pulsed-light output versus current
technique will be discussed in detail in the following (L-I) measurements are taken with a calibrated fast
sections. HgCdTe photovoltaic detector for the laser light or a
cooled InSb detector for the nonlinear light. The cw
QCL Device Processing L-I measurements are taken with a calibrated power
The QCL devices are typically processed into 4 mm detector.
to 16 mm wide, 0.5 mm to 3 mm long, deep-etched In Figure 4, 4a shows the pulsed L-I curves of a
ridge-waveguide lasers (Figure 1b). First, using stan- device with the active region detailed in Figure 2
dard photo-lithography, the laser mesas are defined emitting at 8 ␮m at temperatures between 5 K and
using either chemical wet-etching or dry-etching. 300 K. The peak power exceeds 1 W at cryogenic
Then, the insulator Si3N4 or SiO2 is deposited using temperatures but decreases dramatically with in-
chemical-vapor deposition (CVD). GeSe chalcogenide creasing temperature to 500 mW at room tempera-
has also been used as an insulator, because of its ture. The maximum peak power depends strongly on
superior mid-infrared properties. In the next step, a the number of active stages, as can be seen in the inset
window is opened into the insulator using freon of Figure 4a.
plasma-etching to access the top of the mesa electri- QCLs based on the three-well structure discussed
cally, followed by thermal evaporation and the lift-off so far also operate in continuous-wave (cw) mode at
of the top Ti/Au contact. The wafer is thinned (i.e., low temperatures, as shown in Figure 4b. The maxi-
lapped) to 200 mm before depositing the back Ge/Au/ mum cw operating temperature is limited by the ther-
Ag/Au contact. The laser mirror facets are obtained by mal properties of the package, which consists of the
simply cleaving the device to the desired length. device and the holder. Because of the high laser-
Metallic high-reflectance (HR) coatings on the back threshold currents and poor thermal conductivity
facets are also used to reduce the mirror losses. Finally, of the deep-etched ridge-waveguide geometry, the
the as-cleaved or HR-coated devices are mounted with temperature of the active region exceeds the heat-
In on copper heat-sinks and wire-bonded to Au-coated sink temperature by more than 100 K. As a result of
alumina contact pads (Figure 1a). this temperature run-off, the laser can never reach
DFB laser processing differs only slightly from this its threshold current above a certain heat-sink tem-
procedure. Before all the steps outlined above, the DFB perature. The next section will discuss in detail on-
grating is wet-etched across the entire sample; then the going efforts in our group to overcome this effect.
sample is processed in the normal way. The DFB grat- Using metal-buried heterointerfaces, Razeghi and her
ing is visible on the top of the ridge in Figure 1b. QCLs collaborators have achieved QCL cw operation at
have also been processed into more sophisticated res- 6 mm up to 333 K and a power of 424 mW at 293 K
onator designs, including micro-cavity lasers [12, 16] with a plug efficiency of 3.2% [43, 8].
and photonic crystal lasers [4].
Ongoing Research at Bell Labs
Device Characterization The QCL research at Bell Labs is a multi-pronged
The samples are measured on the temperature- effort focused on enhancing and expanding the mid-
controlled cold-finger of an He-flow cryostat at tem- infrared functionality of these devices. This section
peratures from 10 K to 300 K. The lasers are operated briefly summarizes three directions of research and
in either pulsed mode, with 50 ns to 100 ns current- development. First, we are working to extend QCL
pulse widths at repetition rates of 4 kHz and 84.2 kHz, functionality by means of nonlinear light-generation.
or continuous-wave (cw) mode. The laser spectra are Then, building on Bell Labs’ world-renowned ex-
taken using a Nicolet Fourier Transform Infrared perience with material growth, we are exploring new

204 Bell Labs Technical Journal

 1.5 1.5
1.5
5–50 K 250

Max. power (W) 1.0

10 K 100 K 30 K
150 K 80 K
200
0.5
Peak power (W)

Power (mW)
1.0 300 K 200 K 1.0 120 K
0 150
0 20 40 60 80 145 K
Number of stages 250 K
100
0.5 0.5
300 K
50
N = 75
0 0 0
10 0 10 20 0 300 600 900 1200
2 Current (mA)
Current density (kAcm )
(a) (b)

L-I—Output versus current

QCL—Quantum cascade laser

(a) Pulsed L-I curves of an 8 m QCL with 75 stages of the active region shown in Figure 2 at temperatures from
5 K to 300 K. The inset shows the dependence of the maximum power on the number of stages at 10 K and 300 K.
(b) Continuous-wave L-I plot of a 6 m QCL at temperatures from 30 K to 145 K.

Figure 4.
QCL performance characterization in pulsed and continuous-wave operation modes.

material systems that would enable new QCL func- phase-matching techniques. The first obstacle was
tionality. Finally, this exploration into new device and recently overcome by monolithically integrating the
materials concepts is going on in parallel with a con- nonlinear intersubband transitions with the structure
tinuous effort to improve the performance of the of a QCL [32]. The idea of monolithically integrating
more mature InP-based technology. nonlinear intersubband transitions with the QCL
structure is natural, because they both use intersub-
Nonlinear Quantum Cascade Lasers band transitions to generate light. However, from a
QCLs operate in a broad mid-infrared spectral practical point of view, the strict current-injection re-
range, but, as pointed out above, for each material quirements of these devices raise significant design
system, this range is limited by some fundamental challenges. Sum-frequency and second-harmonic
property of the material (e.g., the conduction-band (SH) generation were the first nonlinear processes
offset). Nonlinear light generation offers the poten- demonstrated in QCLs [14, 32]. Third-harmonic gen-
tial of extending the functionality of QCLs beyond eration has also since been reported [31].
the limits set by the intrinsic properties of materials. We are currently investigating means to improve
Of particular technological interest is the low- second-harmonic generation in QCLs [27, 28]. The
wavelength spectral range below 5 mm, in which no first step was to design an active region that acts
clearly advantageous semiconductor laser technology simultaneously as a pump source and a nonlinear
has yet been identified. Nonlinear QCLs use the giant mixing region. In Figure 5, 5a shows the detailed
nonlinear properties of resonant intersubband transi- band structure of the optimized design. This structure
tions [3, 21]. These properties have been studied exhibits a record-high nonlinear susceptibility of
intensively over the past 20 years, but practical |
(2)|  4 105 esu (2 104 pm/V), a value that is at
applications of them have been limited by the lack least three times higher than the values measured for
of compact, powerful pump sources and versatile any other material. The second step was to develop

Bell Labs Technical Journal 205

 500 4 2.5

450
SH refractive index

Mode profiles
2 2.0

Nonlinear power PNL (mW)

400 Laser TM00

Linear power PL (mW)

5 350
I 0
4 300 1.5
SH TM02
250 2
3 10 5 0
200 Thickness (m) 1.0
2 150
1
Injector
E 100 0.5
Active
region 50
Injector
z 0 0.0
0 1 2 3 4
Current (A)
(a) (b)

QCL—Quantum cascade laser SH—Second harmonic

L-I—Output versus current TM—Transverse magnetic

(a) Conduction band diagram of one QCL active region stacked between two injector regions of a structure optimized for
second-harmonic generation and the moduli squared of the essential wavefunctions. The laser transition occurs between
levels 3 and 2. The resonant intersubband cascades are 2 - 3 - 4 and 3 - 4 - 5. The dashed rectangles indicate the extent of
the minibands inside the injector regions. The layer thicknesses of one period of active region-injector are from left to right
starting at injection barrier “I”: 4.1/8.3/1.3/5.2/2.6/4.1/2.1/3.9/2.3/3.7/2.5/3.5/2.6/3.3 nm. The barriers are indicated by bold
font, and the underlined layers are doped to 3.0 1017 cm3.
(b) Pulsed L-I curves for the laser and nonlinear light at 5 K for a device with the active region shown in (a). The inset shows
the profiles of the SH refractive index and the laser TM00 and SH TM02 modes for a phase-matching waveguide. The
calculated refractive indices of the fundamental TM00 and SH TM02 modes for an infinitely wide waveguide are 3.2079 and
3.1689, respectively.

Figure 5.
Structure and performance of a nonlinear QCL emitting fundamental light at 9 ␮m and second-harmonic light at
4.5 ␮m.

a phase-matching technique between the fundamen- of the waveguide, a record second-harmonic power of
tal and nonlinear light. We found that, due to the 2 mW was achieved at 4.55 ␮m (Figure 5b) [28]. This
flexibility in the design of the QCL waveguide, modal- result was made possible by the use of an InP top-
phase matching is the most promising candidate. cladding waveguide and of HR coating on the laser
Modal-phase matching consists in matching the back facets. We are currently working on lowering
effective refractive indices of the laser and nonlinear the SH emission wavelength below 4 ␮m and im-
light of different order transverse modes. For the proving the thermal characteristics of the devices. We
second-harmonic process, the laser TM00 mode and are also investigating other nonlinear processes, such
the SH TM02 mode offer the best choice in terms of as frequency down-conversion and coherent inver-
refractive indices, lowest losses, and highest overlap sionless lasing [29].
with each other and with the active region. The inset
of Figure 5b shows the profiles of these two modes Novel Materials
together with the profile of the SH refractive index. Another approach to expanding QCL functional-
An additional degree of freedom is given by the strong ity is to explore new material systems. The GaAs/
dependence of the refractive indices on the ridge- AlGaAs system has been shown to be a potential com-
width. Using the active region of Figure 5a and in- petitor of InP lattice-matched materials [37]. This
cluding phase-matching considerations in the design system is particularly attractive because it uses the

206 Bell Labs Technical Journal

most mature III-V growth technology on a relatively on (100) GaAs substrates. In order to reduce the
cost-effective GaAs substrate. Unfortunately, GaAs interface roughness and to control the well dimen-
QCLs suffer from the fact that AlGaAs is an indirect- sions with 1% accuracy, the QW widths were chosen
gap semiconductor above about 45% Al composition. to be exact integers of the monolayer spacing (i.e.,
Once the minimum of the X-valley in the injection 25.5, 31.13, 36.79, and 45.28 Å, corresponding to 9,
barrier becomes lower than the -point upper-laser 11, 13, and 16 monolayers, respectively). The high
level, the device performance degrades rapidly [42]. Al-composition AlGaAs is a digital alloy formed by
This effect limits the operation wavelength of a GaAS a short-period superlattice of 11.3 Å AlAs and 8.5 Å
QCL to the 8 mm to 20 mm range. In contrast to the GaAs. The QWs were modulation-doped with car-
conduction band, the valence band of the AlGaAs al- bon, using a unique custom-made solid carbon
loys has a maximum at the -point throughout the source. Carbon is preferable to beryllium for p-type
compositional range. Therefore, a device using hole doping, because it has a lower diffusion constant at
intersubband transitions in the valence band rather the MBE growth temperatures. The C-source we
than electron intersubband transitions in the con- employed is compatible with very high purity GaAs
duction band could take advantage of the full valence- material growth, as confirmed by mobility measure-
band offset between pure GaAs and AlAs compounds. ments of two-dimensional electron gasses. Using Hall
This valence-band offset (i.e., 0.51 eV) is comparable measurements, the two-dimensional hole gas for the
to the conduction-band offset in the InGaAs/InAlAs 31 Å QW sample was estimated to have a density of
material system and would enable emission in a similar 1.6 1012 cm2 and mobility of 8000 cm2/Vs at 5K.
5 mm to 20 mm mid-infrared range. Moreover, hole Figure 6 shows the in-plane (s-polarized) and
QCLs have the potential to allow the design of devices out-of-plane (p-polarized) intersubband absorption
with new properties, such as surface-emitting QCLs and of the 31 Å QW sample. For QWs 25 Å to 45 Å wide,
vertical cavity surface-emitting lasers (VCSELs).
Because of the complexities of the valence band,
the hole intersubband transitions have not been stud- T = 17 K
ied as intensively as their electronic counterparts.
Transmission (arb. units)

Most of the investigations of intersubband transitions 0.4

in the valence band have been done in p-type Si-SiGe s-polarized

HH hole transition energy (meV)

QWs, in which intersubband absorption, electrolumi-

200
nescence, and photocurrent measurements have been
reported [2, 5–7]. However, the Si-SiGe material sys- 0.2
tem presents some significant theoretical and growth 150
challenges associated with the built-in strain of these
p-polarized
lattice-mismatched materials. The GaAs-AlGaAs sys- 25 30 35 40 45
QW width (Å)
tem has the advantage of being virtually strain-free, 0.0
200 300
and it is well understood from the material-growth
Energy (meV)
point of view, but most of the research on this sys-
tem has focused on infrared photodetectors [39]. QW—Quantum well

As a first step toward a hole-based mid-infrared The spectra for p-polarized (circles) and s-polarized
(diamonds) light are shown for measurements at 17 K.
emitter, we are exploring the properties of inter- The inset shows the heavy-to-heavy hole intersubband
valence band transitions in GaAs QWs with high Al transition energies for 4 QW widths.
composition barriers. Bound-to-bound intersubband
absorption in C-doped GaAs QWs was measured Figure 6.
Mid-infrared intersubband transmission spectra of
in the mid-infrared range. Taking advantage of a high-
a sample containing 31 Å wide, carbon modulation-
purity MBE facility at Bell Labs, GaAs QWs with doped GaAs QWs with digital Al0.57Ga0.43As alloy
modulation-doped Al0.57Ga0.43As barriers were grown barriers.

Bell Labs Technical Journal 207

the absorption peak covers a broad mid-infrared range
(i.e., 206 meV to 126 meV, 6 mm to 10 mm wave-
length) that is promising for application in quantum
cascade emitters (see inset of Figure 6). Accurate sim- MOCVD MBE MOCVD
ulations of valence-band structure are essential for
InP substrate
the design of complex multi-QW QCL structure. We
performed 6-band k  p calculations of the heavy-to- (a) (b)
heavy hole transition energies considering the full MBE—Molecular-beam epitaxy
band structure of the digitally alloyed barriers and MOCVD—Metal-organic chemical-vapor deposition

found them to be in very good agreement with the The lighter regions in (b) are the InGaAs (bottom) and
experimental results for wider wells. The model is less InGaAs/InAlAs (top) MBE-grown QCL structure etched
into a ridge-waveguide.
accurate in predicting the out-of-plane absorption for
the narrowest wells and the in-plane absorption fea- Figure 7.
tures. For all practical purposes, however, the model Schematics of the InP MOCVD overgrowth process
is accurate enough to be adequate for band-structure (a) and side image of a deep-etched ridge-waveguide
(b) on which InP (darker regions) was grown by
calculations of hole QCL structures.
MOCVD.
In this subsection we have been discussing only
the GaAs system, but research on new QCL materials
at Bell Labs also includes work on strained InGaAs/ DFB lasers. However, both the thermal and optical
InAlAs on InP to access the low-wavelength mid- properties of the deep-etched ridge waveguides can
infrared range as well as investigations of the novel be improved dramatically by using a buried hetero-
GaN system. Mid-infrared absorption measurements junction design. In this case, the mesa patterning is
have been reported on GaN/AlGaN quantum wells done using an SiO2 mask instead of photoresist. After
[17, 18], and work is currently being done to study the ridges have been wet-etched into the MBE-grown
photocurrent and resonant tunneling. material, the wafer is transferred to an MOCVD reac-
Advanced Fabrication of InP-Based QCLs tor and a thick layer of InP is grown on the sides of
In the section on device characterization, we in- the ridges (see schematics in Figure 7). MOCVD is
troduced some of the main requirements for QCL ap- the technique of choice for this process, because it
plications that are still challenging for the technology: allows selective-area overgrowth on the exposed semi-
high-power operation, cw operation, and room- conductor but not on the SiO2 mask. The main bene-
temperature operation. Our ongoing development to fit of this design is that it increases the heat dissipation
overcome these challenges involves designing high- out of the active region, allowing operation at higher
performance structures, improving the quality of MBE currents and, consequently, higher maximum cw
material by improving the MBE process, and per- temperature. InP also has the advantage of a low re-
forming advanced device fabrication. MBE process fractive index that defines the lateral waveguide con-
improvement includes optimization of the doping finement. The losses of this geometry are lower than
level, growth parameters (e.g., temperature, growth they are for deep-etched ridge waveguides, because
rate, and arsenic overpressure), and a systematic effort the overlap of the wave with the metal top-contact
to lower the defect level and unintentional back- is minimal. Figure 7 shows a cross-section of a device
ground concentration. Advanced QCL fabrication with InP overgrown on the sides of the ridge. The
includes epi-side-down mounting of the samples, maximum cw operating temperature is increased by
HR-coating of the back facets, and InP overgrowth by 25 K over that of a device without the buried hetero-
either MBE or MOCVD. junction, and there is a total maximum cw tempera-
MBE overgrowth has mainly been used to grow ture increase of 50 K when HR coatings are used on
the top InP claddings or to create a buried grating for the back facet.

208 Bell Labs Technical Journal

Applications amplitude modulation that is detected by a lock-in
The operating range of QCLs makes them ideal technique. Recently, a lot of interest has been shown
for chemical and biological sensing applications [25]. in photoacoustic trace-gas sensing. In the technique,
Most of the chemical compounds of interest for en- the optical beam is periodically modulated in ampli-
vironmental or defense applications have absorption tude before illuminating the cell containing the
features in the mid-infrared. Their transmission absorbing chemical. The periodic heating of the
through free atmosphere exhibits two windows chemical creates an acoustic wave that is detected
(between 3 mm and 5 mm and between 8 mm and with a microphone. This is a low-cost and sensitive
13 mm) clear of water absorption. In these windows, technique that avoids the use of cooled mid-infrared
strong absorption peaks have been identified for detectors, but its sensitivity is typically limited by the
trace gasses, such as NO, NH3, CH4, and CO. Several laser power [23].
spectroscopic techniques have been used to detect QCLs have so far been used in numerous spec-
these gasses. The simplest method is direct absorp- troscopic studies for environmental purposes, health
tion. In this setup, the change in intensity of a laser monitoring, and military applications. The first
beam passing through a cell containing the target application of a thermoelectrically cooled DFB-QCL
chemical is measured. The most commonly used to continuous spectroscopic monitoring of CO in am-
optical method for trace-gas sensing is tunable in- bient air achieved a noise equivalent detection limit
frared laser diode absorption spectroscopy (TILDAS). of 12 parts per billion (ppb) using a 102 cm optical
The high sensitivity and specificity of this technique path-length and a 2.5-minute data acquisition time at
is achieved by modulating the frequency of the laser a 10 kHz pulsed-laser repetition rate [26]. Figure 8
so that it periodically scans through the absorption shows the CO concentration monitored in time over
peak of the chemical of interest. The absorbing 14 hours with 10 ppb sensitivity. A spectroscopic gas
material converts the frequency modulation into an sensor for nitric oxide (NO) detection based on a

0.60

0.50
Sensitivity ~10 ppb
CO concentration, ppm

0.40

0.30

0.20

0.10
22:13:20 01:00:00 03:46:40 06:33:20 09:20:00 12:06:40
Time, hrs:min:sec

Figure 8.
CO concentration recorded over time in Houston, TX, with QCL spectroscopy.

Bell Labs Technical Journal 209

cavity ringdown technique and a cw QC-DFB laser been achieved with QCLs and subpicosecond pulses
operating at 5.2 mm measured ppb NO concentra- have been produced with a repetition rate of 100 MHz
tions in N2 with a 0.7-ppb standard error [24]. NO to 10 GHz.
detection is important for a number of applications,
such as atmospheric pollution and toxic process mon- Conclusions
itoring, vehicle exhaust control, and noninvasive In the ten years since their invention, QCLs have
medical diagnostics. NO in exhaled breath is a possible evolved tremendously to reach all wavelengths from
biomarker for various lung diseases. Similarly, 3.7 mm to 24 mm. The success achieved with InP
carbonyl sulfide (COS) in human breath has been devices in the mid-infrared range was recently repro-
related to liver and lung diseases. A detection sensi- duced in the terahertz range with GaAs-based devices
tivity of 30 parts in 109 of COS and the selectivity of using the same general principles of intersubband
two stable isotopes has been demonstrated [36]. light emission and electron transport [22]. The
Other QCL spectroscopy studies include ammonia, challenges for the technology now lie in the short-
methane, CO2, and ethylene [41], to name just a few. wavelength range. In the near-infrared range, several
Recently, a trace gas sensor based on quartz- technologies are competing. We are exploring
enhanced photoacoustic spectroscopy with a QCL nonlinear light generation to extend the operation
operating at 4.55 ␮m was developed for the detection range of InP lattice-matched materials. In parallel with
of N2O and CO and was found to have a sensitivity of development to improve the performance of InP-
4 ppb for N2O [23]. based devices, we are investigating new materials and
In addition to their use in in situ trace-gas sensing, device concepts, including strained InGaAs/InAlAs on
QCLs could have an impact on remote sensing with InP and hole QCLs in GaAs/AlGaAs. Completely novel
LIDAR. Gittins et al. [13] used a multimode Fabry- material systems, such as GaN, may have to be de-
Perot QCL operating at 8 mm for backscatter absorp- veloped in order to bridge the gap to telecommunica-
tion measurements on isopropanol vapor. They tions wavelengths.
explored the use of QCLs for differential-absorption The QCL devices operate either single- or mult-
LIDAR (DIAL) and found that even a simple configu- mode at temperatures well above 300 K. Multi-
ration resulted in a detection limit of 12 ppm. wavelength operation has also been achieved using
Another class of applications for which QCLs have either heterogeneous active regions or single cascades
strong potential to be of use is free-space optical com- with giant nonlinear properties. Moreover, ultra-
munications. Because of lower Rayleigh and Mie scat- broadband emission in the 6 mm to 8 mm range has
tering, the attenuation of mid-infrared light through been demonstrated [19]. Currently there is a lot of
open atmosphere is much lower than it is for visible or interest in expanding these concepts to a broadband
near-infrared light. Losses nearly one hundred times emitter that would simultaneously cover the entire
lower than those for shortwave infrared transmission mid-infrared range for sensing applications.
can be expected in the second atmospheric window This paper has focused on Fabry-Perot and DFB
(i.e., 8 mm to 13 mm) in clear weather conditions. For QCL waveguides, but a significant amount of research
lower visibility conditions, the advantage is even has been done on other resonator cavities, such as
greater. An optical link based on QCL technology has microcavity and photonic crystal devices. Microcavity
been demonstrated to transmit audio, video, and data resonators are interesting because they are spectrally
signals over 200 m at speeds up to 2.5 Gbit/s [30]. A narrow, low-threshold, and low-noise laser sources.
unique feature of QCLs is their ultrafast carrier- Whispering-cavity microdisc lasers, bow-tie lasers,
relaxation lifetime (1 ps), which is dominated by and asymmetric cavity resonators have been stud-
electron-optical phonon scattering. This property ied for their unusual light-emitting patterns [1].
makes QCLs ideally suited for high-speed modulation. Most recently, a photonic crystal QCL has been
Active [33, 38] and passive [34] mode-locking has demonstrated [4].

210 Bell Labs Technical Journal

 QCL technology has moved in a short time from Si0.2Ge0.8/Si Quantum-Cascade Structures Based
research to production and QCLs are now a viable on a Bound-to-Continuum Transition,” Appl.
commercial alternative to other solid-state lasers. As Phys. Lett., 81 (2002), 4700–4702.
[7] L. Diehl, H. Sigg, G. Dehlinger, D. Grützmacher,
compact, rugged, low-power-consumption light
E. Müller, U. Gennser, I. Sagnes, T. Fromherz,
sources, QCLs are ideal candidates for applications, in Y. Campidelli, O. Kermarrec, D. Bensahel, and
particular for mobile field applications such as remote J. Faist, “Intersubband Absorption Performed
trace-gas sensing. The recent demonstration of high- on P-Type Modulation-Doped Si0.2Ge0.8/Si
power, room-temperature, cw QCL operation [8, 43] Quantum Wells Grown on Si0.5Ge0.5
promises to significantly impact the sensing tech- Pseudosubstrate,” Appl. Phys. Lett., 80 (2002),
3274–3276.
nologies. Moreover, QCLs also have potential for free-
[8] A. Evans, J. S. Yu, S. Slivken, and M. Razeghi,
space optical communications, a field that has not yet
“Continuous-Wave Operation of Lambda
been fully explored. Similar to 4.8 ␮m Quantum-Cascade Lasers
at Room Temperature,” Appl. Phys. Lett., 85
Acknowledgments (2004), 2166–2168.
The authors would like to thank Jianxin Chen [9] J. Faist, M. Beck, T. Aellen, and E. Gini,
and Liming Zhang of Lucent Technologies for InP “Quantum-Cascade Lasers Based on a Bound-
overgrowth by MBE and MOCVD, respectively. We to-Continuum Transition,” Appl. Phys. Lett.,
also thank F. K. Tittel and A. A. Kosterev of Rice 78 (2001), 147–149.
[10] J. Faist, F. Capasso, C. Sirtori, D. L. Sivco, A. L.
University for providing the CO spectroscopic data
Hutchinson, and A. Y. Cho, “Laser Action by
shown in Figure 8. Tuning the Oscillator Strength,” Nature, 386
References (1997), 777–782.
[1] Y. Baryshnikov, P. Heider, W. Parz, and [11] J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L.
V. Zharnitsky, “Whispering Gallery Modes Hutchinson, and A. Y. Cho, “Quantum Cascade
Inside Asymmetric Resonant Cavities,” Phys. Laser,” Science, 264 (1994), 553–556.
Rev. Lett., 93 (2004), 133902-1–133902-4. [12] J. Faist, C. Gmachl, M. Striccoli, C. Sirtori,
[2] I. Bormann, K. Brunner, S. Hackenbuchner, F. Capasso, D. L. Sivco, and A. Y. Cho,
G. Zandler, G. Abstreiter, S. Schmult, and “Quantum Cascade Disk Lasers,” Appl. Phys.
W. Wegscheider, “Midinfrared Intersubband Lett., 69 (1996), 2456–2458.
Electroluminescence of Si/SiGe Quantum [13] C. M. Gittins, E. T. Wetjen, C. Gmachl,
Cascade Structures,” Appl. Phys. Lett., 80 F. Capasso, A. L. Hutchinson, D. L. Sivco,
(2002), 2260–2262. J. N. Baillargeon, and A. Y. Cho, “Quantitative
[3] F. Capasso, C. Sirtori, and A. Y. Cho, “Coupled Gas Sensing by Backscatter-Absorption
Quantum Well Semiconductors with Giant Measurements of a Pseudorandom Code
Electric Field Tunable Nonlinear Optical Modulated   8 ␮m Quantum Cascade
Properties in the Infrared,” IEEE J. Quantum Laser,” Appl. Opt., 25 (2000), 1162–1164.
Electron., 30 (1994), 1313–1326. [14] C. Gmachl, A. Belyanin, D. L. Sivco, V.
[4] R. Colombelli, K. Srinivasan, M. Troccoli, Koncharovsky, M. L. Peabody, N. Owschimikow,
O. Painter, C. F. Gmachl, D. M. Tennant, A. M. A. M. Sergent, F. Capasso, and A. Y. Cho,
Sergent, D. L. Sivco, A. Y. Cho, and F. Capasso, “Optimized Second-Harmonic Generation in
“Quantum Cascade Surface-Emitting Photonic Quantum Cascade Lasers,” IEEE J. Quantum
Crystal Laser,” Science, 302 (2003) 1374–1377. Electron., 39 (2003), 1345–1355.
[5] G. Dehlinger, L. Diehl, U. Gennser, H. Sigg, [15] C. Gmachl, F. Capasso, D. L. Sivco, and
J. Faist, K. Ensslin, D. Grützmacher, and A. Y. Cho, “Recent Progress in Quantum
E. Müller, “Intersubband Electroluminescence Cascade Lasers and Applications,” Rep. Prog.
from Silicon-Based Quantum Cascade Phys., 64 (2001), 1533–1601.
Structures,” Science, 290 (2000), 2277–2280. [16] C. Gmachl, J. Faist, F. Capasso, C. Sirtori,
[6] L. Diehl, S. Mentese, E. Müller, D. Grützmacher, D. L. Sivco, and A. Y. Cho, “Long-Wavelength
H. Sigg, U. Gennser, I. Sagnes, Y. Campidelli, (9.5–11.5 mm) Microdisk Quantum-Cascade
O. Kermarrec, D. Bensahel, and J. Faist, Lasers,” IEEE J. Quantum Electron., 33 (1997),
“Electroluminescence from Strain-Compensated 1567–1573.

Bell Labs Technical Journal 211

[17] C. Gmachl and H. M. Ng, “Intersubband “Improvement of Second-Harmonic
Absorption at l  2.1 mm in A-Plane GaN/AlN Generation in Quantum Cascade Lasers with
Multiple Quantum Wells,” Electron. Lett., 39 True Phase Matching,” Appl. Phys. Lett., 84
(2003), 567–569. (2004), 2721–2723.
[18] C. Gmachl, H. M. Ng, and A. Y. Cho, [28] O. Malis, A. Belyanin, D. L. Sivco, J. Chen, A.
“Intersubband Absorption in Degenerately M. Sergent, C. Gmachl, and A. Y. Cho,
Doped GaN/AlxGa1xN Coupled Double “Milliwatt Second-Harmonic Generation in
Quantum Wells,” Appl. Phys. Lett., 79 (2001), Quantum-Cascade Lasers with Modal Phase-
1590–1592. Matching,” Electron. Lett., 40 (2004),
[19] C. Gmachl, D. L. Sivco, R. Colombelli, 1586–1587.
F. Capasso, and A. Y. Cho, “Ultra-Broadband [29] O. Malis, A. Belyanin, D. L. Sivco, J. Chen,
Semiconductor Laser,” Nature, 415 (2002), A. M. Sergent, C. Gmachl, and A. Y. Cho,
883–887. “Recent Progress in Nonlinear Quantum
[20] R. P. Green, A. Krysa, J. S. Roberts, D. G. Revin, Cascade Lasers,” SPIE Proceedings of Photonics
L. R. Wilson, E. A. Zibik, W. H. Ng, and J. W. West, forthcoming in 2005.
Cockburn, “Room-Temperature Operation of [30] R. Martini, C. Gmachl, J. Falciglia, F. G. Curti,
InGaAs/AlInAs Quantum Cascade Lasers C. G. Bethea, F. Capasso, E. A. Whittaker,
Grown by Metalorganic Vapor Phase Epitaxy,” R. Paiella, A. Tredicucci, A. L. Hutchinson, D. L.
Appl. Phys. Lett., 83 (2003), 1921–1922. Sivco, and A. Y. Cho, “High-Speed Modulation
[21] M. K. Gurnick and T. A. De Temple, and Free-Space Optical Audio/Video
“Synthetic Nonlinear Semiconductors,” Transmission Using Quantum Cascade Lasers,”
IEEE J. Quantum Electr., 19 (1983), 791–894. Electron. Lett., 37 (2001), 192–193.
[22] R. Kühler, A. Tredicucci, F. Beltram, H. E. [31] T. S. Mosely, A. Belyanin, C. Gmachl, D. L.
Beere, E. H. Linfield, A. G. Davies, D. A. Sivco, M. L. Peabody, and A. Y. Cho, “Third
Ritchie, R. C. Iotti, and F. Rossi, “Terahertz Harmonic Generation in a Quantum Cascade
Semiconductor-Heterostructure Laser,” Lasers with Monolithically Integrated
Nature, 417 (2002), 156–159. Resonant Optical Nonlinearity,” Optics
[23] A. A. Kosterev, Y. A. Bakhirkin, and Express, 12 (2004), 2972–2976.
F. K. Tittel, “Ultrasensitive Gas Detection by [32] N. Owschimikow, C. Gmachl, A. Belyanin,
Quartz-Enhanced Photoacoustic Spectroscopy V. Kocharovsky, D. L. Sivco, R. Colombelli,
in the Fundamental Molecular Absorption F. Capasso, and A. Y. Cho, “Resonant Second-
Bands Region,” Appl. Phys. B, 80 (2005), Order Nonlinear Optical Processes in Quantum
133–138. Cascade Lasers,” Phys. Rev. Lett., 90 (2003),
[24] A. A. Kosterev, A. L. Malinovsky, F. K. Tittel, 043902-1–043902-4.
C. Gmachl, F. Capasso, D. L. Sivco, J. N. [33] R. Paiella, F. Capasso, C. Gmachl, H. Y. Hwang,
Baillargeon, A. L. Hutchinson, and A. Y. Cho, D. L. Sivco, A. L. Hutchinson, A. Y. Cho, and
“Cavity Ringdown Spectroscopic Detection of H. C. Liu, “Monolithic Active Mode Locking of
Nitric Oxide with a Continuous-Wave Quantum Cascade Lasers,” Appl. Phys. Lett.,
Quantum-Cascade Laser,” Appl. Opt., 40 77 (2000), 169–171.
(2001), 5522–5529. [34] R. Paiella, F. Capasso, C. Gmachl, D. L. Sivco,
[25] A. A. Kosterev and F. K. Tittel, “Chemical J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho,
Sensors Based on Quantum Cascade Lasers,” and H. C. Liu, “Self-Mode-Locking of
IEEE J. Quantum Electron., 38 (2002), 582–591. Quantum Cascade Lasers with Giant Ultrafast
[26] A. A. Kosterev, F. K. Tittel, R. Köhler, Optical Nonlinearities,” Science, 290 (2000),
C. Gmachl, F. Capasso, D. L. Sivco, A. Y. Cho, 1739–1742.
S. Wehe, and M. G. Allen, “Thermoelectrically [35] J. S. Roberts, R. P. Green, L. R. Wilson,
Cooled Quantum Cascade Laser-Based Sensor E. A. Zibik, D. G. Revin, J. W. Cockburn, and
for the Continuous Monitoring of Ambient R. J. Airey, “Quantum Cascade Lasers Grown
Atmospheric Carbon Monoxide,” Appl. Opt., by Metalorganic Vapor Phase Epitaxy,” Appl.
41 (2002), 1169–1173. Phys. Lett., 82 (2003), 4221–4223.
[27] O. Malis, A. Belyanin, C. Gmachl, D. L. Sivco, [36] C. Roller, A. A. Kosterev, F. K. Tittel, K. Uehara,
M. L. Peabody, A. M. Sergent, and A. Y. Cho, C. Gmachl, and D. L. Sivco, “Carbonyl Sulfide

212 Bell Labs Technical Journal

 Detection with a Thermoelectrically Cooled semiconductor lasers. Her professional interests
Midinfrared Quantum Cascade Laser,” Optics include the characterization of the optical
Lett., 28 (2003), 2052–2054. and electronic properties of novel semiconductor
[37] C. Sirtori, H. Page, C. Becker, and V. Ortiz, materials.
“GaAs-AlGaAs Quantum Cascade Lasers:
Physics, Technology, and Prospects,” IEEE J.
CLAIRE GMACHL is an associate professor in the
Quantum Electron, 38 (2002), 547–558.
Electrical Engineering Department at
[38] A. Soibel, F. Capasso, C. Gmachl, M. L.
Princeton University in New Jersey. Before
Peabody, A. M. Sergent, R. Paiella,
joining the faculty at Princeton, she was a
H. Y. Hwang, D. L. Sivco, A. Y. Cho, H. C. Liu,
distinguished member of technical staff at
C. Jirauschek, and F. X. Kartner, “Active Mode
Bell Labs in Murray Hill, New Jersey. She has
Locking of Broadband Quantum Cascade
an M.S. degree in physics from the University of
Lasers,” IEEE J. Quantum Electron., 40 (2004),
Innsbruck in Austria and a Ph.D. degree in electrical
844–851.
engineering from the Technical University of Vienna,
[39] F. Szmulowicz and G. J. Brown, “Calculation
also in Austria. Dr. Gmachl has demonstrated many
and Photoresponse Measurement of the
innovative QCL concepts, such as bidirectional,
Bound-to-Continuum Infrared Absorption in
multiwavelength, and broadband QCLS, and, recently,
P-Type GaAs/AlxGa1xAs Quantum Wells,”
nonlinear light generation in QCLs. Other key
Phys. Rev. B, 51 (1995), 13203–13220.
contributions include the development of single-mode
[40] A. Tredicucci, C. Gmachl, F. Capasso, D. L.
and tunable-distributed-feedback QCLs and chaotic
Sivco, A. L. Hutchinson, and A. Y. Cho, “Long
micro-cavity lasers with high optical power and
Wavelength Superlattice Quantum Cascade
directionality. Her professional interests include
Lasers at 17 ␮m,” Appl. Phys. Lett., 74 (1999),
research and education in electrical engineering, with a
638–640.
focus on optics and optical electronics. Dr. Gmachl is a
[41] D. Weidmann, A. A. Kosterev, C. Roller,
member of the 2002 TR100 and is a 2002/03 IEEE/LEOS
R. F. Curl, M. P. Fraser, and F. K. Tittel,
Distinguished Lecturer. She is a recipient of the 2003
“Monitoring of Ethylene by a Pulsed
Snell Premium award of the IEEE, the 2000 NASA
Quantum Cascade Laser,” Appl. Opt.,
Group Achievement Award, the 1996 Solid State Physics
43 (2004), 3329–3334.
Award of the Austrian Physical Society, and the 1995
[42] L. R. Wilson, D. A. Carder, J. W. Cockburn,
Christian Doppler Award of Austria. She is a senior
R. P. Green, D. G. Revin, M. J. Steer,
member of the IEEE and LEOS.
M. Hopkinson, G. Hill, and R. Airey,
“Intervalley Scattering in GaAs-AlAs
Quantum Cascade Lasers,” Appl. Phys. Lett., DEBORAH L. SIVCO is a member of technical staff in
81 (2002), 1378–1380. the Semiconductor Physics Research
[43] J. S. Yu, A. Evans, J. David, L. Doris, Department at Bell Labs in Murray Hill,
S. Slivken, and M. Razeghi, “High-Power New Jersey. She has a B.S. degree in
Continuous-Wave Operation of Quantum- chemistry from Rutgers University in
Cascade Lasers up to 60 C,” IEEE Photon. New Brunswick, New Jersey, and an M.S. in
Tech. Lett., 16 (2004), 747–749. materials science from Stevens Institute of Technology
in Hoboken, New Jersey. She has been working on
(Manuscript approved May 2005) the molecular beam epitaxy (MBE) growth of III–V
compounds since 1981. In 1994, she prepared the
OANA MALIS is a research consultant in the world's first quantum cascade laser, in collaboration
Semiconductor Physics Research with Faist, et. al. Her professional interests include the
Department at Bell Labs in Murray Hill, fabrication of advanced mid-infrared light sources and
New Jersey. She has an M.S. degree in detectors, in particular quantum cascade lasers. She is a
physics from the University of Bucharest co-recipient of the 1994 Newcomb Cleveland Prize,
in Romania and a Ph.D. degree in the 1995 British Electronic Letters Premium Award, the
physics from Boston University in Massachusetts. 1996 Technology of the Year Award of Industry Week
Dr. Malis is currently leading the Lucent research magazine, and the 2000 NASA Group Achievement
and development in mid-infrared intersubband Award.

Bell Labs Technical Journal 213

LOREN N. PFEIFFER is a distinguished member of
technical staff in the Semiconductor Physics
Research Department at Bell Labs in Murray
Hill, New Jersey. He received his B.S. degree
in physics from the University of Michigan
in Ann Arbor and his Ph.D., also in physics,
from The Johns Hopkins University in Baltimore,
Maryland. A world-renowned leader in the field of
molecular beam epitaxy (MBE) technology, he is
currently focusing on high-purity growth of
GaAs/AlGaAs quantum well heterostructures. His
devices have been proven to exhibit the highest
electron mobility ever reported for any semiconductor.
His professional interests include ultra-clean
semiconductor growth using MBE, as well as studies of
quantum wires, quantum dots, and other nanoscale
structures with novel properties. Dr. Pfeiffer is a fellow
of both the American Physical Society and the Johns
Hopkins Society of Scholars, and he is the recipient of
the 2004 McGroddy Prize of the American Physical
Society.

A. MICHAEL SERGENT is a member of technical staff in

the Condensed Matter Physics Research
Department at Bell Labs in Murray Hill, New
Jersey. He is currently involved in the
development of new processing techniques
for quantum cascade lasers and contacting
Pentacene single crystals for defect studies. He also
supports the GaN molecular beam epitaxy group at Bell
Labs, where his efforts focus on improving MBE
technology and developing insulated-gate Hall bar
structures for low-temperature transport studies of
AlGaN/GaN heterostructures.

KENNETH W. WEST is a member of technical staff in the

Semiconductor Physics Research
Department at Bell Labs in Murray Hill, New
Jersey. He is currently working on the
growth and characterization of novel
semiconductor structures. His professional
interests include development of new molecular beam
epitaxy processes for high-purity semiconductor
structures. He holds a B.S. degree in physics from
Bucknell University in Lewisburg, Pennsylvania. ◆

214 Bell Labs Technical Journal