Micro-integrated Diode Laser Modules for
Quantum Optical Sensors in Space
vorgelegt von
Anja Kohfeldt,
M.Sc.
geb. in Malchin
von der Fakultät IV – Elektrotechnik und Informatik
Technische Universität Berlin
zur Erlangung des akademischen Grades
Doktor der Naturwissenschaften
– Dr. rer. nat. –
genehmigte Dissertation
Promotionsausschuss:
Vorsitzender: Prof. Dr. Klaus Petermann
Gutachter: Prof. Dr. Günther Tränkle
Gutachter: Prof. Achim Peters, Ph.D.
Gutachter: Prof. Dr. Klaus Brieß
Tag der wissenschaftlichen Aussprache: 20.07.2017
Berlin 2017
Abstract
Quantum sensors are rely on laser technology for preparation and for manipulation
of their quantum probes. With the endeavour of operating quantum sensors in micro-
gravity in order to gain measurement precision, laser technology has to be developed
further to meet the new requirements. In addition to high frequency stability at specific
wavelengths and sufficient optical output power, the laser modules have to be small in
size, lightweight, mechanically robust and energy efficient to be suitable for operations
outside the laboratory, even outside Earth‘s mesosphere.
With FOKUS, the first optical clock in space, and MAIUS, the first atom interferom-
eter in space, two ambitious sounding rocket experiments were initiated that push the
boundaries of both quantum technology and laser technology further.
In this thesis, the laser modules for both quantum sensor missions were developed,
assembled, characterized, and validated. Semiconductor laser diodes are best suited
for small, robust, and energy efficient laser sources. However, the optical output power
of a single diode is limited, and they suffer from frequency stability issues, such as
distortions induced by feedback reflection, and thermal drift.
To overcome these issues, a master oscillator power amplifier (MOPA) configura-
tion, embedded in a hybrid micro-integrated design was chosen. As optical master
oscillator (MO) a distributed feedback (DFB) diode optimized for narrow linewidth
emission at the designated wavelength is used. The MO is shielded against external
feedback by optical isolators. The optical emission of the DFB diode is guided to a sep-
arate optical amplifier chip, boosting the optical output power. All optical components
are hosted on a micro-optical bench (MIOB) made of aluminium nitride for mechanical
stiffness and good thermal conductivity. In addition to the optical and electro-optical
components, it contains temperature sensors for monitoring and thermal stabilization,
as well as an electrical interface supporting the frequency stabilization by allowing the
on-board modulation of the laser diodes injection current. The MIOB has a footprint of
80mmx25mm. It omits movable parts to increase mechanical stability, and it provides
space for an optical fibre coupler on board the MIOB.
To ensure that the requirements on the laser modules can be fulfilled components
are run through a qualification process as described in this thesis before being inte-
grated onto the MIOB. Further, an assembly procedure with active alignment and in-
assembly-characterization is presented.
After integration the laser modules undergo an electro-optical characterization, re-
vealing that a FWHM linewidth of <400kHz and an intrinsic linewidth of <15kHz
(3mm long DFB diode at 100mW DFB output power) can be achieved. The laser
modules provide a spectral single-mode tuning range of >1.5nm, an optical free space
output power of >1W, and an electrical-to-optical efficiency of almost 30%.
The robustness against environmental influences was validated with random-
vibration tests up to 29gRMS, a half-sine shock test at 1500g, and thermal cycles up
to a temperature range of –55°C to +85°C. The laser modules still operated according
to the requirements after these stress tests.
The suitability of these laser modules for spaceborne quantum-optical sensors is
proven in the FOKUS and MAIUS experiments. FOKUS was launched successfully in
April 2015, hosting a DFB module based on the technology developed in this thesis. The
FOKUS apparatus was still operational after returning to Earth. MAIUS was launched in
January 2017, hosting 5 MOPA modules and a DFB module. All laser modules operated
as expected during and after the mission.
The laser modules developed in this thesis are suitable for Earth-bound experiments
as well, e.g. in mobile measurement setups or to save space in laboratories. The concept
is adaptable to other wavelengths, not only increasing the number of atom species that
can be manipulated but also enabling usage in other applications, such as in optical
communications.
Zusammenfassung
Für Quantensensoren spielen Laser eine entscheidende Rolle, sowohl bei der Auf-
bereitung als auch bei der Manipulation der zu untersuchenden Quantenproben. Mit
den Bestrebungen Quantensensoren in Schwerelosigkeit zu betreiben um die Messge-
nauigkeit zu erhöhen, muss sich auch die Lasertechnologie weiterentwickeln. Neben
den Anforderungen nach hoher Frequenzstabilität bei bestimmten Wellenlängen und
ausreichender optischer Ausgangsleistung müssen Lasermodule kompakt, leicht, ro-
bust und energieeffizient sein um aus außerhalb des Labors, gar außerhalb der Erdme-
sosphäre betrieben werden zu können.
Mit FOKUS, der ersten optischen Uhr im Weltraum, und MAIUS, dem ersten Atom-
interferometer im Weltraum, sind zwei ambitionierte Höhenforschungsraketenmissio-
nen ins Leben gerufen worden, welche sowohl die Quantensensortechnologie, als auch
die Lasertechnologie voranbringen. Im Rahmen dieser Dissertation wurden die Laser-
module für diese beiden Missionen entwickelt, integriert, charakterisiert und validiert.
Halbleiterlaserdioden sind sehr kompakt und robust, sowie energieeffizient. Jedoch
ist die optische Ausgangsleistung einzelner Dioden begrenzt und sie sind anfällig für
spektrale Störungen, wie optisches Feedback und thermische Drifts.
Um diesen Limitierungen entgegenzuwirken wurde das Konzept eines hybriden,
mikro-integrierten Masteroszillator-Poweramplifier (MOPA) gewählt. Als Masteroszil-
lator (MO) dient ein Distributed-Feedback (DFB) Laser, welcher für die Emission mit
schmaler Linienbreite optimiert wurde. Der MO wird mit Hilfe von optischen Isola-
toren vor externem Feedback abgeschirmt. Der optische Beam des MO wird zu einer
separaten Verstärkerdiode geführt, welche die optische Ausgangsleistung erhöht. Alle
optischen und elektro-optischen Komponenten des Lasermoduls sitzen auf einer mikro-
optischen Bank (MIOB) aus Aluminiumnitrid mit hoher mechanischer Steifigkeit und
guter Wärmeleitfähigkeit. Neben den optischen Komponenten sitzen auf der MIOB
Temperatursensoren zur Überwachung und Stabilisierung der Modultemperatur. Zu-
dem bietet die MIOB ein elektrisches Interface mit Coaxialanschlüssen und der Op-
tion, den Injektionsstrom für die Frequenzstabilisierung der Halbleiterdioden auf der
MIOB direkt zu modulieren. Die MIOB hat eine Grundfläche von 80 mm x 25 mm und
verzichtet komplett auf bewegliche Teile um die mechanische Stabilität zu erhöhen.
Um sicherzustellen, dass die Lasermodule die Missionsanforderungen erfüllen kön-
nen, werden die optischen und elektro-optischen Komponenten vor der Integration
einem Qualifizierungsprozess unterzogen, welcher in dieser Arbeit beschrieben ist. Des
Weiteren wird der Integrationsprozess der Komponenten in die MIOB ausgeführt.
Nach Fertigstellung der Integration werden die Lasermodule einer elektro-optischen
Charakterisierung unterzogen. Diese zeigt auf, dass die Lasermodule FWHM-
Linienbreiten von <400kHz und intrinsische Linienbreiten von <15kHz (3mm lange
DFB-Diode, bei je 100 mW Ausgangsleistung) erreichen können. Die Lasermodule sind
spektral single-mode über einen Wellenlängenbereich von 1,5nm durchstimmbar, erre-
ichen eine optische Ausgangsleistung von >1W und weisen eine Effizienz von nahezu
30% bei der Wandlung von elektrischer zu optischer Energie auf.
Die Widerstandsfähigkeit gegen externe Umwelteinflüsse wird mit Random-
Vibration-Tests mit bis zu 29 gRMS, Halb-Sinus-Schocktest mit Amplituden bis zu 1500 g,
und Thermalzyklentests im Temperaturbereich von –55°C bis +85°C validiert. Die Per-
formance der Lasermodule entsprach auch nach den Umwelttests noch den Missionsan-
forderungen.
Die Eignung der Lasermodule für den Einsatz in Quantensensoren im Weltraum
wird durch die FOKUS und MAIUS-Missionen unter Beweis gestellt. Die FOKUS-
Apparatur, welche ein DFB-Modul basierend auf der Technologie dieser Arbeit enthielt,
wurde im April 2015 erfolgreich gestartet. Die Apparatur war auch nach Rückkehr auf
die Erde noch voll funktionsfähig. Die MAIUS Mission wurde im Januar 2017 erfol-
greich durchgeführt, mit 5 MOPA-Modulen und einem DFB-Modul an Bord. Auch bei
MAIUS funktionierten die Lasermodule während und nach dem Raketenflug einwand-
frei.
Die Lasermodule, die während dieser Dissertation entwickelt wurden, eignen sich
auch für den Einsatz in erdgebundenen Experimenten. Das Konzept ist für andere
Wellenlängen adaptierbar. Damit erweitern sich nicht nur die Zahl der manipulierbare
Atomspezien in Quantenexperimenten, auch der Einsatz der Lasermodule für andere
Anwendungen, wie etwa optische Kommunikation, wäre denkbar.
Acknowledgements
First, I would like to thank Professor Günther Tränkle who gave me the opportunity
to work in the motivating environment of the Ferdinand-Braun-Institut, Leibniz
Institut für Höchstfrequenztechnik (FBH). I would like to acknowledge his support,
professional supervising, and helpful discussions.
I am deeply grateful to Dr. Andreas Wicht for his support and his valuable
contributions to this work. His enthusiasm and commitment to research was very
inspiring. Further, I would like to thank him for sharing his wide experience in the
field of quantum optical metrology.
I would also like to acknowledge Professor Achim Peters, for his insights into
applied physics, and to thank him for using my laser modules in the laser systems
designed in his group.
I would like to thank Max Schiemangk for introducing me to the Hexapod micro
integration environment and the linewidth measurement, and for his helpful support
in the lab and for the proofreading of almost all of my abstracts.
This work would not have been possible without everybody at FBH who was
involved in diode laser processing and production. Especially I want to acknowledge
Sabrina Kreutzmann and Arnim Ginolas for the diode assembly, but also for all dis-
cussions about bonding, wire bonding, and for the help to remove faulty components.
Maik Erbe and Robert Smol did a great job in supporting me with all the electro-optical
characterizations performed over the years. Thank you!
Further, I would like to thank the entire Laser Metrology Lab at FBH for establishing
and running a fantastic lab, but also for all the lunch and in-between discussions, and
for the beer and pizza from time to time.
I am grateful to Markus for the motivation that kept me working and for the
discussions about cold atoms in general.
Special thanks go to Aline, Hanna, Sunti, Belkeys, Fredi, and Heike for reading and
commenting on this thesis.
I would like to thank the DLR for funding under grant number: 50WM1134.
Finally, I would like to thank my family, first Benno for always being around when
I need him, for all the motivation, dissipation, and bringing me back to focus again.
And of course I have to thank my parents for all the support and trust they put into me.
This thesis would have been impossible without you. <3
Contents
1 Lasers and Quantum Optical Sensors in Space . . . . . . . . . . . . . . . . . 1
1.1 Lasers in Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Quantum Optical Metrology in Space . . . . . . . . . . . . . . . . . . . . . 2
1.3 Outline of this Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Quantum Sensors and Laser Module Requirements . . . . . . . . . . . . . . 5
2.1 The Sounding Rocket Experiment FOKUS . . . . . . . . . . . . . . . . . . 5
2.1.1 FOKUS Mission Goals . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.3 Flight Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.4 Requirements of Laser Modules in FOKUS . . . . . . . . . . . . . 10
2.2 The Sounding Rocket Mission MAIUS . . . . . . . . . . . . . . . . . . . . . 11
2.2.1 Mission Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.2 Scientific Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.3 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.4 Requirements on Laser Modules . . . . . . . . . . . . . . . . . . . 16
3 Diode Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1 Concept and Functionality of Diode Lasers . . . . . . . . . . . . . . . . . . 19
3.1.1 Radiative Transitions and Laser Conditions . . . . . . . . . . . . . 20
3.1.2 Advantages and Application of Diode Lasers . . . . . . . . . . . . 21
3.1.3 Basic Properties of Semiconductor Lasers . . . . . . . . . . . . . . 21
3.1.4 Example: Distributed Feedback lasers . . . . . . . . . . . . . . . . 28
3.1.5 Example: Power Amplifier Chips . . . . . . . . . . . . . . . . . . . 28
3.2 Monolithic and Hybrid MOPAs . . . . . . . . . . . . . . . . . . . . . . . . . 31
4 Concept and Assembly of Micro-integrated Laser Modules . . . . . . . . . 33
4.1 Concept and Design of the Hybrid Integrated Laser Modules . . . . . . 33
4.1.1 Electro-Optical Design of the Laser Modules . . . . . . . . . . . . 34
4.1.2 Structural Design of the Laser Modules . . . . . . . . . . . . . . . 38
4.1.3 Features of the Laser Modules . . . . . . . . . . . . . . . . . . . . . 44
4.2 Laser Module Assembly Process Flow . . . . . . . . . . . . . . . . . . . . . 45
4.3 Integration of Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3.1 Assembly Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3.2 Integration of Master Oscillator Optics . . . . . . . . . . . . . . . 49
4.3.3 Integration of Amplifier Optics . . . . . . . . . . . . . . . . . . . . 53
ix
5 Pre-Integration Component Characterization . . . . . . . . . . . . . . . . . . 57
5.1 Laser Diode Pre-Characterization . . . . . . . . . . . . . . . . . . . . . . . 58
5.1.1 Diode Selection Procedure . . . . . . . . . . . . . . . . . . . . . . . 61
5.1.2 Pre-characterization and Burn-In Facility . . . . . . . . . . . . . . 62
5.1.3 Qualification Criteria for Laser Diodes . . . . . . . . . . . . . . . . 64
5.1.4 Results of DFB Diode Pre-Integration Characterization . . . . . . 66
5.1.5 Results of Power Amplifier Pre-Integration Characterization . . 71
5.2 Passive Component Pre-Integration Characterization . . . . . . . . . . . 77
5.3 Conclusion of Pre-Integration Qualification . . . . . . . . . . . . . . . . . 79
6 Characterization and Qualification . . . . . . . . . . . . . . . . . . . . . . . . 81
6.1 Electro-optical Characterization . . . . . . . . . . . . . . . . . . . . . . . . 81
6.1.1 Optical Output Power . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.1.2 Optical Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.1.3 Linewidth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.1.4 Modulation Capability of the MIOB . . . . . . . . . . . . . . . . . 89
6.2 Environmental Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.2.1 Definition of Test Parameters . . . . . . . . . . . . . . . . . . . . . 91
6.2.2 Results of Environmental Tests . . . . . . . . . . . . . . . . . . . . 93
6.3 Conclusion of Characterization and Stress Tests . . . . . . . . . . . . . . 97
7 Evaluation of Sounding Rocket Missions . . . . . . . . . . . . . . . . . . . . . 99
7.1 FOKUS Mission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
7.1.1 Qualification and Ground Operation . . . . . . . . . . . . . . . . . 99
7.1.2 The Flight Campaign . . . . . . . . . . . . . . . . . . . . . . . . . . 101
7.1.3 Post-Flight Characterization of Laser Module . . . . . . . . . . . . 103
7.2 MAIUS Mission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
8 Summary and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
8.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
8.2 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
A List of diodes used in this work . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
B Schematics of MIOB PCBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
C Examples of Facet Inspection Ratings . . . . . . . . . . . . . . . . . . . . . . . 121
C.1 Qualified Facets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
C.1.1 Blameless Facets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
C.1.2 Qualified Facets with Non-critical Defects . . . . . . . . . . . . . . 122
C.2 Facet Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
D Terms and Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
E Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
E.1 Print . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
x
E.2 Oral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
E.3 Poster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
xi
xii
Chapter 1
Lasers and Quantum Optical Sensors
in Space
1.1 Lasers in Space
Ever since the first laser, an acronym for light amplification by stimulated emission of
radiation, was constructed by Mainman in 1960 [1], the monochromatic and coherent
light sources gained rapidly in importance and applications. One year later, in 1961,
a ruby laser was used by Campbell and Koester [2]to destroy a tumour. With that,
the laser was introduced to medicine. Nowadays it is an indispensable part of surgery.
Today, lasers have widely spread applications in industry, everyday life, spanning from
material processing [3]to optical communication [4,5], medicine and entertainment.
Lasers also play an important role in metrology and science. A short wavelength source
of coherent radiation not only allows for high precision measurements [6], it can also
be used to manipulate atom states and hence constitute a basic tool for atom physics
and quantum optical metrology (see section 1.2).
With this wide spectrum of applications it is not surprising that lasers are operated
in very different environments. One of the most demanding operation environments
is space1. Not only that lasers have to be small in size, low in weight, and power
efficient in order to minimize launch costs and required resources, they have to survive
vibrations and shocks, radiation, and temperature gradients up to some 100 K. When
operated on a satellite, they have to be extremely reliable because malfunctioning parts
can typically not be replaced.
In space, lasers are mostly used in measurement and communication devices. A
popular measuring device is a light detection and ranging (LIDAR) system, which can
be used for supporting docking manoeuvres and to analyse atmospheric compositions.
The first airborne LIDAR instrument was reported in 1985 [7], followed by the first
LIDAR instrument in space, the Lidar In-space Technology Experiment (LITE), con-
ducted by "National Aeronautics and Space Administration" (NASA) in 1994 at the
STS-64 space shuttle mission. It used a pair of flash light pumped Nd:YAG lasers, con-
1There are a lot of definitions of "space". One of the most accepted definitions is given by the Fédéra-
tion Aéronautique Internationale (FAI) with everything outside the Kármán line, 100 km above sea level.
2CHAPTER 1. LASERS AND QUANTUM OPTICAL SENSORS IN SPACE
suming 1865W when transmitting [8], the total instrument mass was 990kg. Modern
spaceborne LIDAR missions, such as the EarthCARE mission, scheduled for 2018, are
still based on Nd:YAG lasers, however, the instrument’s consumption and mass could
be reduced to around 300W and 230kg, respectively [9].
Optical satellite communication with lasers gains importance, building a network
around Earth to cope with the increasing demand for higher data rates and availability
even in remote areas [10]. The first inter-satellite link was demonstrated 2001 with
the Semiconductor-laser Inter-satellite Link EXperiment (SILEX) by European Space
Agency (ESA) [11]using semiconductor lasers with an optical output power of 60mW,
transmitting up to 50Mbit/s. The total mass of the optical terminal was 160kg [12].
Depending on the application, spectral requirements on the lasers can be very strict.
In communication, dependent on the modulation procedure, linewidths smaller than
8MHz2, or even down to 6.9 kHz3are necessary. Emission output power levels larger
than 1W are desirable to bridge long distances in space communication.
The requirements on laser sources for space communications are comparable to
requirements that apply to certain quantum optics applications. Natural linewidths of
alkali atom transitions may be, as well, in the range of 6 MHz [15, 16], the linewidth
requirement for the coherent manipulation of atomic ensembles may be in the range
of less than 100kHz [17]. Also, more than 1W output power per laser source may be
required. These quantum optical sensors also reach out into space to gain precision, as
will be described in section 1.2. Hence, stable, compact, and reliable laser sources will
soon be required for quantum metrology applications in space.
Commonly, fibre lasers or optically pumped solid-state laser technologies are used to
fulfil the requirement combination of narrow linewidth and high output power. How-
ever, solid-state lasers are quite energy inefficient [8], large, and are limited to very
specific emission frequencies. By developing all-semiconductor based laser systems for
space applications, it is possible to minimize the size and weight as well as the power
consumption of the instruments. Semiconductor lasers also enable a wide range of
emission frequencies and are, therefore, the best technology choice for quantum opti-
cal sensors in space.
1.2 Quantum Optical Metrology in Space
The discoveries in the field of quantum physics in the past century [18–21]laid the
basis for the construction of measurement devices with a precision beyond what is
possible with classical approaches. Quantum metrology makes use of the wave-particle
dualism [22]and is in theory ultimately limited by the uncertainty principle. High-
precision quantum sensors can be used to detect gravitational waves [23]. They are
used to precisely determine physical constants [24–26]or acceleration [17, 27]. The
most precise, man-made reference systems (clocks) are based on atomic transitions
[28,29]. Applications for quantum technology can be found not only in fundamental
2with binary phase-shift keying (BPSK) [13]
3with quadrature amplitude modulation (QAM) [14]
1.2. QUANTUM OPTICAL METROLOGY IN SPACE 3
physics, but also in time keeping and navigation, and in Earth exploration, as well as
in telecommunication.
Quantum optical technologies are based on single atoms or ions, or on precisely
controlled ensembles of atoms or ions. They owe their potential to the fact that the
properties of an atom are well known and the interactions with the environment can
be described very accurately. This is important for understanding the measurement
environment and to guarantee accuracy and reproducibility of the measurement. A
milestone for quantum optical sensors was the first realization of a Bose-Einstein con-
densate (BEC) [30,31], a macroscopic quantum object that can be described by a single
wave function and allows for the implementation of high-resolution atom interferom-
eters [32,33], for example to determine accelerations.
The sensitivity of these experiments, however, increases with the measurement
time. To enhance the sensitivity, atom interferometers have been realized with atomic
ensembles launched in an atomic fountain apparatus [27,34,35]or dropped in drop-
towers [36]. Both approaches provide microgravity environments on a limited scale
on Earth. However, in Earth-bound experiments the duration of the microgravity pe-
riod and hence of the measurement time is limited. Deploying atom interferometers in
space is an approach to overcome this limitation. A first step towards space is the de-
ployment on a sounding rocket, extending the measuring time from seconds to several
minutes.
Measurement apparatus, leaving the laboratory in order to be operated on mobile
platforms or even in space, must fulfil the environmental requirements, given by the
platform. Hence, new technologies have to be developed.
Quantum sensors rely on ultra-precise oscillators, lasers for atom cooling, stimu-
lation, and manipulation. In most laboratory based experiments the precise oscilla-
tors are realized with solid-state lasers, such as the Nd:YAG lasers [23], Ti:Sapphire
laser [37]or Er:Glass with a crystal cavity [38,39]. These laser systems require an ac-
tive, optical medium for pumping, either realized with flash lights or with diode lasers.
This reduces the efficiency of the laser, adds complexity to the system and increases
the weight. The wavelength is defined by the doping, hence the spectral tuning range
is very limited for most solid-state laser types. With renouncing the solid-state cavity
and operating with the semiconductor laser directly, the laser system of quantum sen-
sors can be designed in a more compact and energy efficient way. However, frequency
stability of the diode emission has to be ensured.
The focus of this work is on the development and implementation of semiconductor
based laser modules for quantum sensors on board sounding rockets.
4CHAPTER 1. LASERS AND QUANTUM OPTICAL SENSORS IN SPACE
1.3 Outline of this Thesis
As explained above, laser systems that are reliable, compact, robust, and energy effi-
cient are key components of quantum optical sensors, specifically for sensors operated
in space. The aim of this thesis is to develop laser modules for quantum optical exper-
iments on BECs of rubidium and potassium on board of sounding rockets.
Chapter 2 introduces two quantum optical experiment missions that are operated in
space on board of sounding rockets, the FOKUS mission and the MAIUS mission. Based
on the experiment’s requirements and on the operation environment, the requirements
and specifications that apply to the lasers for these experiments are derived.
In chapter 3 the basic theory of semiconductor lasers is presented. Laser module
configuration concepts such as the MOPA concept and the hybrid integration approach
are discussed.
Based on the requirements identified in chapter 2 and the technology introduced
in chapter 3, the design of the laser modules is described in chapter 4. The chapter
also outlines the integration process of the semiconductor laser modules, including the
description of a specially developed assembly station for active alignment of all optical
components, as well as the description of the micro-integration procedure.
Single components, such as laser diodes, amplifiers, as well as passive components
have to be qualified before integration to ensure stable and reliable performance. Chap-
ter 5 focusses on the preparation and characterization of these components. Qualifica-
tion criteria of the components are defined and the description of a specially developed
pre-integration measurement setup is presented. Furthermore, typical reasons of fail-
ure as well as the total yield of the component qualification process are described and
discussed.
Chapter 6 presents a detailed characterization of the laser modules after integration.
The electro-optical performance, such as spectral behaviour, optical output power, and
frequency stability, is investigated and compared to the mission critical requirements
given in chapter 2. Environmental stress tests are carried out and the laser module
performances before and after the stress test campaign are compared.
The validated laser modules are delivered for integration into the mission’s experi-
ment apparatus. In chapter 7 the results of the FOKUS experiment and the laser module
behaviour during the MAIUS mission are presented. The FOKUS experiment aims at
the demonstration of the suitability of the newly developed laser module technology
for the employment in quantum optical sensors in space. Further, this chapter presents
a post-flight characterization of the FOKUS laser module, validating that it is still oper-
ational.
Chapter 8 gives a brief summary of lessons learned during the fabrication of the
laser modules and an outlook for future work.
Chapter 2
Quantum Sensors and Laser Module
Requirements
A new generation of quantum optical sensors is being deployed in microgravity in order
to improve measurement accuracy [28, 40]. To move an experiment out of the labo-
ratory and into microgravity (µg), most of the hardware, including the laser modules,
have to be re-designed to be able to cope with the new operational environment.
This chapter introduces two experiments on board of sounding rockets that repre-
sent quantum sensors in space. In addition to the mission goals and a brief description
of the technical realization of the experiments, the requirements of the laser modules
are outlined. The mechanical requirements are defined by the environment on board
a sounding rocket, the electro-optical requirements are determined by the experiment
design and the task that shall be performed by the lasers in the experiment. The sum-
mery of these requirements is the baseline of the design of the laser modules that will
be discussed in this thesis.
2.1 The Sounding Rocket Experiment FOKUS
Precision frequency control of light is a key requirement for quantum sensors. In or-
der to allow for ultra-precise measurements in microgravity, suitable technology for
frequency stabilization has to be provided. The experiment "Faserlaserbasierter Optis-
cher Kammgenerator unter Schwerelosigkeit" (FOKUS)1, led by Menlo Systems GmbH,
a company specialized on frequency combs [41], is the first mission ever to operate op-
tical frequency stabilization technology for semiconductor lasers in space. The FOKUS
experiment is described by Lezius in [42].
The following paragraphs present the mission goals and experimental setup of
FOKUS and summarize the requirements for the diode laser modules employed in this
experiment. The results of the FOKUS experiment, launched in April 2015, are de-
scribed in chapter 7.
1The FOKUS activities at the FBH were funded by the Federal Ministry for Economic Affairs and
Energy (BMWi) under the grant number 50WM1240
6CHAPTER 2. QUANTUM SENSORS AND LASER MODULE REQUIREMENTS
2.1.1 FOKUS Mission Goals
The mission goals of the FOKUS project can be summarized as:
• The development and validation of frequency stabilization technologies of laser
systems on board a sounding rocket
• The deployment of the first ever diode laser module with its frequency stabilized
to an optical transition, and the first ever frequency comb in space
• A preliminary test of the master laser module and frequency reference system
that should also be provided for the "Materiewelleninterferometrie unter Schw-
erelosigkeit" (MAIUS) mission
• A proof-of-principle for a spaceborne local position invariance (LPI) experiment
that includes an optical clock.
The LPI results from the Einstein equivalence principle (EEP) and signifies that two
clocks of different internal structure should see the same red-shift as they move together
through a changing gravitational field. The experimental setup allows the comparison
of two clocks that are based on different oscillators and physical implementations and
can hence be considered clocks of different type.
2.1.2 Experimental Setup
The experiment can be divided into two optical subsystems: the frequency comb system
and the rubidium (Rb) reference system, as illustrated in figure 2.1.
Figure 2.1: Schematic of FOKUS experiment: the RF atomic CSAC is compared to
the optical atomic clock, formed by the Rb reference system, via the frequency comb
system. Optical signals are illustrated in red, electrical signals in blue.
The Rb reference system (together with the counter) can be considered as optical
atomic clock, delivering a frequency stabilized to an optical transition of Rb. The fre-
quency comb system is referenced on a chip scale atomic clock (CSAC) (SA.45s CSAC,
Symmetricon), based on microwave oscillation, and allows the comparison of the CSAC
with the optical atomic clock. A frequency doubling of the output of the frequency
comb system, done by second harmonic generation (SHG), is required to adapt the
2.1. THE SOUNDING ROCKET EXPERIMENT FOKUS 7
output of the frequency comb to the range of the frequency of the Rb reference sys-
tem. Both frequencies, the Rb reference system output and the corresponding needle
of the frequency doubled comb system, are superimposed on a photodiode (PD) and
the difference of both frequencies is detected by a counter.
Figure 2.2: FOKUS Rb reference system, hosting a spectroscopy board (right) with
master laser (left, underneath semi-transparent lid) in flight model housing
The Rb reference system is depicted in figure 2.2. It consists of a semiconduc-
tor laser module and a spectroscopy module. The semiconductor laser module is a
distributed feedback (DFB) diode laser micro-integrated on an aluminium nitride (AlN)
based micro-optical bench (MIOB). It emits approximately 19mW at 780nm. More de-
tails of the design of the laser module will be given in chapter 4. The free-space laser
beam is collimated with micro-lenses and coupled into a polarization maintaining fibre
with a Zerodur fibre coupler on board the MIOB. The beam is guided to the spec-
troscopy module. The spectroscopy module rests on a 50mm thick Zerodur base plate
in order to maintain the positions of the optically aligned components, independent of
the environmental temperatures and mechanical loads. It contains fibre coupled input
and free-space optics and is described by Duncker in [43]. The incoming light is first
split into two beams with a polarizing beam splitter (PBS). Both beams are guided
through a vapour gas cell filled with Rb. While one beam passes the cell only once, the
second beam passes the Rb cell twice, producing a Doppler broadened and a Doppler
free spectroscopy signal, respectively. Each beam is recorded on a separate PD. The
AC part of the Doppler free PD signal is mixed down with a 10MHz reference signal,
provided by the CSAC and serves as a frequency modulation spectroscopy (FMS) error
signal for the feedback loop. The FMS error signal is then processed by the servo elec-
tronics and fed back to the laser’s injection current in order to lock the laser emission
frequency to the frequency of the D2 transition (52S1/2→52P3/2) of Rb [15]. A fibre-
integrated splitter between laser module and spectroscopy module is used to guide a
8CHAPTER 2. QUANTUM SENSORS AND LASER MODULE REQUIREMENTS
fraction of the light to another photodiode, allowing for beat measurements with the
light from the frequency comb.
The comb system contains a mode-locked erbium (Er)-fibre laser [44,45]emitting
approximately 3mW at 1560nm with a bandwidth of 30 nm [46]. The output of this
oscillator is amplified and split with a 1:2 fibre splitter. One part is used in carrier-
envelope-offset (CEO) beat detection, the other part is frequency doubled in a SHG
stage in order to facilitate beat note measurements with the light from the Rb stabilized
diode laser.
The chip scale atomic clock, providing the 10MHz reference signal of the FOKUS
scientific payload, is also used for phase-locking of the repetition rate signal and the
carrier envelope offset signal. The CSAC also references the counter that measures the
phase of the beat note signal between the Rb system and the frequency comb, as shown
in figure 2.1.
2.1.3 Flight Configuration
Figure 2.3: Flight configuration of a VSB-30 sounding rocket
The German sounding rocket program "Technologische Experimente unter Schwere-
losigkeit" (TEXUS) serves as launch vehicle for the FOKUS experiment. The rocket mo-
tor of the TEXUS rocket is a "Veículo de Sondagem Booster – 30" (engl.: Booster Sound-
ing Vehicle) (VSB-30) type booster. It consists of two solid propellant rocket stages,
which are depicted in figure 2.3. The rocket was developed at "Instituto de Aeronáutica
e Espaço", Brazil (IAE) in cooperation with "Mobile Raketenbasis des Deutschen Zen-
trums für Luft- und Raumfahrt" (DLR-MORABA) and is launched from Esrange Space
Center near Kiruna, Sweden.
The design of the launch vehicle and environment provided by the launch vehi-
cle set limits on the mechanical design of the payload and define the environmental
requirements of the payload design. This also affects the components and modules
integrated into the scientific payload.
The payload is mounted on top of the rocket engines with a motor adapter ring
adjusting the diameter differences between engines and payload. The entire payload
2.1. THE SOUNDING ROCKET EXPERIMENT FOKUS 9
(a) (b)
Figure 2.4: Acceleration load (a) and vibration level (b) of a typical launch. Burnout
of 1st stage at 13.5s, maximum load is reached before burnout of 2nd stage at 36 s.
Data from TEXUS 45 mission [47,48]
is limited to 400kg and has to fit into an approximately 4500 mm x 440 mm drum
shaped envelope [49]. Since the envelope dimensions also include thickness of the
outer structure and secondary mounting structure, the space for the actual experiment
is even more limited.
With an apogee of about 260km approximately 360s of microgravity with a µg
level of 10−4g[49]can be achieved. During ascent the engines of the VSB-30 will cause
significant acceleration and vibration. Figure 2.4 illustrates a typical acceleration load
during launch. The z-acceleration peaks shortly before burnout of the second stage
(S30 engine) with 12.1g at 13s. Vibration levels peak at ignition of each stage at 0s
and 18s time of flight. The integrated RMS value should not exceed 2 gRMS. However,
it is recommended for TEXUS flights that payload components as well as experiment
modules are considered with 8.1 gRMS qualification level [47]in all three axes as design
parameter. At re-entry even higher accelerations are expected, though they might differ
due to the re-entry conditions at 40km altitude and the centre of gravity of the payload.
Friction of the atmosphere causes 20 – 25g while touch down shocks are in the range
of 50g to 250g [47]. Depending whether the experiment shall be operational after
touch down, these acceleration loads have to be taken into account as well.
During launch preparation the VSB-30 is connected to the ground support equip-
ment (GSE), providing the payload with electricity, data link and fluids like cooling
water. At launch the connection to the GSE is cut and the experiments have to perform
autonomously. Since the service module does not provide sufficient electricity for most
experiments, a battery with additional capacity to supply the experiment’s electronics
have to be included in the scientific payload, on the cost of space and weight for the
experimental apparatus. Additionally this limits the power consumption of the exper-
iment. Therefore energy efficient components for the experimental hardware have to
be chosen.
For frequency stabilized components heat management is an important issue as
well. Since the operation during flight only lasts several minutes, an active cooling of
10 CHAPTER 2. QUANTUM SENSORS AND LASER MODULE REQUIREMENTS
requirements value comment
electro-optical
wavelength (λ) 780.241 nm 87Rb D2 line [15]
FWHM linewidth ≤6MHz 87Rb D2 line [15]
tunability at WP 100GHz 87Rb D2 [15] + margin
free-space optical power >10mW (CW) system design
max. WP current 200mA current driver limitation
mechanical/environmental
volume <(200 x 100 x 50)mm3system design
vibration loads, qualification 8.1gRMS VSB-30 rocket [47]
shock loads, qualification ≥250g VSB-30 rocket [47]
WP temperature ≥35° C passive temp.control
storage temperature –30°C ... +45°C Swedish winter [47]
Table 2.1: Requirements on laser module of the Rb reference system due to experimen-
tal needs and microgravity platform environment
the experiment’s apparatus is not required. Still, the heat distribution and passive cool-
ing structures are critical design parameters. Due to the electronics the environment
inside the envelope will reach temperatures above 35° C. The laser module, however,
requires active thermal stabilization to maintain frequency stability. Ideally the working
point (WP)2temperature of the laser module is in the range close to this environmental
temperature in order to reduce the load on the cooling system.
2.1.4 Requirements of Laser Modules in FOKUS
Table 2.1 summarizes all requirements on the laser of the rubidium reference system
for the FOKUS experiment. Electro-optical requirements refer to the operation of the
laser and influence the choice of the laser diode. The mechanical requirements make
demands on the package and physical interfaces of the laser module.
Spectral requirements, such as wavelength and linewidth, are determined by the
87Rb transition of the D2 line. In order to stimulate atoms in this state, the laser has
to emit at the transition wavelength, which is 780.241 nm (384.230 484 THz) for the
D2 line [15], and the lasers full width at half maximum (FWHM) linewidth should be
well below the natural linewidth of the transition. Since the natural linewidth of the
87Rb D2 line transition is 6.07 MHz [15], the laser’s emission linewidth must not exceed
this value. In order to be able to address all allowed transitions of the D2 line, a range
of emission frequency of λD2- 2.369GHz to λD2+4.198GHz is necessary, resulting
in a tuning range of 6.567GHz. In addition, the tunability of the laser’s frequency
has to exceed the frequency drift of an unstabilized laser, which also can reach several
GHz. With that, a minimum tuning range of 10 GHz would be required to achieve
all D2 lines. However, the wider the tuning range, the easier and more stable the
2The working point is the set of physical parameters a diode laser is operated at.
2.2. THE SOUNDING ROCKET MISSION MAIUS 11
frequency stabilization algorithms can perform. Therefore a tuning range of more than
100GHz seems appropriate. Also this would guarantee a single-mode emission at the
desired wavelength not only for one but for various working points and would allow
experimental shifting the wavelength in the GHz range.
Several mW of optical output power are required due to losses at the fibre couplers
and free-space components located on the spectroscopy module. In order to provide
sufficient optical power for the experiment, at least 10 mW optical free-space output
power of the laser module should be provided. The maximal injection current of the
DFB laser is limited by the electronics of the experimental setup, which only delivers a
maximum current of 200mA. Since the electrical power for temperature stabilization
is limited as well, the WP temperature of the laser should be around 35°C.
Mechanical requirements, such as vibration and shock loads, are given by the
launch vehicle, as described above, and are defined in the TEXUS-interface control
document [47]. The physical size of experiments operated on sounding rockets is
limited. In FOKUS, only a box with a volume of 200x10050mm3is available for the
Rb reference system, containing both the laser module and the spectroscopy module.
Consequently, the laser module has to be smaller than this. The rocket launch side is
situated near Kiruna in Sweden. Since the campaign can be in winter and last several
weeks and recovery may last several days, the experiment may be exposed to –30 ° C
outside temperature.
The Rb reference system with DFB laser and spectroscopy module has the same
design and specification as the master laser and reference system of the MAIUS mission,
which is described in the following section.
2.2 The Sounding Rocket Mission MAIUS
Figure 2.5: Official logo
of MAIUS-1 mission
The first mission to operate an atom interferome-
ter [50]on board a sounding rocket is the MAIUS
mission. MAIUS is executed by a project group lead
by the "Leibniz-Universität Hannover", Institut für
Quantenoptik (LUH), and is part of the project series
"Quantengase unter Schwerelosigkeit" (QUANTUS).
The main objective of QUANTUS is the production
of and research on Bose-Einstein condensates (BEC)
in a microgravity (µg) environment. Predecessor ex-
periments of MAIUS as described in [51]and [36]
are operated at the droptower in Bremen, Germany.
The droptower only provides 4.7s (in drop config-
uration) and 9.4s (in catapult configuration) of µg
operation time, thus limiting measurement time and accuracy.
This section will give a brief overview of the MAIUS goals and experimental appa-
ratus, and determine the requirements for laser modules necessary for Rb-BEC based
atom interferometry.
12 CHAPTER 2. QUANTUM SENSORS AND LASER MODULE REQUIREMENTS
2.2.1 Mission Goal
In order to enhance observation time of the experiment and to prepare for future mis-
sions of spaceborne interferometry setups the sounding rocket mission MAIUS-1 was
initiated. Based on the goals of the QUANTUS experiments the MAIUS mission goals
were defined as follows:
• The creation of the first BEC using rubidium-87 in space
• BEC expansion velocity of less than 0.3 mm/s, corresponding to a temperature
below one nanokelvin
• BEC observation time of several seconds
• The demonstration of matter-wave interferometry based on a BEC in space.
The MAIUS-1 apparatus is a technical demonstrator and shall prove the ability to per-
form matter-wave interferometry on board a sounding rocket. This shall serve as a door
opener for future space missions, e.g. on board the International Space Station (ISS)
or other satellites.
2.2.2 Scientific Methods
In order to understand the requirements of the laser modules that are needed to achieve
the mission goals of MAIUS, this section gives a brief introduction in the scientific meth-
ods used in this experiment. A more detailed description of the physics and techniques
behind matter-wave interferometry can be found in [52].
2.2.2.1 Cooling of Atoms and BEC Creation
Figure 2.6 illustrates the hyperfine structure of 87Rb, the element used in MAIUS.
In order to prepare a BEC, the corresponding atoms have to be cooled. This
is accomplished by a combination of optical Doppler cooling with laser beams and
trapping of atoms in a magnetic field [53, 54]. Three orthogonal laser beams with
their frequency red-shifted to one of the absorption lines (here: D2 line of 87Rb,
|52S1/2,F=2〉→|52P3/2,F0=3〉, see figure 2.6) are directed onto the Rb atoms. While
absorbing the photon energy the momentum of the photon is transferred to the atom
effectively causing a friction force on the atom in the direction opposite to the laser
beam. Due to the orthogonal directions of the tree laser beams and the fact that the
photons, which are emitted spontaneously during the relaxation of the excited atom,
are emitted isotropically, the atom loses kinetic energy. Since the temperature of an
ensemble of atoms is a measure of the random internal kinetic energy, reducing the
kinetic energy of the atoms is equivalent to reducing the temperature of the atoms.
With a low probability, an atom relaxes into the lower hyperfine state |52S1/2,F=1〉,
as depicted in blue in figure 2.6, which removes the atom from the cooling cycle. To re-
excite these atoms and to bring them back into the optical cooling cycle, an additional
laser with a wavelength tuned to the |52S1/2,F=1〉→|52P3/2,F0=2〉transition of 87Rb
is required. This laser is called re-pump laser and has a wavelength offset of +6.5 GHz
with respect to the cooling laser [15]. The re-pumper needs three orders of magnitude
2.2. THE SOUNDING ROCKET MISSION MAIUS 13
52P3/2
52S1/2
F’=3
F’=2
F’=1
F’=0
F=2
F=1
6.83 GHz
72 MHz
157 MHz
267 MHz
384.230 THz
780.241 nm
(frequency red shifted)
(virtual level)
cooling transition
relaxation
repumping transition
Bragg pulses
detection
Figure 2.6: Hyperfine structure of 87Rb D2 transition line and frequencies differences.
Data from [15]
less optical output power than the cooling lasers due to the low probability of the atoms
relaxing to this state.
Cooling, as described above, is limited by heating through spontaneous emission
that leads to the Doppler-limit, which corresponds to a temperature in the range of
100µK[55]. In addition, the atoms will diffuse away from the volume in which the
six laser beams intersect because the frictional force is not dependent on position. A
restoring force can be introduced by adding a magnetic quadruple field to maintain a
defined position of the atom cloud [56]. This technique is called magneto-optical trap
(MOT).
In order to create a BEC, temperatures in the nK regime are necessary [57]. The
final cooling step is performed with evaporative cooling, firstly described by Hess in
1986 [58]. The atoms which were cooled in the magneto-optical trap (MOT) are trans-
ferred into a harmonic magnetic trap. A radio frequency (RF) field, appropriately ap-
plied to the atom chip, selectively removes atoms with the highest kinetic energies in
the ensemble and the remaining atoms re-thermalize at a lower temperature. When
sweeping down the frequency of the RF signal, the remaining atom cloud gets colder
until it reaches the critical temperature for phase transition to a BEC.
A more detailed description of the physics and techniques used in MAIUS can be
found in [48].
14 CHAPTER 2. QUANTUM SENSORS AND LASER MODULE REQUIREMENTS
2.2.2.2 Matter Wave Interferometry
BECs can serve as matter wave source for interferometry experiments. Ultra-cold atoms
in BECs lose their individual identity and form a macroscopic wave function, as de-
scribed in the Nobel lectures of Ketterle and Cornell [30, 31]. Very low momentum
widths and a macroscopic de Broglie wavelength allow for new levels of interferometric
sensitivity, large interrogation times [36]and an enhanced signal to noise ratio (SNR).
A popular method for determination of the gravitational acceleration of the atoms
is the Mach-Zehnder-type interferometer [32, 36]. The wave is separated into two
paths, travelling for a time 2T before both paths are superimposed again. Interfer-
ometric fringes are caused by the accumulated phase difference between both paths
when travelling separately. The matter-wave manipulation is realized with a pair of
counter-propagating laser pulses that are detuned either by the recoil energy or by the
energy difference of the two ground states. This causes the atoms to oscillate via a
virtual transition, which is detuned to a real transition by several 100 MHz in order
to avoid spontaneous emission. The length of the laser pulse defines the effect of the
laser pulse. A full πpulse causes a population inversion, also called "mirror", a π/2
pulse creates a superposition of two equal distributed states, a "splitter". A splitter that
only shifts the momentum of the atoms is called Bragg splitter. If the atoms oscillate
between the internal ground state and momentum state, the splitter is called Raman
splitter. In MAIUS a Bragg splitter is realized.
The Mach-Zehnder interferometer is then realized with a sequence of π/2 – π–π/2
laser pulses. The Bragg splitter coherently transfers half of the atoms to another state
and gives them a momentum kick. The superposition is done by exchanging momentum
states (mirror) after a time T and by the coherent recombination after time 2T with a
second splitter pulse.
The number of atoms in each state is determined by absorption imaging. The light
of a laser pulse tuned to the |F=2〉 →|F0=3〉transition, as depicted in figure 2.6, is
captured by a charge-coupled device (CCD) camera, revealing the atoms that absorbed
the detection light. From the intensity and location of the atoms on the camera the
number of atoms and the phase difference can be estimated. With that, acceleration
through gravitational forces can be calculated.
2.2.3 Experimental Setup
The apparatus of the MAIUS experiment is divided into the following subsystems:
• The laser system providing frequency stabilized laser beams tuned for laser cool-
ing, interferometry, and imaging,
• The physics package hosting the vacuum chamber in which the BEC is prepared
and the experiments are performed,
• The electronics hosting the laser drivers and controls, and the experiment com-
puter,
• The battery module providing electrical power during flight,
• The rocket including the rocket service modules and umbilical.
2.2. THE SOUNDING ROCKET MISSION MAIUS 15
Figure 2.7: Sketch of VSB-30 rocket with
MAIUS payload, modified version in [59]
Figure 2.7 depicts the subsys-
tems and the MAIUS rocket. The
scientific payload fits into a struc-
ture of 2790mm length, 500mm
diameter, and 5mm wall thickness.
The rocket engine is the VSB-30,
the same rocket engine as used in
FOKUS, see section 2.1. The envi-
ronmental stress applied on the sci-
entific payload by the launch vehi-
cle therefore is comparable.
The complete MAIUS payload
is described in the work of Seidel
[48]. Here, we only focus on the
laser system.
The laser system itself consists
of several subsystems. It contains
the laser modules, the frequency
stabilization unit, and the distribu-
tion and switching units. The sub-
systems are connected through fi-
bres and optical fibre splitters. A
schematic of the laser system is shown in figure 2.8.
The compact housing of the laser system is designed to maintain the temperature in-
side as stable as possible. On ground water cooling is provided, during flight a massive
aluminium heat sink absorbs the heat produced by the laser modules and their Peltier
elements, resulting in a rise of temperature <4 K. Each laser module is individually
temperature controlled to allow for individual WP settings and to maintain frequency
stability.
Figure 2.8: Schematic of laser system with four experiment lasers, spectroscopy mod-
ule, and switching and distribution module, as presented in [60]
16 CHAPTER 2. QUANTUM SENSORS AND LASER MODULE REQUIREMENTS
There are four laser modules: three of them are science lasers, providing the
beams for two-dimensional magneto-optical trap (2D-MOT) cooling and interferome-
try, three-dimensional magneto-optical trap (3D-MOT) cooling and detection, and the
re-pumping during the cooling process. The fourth laser is part of the frequency stabi-
lization unit, comparable to the FOKUS setup, and serves as frequency reference. Two
redundancy lasers are provided additionally. If a science laser module fails during the
testing on ground, the fibre outputs can be switched without re-assembling the laser
system. The science and redundancy laser modules are master oscillator power ampli-
fier (MOPA) modules and are the subject of this thesis. All free-space laser beams are
coupled into a polarization-maintaining fibre with a Zerodur-based fibre coupler and
are monitored with in-line photodiodes.
The frequency stabilization unit consists of a DFB optical master oscillator (MO) and
a Zerodur-based spectroscopy module. Its design is identical to that of the Rb reference
system used in FOKUS as described in section 2.1. For frequency stabilization of the
three science lasers a fraction of the light of each MOPA is superimposed with light from
the DFB master laser on a fast photodetector. The corresponding beat note signals are
then used for frequency offset-locking of each of the MOPA laser modules.
The main portion of the MOPA light is tailored to provide the various continuous
wave (CW) fields and light pulses required for laser cooling, preparation of a BEC and
implementation of an atom interferometer. The switching unit, again, is based on a
5cm thick Zerodur base plate and contains fibre coupled inputs and Zerodur based free
space optics, as described by Duncker in [43]. Several acousto-optic modulators (AOM)
and mechanical shutters shape the laser pulses and switch the light from various inputs
to various outputs. The power levels of the laser beams in the optical fibres of the
switching and distribution module are monitored with in-line photodiodes to identify
system malfunctions. Using fibre splitters, the light is then distributed to 11 polarization
maintaining fibres guiding the light to the physics package.
Although each MOPA delivers more than 1 W free-space output power the over-
all laser system provides 115mW and 110mW cooling light for the 2D-MOT and the
3D-MOT, respectively. The losses can be explained by imperfect fibre coupling and in-
line fibre optics. Still, the provided output power is sufficient for experiment operation.
2.2.4 Requirements on Laser Modules
A summary of requirements of the MAIUS laser modules is given in table 2.2, based on
the experimental needs and operation environments.
As in FOKUS, MAIUS targets the 87Rb D2 line transition for atom manipula-
tion. This defines the emission wavelength of the laser modules to be 780.241nm
(384.230484THz), as well as the FWHM linewidth of the laser module to be less than
the natural linewidth of the transition (∆ν≤6.07MHz). The applications targeted
for laser light in MAIUS are more versatile than in FOKUS. Still, as explained in sec-
tion 2.2.2, all optical frequencies required for Doppler cooling and the re-pumping are
within the range of the Doppler-broadened Rb D2-line. For the implementation of an
atom interferometer a laser linewidth well below the natural linewidth is sufficient.
Since the lasers serve multiple purposes (cooling, interferometry, detection), as shown
2.2. THE SOUNDING ROCKET MISSION MAIUS 17
requirements value comment
electro-optical
wavelength (λ) 780.241 nm 87Rb D2 line [15], as in FOKUS
FWHM linewidth ≤6MHz 87Rb D2 line [15], as in FOKUS
tunability around WP 100GHz as in FOKUS
optical power per science laser >1W (CW) complex distribution
optical power of reference laser >10mW (CW) as in FOKUS
max. power consumption per MOPA 7.2W current driver limitation
mechanical/environmental
vibration loads 8.1gRMS VSB-30 rocket, as in FOKUS
shock loads ≥250g VSB-30 rocket, as in FOKUS
WP temperature ≥35° C system design, as in FOKUS
storage temperature –30° C ... +45° C as in FOKUS
Table 2.2: Requirements on laser modules in the MAIUS laser system according to the
scientific and environmental needs
in figure 2.6, the tuning range of a single laser has to cover different wavelengths. All
wavelengths are in the range of some GHz, resulting in the demand of a multiple GHz
tuning range. Still, this is covered by the tuning range defined in section 2.1.4. With
that, there are no further requirements on the tuning range and linewidth of the laser
modules compared to the spectral requirements as defined in FOKUS.
An optical output power of 10mW for the frequency reference laser is sufficient, as
explained in section 2.1.4. According to section 2.2.3, an optical output power in the
1W range is required for the science lasers, due to the more complex laser switching
and distribution concept of the MAIUS experiment. The power consumption per laser,
however, is limited to 7.2W due to limited resources during flight and the limitations
of the current driver electronics. In order to relieve the thermal stabilization system
the working point temperature of the laser modules should match the temperature
of the environment in the payload container, which will be around 35 ° C, as in
the FOKUS mission. As for FOKUS, the rocket launch side is situated near Kiruna in
Sweden, resulting in a possible exposure to –30 ° C outside temperature in winter times.
The rocket motor of the MAIUS rocket is of the same type as the motor used for the
TEXUS rocket in the FOKUS experiments. Shock and vibration loads on the MAIUS pay-
load therefore are comparable to the FOKUS mechanical environmental requirements.
18 CHAPTER 2. QUANTUM SENSORS AND LASER MODULE REQUIREMENTS
Chapter 3
Diode Lasers
The previous chapter introduced quantum optical experiments that depend on compact
and efficient lasers. This chapter picks up the requirements of the laser modules and
presents semiconductor laser technology that is suitable for mobile high-precision op-
tical measurement devices. Semiconductor light amplification by stimulated emission
of radiation (laser) diodes are small in size, easy to use, and highly efficient. Com-
bined with other semiconductor diodes and discrete optics, powerful, frequency stable,
and frequency tunable laser modules can be created. In this chapter semiconductor
diode lasers are introduced, their functionality is explained and some configurations
for hybrid laser modules are discussed.
3.1 Concept and Functionality of Diode Lasers
Even before the concept of a photon was introduced by Einstein in 1905 [19], the con-
cept of a powerful "fiery beam" was present in fiction novels, such as in H.G. Wells "The
War of the Worlds", 1898 [61]. In 1916 Albert Einstein described the effect of stimu-
lated emission [20], a reversal to photon absorption triggered by an external photon
and resulting in the release of a copy of the trigger photon. However, only in 1958
the realization of a laser became feasible by the theoretical description of an "optical
microwave amplification by stimulated emission of radiation (MASER)" by Schawlow
and Townes [62], followed by the first demonstration of optical lasing by Maiman
1960 [1]where a ruby laser was pumped with a flash-light. Since then, many other
lasers have been developed with varying gain materials, such as gas, organic materials,
other solid crystals, or semiconductor materials, achieving a wide variety of different
wavelengths. They can be pumped with different pump mechanisms, such as coherent
or non-coherent optical waves, or electrical current.
The idea of semiconductor laser diodes was first articulated in 1953 by von Neu-
mann, although his manuscript was not published before 1987 [63, 64]. In 1962 the
first semiconductor laser, a pulsed GaAs laser at 850 nm, was operated by the group of
Hall [65]. The introduction of heterostructures [66, 67]allowed for lower threshold
current density and room temperature operation [68]. Today, diode lasers have become
a cheap and available component, used amongst others in optical storage units and for
20 CHAPTER 3. DIODE LASERS
optical data transmission, as well as in science.
Laser diodes can be designed in various ways, such as surface emitters and edge
emitters. In this thesis, only edge emitters are discussed. Figure 3.1 defines the spatial
orientation of an edge emitting laser diode as used in this thesis in order to be able to
describe wave propagation in the semiconductor material according to the optical axis
and the p-n-junction. The longitudinal propagation is defined to be in parallel with the
optical axis. A lateral wave propagates perpendicular to the optical axis, but still is in
plane with the p-n-junction, whereas vertical propagation is perpendicular to both the
optical axis and the p-n-junction plane.
In the following section, the functionality and main properties of semiconductor
lasers are discussed.
Figure 3.1: Definition of the spatial orientation of an edge emitting laser diode in this
thesis
3.1.1 Radiative Transitions and Laser Conditions
Laser devices are based on the principle of stimulated emission [20]. An electron in
an excited energy state Emrelaxes to a lower energy state or ground state Enwhen
interacting with a photon fulfilling the equation
ν=Em−En
h(3.1)
where νis the frequency of the photon and his the Planck constant. In contrast to
spontaneous emission, the photon released to overcome the energy difference between
Emand Enis coherent with the stimulating photon and cannot be distinguished from
the original one. A photon can be absorbed as well, lifting an electron from Ento Em
state.
The concept of a laser implies two conditions: first, the emission of photons has to
exceed the absorption to guarantee constant supply of photons, and second, stimulated
emission has to dominate over spontaneous emission to achieve a coherent photon
beam. The first condition can be fulfilled when the gain material is set into the state
of population inversion by a pump. For the second condition coherent photons have to
be kept inside the system while non-coherent photons have to be filtered out. This is
realized with an optical resonator that contains the gain medium [69].
3.1. CONCEPT AND FUNCTIONALITY OF DIODE LASERS 21
3.1.2 Advantages and Application of Diode Lasers
Diode lasers have a wide variety of applications. Since one semiconductor wafer can
host thousands of laser diodes, the production costs can be very low. The laser diodes
are compact in size, typically some millimetres long and a fraction of a millimetre wide.
For population inversion, only electrical pumping is required, although optical pumping
is also possible. Internal structures of the semiconductor material, as the facets or in-
ternal gratings, can form a resonator. Material combinations of gallium nitride (GaN),
gallium arsenide (GaAs), and indium phosphide (InP) allow for wavelengths ranging
from 400nm to 3500nm [70]and continuous wavelength tuning up to some nanome-
tres is possible. With that, a semiconductor laser is an easy-to-use device, suitable for
mobile applications in everyday life, such as in optical communication and laser point-
ers. Due to its simplicity, compactness, and robustness, they also lend themselves to a
wide range of applications in harsh environments such as spaceflight.
Monolithic semiconductor lasers provide an output power up to some ten Watts
CW [71]and typically about 100kHz – 10MHz FWHM linewidth [72]. Setups with
external optical feedback can reach FWHM linewidths in the range of 1kHz [73]. Still,
there are other laser concepts, such as gas lasers and solid state lasers that are superior
in terms of output power (CO2laser with 2 – 6kW [74]) or linewidth (ND:YAG laser
with sub Hertz linewidth [75]) which are also used in space [76,77]. These systems
are limited in wavelength availability, and they are more complex, bigger, less robust in
terms of mechanical stability, and less energy efficient than semiconductor based laser
modules. Semiconductor lasers therefore are a future alternative for applications with
demands on specific wavelengths on ground and in space.
3.1.3 Basic Properties of Semiconductor Lasers
Semiconductors are solid materials that allow for the movement of charge carriers due
to the material composition and doping. Typically, these materials are compound III-V
materials as GaAs or InP and provide a p-n-junction, where injected carriers combine
permanently. This region of permanent recombining is called active region. Electron -
hole recombination can be illustrated with the energy band model [78]. Energy bands
are discrete energy levels that an electron is allowed to have, whereas band gaps are
ranges of energies that are forbidden by the laws of quantum mechanics. At absolute
zero temperature the valence band is the highest energy level occupied by electrons,
while the conduction band is the lowest unoccupied energy level. In semiconductor
materials these energy bands are separated by the band gap energy Eg, the minimal
difference between valence and conduction band, as depicted in blue in figure 3.2.
The Fermi level, defining the probability of occupation to be 50%, is in the middle of
the band gap. Egis defined by the semiconductor material itself.
For improving recombination efficiency, vertical carrier confinement, as in a dou-
ble heterostructure, is common. In a double heterostructure a low band gap material
is sandwiched between higher band gap materials, as shown on top of figure 3.2b,
in order to form a carrier trap. Differences in the refraction index of the sandwiched
material also result in a vertical optical waveguide, allowing for fundamental mode
22 CHAPTER 3. DIODE LASERS
enforcement in vertical direction, like in the red curve in figure 3.2b. The threshold
current decreases with the thickness of the active area. Modern epitaxy allows for an
active area thickness in the range of the electron’s De Broglie wavelength. In this case a
quantum well region is generated increasing the internal efficiency of the laser. Lateral
carrier confinement can be achieved with a ridge-waveguide structure (index guiding),
changing the refraction index profile in lateral direction, and restricted electrical pump-
ing of the waveguide region (gain guiding).
Figure 3.2: Emitting Junctions a) Homojunction. b) Double-heterojunction. In the case
of a heterojunction, a better confinement of the photon carriers can be obtained. [79],
licence: CC-PD
In longitudinal direction, optical feedback is essential to support stimulated emis-
sion. For some types of diode lasers the cleaved facets of the semiconductor material
form the resonator mirrors so that the coating of the facets influences the laser proper-
ties. Refraction coatings at front and rear facet allow for controlling the ratio of optical
output power emitted on both facets. In addition, they also increase the resistance to
facet damages. Commonly a low refraction coating at the front facet and high reflec-
tion (HR) coating at the rear facet are used to concentrate the optical output on one
side of the diode.
The optical feedback in the active region forms a Fabry-Pérot resonator, typically
supporting several longitudinal modes due to the macroscopic size of the laser diode.
Often, single mode operation is aspired and a wavelength selective element, as will be
described in section 3.1.4, is required.
Lattice mismatches between active region and waveguide or cladding lasers can
also be used for adjusting the emission wavelength. The emission wavelength also de-
pends on the temperature of the device. With an increase of the temperature T, both
the resonator length L and group refraction index nR,g1increase, resulting in a higher
wavelength, as the modal emission wavelength is proportional to 2nr,gL[80]. Also the
band gap Egdecreases with higher temperature, and, with respect to equation 3.1, re-
1relation between vacuum velocity c0and velocity of the wave packet cG(λ)(all emitted modes)
3.1. CONCEPT AND FUNCTIONALITY OF DIODE LASERS 23
sults in a higher emission wavelength [80]. Fine tuning of the emission wavelength is
commonly realized by tuning of the injection current. The injection current defines the
carrier density in the active region. A change of the carrier density results in a change of
the refraction index nR,gand the temperature in the active region, which is followed by
the change of the emission wavelength as described above. In addition, an increasing
number of carries results in an increasing number of photons in the resonator which is
equivalent to the optical output power.
These effects also can be used for modulating the frequency and amplitude of the elec-
trical field of the output beam. The following sections introduce the mathematics be-
hind semiconductor laser emissions.
3.1.3.1 Mathematical Description
The behaviour and interaction of electrons and photons in a laser diode can be de-
scribed with rate equations. Common are the rate equations for carrier density ncand
for the photon number S, as described by Petermann [81]. The change of carrier den-
sity per unit time is given by the difference between the injection rate of electrons Ri
and the electron recombination rate Rrec per unit volume in the active region:
dnc
dt =Ri−Rrec (3.2)
Only a fraction of the injected current Icontributes to carrier recombination in
the active region, thus, ηireflects this injection efficiency. We can state for Rithat there
are ηiI/qelectrons per second injected in the active region with a volume V, where q
is the elementary electrical charge [82].
Rrec is a combination of the spontaneous recombination rate Rsp, the non-radiative
recombination rate Rnr, the leakage rate Rl, and the stimulated recombination rate Rst .
Rsp,Rnr , and Rlcan be summarized as natural carrier decay. They are defined as num-
ber of carriers Nper carrier lifetime τ.
The rate of stimulated emission Rst describes the electron-hole recombination that is
stimulated by photons, generating even more photons. The number of photons Sin-
creases by travelling across a pumped gain medium with the length ∆z. For sufficient
small ∆zthis can be expressed with
S+∆S=Seg∆z≈S(1+g∆z)(3.3)
where gis the gain of the increasing photons. Taking into account that the group
velocity of the photons vgdepends on ∆zby vg=∆z/∆t, the rate of increasing photon
number Rst =∆S/∆tcan be defined as number of photons Smultiplied with the gain
gand the group velocity vgof the photons in the gain medium [82]. Combined, that
results in a rate equation for electrical carrier density as follows:
dnc
dt =Ri−Rsp −Rnr −Rl−Rst (3.4)
=ηiI
qV −N
τ−Sgvg(3.5)
24 CHAPTER 3. DIODE LASERS
The rate equation of photon density Sis also dependent on the spontaneous and
stimulated emission rate, and introduces the spontaneous emission factor βsp, as well
as the confinement factor Γ=V/Vp, which is the ratio between the volume of the active
area and the volume of the optical mode [82]:
dS
dt =ΓβspRsp +ΓRst −S
τp
(3.6)
=ΓβspRsp +ΓvggS −S
τp
(3.7)
The photon density increases with stimulated and spontaneous emission, although
just a small fraction of spontaneous emitted photons contribute to the lasing mode, as
expressed by the spontaneous emission factor βsp which is in the order of 10−5... 10−4
[82]. The photon density in the resonator is reduced by photons leaving the resonator
or being absorbed internally. τpis the average lifetime of a photon in the resonator.
Comparing both rate equations from equation 3.5 and 3.7 it is shown that stimulated
emission reduces the number of carriers and increases the number of photons.
3.1.3.2 Output Power and Efficiency
The steady state solution of equation 3.5 above threshold (g=gthr ) can be written as
S=ηi(I−Ithr)
qνggthr V(3.8)
As a steady state solution to equation 3.7 and for quite small values for βsp we find the
threshold modal gain Γgthr to be
Γgthr =1
νgτp
(3.9)
With the optical energy been stored in the cavity, EMode equals Stimes the photon
energy hνtimes the volume of the optical mode Vp. The optical power can be calculated
by multiplying EMode with the energy loss rate through the cavity facets (mirrors). The
mirror loss rate can be expressed by 1/τm[82].
Po=EMode
τm
=hνVpS
τm
(3.10)
=hν
qηd(I−Ithr)for I<Ithr . (3.11)
Ithr is the threshold current, and ηdthe differential quantum efficiency, defined by the
number of photons emitted per injected electron [82]that takes the injection efficiency
ηi, the lifetime of a photon τp, and mirror losses 1/τminto account.
Below threshold current, the laser just emits spontaneously, like a light emitting
diode (LED). For an injection current I>Ithr the carrier density ncis basically constant
3.1. CONCEPT AND FUNCTIONALITY OF DIODE LASERS 25
[83], resulting in an increasing number of photons Sand a linear relation between
injection current and output power.
The dependency of the threshold current on the temperature can be expressed with
[82]
Ithr ∝eT/T0(3.12)
T0is the characteristic temperature, given by the material system. T0depends on the
barrier height, defined by the wavelength of the diode, and the necessary threshold
current density. For GaAs diodes with wavelength between 760...800nm, the value of
T0is around 100...140 K [71].
The conversion efficiency ηof a laser can be described with the ratio between optical
output power Poand the electrical input power Pin, whereas the difference between Pin
and Pois the dissipated power Pd:
Pd=Pin −Po=Pin(1−η)(3.13)
= (I2Rs+IVd+IVs)(1−η)(3.14)
As reflected in equation 3.14, the input power Pin can be expressed depending on the
serial resistance Rsof the laser diode, the current-independent serial voltage Vsof the
laser diode, and the ideal diode voltage Vd[82].
The dissipation power Pd, in turn, depends on the thermal resistance Rth of the
diode, causing a temperature rise ∆Tin the laser [82]. A simplified description of the
thermal resistance takes the thermal conductivity ρth, the footprint of the laser diode
ALD, and the distance of the heat source (active region) to the heat sink Hver t into
account.
Rth =∆T
Pd
(3.15)
=Hvert
ρth ALD
(3.16)
With an expression for the thermal resistance, the temperature in the active region
can be described with
T(t) = ∆T+TW P =Rth ·Pd+TW P (3.17)
where TW P is the temperature to which the laser is exposed externally.
With rising temperature, the optical output power decreases beyond a certain
threshold. This effect is called thermal roll-over [71]. The optical power, as given
in equation 3.11, can be complemented with its temperature dependency:
Po=hν
qηde−∆T
T1[I−Ithr e−∆T
T0](3.18)
T0and T1are characteristic temperatures for the device. T0was already introduced
in equation 3.12. T1typically is three to five times the value of T0. The parameter
set (T0,T1,Ithr ,ηd) is defined for operation at room temperature. As shown in equa-
tion 3.17 and 3.18, higher operation temperatures cause a higher thermal resistance,
26 CHAPTER 3. DIODE LASERS
which limits the optical output power and increases strongly the temperature in the ac-
tive region [71]. This underlines the necessity of a good thermal conduction of the high
power diodes to ensure low operation temperatures in order to maximize the optical
output power.
3.1.3.3 Modulation Behaviour
The stability of a laser diode in terms of optical power and frequency can be influenced
by the injection current and the temperature in the environment. Direct modulation
of the diode laser is a technique not only used in optical communications but also
relevant for frequency stabilization of a laser. In direct modulation techniques, the
injection current of the laser diode is modulated, resulting in variation of the carrier
density, which causes a modulation of the refractive index, of the temperature in the
active region, and of the number of photons in the resonator [84]. A modulation of
the photon number Sin the resonator influences the optical output power directly,
as shown in equation 3.10. Hence, the modulation of the injection current causes a
modulation of the output power, and the frequency and phase of the optical field.
The modulation of the injection current also affects the temperature of the active
region. We assume the quasi-stationary case
I(t) = Is+Im,U(t) = Us+Um→T(t) = Ts+Tm(3.19)
where the subscript "s" symbolizes the static bias value whereas the subscript "m"
indicates the modulation variation. According to equation 3.13 and 3.17 and the quasi-
stationary assumption that Um=ImRs, in the first order of current modulation, the
temperature modulation can be described by
Tm≈Zt·Im·(I0Rs+U0)·(1−η)(3.20)
where Ztis the thermal impedance of the laser diode, Rsis the serial resistance of
the laser diode, U0is the bias voltage, and ηis the conversion efficiency of electrical
power into optical power.
Current modulation is directly proportional to temperature changes. The change in
temperature changes both, the resonator length and the refraction index as described
in [85]:
Lm∝L·Tm(3.21)
nr,m∝nr·Tm(3.22)
With rising temperature, the resonator length and the refraction index increase, and
hence the emission frequency decreases, since
fm∝c
2nrL∝−Tm. (3.23)
In summary, it can be stated that the modulation of the injection current causes a
modulation of the emission frequency due to modulation of the temperature and of the
carrier density, as well as a modulation of the output power.
3.1. CONCEPT AND FUNCTIONALITY OF DIODE LASERS 27
However, modulating the injection current Iand thereby inferring a modulation
of the carrier density nccauses a modulation of the refraction index nrin the active
region. The linewidth enhancement factor (or Henry factor) αdescribes the inverse
proportionality between the change in carrier density ncand the change in the refrac-
tion index nr:
α=−4π
λa
dnr
dnc
(3.24)
Therefore, one has to differentiate between the thermally induced and the carrier
density induced modulation on the emission frequency. Thermal effects can only be
observed at sufficiently low modulation frequencies. According to Kobayashi [85]max-
imum modulation frequencies at which thermal effects dominate are approximately at
10MHz, due to the slow thermal reaction time of the system. For larger modulation
frequencies the carrier density influence dominates.
3.1.3.4 Linewidth
The intrinsic spectral stability is limited by the emission of incoherent photons, gen-
erated by spontaneous emission, which add phase and power noise to the coherent
part of the optical field. This effect was first described by Schawlow and Towns in
1958 [62]and extended by Petermann [86]and Henry [87]to the description of diode
lasers. According to them, the linewidth is proportional to the spontaneous emission
enhancement factor K, to (1+α2)(αis the Henry factor, see equation 3.24), the thresh-
old gain gthr , and carrier losses αm, as well as the inverse optical output power Po:
∆νsp =
v2
ghνβsp
8π
gthrαm
Po
K(1+α2)(3.25)
vgis the group velocity, hνis the photon energy, and βsp is the spontaneous emission
factor. αmcan be estimated with [81]
αm=1
2Lln1
RfRr(3.26)
As seen in the equations above 3.25 and 3.26, the linewidth can be decreased by im-
proving the quality of the resonator, e.g. by increasing the resonator length L(an
alternative argumentation is delivered by Kojima [88]), the reflectivity of the front and
rear facet Rfand Rrand by increasing the output power. Increasing the reflectivity of
both facets would reduce the output power, since less photons are able to leave the
cavity. Increasing the cavity length does decrease the linewidth proportionally but this
comes at the cost of lower efficiency due to decreased carrier density in the cavity [89]
for constant injection currents. For applications which demand for excellent spectral
stability, laser concepts with an extended or external resonator, e.g. an extended cavity
diode laser (ECDL), as described in [90]and [73], can be applied. Still, ECDL setups
are more complex and the spectral tuning range typically is limited to the free spectral
range (FSR).
28 CHAPTER 3. DIODE LASERS
3.1.4 Example: Distributed Feedback lasers
For many applications, such as laser cooling or atom interferometry, single frequency
emission of lasers is mandatory and a mode selection of longitudinal modes is required.
In this work, we focus on distributed feedback (DFB) diode lasers, a widely used laser
type, first realized by Kogelnik and Shank in 1971 [91]. In DFB lasers, a longitudinal
mode-selective grating is grown along the entire length of the active region, typically
implemented in the epitaxial structure of the diode. That results in a periodic modula-
tion of the effective refraction index in the active region allowing only the propagation
of two modes adjacent to the Bragg wavelength. The Bragg condition itself forms a
stop band where light with this wavelength cannot propagate in the active region.
Distinctive irregularities in the grating and unsymmetrical reflectivity at the facets
of the diode are used to prefer only one of the two modes located next to the stop band
in the laser. Since the integrated grating behaves like a distributed reflector, discrete
mirrors at the facets are not required, the front facet can be anti-reflection (AR)
coated. However, the rear facet is commonly high-reflection coated to support only
one mode and to direct the output power mainly to the front facet.
The DFB diodes, used in this thesis, are optimized for low linewidth emission and
high output power. This is done by optimizing the epitaxial structure of the device, as
well as the lateral confinement of the ridge waveguide [92].
3.1.5 Example: Power Amplifier Chips
There are applications which require a CW optical output power in the range of 1 W
or more, for example in inter-satellite free space communication [5]or in laser cooling
[56], as described in chapters 1 and 2. In order to provide high optical output power
in combination with spectrally stable single mode emission, light has to be generated
by means of a master-oscillator-power-amplifier concept (MOPA). The output power
of the master oscillator (MO), which is optimized for spectral stability rather than for
high output power, is amplified by a power amplifier (PA) without the loss of spectral
stability. The spectral behaviour of the original MO beam shall not be influenced by the
amplification.
This demand can be satisfied with a semiconductor optical amplifier (SOA) [93], a
semiconductor chip having a similar structure as a Fabry-Pérot laser. The population
inversion is, again, created with electrical pumping, but both facets of a SOA are AR
coated so they do not serve as resonator. The MO laser beam is coupled into the rear
SOA facet and triggers stimulated emission while travelling through the gain material
of the SOA. By doing so, the output power of the beam increases by a gain factor of
gusat . However, with increasing input power of the signal beam, the gain amplification
decreases. This effect is called gain saturation and is caused by depletion of the carriers
in the active region at the signal’s saturation power. The gain of a saturated amplifier
can be described with
gamp =gusat
(1+Po
Po,sat )(3.27)
3.1. CONCEPT AND FUNCTIONALITY OF DIODE LASERS 29
with gusat as the unsaturated small signal gain, Poas the output power of the amplifier
and Po,sat as the saturation power. The saturation power of the output beam is defined
as output power where the gain of the amplifier has reduced by 3dB compared to small
signal gain [93]and is marked in figure 3.3 with Po,sat .
Figure 3.3: Example of amplifier gain vs. output power
The saturation power Po,sat is limiting the output power of the amplifier and is given
by [93]
Po,sat =A
Γ
hν
τ
(nc−nc,0)
gm
(3.28)
with A/Γdenoting the amplifier mode cross-section area, with hνdescribing the photon
energy, τgiving the lifetime of the carriers, the material gain coefficient gm, and the
difference of the carrier density ncand the transparency carrier density nc,0.
Since the saturation power is limiting the output power, a large saturation power
is desirable. According to equation 3.28 the saturation power can only be increased
by increasing A/Γat a given wavelength and material system, since the photon energy
hνincludes the wavelength, and the gain coefficient gm, and carrier densities ncand
nc,0 depend on the material system of the laser diode. τsaturates for high injection
currents. Only A/Γis given by the design of the active region of the amplifier. An
approach to increase the saturation power by design of the amplifier will be described
in paragraph 3.1.5.2.
To every output saturation power Po,sat an input saturation power Pi,sat can be as-
signed. For a signal input Pin >Pi,sat the amplifier is saturated, the gain of the amplifier
is reduced. Pi,sat is dependent on the coupling efficiency of the input power into the
amplifier and can be influenced to a certain degree by the design of the laser module.
Operation of the amplifier in saturation reduces the rate of spontaneous emission
and thereby reduces amplified spontaneous emission (ASE) noise. In addition, the satu-
ration of gain causes the suppression of small fluctuations in input signals power, which
improves the power stability. However, this is a disadvantage when the amplitude of
30 CHAPTER 3. DIODE LASERS
the master oscillator is modulated on purpose to implement a power modulation. The
amplitude modulation will be suppressed as well.
At large population inversion the emission of spontaneous photons, the ASE, in-
creases as well. The incoherent and non-polarized emission is amplified and adds noise
to the power and phase of the optical field. This limits the number of amplifiers that
can be cascaded, e.g. in optical communication applications. In high power operation,
high power density, both in the gain medium and at the facets, can exceed the damage
threshold of the SOA [94], causing catastrophic optical damage (COD). With passiva-
tion of facets [95]these damages are less probable nowadays. However, the optical
output power of the SOA is limited by design as shown in equation 3.28.
As described in section 3.1.3.2 and shown in equation 3.18, the output power
decreases with increasing temperature in the active region. Due to high SOA injection
currents, which contribute quadratically to the dissipated power over an ohmic resistor
as shown in equation 3.14, thermal roll-overs [93]often can be observed when
working with SOAs. Sufficient thermal conduction of the diode therefore is necessary
to dissipate the accumulated heat in order to defer the thermal roll-over effect.
In the following sections two designs of a SOA are discussed, addressing both the
question of output power and of beam quality.
3.1.5.1 Ridge Waveguide Amplifiers
The ridge waveguide (RW) amplifier is a typically 4mm to 6mm long double het-
erostructure chip with a mode selecting grating and with AR coated front and rear
facets. The waveguide can be tilted [96]or partially bent [97]to suppress the effect of
facet reflection and prevent the chip from starting to act like a laser itself.
Due to the strict confinement of the optical wave in the chip, the output beam char-
acteristics are expected to be nearly Gaussian. This is an advantage, when the output
beam has to be coupled into a fibre. With a Gaussian beam profile and compatible beam
widths most of the beam power can be coupled into the fibre, and thus result in a good
coupling efficiency. Thermal exposure of the coupling components, which is caused by
non-coupled beam power, is reduced the better the beam profile of the laser matches
the in-coupling interface of the fibre. The good beam parameters are achieved with the
mode confining by the RW, which is typically only few micrometres wide. However,
this comes at the cost of limited saturation power Po,sat , as the size of the active area is
reduced, and with that A/Γ, as denoted in equation 3.28. The achieved output power
of a RW amplifier therefore is reduced as well.
3.1.5.2 Tapered Amplifiers
Expanding the gain region towards the output facet as described in [98]is a common
approach to increase the saturation power Po,sat of an SOA as it increases A/Γ, shown
in equation 3.28. Figure 3.4b illustrates a tapered amplifier with a RW pre-amplifier
on the right side of the optics-block. The RW shaped pre-amplifier can be included
3.2. MONOLITHIC AND HYBRID MOPAS 31
for filtering the incoming modes, providing a Gaussian input for the tapered amplifier
section and thereby improving the beam quality of the amplifier.
Still, a tapered gain section results in the deterioration of carrier confinements and
potentially allows for other modes than the fundamental mode, because the lateral
carrier confinement decreases with increasing width of the active area. That results
in a decreased beam quality, and decreases the fibre coupling efficiency. In addition,
the difference in divergence of the output beam between the fast axis and slow axis
increases due to the elliptical shape of the active area at the output facet, which makes
beam collimation more complex.
3.2 Monolithic and Hybrid MOPAs
(a) Monolithic MOPA (b) Hybrid MOPA
Figure 3.4: Concept of monolithic and hybrid semiconductor laser MOPA
Compact design is a key benefit of diode lasers and qualifies them for mobile ap-
plications. As discussed in section 3.1 single mode, narrow linewidth lasers can be
processed and diode lasers can achieve an output power larger than one watt CW, ful-
filling the requirements listed in table 2.2. Combining a mode selective section with a
gain section on one chip, as shown by Fiebig in [99]and explained in [100], will reduce
complexity of the system and decrease the production costs radically. This approach is
called monolithic integration and is depicted in figure 3.4a. However, this system suf-
fers from poor spectral stability due to reduced lateral carrier confinement in the gain
section and internal feedback, which disturbs the mode selection of the chip. Although
diodes with high CW output power of 1 W and a narrow linewidth of 1.4 MHz were
reported in [101], the continuous tunability is reduced by spectral unstable behaviour
for low injection currents.
An alternative approach to monolithic integration is the hybrid integration of laser
diodes, as shown in figure 3.4b. A master oscillator (MO) is decoupled from the ampli-
fying section, optical feedback from facets or internal section transition is prevented by
an external optical isolator. The laser beam emitted by the MO is coupled into the ac-
tive region of a semiconductor amplifier, keeping the spectral behaviour but increasing
the optical output power. This assembly increases the complexity and costs of the laser
module. However, it preserves the spectral properties of the MO, providing a laser sys-
tem with high output power, narrow linewidth, and large continuous frequency tuning
range.
For that reason, we chose a hybrid MOPA concept for our laser sources.
32 CHAPTER 3. DIODE LASERS
Chapter 4
Concept and Assembly of
Micro-integrated Laser Modules
In order to obtain laser modules with an optical output power larger than 1W in com-
bination with an emission frequency stability in the lower MHz range, as demanded in
table 2.2, we combine the spectral properties of DFB laser diodes with semiconductor
power amplifiers. In addition to the electro-optical properties, the mechanical stabil-
ity and compactness of the laser modules are important to guarantee operation of the
laser module in the harsh experiment environment. In the first section of this chapter
the concept and design of the laser modules, which addresses all requirements given
above, will be introduced. The second section tackles the demands for reproducibil-
ity and documentation of the assembly process by explaining the overall processes in
general and the process of integration of optics in the laser modules in particular.
4.1 Concept and Design of the Hybrid Integrated Laser
Modules
This section introduces the design of the laser module that shall be used in the FOKUS
and MAIUS missions. In the design two aspects were taken into account, the functional
aspect, reflected in the electro-optical design, and the physical aspect, presenting the
mechanical structure and interfaces of the laser module.
The electro-optical design discusses the choice and arrangement of the semiconduc-
tor diodes and optical components, such as optical isolator and lenses. This design is
chosen to fulfil the electro-optical requirements of the laser modules.
The mechanical structure satisfies the mechanical and environmental requirements
of the laser modules, hosts the optics and provides the electrical and mechanical in-
terfaces. A fully integrated laser module is shown in figure 4.1. The paper clip next
to the laser module gives an idea of the size of the module, which has a footprint of
80x25mm2.
34 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
Figure 4.1: Hybrid integrated MOPA module on AlN MIOB with integrated electrical
interface. A modified version published in [102]
4.1.1 Electro-Optical Design of the Laser Modules
Laser modules with a separate optical master oscillator (MO) and optical power ampli-
fier (PA) are called MOPAs. The optical design of the MOPA module is shown schemat-
ically in figure 4.2, featuring, from left to right, an optical isolator, collimation lenses
for the rear output of the MO, the MO-DFB diode itself, collimation lenses for the front
output of the MO, a second optical isolator, coupling lenses for the rear input of the PA,
the PA itself, and collimation lenses for the PA front output. The upper part of figure 4.2
depicts the lateral view of this ensemble, meaning the view from above. The vertical
view in the lower part of the figure shows the projection from the side of the optical
ensemble.
As described in section 3.2, the hybrid integration approach allows the integration
of optical isolators, protecting the DFB laser from optical feedback and thus from distor-
tion of the propagating wave. In order to guide the beam through the optical isolator,
the beam has to be collimated to be able to propagate through the aperture of the isola-
tor device. To couple the beam into the amplifier chip the beam has to be focused onto
the active area of the chip, meeting the divergence properties of the chip. The output of
the amplifier is collimated to compensate for the PA chip’s divergence and prepare the
output beam for further usage. The beam collimation and isolation at the rear output
of the DFB laser make an additional output available that can be used for monitoring of
the condition of the DFB laser, or as an additional light source with low output power.
The beam guiding, such as collimation and focussing, is done by the miniaturized glass
lenses. The choice of the lens parameters, such as the numerical aperture and the focus
length, depends on the aimed beam diameter and the divergence angles of the diodes.
In the following section the laser and amplifier diodes are introduced, and the cho-
4.1. DESIGN OF THE HYBRID INTEGRATED LASER MODULES 35
Figure 4.2: Lateral and vertical view of optical configuration on MOPA
sen lenses and optical isolators are presented.
4.1.1.1 Laser Diodes
In this work, we use DFB diodes optimized for narrow linewidth emission at 767nm and
780nm, designed and processed at the "Ferdinand-Braun-Institut, Leibniz-Institut für
Höchstfrequenztechnik" (FBH). Emission wavelengths below 870 nm are demanding
for the designing and the epitaxial processes of GaAs based semiconductor lasers. The
high aluminium (Al) content in the cladding results in oxygen contamination during the
growing steps of the grating, as described in [92]. In addition, the material composition
and thickness of the grating layer has to be adapted to obtain non-absorbing gratings
for wavelengths below 870nm.
MO diodes with an optimized grating are fabricated as DFB diodes with a RW for
better lateral wave confinement. DFB diodes with two different lengths, 1.5 mm and
3mm, were tested. The 1.5 mm long diodes fulfil the linewidth requirements for laser
cooling. 3mm diodes, however, can achieve an even narrower linewidth [89]at the
cost of higher threshold currents and decreased efficiency.
The facets of the DFB lasers are coated with AR coating on the front facet and 95 %
reflection coating on the rear facet. That results in a higher optical output power at the
front facet compared to the rear facet, because 95 % of the photons are reflected at the
rear facet. The asymmetrical coating of the facet also supports the mode selection of
one of the two competing DFB modes.
In addition to the DFB diodes amplifier chips for 767 nm and 780 nm with a RW pre-
amplifier section, serving as mode filter, and a tapered section of an optical amplifier
(TA) for power boost, are fabricated. Both facets of the amplifier chip are AR coated,
because photons are supposed to pass the amplifier and shall not be reflected at any
facet.
In order to be able to perform an optical simulation of the MOPA design it is neces-
sary to know the divergence angle and the beam waist at the chips facet for both DFB
lasers and PAs. The simulation verifies the design and choice of lenses, as described in
36 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
the next section.
The beam propagation properties of the diode lasers used in this work are listed in
table 4.1. With these properties it is possible to estimate the lens parameters, required
to manipulate the beam propagation in a hybrid laser module, as shown in figure 4.2
and described in the following section.
Table A.1, located in in appendix A, lists the laser and amplifier diodes integrated
into laser modules and used in this thesis.
Type Θver tical Θhorizontal d0,ver tical d0,horizontal
front/rear front/rear front/rear front/rear
DFB 34.8° /34.8° 17.6 °/17.6° 1.6 µm/1.6µm 3.2 µm/3.2µm
RW-TPA 38.8 °/40.8 ° 13.0 °/23.2 ° 1.6µm/1.4 µm 4.4 µm/2.5µm
Table 4.1: List of beam propagation properties of DFB lasers and amplifier diodes used
in FOKUS and MAIUS modules. Values behind "/" refer to rear facet properties.
4.1.1.2 Optical Components and Design
In hybrid laser modules, beam guiding optics are required to couple the beam, which
is emitted by the MO, into the amplifier chip, and to collimate the beam outputs. The
beam propagation, illustrated in figure 4.2, was simulated beforehand with the soft-
ware WinABCD [103]. The simulation helps to identify the lens types and lens proper-
ties that qualify for micro-integration and to find the approximate positions of the lenses
along the optical axis within the spatial limitations of the MIOB. The MIOB, described
in the next section, provides a 2.2 mm wide and a maximal 30 mm long channel for the
placement of the optical components. This limits the size of the lenses and the size of
the beam. Other frame conditions for the simulation are the laser diode divergences,
as listed in table 4.1, and the aimed beam diameter of 0.6mm providing a power con-
tent of 95%, resulting in a free aperture of the lenses of >1.2 mm. A free aperture
of more than the double of the beam diameter is recommended to avoid clipping and
deformation of the beam at the edges of the optics.
The simulation revealed that cylindrical lenses are suitable to compensate of the
different horizontal and vertical divergence angles of the semiconductor diodes, when
aiming for a round beam shape, because the horizontal axis and the vertical axis can
be collimated separately. Another advantage of cylindrical lenses, compared to round
lenses, is, that they provide a degree of freedom (DOF) in positioning of the lens. The
vertical position of a lens forming the horizontal beam propagation can be chosen freely
as long as the beam is within the free aperture of the lens. This DOF can be used for
mounting of the lenses in the MIOB. Shrinkage in the direction of the DOF of the
adhesive that is used to fasten the lens into the MIOB will not affect the beam shape
as long as the lens is not tilted. Since there are no adhesives with zero percent of
shrinkage, this aspect has to be taken into account in the design as well.
In this work the beam guiding is accomplished by cylindrical plano-convex micro-
lenses, provided by the manufacturer Ingeneric. The lenses feature a numerical aper-
ture of 0.8 and an aspheric design to avoid optical aberrations. The front and the rear
4.1. DESIGN OF THE HYBRID INTEGRATED LASER MODULES 37
facet of the lenses are AR coated for a wavelength of 780nm. The fast axis collima-
tor (FAC) takes over beam shaping in guiding in the vertical plane, while the slow axis
collimator (SAC) is guiding the horizontal plane of the beam. To shape the beam, there
is always a combination of FAC and SAC required. The output of the amplifier is col-
limated with a three lens system, including a FAC and an SAC-SAC-telescope in order
to form a round, non-astigmatic beam. The rear output and guiding within the MOPA
is realized with a FAC-SAC pair. Figure 4.3 shows a detailed picture of the MO main
beam collimation and isolator. In this picture, the rear output is not collimated.
The simulation shows a vertical beam diameter of 0.583 mm behind the DFB front
output, formed with the FAC lens FAC-08-900. However, for the slow axis collimation
of the MO, the nominal beam diameter could not be realized with the available lenses.
Since both the FAC and SAC lens have a thickness of 1.5mm, the minimal focal length
of the SAC should be >2mm, since the two lenses cannot overlap physically. A fo-
cal length >2mm results in a beam diameter of 0.8 mm. The horizontal diameter of
0.814mm was generated with the CLY-PL-CX-2.67 lens with f=2.6mm and coupled
into the amplifier with an ACYL-F2.1 lens with a f=2.1mm. For output collimation
we again chose the FAC-08-900 for FAC collimation and pick a telescope configuration
with ACYL-F2.1 and ACYL-F2.5 lenses in order to form a symmetric, round beam with a
simulated diameter of 0.628mm. The beam diverges to 0.8 times 0.83mm in 300mm
distance to the amplifier.
Figure 4.3: Detailed picture of a MOPA: DFB on submount with temperature sensor,
coax connector for temperature sensor, collimating lenses and optical micro-isolator
The optical isolators are placed at the rear and at the front output of the DFB diode,
as illustrated in figure 4.2. In the front output path, we use a semi-double stage micro-
isolator, type I-78-LM-SD-1.4-4, provided by Isowave. It has a specified isolation of
>55dB and a specified insertion loss of <7dB. Pre-integration characterization of the
isolators confirmed these specified values, revealing an average isolation of 56.8 dB and
an average insertion loss of 6.3dB. The components inside the isolators are tilted by
38 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
4°to prevent back reflection of the beam. The clear aperture is 1.4mm, fitting twice
the size of the nominal beam diameter.
Due to high insertion losses of the micro-isolators and low output power available
at the rear output of the MO, a single stage isolator I-780-MM-1.4-4-WP-0 provided by
Isowave with only <4dB specified insertion loss but also only >35dB isolation was
chosen for the auxiliary beam output.
An issue with the optical micro-isolators is that the power which can pass the isolator
without thermal degradation of the isolation is limited due to heating of the thin film
components inside the isolator. The manufacturer recommends a maximal power of
<50mW per 1mm2aperture per isolation stage, allowing for only 30 mW input power
with a beam diameter of 0.6mm (and 57mW with d=0.8mm) in the isolator. Our DFB
diodes provide an output power up to 120 mW. The optical linewidth decreases with
high output power and the tuning range increases. Since we are aiming for a large
tuning range and narrow linewidth emission a limitation of the injection current of our
DFB lasers in order to meet the recommended maximal optical power input into the
isolators would restrict the performance of our laser system. However, the experimental
verification of the isolation with an input power in the range of 100 mW has shown that
the isolators maintain their performance beyond the specified input power range.
4.1.2 Structural Design of the Laser Modules
The optics described above are mounted on a ceramic micro-optical bench which is
clamped onto a copper (Cu) mount adapter that provides mounting holes for system
integration. A MIOB with a mount adapter is shown in figure 4.4. Both MIOB and Cu
mount adapter are described in this section.
Figure 4.4: MIOB with integrated DFB and fibre coupling on a Cu mount adapter (Im-
age courtesy of V. Schkolnik [104])
4.1. DESIGN OF THE HYBRID INTEGRATED LASER MODULES 39
4.1.2.1 Micro-optical ceramic bench and Submounts
In order to design a laser module that fulfils the mechanical requirements, listed in
table 2.2, a structure is chosen that hosts the electro-optical and optical components
by omitting movable parts and being small in size. A good thermal conductivity of
the structural material is required for thermal stabilization of the laser diodes and to
dissipate heat generated by the thermal resistance of the diodes, as explained in sec-
tion 3.1.3.2.
High stiffness is very important for the structure hosting the optics and the laser
diodes. A deformation of the MIOB may result in a degradation of the MO-to-PA cou-
pling. This in turn may cause a reduction of the output power, excess ASE or a degra-
dation of the PA output beam quality. Further, a potential deformation will deflect
the PA output beam which may result in a degradation of the fibre coupling efficiency
(see section 4.1.3.2). All the effects mentioned above reduce the available optical out-
put power and potentially degrade the spectral performance of the laser module. A
deformation-induced variation of the output power could be observed with previous
modules described in [105], when applying asymmetric pressure on the only 1 mm
thick AlN MIOB.
To overcome the issues mentioned above, the MIOB, designed for this work, is based
on a 4.5mm thick AlN base plate which is 25 mm wide and 80 mm long. AlN was cho-
sen for its good thermal conductivity of close to 200W/mK and its high stiffness with
a Young modulus of >300GPa. On top of this AlN base plate several functional AlN
ceramic plates are mounted, proving frames for isolators, mount structures for lenses,
and the electrical interface. The ceramic plates are soldered on top of each other in
sequenced processing steps. Some layers, especially the ones hosting the electrical
interfaces, are lithographically structured in thin film processes, allowing for the im-
plementation of printed circuit board (PCB)-like electrical networks and the mounting
of discrete electrical components directly on board the MIOB ceramic.
The laser and amplifier diodes are mounted on AlN submounts, which also are
lithographically structured so they can host a temperature sensor (see section 4.1.3.1)
and allow for separate control of different sections of the diodes. In addition, the
submounts come with a layer of gold tin (AuSn) solder to support p-down mounting
by minimizing the risk of shortening p- and n-contact of the flipped diode by seeping out
of surplus solder. DFB diodes are mounted with the p-side pointing up, as illustrated
on the left in figure 4.5, in order to reduce mechanical stress close to the active region
that could disturb the stable single mode propagation and may reduce the polarization
purity. The amplifiers, in contrast, are mounted with the p-site facing the submount,
also called p-down mounting and illustrated on the right in figure 4.5, to reduce the
thermal resistance between the active volume and the heat sink and with that to allow
for maximum heat extraction.
There are individual submounts for each chip length allowing to interface both main
and rear output of a diode. Submounts for p-up mounting have a height of 1.0mm.
P-down submounts have a height of 1.125 mm to compensate for the bulk height of the
semiconductor diode itself so the beam height on the MIOB can be maintained.
The MIOB provides a channel for optics integration, formed by rails to hold the FAC
40 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
Figure 4.5: Concept of p-up and p-down mounting
lenses in lateral direction, and a plane bottom to host the SAC lenses. The submounts
are adhesively bonded to the bottom of this optical channel as well.
The MIOBs and the submounts are manufactured by an external manufacturer. In-
tegration of submounts and optics as well as wire bonding are carried out at FBH.
4.1.2.2 Electrical Interface on MIOB
The MIOB hosts several electrical interfaces to address multi-section laser and ampli-
fier diodes, as well as several sensor read-out interfaces. The electrical interfaces and
sensor read-outs are organized on AlN based PCBs, they are highlighted in figure 4.6.
The electrical interface PCBs provide gold finished bonding pads to be able to connect
the laser and amplifier diodes, mounted on submounds. The submounts also provide
bonding pads, they are connected to the interface PCBs by wire-bonding. The bond
wires, connecting the submount with the electrical interface are clearly visible in fig-
ure 4.3 on the right hand side of the lenses. The short bonding wires left to the beam
axis connect the temperature sensor. The longer bonding wires right to the beam axis
connect the laser diode to the electrical interface of the MO.
The MIOB can be connected to supply electronics with miniature, UHF coaxial con-
nectors. Most of the coaxial sockets are Radiall micro-miniature coaxial (MML) in-
terfaces, as depicted in figure 4.3, left to the lenses. An exception is the TA injection
current interface, where micro-miniature coaxial (MMCX) sockets for higher current
capability were chosen.
There are in total five electrical interfaces on two interface PCBs. This enables the
supply of a maximum of three sections of a MO chip, and two sections of an amplifier
chip. Two of these interfaces also provide a modulation capability, allowing modu-
lating the injection current, one for the MO injection current, and the other one for
modulating the amplifier section of the PA diode.
Figure 4.7 depicts the modulation interface of the MO, featuring a DC port for mod-
ulation operation (MDC), and two ports for a modulation signal, the RF modulation
port (MRF), capable of radio frequency (RF) modulation, and the transistor modulation
port (MMOD), suitable for low frequency (LF) modulation frequencies. The MMOD
port delivers the gate voltage to a common source transistor circuit, reducing the injec-
tion current, supplied by MDC, for the laser every time the transistor opens. The transis-
4.1. DESIGN OF THE HYBRID INTEGRATED LASER MODULES 41
Figure 4.6: Electrical interface provided by a MIOB. TADC: DC injection current of PA
gain section, TAMOD: modulation of PA gain section, RWDC: DC injection current of PA
pre-amplifier section, DC: DC injection current of MO (bypassing modulation electron-
ics), MDC: DC injection current of MO (to be modulated by modulation electronics),
MRF: RF modulation port, MMOD: LF modulation port, TSMO: MO temperature sensor
port, TSMIOB: MIOB temperature sensor port, TSPA: PA temperature sensor port
tor T1 is a N-channel junction gate field-effect transistor (JFET), normally open, when
no distinctive gate voltage is applied. The modulation method of reducing the injection
current allows very low modulation frequencies down to direct current (DC) range and
is therefore capable of LF modulation. In addition, the laser diode is protected against
reverse biasing, since the current direction cannot be inverted accidentally. Still, the
modulation bandwidth is limited by the bandwidth of the transistor which corresponds
to about 500MHz. The MMOD input provides a pull-down resistor R1 that defines the
gate input of the transistor to be low when there is no signal connected to the MMOD
port.
The MRF port adds an attenuated modulation voltage onto the DC injection signal.
The input signal is attenuated by a factor of 10 to minimize the risk of accidentally
reversing the bias of the laser diodes injection current. Over-driving the input at MRF
can cause a reverse bias at the laser diode, which may lead to fatal damage of the diode.
However, the direct combination of the modulation signal with the DC signal allows
for modulation frequencies in the GHz range. Both, the MDC and MRF port together
with the output interface for the laser diode, form a simple bias-tee, including the
inductor L1 and capacitor C1. L1 protects the current source attached to the DC current
port MDC against disturbance caused by the applied modulation frequencies. Whereas
the capacitor C1 protects the signal generator connected to the MRF modulation port
against the DC loads. In addition there is the inductor L2, filtering RF frequencies,
supplied by the MRF port, from the drain of the LF-modulation transistor T1. With
these protection components, the MRF port contains a high pass filter formed by C1
and the resistance of the laser diode, opening at 1 MHz. But since the MMOD port is
available for LF modulation, the high pass filter does not minimize the functionality of
42 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
Figure 4.7: Electrical interface on MO-MIOB: the DC port supply can either be modu-
lated with a common source circuit (MMOD with T1) or with a bias-tee path (MRF).
the interface.
The electrical interface PCB for the MO also provides an additional port that is not
connected to the modulation ports. This port simply is called DC (in contrast to MDC)
and guides the applied signal directly to the laser diode. This port does not contain any
filters and therefore is used when the injection current is modulated externally or if a
multi-section chip requires the excitation of an additional amplifier section.
The third supply input of the MO interface PCB is marked with "HT". It can be used
to supply another section of the MO, act as redundant input, or drive a heater section,
if the MO device provides one.
The current supply PCB of the amplifier hosts a DC port for the RW pre-amplifier
section (RWDC) and the high current interface TADC for the gain section of the ampli-
fier, tolerating several Ampere of current. The DC port for the tapered section (TADC)
interface provides two MMCX connectors to share the current load and to increase the
effective diameter of the current guiding pins and cables. The injection current of the
gain section of the amplifier can be modulated by applying a signal to the modulation
port for the PA gain section modulation (TAMOD). The current modulation is again
realized with a common source transistor circuit, comparable to the MMOD port on
the MO interface PCB as illustrated in figure 4.7.
The complete schematic of the electrical interface PCBs and a component list is
provided in appendix B.
In addition to the electrical interfaces of the laser diodes the MIOB provides three
temperature sensor read-outs for temperature monitoring close to the laser diodes and
on the MIOBs optical bench. The temperature read-out interfaces host a MML con-
nector and two filter capacitors to read out the value of the temperature sensors. In
addition, they allow for mounting of an extra sensor measuring the temperature of the
MIOB close to, but not on the submount, in case there is no temperature sensor avail-
able on the submout. The temperature read-out PCBs of the MO, the PA, and the MIOB
are identical in design.
4.1. DESIGN OF THE HYBRID INTEGRATED LASER MODULES 43
4.1.2.3 Mechanical Adapter
To be able to mount the ceramic MIOB in the assembly station, characterization as-
sembly, or users system, a mechanically stable and thermally conductive adapter is re-
quired. This adapter is referred to as conductively cooled package (CCP) and is made
of Cu for its high thermal conductivity of 400 W/mK. The surface is protected against
oxidation and minor mechanical damage with a 3µm thick layer of galvanically applied
nickel (Ni). A CCP is depicted in figure 4.8. To ensure thermal contact and to avoid
bending of the ceramic body upon mounting, the top and bottom surfaces of the CCP
have a specified planarity of 10 µm maximal bow over full length and width – a chal-
lenge in production of the CCP. Therefore, the CCPs were manufactured with electrical
discharge machining (EDM) and individual characterized with a 3-dimensional optical
surface profiler. Each CCP has a serial number, which is engraved by laser, highlighted
in red in figure 4.8.
The thermal expansions of Cu (16.5µm/mK) and AlN (5.3µm/mK) do not match.
Therefore the MIOB cannot be mounted on the CCP by adhesive bonding, since thermal
expansion differences e.g. between storage and operation at WP temperature, would
add mechanical stress to the system, causing deformation or even loosening of the
MIOB. The AlN-MIOB therefore is mounted on top of the CCP by three vertical clamps,
positioned on the long edges with dowel pins and screws to with a well-defined torque
is applied. The clamps are also manufactured by EDM, guaranteeing a chip free, well-
defined round contact surface where the clamp presses on the MIOB. Loosening of the
screws, both of the ones holding the clamps and of the screws attaching the adapter to
the experimental system, is prevented by SCHNORR safety spring washers.
Figure 4.8: CCP with serial number (marked red) and imprints of the safety spring
washers (marked blue)
4.1.2.4 Housing with Electrical Interface
Although the MIOB already hosts coaxial interfaces, it is recommended to convert the
MML and MMCX connectors to SubMiniature version A (SMA) connector or some other
standards common in laboratories. Since most supply electronics neither provide a
MML nor a MMCX interface, an adaptor is required. Apart from the availability of com-
mon connectors like SMA, a conversion on the MML interface is useful due to the limited
44 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
length of available MML cables, which are not longer than 100 mm. Another reason
is the fact that MML sockets are specified only for approximately 10 re-connections.
To limit the chance of damage to the laser module or degrading performance due to
exceeding the number of re-connects, it is recommended to keep the MML cables at-
tached to the MIOB and mount them to an electrical adapter. In addition, long levers,
formed by cables, can cause the sockets to break off from the MIOB.
In addition to providing an electrical adaptor within <100mm of the sockets of the
MIOB, the optical setup needs to be protected against dirt, physical impact, e.g. induced
by user, as well as for streams of air that may cause frequency distortion. To protect
from both, the electrical adaptor is integrated into a housing package, as shown in
figure 4.9. In our package design, we are able to integrate the adaptor PCBs directly on
the sides of the CCP, forming the long side covers of the housing. The electrical adaptor
side panels of the housing are depicted in figure 4.9. The adaptor PCBs provide various
SMA interfaces and a Sub-D interface, compatible to ILX temperature controllers used
in our laboratory.
The cover lid and side panels of the short side of the housing are formed by Al parts
and provide holes for the optical beam, as well as screw threads for the mounting of
the long side panels. The Al parts have a black anodized finish in order to prevent
reflection and reduce the scattering of light inside the housing, which may disturb the
laser.
In the MAIUS and FOKUS laser system, this housing is not necessary since the laser
systems in both experiments are optimized for low weight and are designed to provide
one housing solution for all laser system components.
(a) (b)
Figure 4.9: Housing package with integrated electrical adaptor (a) on left side, and (b)
on the right side of the optical axis
4.1.3 Features of the Laser Modules
4.1.3.1 Temperature Monitoring
Since the emission wavelength of the DFB laser depends on the temperature, the tem-
perature of the DFB laser has to be stabilized in order to minimize drifts in wavelength.
4.2. LASER MODULE ASSEMBLY PROCESS FLOW 45
A feature of the MIOB is the integration of multiple temperature sensors, both on the
submount of the MO, on the submount of the PA, and on the optical bench of the
MIOB. The MO sensor close to the DFB chip and the MIOB sensor left to the isolator
are depicted in figure 4.3.
The output of the sensors can be used as control value for temperature stabilization.
The sensor on the MO submount will deliver a temperature value very close to the
actual temperature of the MO diode, enabling a more precise temperature control of
the diode.
The sensor at the PA submount also is an indicator for the coupling efficiency of the
to-be-amplified beam into the PA. Since the input energy is constant, the lesser photons
are injected into the optical amplifier, the more of the incoming photon energy has to
be converted to heat, as long as the amplifier is not saturated.
With the MIOB sensor the thermal resistance of the MIOB can be estimated. In
addition, it serves as a redundancy sensor for temperature controlling and offers clues
on the temperature gradient between MIOB and the diodes on the submounts.
4.1.3.2 Fibre Coupling
To be able to use the laser beams in the experiments described in section 2.1 and 2.2,
fibre coupling of the modules output beam is required.
To minimize fluctuation of the power coupled into the fibre due to thermal expan-
sion of the mounting structures, the fibre coupler and the laser optics are placed on
the same mount, the MIOB. With that, miss-alignment due to thermal expansion of
the integrated materials are reduced to the expansion of the miniaturized optical com-
ponents themselves, and the system will not suffer from different expansion rates of
different mounts. This increases the stability of the system.
In addition, an on-board coupling is a more compact solution than external cou-
pling. For this reason the MIOB provides a 20mm long and 25mm wide section behind
the optical bench on the base plate. The fibre coupler itself was designed and is inte-
grated by the project partner "Universität Hamburg", Institut für Laserphysik (UHH),
details are described in [43]. A fibre coupled MIOB is shown in figure 4.4.
4.2 Laser Module Assembly Process Flow
In the previous section all components of a laser module were introduced. The fol-
lowing section deals with the assembly of the components in order to finalize the laser
module.
A reproducible integration of mobile optical setups is a complex process with several
in-between-characterizations and qualification tests as shown in figure 4.10. Before in-
tegration all active and passive optical components pass through a qualification process,
as described in chapter 5. The most complex one is the semiconductor diode chip quali-
fication, see step 1.1 in the general process flow chart in figure 4.10, including spectral
and power characterization as well as a burn-in phase for 120 h. The optical micro-
isolators (step 1.2) are characterized optically for isolation and transmission, and their
46 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
facets are controlled for defects. The MIOBs (step 1.3) are cleaned and undergo a vi-
sual and electrical inspection, and for the CCPs (step 1.4) the surface condition and
deformation (bow) of the surface are determined and evaluated. When all necessary
components are qualified, the integration process can be initiated, starting with the
optical master oscillator integration (step 2) before continuing with the optical power
amplifier integration (step 3).
The integration of the optical components into the MIOB takes place at an espe-
cially designed assembly station, described in section 4.3.1. After the selected DFB
laser is integrated into the MIOB, the module is placed on the assembly station and
connected to the current source. The laser diode is tested electrically before the colli-
mation lenses and the isolators are integrated. The spectrum and optical power of the
laser is recorded, which serves as reference of the performance of the laser during the
assembly of the optical components close to the DFB laser. After integration of each
optical component group the output power and the optical spectrum is recorded. If
the spectrum of the DFB laser shows any deviation of the previous spectral behaviour
during the MO assembly, the integration process is aborted and the chip is replaced.
After the integration of the collimation lenses of the MO and the optical isolators,
the selected PA chip is placed into the module. Since the PA mounting is done in another
laboratory, the module has to be temporarily removed from the assembly station. After
re-installation into the assembly station, the electrical connections of the chip are tested
before the coupling and collimation optics are integrated.
After integration of all required optics, the module undergoes extensive electro-
optical tests, step 4 in the process flow in figure 4.10, in order to document that all
electro-optical requirements can be fulfilled. Chapter 6 presents the corresponding
measurement methods and representative results achieved with the laser modules in
this step. If a module does not perform according to the requirements, single optical
components have to be manipulated or replaced.
In a last visual inspection, step 5, the module is controlled for visual defects, such
as deformed or broken bond wires, and dirt. Particles are removed and broken bond
wires may be replaced. Pictures taken in this step belong to the documentation as well
as the measurement results of steps 4.
The module is then mounted into a transport box, the packaging in step 6. The
electric cables remain attached to the module in order to reduce de- and re-connecting
the sockets and to avoid damage of the electrical interfaces. Finally the modules are
delivered to the project partners together with an individual user manual, including
documented performance values (step 7).
4.2. LASER MODULE ASSEMBLY PROCESS FLOW 47
laser module assembly
1.3 MIOB check 1.4 CCP check
all components
qualified?
4. EO characterisation
expected behavior?
manipulation or
exchange of components
5. final inspection
7. final documentation
6. packaging
module faultless? just dirty?
clean module
module delivery
no
yes
yes
yes
yes
no
no
no
1.2 µIso qualification
problems with:
output power,
PER, beam shape,
PA components
problems with:
spectral behavior,
linewidth, MO components
input
input
input
input
1.1 chip qualification
2. MO integration
3. PA integration
Figure 4.10: Overview of laser module integration process flow
48 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
4.3 Integration of Optics
The assembly process is divided into two sub-processes, denoted as "2." and "3." in the
process flow chart in figure 4.10. The first process is the assembly of the optical master
oscillator, including beam collimation and the integration of the micro-isolators. The
second process is the integration of the optical power amplifier and the beam shaping of
its optical input and output. The assembly procedure follows a process definition, defin-
ing the steps to be taken and documentation to be carried out. Each assembly step is
accompanied by visual documentation and characterization of the electro-optical prop-
erties. This enables the detection of problems early in process and allows for analysis
and corrective measures in case of potential malfunctions.
4.3.1 Assembly Setup
The optics assembly setup is depicted schematically in figure 4.11. All optical compo-
nents, except for the submounts that host the semiconductor chips, are integrated onto
the MIOB with this setup. The positioning of the lenses and isolators is carried out
with a Physics Instrument Hexapod F-206.S, a manually controlled robot with a spatial
and angular resolution of 0.1µm and 2 µrad, respectively. The components are kept in
place with a vacuum interface made of Macor, a glass-ceramic with very low thermal
expansion to keep the position when exposed to ultra violet (UV) light or heat. The
Hexapod host an electrical interface to connect the electro-optical components to elec-
trical supply sources. Also a Peltier element in integrated to the setup to stabilize the
MIOBs temperature. The assembly station (ii) also hosts a microscope and a cold-light
source for monitoring the manipulation of the components via the robot. The UV light
source Omnicure 1000 with two light pipes each providing 230 mW of light is curing
the adhesive after active alignment of the optical components.
Figure 4.11: Schematic of assembly station and monitoring
4.3. INTEGRATION OF OPTICS 49
The beam, exiting the assembly station, is monitored with two cameras, positioned
within one Rayleigh range of the collimated beam according to the optical simulation,
described in section 4.1.1.2. A fraction of the beam is separated with beam splitters
and guided to two cameras in order to monitor the beam profile, see (iii) in figure 4.11.
One camera monitors the far field while the other camera, positioned in the focus of a
bi-convex lens with a focal length of 10cm, maps the near field of the laser beam. The
near field camera is used for beam collimation which is done by minimizing the spot
size on the camera, the far field camera can be used for monitoring the actual beam
diameter and shape as well as power content in the central lobe.
Besides the beam monitoring, the assembly setup allows for simultaneous charac-
terization of electro-optical properties of the laser, namely the optical output power,
recorded with a power meter (iv), optical spectrum, recorded with an optical spectrum
analyser (OSA) and the emission wavelength, monitored with a wavemeter (v). In ad-
dition, path (iv) provides pinholes with a diameter of 1 mm at a distance of 1 m from
the laser for adjusting the beam position on the MIOB with an accuracy of 63 mrad. For
this pinhole aiming, the power passing the pinhole is maximized. This beam guiding
facility guarantees a reproducible pointing of the beam on the MIOB.
The rear output of the laser can be collimated using the pinhole and far field camera
installed at the rear site of the assembly station, denoted by (i). Since the rear output
is only used with photodiodes or shall be dumped directly to avoid scattered light,
focussing on a distant point is a sufficient method of collimation. The camera for rear
collimation has a distance of 2m from the chip.
4.3.2 Integration of Master Oscillator Optics
The integration of the optical master oscillator optics is divided into several steps, as
shown in figure 4.12. After burn-in and pre-characterization, as described in section
5.1, a qualified chip on submount is chosen. Again, the facets of the chip are controlled
visually because they could be affected by dirt or damage during pre-selection process.
If the facets passed visual inspection again, the submount is placed into the MIOB,
aligned along the rails framing the optical path of the MIOB. A thermal adhesive with
bonds the submount assembly to the MIOB. The chip on submount is than wire-bonded
to the electrical interface of the MIOB.
With the DFB laser integrated and connected, the MIOB is installed into the as-
sembly station, described in the previous section. The MML cables are attached to the
electric interface ports driving the DFB laser and reading the available temperature
sensors. These cables remain on the module to minimize mating stress on the connec-
tors. During the first electrical test of the module, the voltage required to drive 2 mA
through the chip is documented as reference value. The voltage typically is in the range
between 1V and 2V. If not, the cable connection or bond wires do not work properly
and have to be checked. The minor current of 2 mA was chosen to prevent damage to
the diode in case there is a slack joint in the connection.
50 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
2. MO integration
facet control of DFB
facets faultless and clean?
submount on MIOB,
bond connection
estimate WP
document
beam parameters
MO integration complete
document
beam parameters
is rear ourput collimated?
collimate rear output
collimate main output
yes
yes
no
no
chose DFB on submount chip rejected
MIOB setup @
assembly station
collimate beam
integrate µIso
Figure 4.12: Overview of master oscillator optics integration
4.3. INTEGRATION OF OPTICS 51
With inserting a temporal collimation, the working point (WP), including DFB in-
jection current and operating temperature, is determined with help of the wavemeter,
according to the targeted wavelength. In this thesis the targeted wavelength is either
767nm or 780nm. During the integration process, the DFB laser is operated at the
identified WP to optimize the alignment of the optical components at the same condi-
tions the laser module will be operated at later on.
The collimation process of the beam starts with adjusting the FAC, followed by the
SAC. The lens position is adjusted actively by minimizing the spot size of the beam on
the near field camera. The near field camera, as shown in figure 4.11, is positioned
in the focus of a lens. This lens reverses the effect of the collimation lens, causing
collimated rays to converge to a spot [106]. The spot size will be at its minimum
when the beam rays are collimated because rays diverting from the parallel direction
of propagation will be projected aside the focus point. Since the FAC and SAC lenses
are cylindrical lenses, they only manipulate either the vertical and horizontal beam
propagation, respectively. The pointing of the beam is adjusted by maximizing the
power fed through the pinhole in 1m distance to the chip. Before curing of the FAC,
the vertical size of the near field beam projection is recorded for different FAC-lens
positions along the optical axis, as shown in figure 4.13a, to document the optimal
position. The parabolic shape of the values in figure 4.13a also hints that there are
further distortions in the beam line blocking or bending the light and that the lens is
not tilted.
(a) FAC (b) SAC
Figure 4.13: Variation of near field beam diameter by varying the position of the (a) FAC
and (b) SAC lens. The beam is collimated at the Zero-position in this plot.
FACs are glued laterally to the rails of the MIOB to be able to adjust the vertical
position of the lens and to minimize distortion effects, introduced by the shrinking
of the adhesive under curing. However, non-symmetrical curing may wrap the FAC-
lens, causing miss-alignment. The positioning of the UV-arms for curing therefore has
to be chosen carefully. All optical components are glued with a low out-gassing UV
adhesive that has been used for space applications before. The adhesive is cured with
approximately 120Ws/cm2.
52 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
The SAC lens is placed the same way the FAC was adjusted, only, they are glued
directly onto the bench to allow lateral position adjustment. A sweep of the SAC lens
position along the optical axis with resulting beam diameters on the near field (NF)
camera is shown in figure 4.13b. Variations of the SAC position in the µm range along
the optical axis do not affect the beam diameter as strong as the FAC position. This can
be explained with the longer distance from the source and the larger focal length of the
SAC lens.
After collimation of the DFB, the optical output power, emission spectrum, and
voltage for various injection current settings are recorded. This documents the DFBs
behaviour after collimation and ensures that the DFB has not been damaged or dis-
turbed under the integration process. Additionally, the beam profiles of the far and
near field are saved as reference.
The next step addresses the integration of the micro-isolator. The isolator is placed
into the corresponding cut-out in the MIOB and is coarsely adjusted by rotating it
around the optical axis of the beam in order to maximize the power propagating
through the device. For precise alignment, a PBS, oriented to reflect the nominal po-
larization, is placed on the bench. The power behind the PBS is minimized by rotating
the isolator before the isolator is glued on one side to the bench. The other side of the
isolator is only accessible after turning the MIOB by 180° around the vertical axis. The
optical power behind the PBS is documented after each curing step to ensure that the
isolator’s alignment is maintained throughout the adhesive bonding. After integration
of the isolator, optical power and spectrum are recorded again and compared with the
measurement before isolator integration. The power behind the micro-isolator is re-
duced by 3-4dB compared to the output power of the laser diode, due to the insertion
loss of the isolator. The spectral characteristics should be maintained.
The rear output of the laser is collimated by minimizing the spot size on a far field
camera, placed in a distance of 2 m to the chip and illustrated in figure 4.11. The
output power of the rear exit is documented before and after integration of the rear
micro-isolator. The integration of the rear isolator is comparable to the integration of
the front isolator.
With the front and the rear output collimated, the micro-isolators in place, and the
beam characterized, the MO integration is complete. The following step would either
be the integration of the PA, described in the next section, or the integration of the fibre
collimator, performed by the UHH and mentioned in section 4.1.3.2. A module without
optical amplifier can be used for applications where less output power is required, e.g.
the master laser in MAIUS or the laser source in FOKUS.
4.3. INTEGRATION OF OPTICS 53
4.3.3 Integration of Amplifier Optics
For applications with a demand on output power higher than 20 mW, an amplifier can be
integrated. The sub-processes for integration of the amplifier are shown in figure 4.14.
The amplifiers undergo a burn-in and pre-characterization process and get qualified
with the criteria defined in section 5.1.
The selected amplifier-submount assembly are aligned, adhesively bonded, and
wire-bonded the same way as the oscillator chips. Heating the module up to 100 °C
during the wire-bonding process of the submount does not affect the positioning of the
optical components that have been integrated during the previous integration steps, as
long as no mechanical stress is applied to them.
After re-installing the MIOB onto the assembly station and connecting the MML
cables for the RW pre-amplifier section, for the tapered main-amplifier section, and
for the temperature sensor located on the amplifier submount, an electrical test of the
laser module and its connection to the current drivers is performed. This is done by
logging the voltages that is applied to the diodes while driving the MO diode and the
RW section with 2 mA and the TA section with 20mA. These values are reference values
for electrical tests and should be in the range between 1 V and 2 V. The value of the MO
diode has been recorded earlier (before the MO lenses and isolators were integrated)
and is now compared to this reference value. If the voltage is out of range, the electrical
connections have to be checked.
Before the beam emitted by the MO can be coupled into the amplifier, the output of
the amplifier has to be collimated, in order to be able to monitor the amplifiers output
power. This is required to judge the quality of the coupling.
Since the beam profile of a semiconductor amplifier in ASE operation differs in its
characteristics compared to when a seeding beam is coupled into the amplifier, the po-
sitioning of the collimating lenses cannot be finalized before establishing an in-coupling
into the PA. Therefore, a temporal collimation of the output is required. This is done
with an aspheric round lens and a cylindrical SAC lens to compensate for the differ-
ences in divergence. The round lens has an optimized height, so it can be placed on
the bench but achieves roughly the vertical position of the FAC. Both lenses are placed
loosely on the bench to obtain a fairly good collimation and will be removed for final
collimation after finishing the coupling process.
The next step aims at coupling the MO emission into the PA chip. For an efficient
mode coupling between the MO output and PA input the beam parameters of the MO
output have to be matched to the beam parameters of the PA input. This is done by
focussing the beam with appropriate lenses as determined in the optical simulation (see
section 4.1.1.2) into the active region of the RW section.
For coarse alignment of the coupling lenses the RW section of the amplifier chip is
operated as a photo detector and the photo current, generated by photons stimulating
the doped materials in the active volume, is monitored. The alignment is achieved by
maximizing the photo current.
54 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
3. PA integration
facet control of TPA
facets faultless and clean?
submount on MIOB,
bond connection
temporal
output collimation
PA integration complete
document
beam parameters
yes
no
chose TPA on submount chip rejected
MIOB setup @
assembly station
RW incoupling
collimate beam
remove temporal
output collimation
Figure 4.14: Overview of power amplifier optics integration
4.3. INTEGRATION OF OPTICS 55
The beam in-coupling process starts with the SAC that is put down onto the bench
in order to ensure that the MO beam roughly overlaps with the active area at the input
of the PA chip. The SAC position cannot be finalized until the FAC is in place, since the
FAC interferes in the beam propagation between the laser and the SAC, therefore the
SAC position can only be approximated. The FAC is adjusted coarsely by maximizing
the photo current until a value of several mA is reached, and fine-tuned by maximizing
the output power of the amplifier. The vertical lens position at maximum photo cur-
rent usually differs from the vertical lens position at maximum output power due to
thermally induced distortion of the setup when the PA is driven with a total current of
2200mA during fine alignment of the FAC and SAC focussing lenses.
(a) FAC (b) SAC
Figure 4.15: Variation of amplifier output power by varying the position of the (a) FAC
and (B) SAC coupling lens. The MO-beam is coupled best into the the amplifier at
zero-position in this plot.
Before gluing the lens it must be ensured that the main power contribution is caused
by the seeding beam, not by ASE activity. This is done by checking the spectrum of the
beam at working point parameters. In the spectrum the ASE should be suppressed by at
least 30dB. The output power for various FAC lens positions, as shown in figure 4.15,
are recorded to document the optimal position and determine possible misalignment.
Figure 4.15 shows the development of the output power of the amplifier when the ver-
tical (FAC) and lateral (SAC) lens positions deviates from the optimum. The output
power in the FAC lens adjustment is lower due to the approximated but not jet opti-
mized position of the SAC lens. The maximal output power auf 1.22 W in the FAC lens
adjustment suggests that the SAC lens was miss-aligned by about 40µm. After SAC
lens alignment, an optical output power of 1.45 W could be achieved. Typically, at the
point of maximal output power the ASE background is suppressed most strongly.
After the coupling lenses are integrated, the temporal collimation lenses can be
replaced with the collimation lenses. The collimation FAC is integrated first, the same
way as the collimation FAC of the MO. The two SAC lenses are placed on the bench
according to the simulated position and fine tuned until the near field camera shows
a minimum for the spot size. The far field camera should show a round spot within
56 CHAPTER 4. CONCEPT OF MICRO-INTEGRATED LASER MODULES
the specified beam dimensions and with more than 50% power content in the central
lobe. The higher the power content in the central lobe, the better the fibre coupling
efficiency.
After finishing the integration of the SAC telescope lenses, the module is complete
and can be fully characterized. The results of the characterization of the laser modules
are described in chapter 6.
Chapter 5
Pre-Integration Component
Characterization
A hybrid integrated laser module consists of actively driven components and passive
components that together make up a complex laser system. When integrating multiple
lasers systems, reproducibility of the performance is a measure for process control. With
high demands on the performance, as they are defined in chapter 2, reproducibility
becomes indispensable for fulfilling the requirements. Another measure for process
control is the yield of production. The higher the yield, the shorter the time until the
products can be provided and the production costs are smaller.
For both a high yield and reproducibility of the performance, the components, in-
tegrated in the system, have to provide a stable quality and reproducible performance
themselves. However, this cannot be guaranteed by production. The integrated com-
ponents are highly specialized and thus are produced in rather small numbers, not
allowing for process optimization. In addition, variations in production scatter the per-
formance of the components as well. Therefore each critical component has to be tested
and qualified before integration into the system. Table 5.1 lists the critical components
and their characteristics that had to be verified.
In the following chapter the qualification process, the test criteria and the results of
these tests are described.
component form qualification property
DFB lasers active spectral behaviour, output power, facet condition
tapered amplifiers active spectral behaviour, output power, facet condition
micro-isolators passive optical behaviour, facet condition
micro-lenses passive facet condition
optical bench passive electrical resistance of interfaces
mechanical interface passive planarity of contact areas
Table 5.1: List of components that are qualified before integration, and their qualifica-
tion property
58 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
5.1 Laser Diode Pre-Characterization
The most critical parts in a hybrid laser module are the semiconductor laser diodes
themselves, namely the optical master oscillator diode and the optical power amplifier
diode. Any defect can cause deviation from the specified performance of the diodes,
which reflects directly on the performance of the laser module.
Figure 5.1 shows measurement values of malfunctioning laser diodes that could be
identified before integration. The measurement methods are described in chapter 6.
The DFB spectrum for varying injection currents in figure 5.1a shows strong multi-mode
behaviour and modes jumps. This particular DFB laser cannot be used in a laser module
dedicated to the missions described in chapter 2, because the requirement of single
mode emission cannot be provided for injection currents <200mA. This, however, is
the current limit of the current drivers in the MAIUS mission apparatus. In addition, the
stability of the frequency emission for >200mA of this laser has to be questioned. The
spectral instability can be caused by defects in the material or tensions in the bonds.
Both can lead to a further degradation of the laser performance and to an early loss of
the laser module. With that, the laser diode is not qualified for integration.
The optical output power of a seeded amplifier shown in 5.1b does not exceed
15mW before decreasing for increasing injection currents. Also, the peak optical output
power is reached at 1000 mA. Taking into account that the specified optical output
power should be at least 2000mW, and the thermal roll-over should not set in for an
TA injection current <2000 mA, this PA clearly cannot fulfil the specifications.
(a) (b)
Figure 5.1: Identified in component qualification: multi-mode spectrum of a DFB laser
(a), and thermal roll-over of a PA(b)
There are a number of possible reasons for poor performance of laser diodes. Irreg-
ularities in processing of the doped layers under production, and coating of the facets, a
disadvantageous splitting angle during wafer separation into diode bars, stress applied
to the chip during assembly, or contamination of the facets with particles can cause
disturbance in the propagation of a single optical wave or even can cause a COD.
There are a number of test methods available for identifying devices with defects or
abnormal behaviour as early as possible in the production process. With the increased
5.1. LASER DIODE PRE-CHARACTERIZATION 59
demand on semiconductor chips for optical communication, automated methods were
established to verify the performance of laser devices in mass production [107,108].
However, for research and prototyping fully automated quality tests are too expensive
and too specialized. Tests performed at the FBH during semiconductor processing are
either manual or semi-automated.
Close inspection of the wafer can reveal areas that show irregularities which oc-
curred during the etching, epitaxy, and galvanization, as well as particles or remaining
coatings that nominally should have been removed. These defects lead to an increased
risk of degradation of performance, reduction of lifetime, or even malfunction and
potential restrictions during assembly, such as bondability of the device in further inte-
gration.
The inspection on wafer level is followed by tests on bar level. The wafers are
separated by cleaving rows, so called laser bars, hosting laser diodes that are laterally
still connected to each other. The facets of the laser bars are coated and a bar inspection
is performed. In this measurement the single diodes are electrically contacted with
needles, and the injection current is ramped per diode, revealing the current threshold,
slope efficiency, optical output power, and the optical spectrum of the device. This
test is not suitable for amplifier chips, because they do not provide a resonator. The
optical output power of a PA in this test contains ASE only, amplifier chips will need a
seeding beam for a performance test. However, the electrical test on bar level allows the
detection of optical master oscillators with poor optical resonators, or malfunction of
the facet coating but does not reflect on the actual performance of the chip on submount
since the thermal conditions are not comparable and the recorded spectrum does not
provide sufficient resolution to determine the side mode suppression ratio.
In order to verify the actual performance of a MO chip or an amplifier, the laser
diodes have to be individually tested on submount level. To reach submount level, the
chips are split of the bar, bonded onto a submount and wire-bonded to an electrical
interface. During these steps, the chips are exposed to thermal and mechanical stress
that can interfere with mode confinement in the chip. Therefore, and to identify diodes
that do not perform according to specification (as shown in figure 5.1) pre-integration
characterizations have to be executed.
However, identifying laser chips that do not perform within specification from the
beginning of life might not be sufficient enough for pre-integration qualification. Espe-
cially the spectral behaviour of laser chips may change during the first hours of opera-
tion. Burning-in the chips before integration is therefore necessary to identify chips that
change their behaviour depending on the time they have been operated. A negative ex-
ample is given in Figure 5.2. It shows the optical spectrum of a chip after bonding to a
submount and integration into a laser module. This chip was characterized (measure-
ment on 23rd January 2012) and found to be a single mode chip before integration into
the laser module. During integration a multi-mode behaviour appeared for injection
currents >180mA. The spectrum of 21st February was recorded after approximately
50 hours of operation of the DFB. The spectrum recorded after another >50 operating
hours (12.03.12) verifies that the spectral performance degraded even further. Most
likely the chip hosts defects in the wave-guide structure, propagating further into the
active region with increasing operation time. These defects cannot be detected with a
60 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
Figure 5.2: Degradation in the spectrum of a DFB diode over time
single measurement. However, most chips change their behaviour in the first operat-
ing hours and stay stable after "settling down" [109]. A burn-in phase of several hours
before recording the initial characterization is therefore recommended [110]. This ini-
tial characterization is required to reveal changes in the chips optical behaviour after a
longer burn-in of more than 100 operation hours. A chip without significant changes
in the emission spectrum and optical output power can be considered as stable. The
criteria of qualification are further defined in the following section 5.1.3.
Despite the electro-optical performance of the laser diodes, the facet conditions have
to be qualified as well. The facet coating can already be inspected on bar level, but the
inspection has to be repeated before chip integration because dirt can accumulate on
the facet each time the device is handled.
To fulfil the strict requirements of the laser systems performance, as defined in ta-
ble 2.2, a selection process on submount level was established, as described in the
following sections.
5.1. LASER DIODE PRE-CHARACTERIZATION 61
5.1.1 Diode Selection Procedure
1.1 chip qualification
mounting laser on submount
control front and rear facet
facets faultless and clean?
24h burn-in
EO characterization
expected performance?
120h burn-in
EO characterization
expected performance?
chip qualified
chip rejected
yes
yes
yes
no
no
no
Figure 5.3: Overview of diode qualification
process
Figure 5.3 illustrates the process
flow of the chip qualification pro-
cess. After submount assembly of
the laser and amplifier chips, the
condition of the facets are docu-
mented and evaluated. Devices
passing the visual tests are placed
into temporary mounts with elec-
trical interface and undergo 24h
continuous operation [110]. Af-
ter this initial burn-in, the optical
power and spectral behaviour of
the DFB-device under test (DUT)
is characterized, both at 20 °C for
statistical evaluation of all DUTs
and at the expected working point
temperature for individual verifica-
tion. Chips passing the qualifica-
tion criteria that are given in the
following section, undergo a sec-
ond, 120 h long continuous opera-
tion cycle. The electro-optical char-
acterization after the long burn-in
phase ensures that the performance
of the chips has not changed signif-
icantly during five days of continu-
ous operation. Diodes with a per-
formance deviant from the initial
measurement have to be rejected
and will not be integrated into the
laser modules.
For PA-DUTs the initial charac-
terization is left out, the chips un-
dergo a 120h burn-in directly af-
ter the visual inspection. This is be-
cause the measuring process of PAs
is more complex and yet not accurate enough to reveal degradation of the chips after
120h operation. The burn-in process is nevertheless carried out to identify diodes with
COD.
62 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
5.1.2 Pre-characterization and Burn-In Facility
The electro-optical qualification of the diodes is performed at a facility that was spe-
cially designed for that purpose during this thesis. In order to connect the diodes to a
current driver and temperature controller, the chips on submount have to be temporar-
ily placed into a mount with a specifically tailored electrical interface. This mount,
depicted in figure 5.4a, is made of copper for good thermal conductance and is coated
with a nickel layer against oxidation. On top, a bondable PCB is placed, hosting three
MMCX interfaces, for interfacing various sections of a diode and for the temperature
sensor on board the submount. The submount is fastened with a wedged wrangle
placed in a notch in the mount. The diode is wire bonded to the PCB and can be placed
both into a burn-in box, shown in figure 5.4b and into the pre-characterization setup,
illustrated in figure 5.5.
For burn-in the diode mounts are placed into light-opaque boxes, so the burn-in pro-
cess cannot be disrupted and does not disturb other work in the laboratory. Each box is
capable of hosting four diode mounts and provides thermal stabilization. 4mm-single-
wire-connectors were chosen as electrical interface of the box, to obtain flexibility in
controlling the single sections of each diode. In addition, 4mm-singe-wire-cables are
commonly available in most laboratories and are the common interface to consumer
level voltage- and current sources. The burn-in box supports two sections per diode
that can be operated separately. For thermal stabilization of the box either the internal
temperature sensor or a sensor of one of the submounts inside the box can be used.
Two Hameq HMP4040A are allocated as current sources during burn-in. An 1 MΩre-
sistor and a capacitor in parallel to the diodes avoid voltage peaks during switching of
the current drivers.
(a) (b)
Figure 5.4: Submount holder with electrical interface and a 3mm DFB mounted (a),
and an open burn-in box for four submount holders (b)
The pre-characterization setup, illustrated in figure 5.5, is designed to be compatible
both for MO and PA operation. In MO operation, the main output of the DUT DFB ((ii)
in fig. 5.5) is collimated with an aspheric lens with a focal length of 3 mm on a manual
xyz-stage. A channel in the diode mount and the arm of the lens holder allow the lens
5.1. LASER DIODE PRE-CHARACTERIZATION 63
Figure 5.5: Schematic of pre-characterization setup in PA characterization configura-
tion. (i) seed from PM fibre, (ii) DUT with current driver and temperature control, (iii)
voltmeter for RW photo current measurement, (iv) power meter with Ulbricht sphere,
and (v) fibre coupling for OSA.
to move freely and close in front of the diodes facet for beam collimation. The beam
is guided through an optical isolator and to a 50/50 non-polarizing beam splitter. The
beam splitter guides one half of the beam power to a photo detector in an Ulbricht
sphere for power monitoring ((iv) in fig. 5.5), and the other half of the beam to a
fibre coupling setup, which allows for spectral analysis simultaneously ((v) in fig. 5.5)
with an OSA. The power measured by the Ulbricht detector is proportional to the true
output power of the chip. The integrating sphere of the Ulbricht detector scatters the
input beam uniformly on its inside surface, compensating for varying input angles and
directions, and preventing saturation of the photo detector. This also allows for high
optical input power such as a PA output beam.
For PA testing, a seeding beam and in-coupling optics are necessary, as shown as (i)
in figure 5.5. The seeding beam is provided by one of the laboratories reference lasers
via a polarization maintaining optical fibre with an output power of 20 mW. An optical
isolator prevents the seed laser to receive optical feedback. Seeding an amplifier with a
beam guided through a fibre ensures a Gaussian beam profile that is independent from
the seed laser, and hence increases the reproducibility. Reproducibility of the seeding
beam is important to be able to compare the results of several amplifiers, because the
performance is strongly dependent on the seeding of the DUT. The seeding beam is
guided with adjustable mirrors and a lens on a xyz-stage, comparable to the collimating
lens at the output, into the RW facet of the DUT amplifier. The current through the
tapered section of the amplifier is turned off in coarse seeding beam positioning to
enable the monitoring of the RW photo current. The RW photo current allows for
monitoring of the seed beam coupling into the amplifier and is suitable for coarse beam
adjustment. It is measured with a multimeter, denoted as (iii) in figure 5.5.ARW photo
current of >6mA is aimed at for the amplifiers used in this work. Fine tuning of the
64 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
seeding beam coupling is required when the TA current is switched on due to thermal
deformation of the mounting setup and the diode during full power operation. The final
position of the in-coupling lenses is determined in active alignment by maximizing the
output power under working conditions.
The output beam of the DUT is collimated with the same lens used in MO operation,
but in addition, a cylindrical lens with a focal length of 40 mm for SAC is required
because of the differences in fast axis – slow axis divergence angles. The beam analysis
setup in PA operation is identical to the analysis setup in MO operation, optical output
power and spectrum can be recorded simultaneously.
5.1.3 Qualification Criteria for Laser Diodes
The MO and PA chip qualification procedure contains visual inspections and electro-
optical characterization of the chips performance, including the monitoring of the op-
tical output power at continuous operation and evaluation of the spectrum for various
injection currents. The following qualification criteria for DFB and amplifier chips were
established to reach the given performance within the given power budget of the exper-
iments, as defined in table 2.2. According to their different functions, the criteria for
DFBs and PAs have to be defined individually. The visual criteria are based on decades
of experience with semiconductor chips at the FBH.
5.1.3.1 DFB Qualification Criteria
Visual Inspection:
• Surface: no scratches or other defects in the area 5 times the ridge width around
the wave guide
• Facets: no particles and dirt, terraces, overlapping surface metal, coating releases,
disruptions or other defects
–within the area five times the ridge width in horizontal direction and three
times the ridge width in vertical direction, typically 11 µm×6.6 µm (active
area defect)
–exceeding the size of one fifth of the chip width, typically 80 µm (bulk defect)
as depicted in figures C.3 and C.4.
Power versus Injection Current Characteristics:
Measurement conditions: at 20°C chip temperature, current range 0mA ... 200mA.
• linear behaviour starting at threshold current, deviation from linear approxima-
tion <0.2dB1, jumps shall be related to mode hops
• minimum output power of 60 mW at 150 mA
Spectral Characteristics:
The targeted wavelength has to be reached at 150mA for a working point temperature
1equals a deviation of 2mW at 50mW total output power
5.1. LASER DIODE PRE-CHARACTERIZATION 65
in the range between 15°C and 40°C.
Measurement conditions: at 20°C chip temperature, current range 0mA ... 200mA.
• single mode behaviour, maximum two mode hops allowed
• no mode hops closer than ±50 pm to the nominal operating wavelength at the
WP setting
• the side mode suppression ratio (SMSR) to modes competing with the main mode
has to be >30dB (resolution bandwidth (RBW) 10pm) for currents >25mA be-
hind the threshold current. (no multi-mode behaviour)
• intrinsic linewidth below 100 kHz, FWHM linewidth below 3 MHz (10 µs)
Characteristics after Burn-in Phase
• degradation in power less than 3%
• spectrum or features in the spectrum have shifted less than 15 pm compared to
initial measurement
5.1.3.2 PA Qualification Criteria
Visual Inspection:
• Surface: no scratches or other defects in the area 5 times the ridge width around
the wave guide
• Facets: no particles, dirt, terraces, overlapping surface metal, coating releases,
disruptions or other defects
–within the area five times the ridge width in horizontal direction and three
times the ridge width in vertical direction, typically 11 µm×7µm, valid for
RW facets at the rear side of the chip (active area defect)
–within the area of wave guide width plus double of wave guide thickness
in horizontal direction times double of wave guide thickness in vertical di-
rection, typically 250 µm×5µm, valid for TA facets at the front of the chip
(active area defect)
–exceeding the size of one fifth of the chip width, typically 80 µm(bulk defect)
Power versus Injection Current Characteristics:
A critical value is the output power in seeded condition under realistic working point
temperatures. For PAs that should amplify emission at 780 nm an output power of 1 W
has to be reached while operating the diode at 35 °C ... 40 °C. 767 nm modules are
expected to be operated at room temperature, these amplifiers therefore have less re-
stricted temperature conditions.
Measurement conditions: at 20°C (767nm-PAs) and 30°C (780nm-PAs) mount tem-
perature, seeding power of 20 mW, RW current 200 mA, TA current range of 0..2500mA
• linear behaviour of output power for seeded operation, deviation from linear ap-
proximation <0.2dB2no thermal roll-over for TA injection currents <2200 mA 3
2equals a deviation of 50mW at 1000mW total output power
3PI=2200mA >PI<2200mA
66 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
• minimal output power of 1000 mW for 2000 mA TA injection current, seeded
operation
Spectral Characteristics
Measurement conditions: at 20 ° C (767 nm-PAs) and 30 ° C (780 nm-PAs) mount
temperature, seeding power of 20mW, RW current 200 mA, TA current range of
0mA ... 2500mA
• in seeded operation only spectral features of seeding oscillator observable, no
additional peaks
• SMSR to ASE background larger than 40 dB (RBW 10pm)
5.1.4 Results of DFB Diode Pre-Integration Characterization
In order to select the best suited DFB diodes for integration the qualification process
is applied to all integration candidates. The qualification process contains a visual
inspection of the surface and facets, and electro-optical (EO) characterizations. The
EO characterization includes an optical output power analysis and an optical spectral
analysis at various injection current settings. Optical power and spectrum are recorded
simultaneously. The criteria for qualification are given in section 5.1.3. Measurements
are performed at both 20°C in order to have comparable measurement conditions in
the following qualitatively analysis of the data, and at an expected working point tem-
perature that can be up to 40°C. Although the following analysis only refers to the
20°C-data, diode lasers that do not meet the qualification criteria at the working point,
e.g. they emit multi-mode, are rejected even if they meet the qualification criteria at
20°C.
In total, 130 DFB diode lasers were inspected for integration into laser modules.
The diode lasers were processed on four different wafers, some diodes were cleaved as
1.5mm long diodes, the others were 3 mm long, as listed in table 5.2. One wafer was
optimized for 767nm emission, a wavelength not used in the MAIUS project but its
predecessor QUANTUS II and presumably in subsequent experiments. The DUT-DFBs
not only differ by their wafer but also by their coating process run, which, in this case,
is individual per wafer. When referring to results of a certain wafer, this also includes
the results of a certain coating run. A distinct statement to the performance of a coating
run cannot be given independently of the wafer run.
Table 5.2 lists the number of DUTs and the number of incomplete data sets per
wafer and chip length. Not all data sets could be completed. In 16 cases the facet data
was missing but the EO characterization was complete. For these DUTs the facets were
inspected and no critical defects were detected, however, the facet conditions were not
documented, the diodes therefore are not counted in the facet evaluation. 7 DUTs only
provided facet data. None of these 23 diode lasers passed the qualification, the data set
therefore was not completed. For diodes which failed the facet inspection, the burn-
in process typically was aborted to save time and resources since the diode already
failed qualification. Diodes with facet defect were characterized only when the facet
inspection was performed or re-evaluated after burn-in, or if the defect was caused by
handling the chip during the qualification process, e.g. by touching a facet by accident.
5.1. LASER DIODE PRE-CHARACTERIZATION 67
Wafer Wavelength Chip Length Total Number Incomplete Data
C2610-6-2 767nm 1.5mm 26 2 (0; 2)
3mm 11 0 (0; 0)
C2213-6-2 780nm 1.5mm 6 3 (0; 3)
C2212-6-1 780nm 1.5mm 53 7 (1; 6)
3mm 24 3 (1; 2)
C1152-6-1 780nm 1.5mm 10 8 (5; 3)
Table 5.2: Overview of DFB integration candidates. Denotation in brackets: (incom-
plete facet data, incomplete electro-optical data)
In the following section the results of the visual inspection and the EO characteri-
zation after burn-in are presented.
5.1.4.1 Visual Inspection
The visual inspection included facet inspection and inspection of the galvanic surface
of chips. Scratches close to the surface of the ridge wave-guide or galvanic surface
damages were not observed on any of the DUTs. Therefore this criterion will be ignored
further on.
Appendix C shows a selection of facet photographs, divided into qualified (sec-
tion C.1) and not-qualified (section C.2) chips due to defects. As shown in section C.1.2,
there are facets with defects that are not critical. These facets also count as qualified.
According to section 5.1.3, facet defects can be classified as active-region-defects, as
shown in figures C.3a-b and as bulk-defects, as shown in figures C.4a-b. An example
of a facet defect arisen after a malfunction during handling the DUT is depicted in
figure C.5.
In total 107 chips passed the visual inspection of both the front and rear facet. This
corresponds to 82% of all integration candidates, or 87 % of all chips with available
facet data. The yield of the facet inspection is depicted in figure 5.6.
In 16 cases the facet was rated as ”failed”. The number of defects in the active
region was, with 9 cases, almost twice as large as the number of bulk defects (5
cases). It is noticeable that all bulk defects were found on DUTs of the same wafer, the
C2610-6-2 wafer, as shown in dark violet color in figure 5.7. The visual inspection of all
1.5mm chips with bulk defects (4 in total) showed that the bulk material was broken
on the edge as shown in figure C.4a, resulting in lose parts and large flaking of the
coating. These defects might have been introduced during cleaving of the single chips
from the bar. The 5th bulk defect was found on a 3 mm chip only after burn-in and
characterization, depicted in figure C.5. It is assumed that during handling of this chip
a macroscopic object, such as the collimating lens, touched the facet and destroyed it
completely. The destroyed facet corresponds to the decreased output power (41 mW at
200mA4) of this chip, disturbing the emission in the chip and hindering the photons to
leave the gain material via the damaged facet.
4An expected output power value for 200mA injection current would have been 90mW.
68 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
Figure 5.6: Total yield of facet inspection and failure causes
Figure 5.7: Yield of facet inspection for each wafer and length
The active-region-defects, on the other hand, are mostly distributed on wafer
C2212-6-1, with 6 chips of 1.5mm length and one chip with a length of 3 mm, mostly
caused by particles close to the ridge area. The wafers C1152-6-1 and C2213-6-2 were
only represented with 6 or less candidates in the facet inspection, not allowing for a
statistical analysis.
5.1.4.2 Electro-Optical Pre-Integration Characterization
The EO characterization evaluates output power and spectrum before and after a 120h
burn-in phase. Pass and fail criteria, as well as measurement conditions are listed in
section 5.1.3. 65 DUTs passed the EO evaluation, which corresponds to 50% of all 130
available diodes, and to 57% of all DFB diodes with available EO data.
49 diodes failed the EO characterization, this is 38% of the total number of diodes
and 43% of all diodes tested. The distribution of rejection causes is depicted in fig-
5.1. LASER DIODE PRE-CHARACTERIZATION 69
ure 5.8. Figure 5.9 shows the yield distribution per wafer.
Some of the rejection causes are specific to certain wafers, such as for the wafers
C2213-6-2 and C1152-6-1 with 100% and 43% failing rate, respectively, due to non-
matching wavelength ranges. For both wafers, the emission wavelength at room tem-
perature is too long compared to the wavelength aimed at in the experiment. To com-
pensate the wavelength offset, the diodes have to be cooled down. If the aimed wave-
length cannot be reached at 150mA and 15°C, the diode is rejected for the risk of
damage due to water condensation.5
Figure 5.8: Causes of failure in EO characterization
Figure 5.9: Yield of EO characterization for each wafer and length
However, the wavelength of the diodes of C2212-6-1 are slightly too short compared
to the targeted emission wavelength. The blue-shift of the emission wavelength results
5At 25 °C room temperature and 50% humidity, water condensates on 14 °C cold surfaces. This con-
dition has been present in the lab during autumn and spring. 1 K is a safety margin.
70 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
in the necessity to heat up the diode to reach the designated Rb D2 transition wave-
length. 3 out of 49 of the 1.5 mm diodes have to be heated up to more than 40 °C in
order to emit at the designated Rb wavelength. MIOB operation temperatures higher
than 40°C are an issue for the power amplifier, because the output power decreases
with increasing operating temperature.
Spectral non-stabilities, such as multiple mode hops, multi-mode behaviour and
frequency drifts after burn-in, are the main reason for rejection, with a total of 76%
of all rejected diodes, as shown in in figure 5.8. 26% of all rejected diodes showed a
significantly different emission spectrum after 120 h of burn-in operation, in a way that
available mode hops have shifted to other injection currents, in some cases multi-mode
behaviour appeared for certain injection current ranges. This effect can be observed
for all remaining wafers. For 35% of all rejected diodes it was not possible to iden-
tify a working point that provided a tuning range of ±50pm in single-mode emission
around the targeted wavelength. 15% of all rejected diodes showed current ranges
of 50mA and higher with SMSR smaller than 30dB (RBW 10pm), resulting in multi-
mode behaviour for these current ranges. All of these diodes were processed on wafer
C2212-6-1. Multi-mode behaviour is caused by weak mode selection in the resonator
which may be caused by irregularities in the grating or on the facets. Two chips failed
the qualification process due to power degradation, one from wafer C1152-6-1 and the
other one from C2610-6-2.
5.1.4.3 Summary Pre-Integration DFB Characterization
Figure 5.10: Total yield of pre-characterization of DFB diodes.
The overall yield of the qualification is depicted in figure 5.10. A chip failed the
qualification if either the facet or the EO performance could not be qualified. In total,
43% of all 130 characterized chips passed both evaluation steps, the visual inspection
and the burn-in with characterization of the EO performance. Ignoring the chips with
incomplete data sets, the yield of DFB diode qualification is 47%.
The rejection rate of 53% of the tested diodes underlines the necessity of the pre-
characterization process before integration. Removing diodes from the MIOB comes
5.1. LASER DIODE PRE-CHARACTERIZATION 71
with the risk of damaging the MIOB due to scratches in the metal finish, and in-
creases the production costs. Therefore integrating a non-qualified laser diode should
be avoided.
20 laser modules were integrated during this thesis. Due to this selection process,
only one case occurred where the DFB diode had to be replaced. The replaced diode
was a 3mm C2212-6-1 diode that became multi-mode after integration. Probably, the
thermal stress applied during integration and wire-bonding damaged the diode or trig-
gered the development of internal defects that were undetected until then. The diode
was replaced with a 1.5mm diode from the same wafer. The module operates faultless
ever since.
5.1.5 Results of Power Amplifier Pre-Integration Characterization
Goal of the qualification process of the PAs is the identification of diodes with decreased
life time expectancy, or diodes that are not capable to deliver the required output power
in the given temperature range. The visual inspection and burn-in process is carried
out to find defects that decrease the life time, the EO characterization reveals whether
the performance of the DUT fulfils the requirements. The qualification criteria can be
found in section 5.1.3.2.
In total, 100 power amplifiers were characterized, summarized in table 5.3. The
71 amplifiers for 780nm are all provided by the same wafer, C2358-3. They are 4 mm
long with a 2mm RW mode filter section and a 2mm long tapered section with an
opening angle of 6°. In addition 29 diodes for 767 nm amplification were characterized.
They were processed on two different wafers, C1030-3 and C2850-3. Diodes extracted
from wafer C1030-3 provide only 1mm of RW mode filtering section and 3mm of
tapered amplifying section. The remaining diodes have the same design as the diodes
for 780nm. The diodes were either mounted with p-side facing the submount for better
thermal conduction to the laser module, called "p-down" mounting and depicted on the
right of figure 5.11, or with the p-side pointing away from the submount, called "p-up"
mounting, see left picture in figure 5.11. P-up mounted diodes host an additional heat
spreader on top, made of copper-tungsten (CuW) for a better matching of thermal
expansion with the GaAs diode substrate.
Figure 5.11: Mounting of the PAs: either p-up with heat spreader (left) or p-down with
the submount as heat sink (right)
As denoted in table 5.3, the rate of incomplete data for certain amplifier types is
with 30 of 58 diodes (780nm, p-down) and 7 of 10 (767nm, 2 mm long RW, p-down)
72 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
Wafer Wavelength RW Length Mounting Total Incomplete Data
C1030-3 767nm 1mm p-up 12 1 (0; 1)
p-down 3 1 (0; 1)
C2358-3 780nm 2mm p-up 13 4 (3; 1)
p-down 58 30 (2; 28)
C2850-3 767nm 2mm p-up 4 1 (0; 1)
p-down 10 7 (0; 7)
Table 5.3: Overview of PA DUTs. All testes chips are 4mm long. Denotation in brackets:
(incomplete facet data, incomplete electro-optical data)
very high. In most cases the data of EO characterization is missing. This is due to the
fact that diodes that failed the visual inspection were not characterised electro-optically
to save resources. In addition, the electro-optical characterization was stopped after
enough amplifiers for integration were qualified.
5.1.5.1 Facet Inspection
The visual inspection includes the inspection of the surface of the diodes as well as
the evaluation of the front and rear facet. As for the DFBs, no defects on the surface
of the DUTs could be observed, further discussions therefore are focused on the facet
inspection. Examples of qualified PA facets (figure C.1c-d) and facet defects (figures
C.3c and C.4c) are given in the appendix C. Again, the facet defects are divided into
bulk area defects and active are defects.
65% of all PA-DUTs passed the facet qualification, as shown in figure 5.12. This is
less than the qualification rate of the DFB diodes (82 %). One reason for the decreased
qualification rates for amplifiers is the special mounting. PAs are mounted either p-
side-up with a heat spreader on top of the p-side or p-side-down on the submount,
as shown in figure 5.11. Both mounting concepts include a soldering process on the
p-side close to the active area, combined with a higher risk of damage of the facet
due to overflowing solder, stress close to the active region of the facet and increased
handling of the chips. In addition, the active area of the output facet of the tapered
power amplifier (TPA) is enlarged due to the tapered waveguide, increasing the risk of
active area defects.
However, p-up mounted chips passed the facet inspection with a rate of 85 % of all
diodes with available facet data (76% of all p-up mounted diodes), as shown in figure
5.13. This rate is slightly lower than the visual qualification rare of the DFBs (82%,
all diodes) and indicates that the p-down mounting increases the risk of damaging the
facet.
This tendency can also be observed in the facet failure rates. 15 % of the p-up
diodes failed the facet inspection, compared to 12% failed DFBs, but 38 % of the p-
down mounted amplifiers failed, resulting in a total failure rate of 30 % for amplifier
chips.
5.1. LASER DIODE PRE-CHARACTERIZATION 73
Figure 5.12: Total yield of facet inspection of PA diodes.
Figure 5.13: Distribution of facet defects due to mounting configuration.
Active area defects and large area defects are equally distributed. Active area dis-
turbances were mostly caused by particles or solder splashes at the output facet, as
depicted in figure C.3c. These failures disturb the beam profile and may decrease the
lifetime of the diode due to local heat development on the particles. Main reason for
large area defects is the formation of terraces on the coating on wafer C2358-3. In
small numbers other problems with the coating, namely an inhomogeneous colouring
of the coating of a facet and shadowing of facets during coating occurred, as well as
defects due to handling and processing of the coated chips, such as damages caused
by contact with the facet, lose bulk material that are broken out under cleaving, and
solder and dirt. All these defects can decrease the life time of the chip or the integrated
module.
5.1.5.2 Electro-optical Pre-Integration Characterization
61 out of 100 amplifiers were characterised electro-optically. 77% of the characterized
diodes (47% of all available PA diodes) passed the EO qualification. The remaining
23% of the characterized amplifiers could not provide the requested optical output
74 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
power at the designated working point temperature. For these diodes thermal roll-over
appears for TA injection currents <2000 mA, as shown in figure 5.1b.
(a) (b)
Figure 5.14: PA submounts after shearing of the diode. Areas where the solder did not
wet the diode surface are marked in red. On the left: Bonding without Ar-cleaning pro-
cess, wettability: 40 %. On the right: submount was cleaned with Ar before bonding,
wettability: 95%.
As reason for low optical output power and thermal roll-over for relatively small in-
jection currents bad thermal coupling between amplifier chip and submount could be
identified. When shearing a not qualified PA laterally from the submount, wettabilities
of the solder down to 40% of the diodes surface were detected. For good thermal con-
tact a wettability of >90% of the surface is required. An example of poor wettability is
shown in figure 5.14a. The figure shows a submount and the remains of the diode after
shearing off the diode. Areas where the solder wetted the submount surface remained
on the submount and can be detected either as shiny golden areas (golden galvanic
finish of the diodes) or black areas (bulk material of the diode). In figure 5.14a the
areas with no remains of the diode are marked red. In order to increase the wettability
of the solder, the bonding temperature was increased by 10 ° C and the surface of the
submount was cleaned in an ultrasonic bath with argon directly before bonding of the
diode. The argon treatment removes organic material and oxidations that constrains
the wetting of the solder on the diodes surface. The increased temperature increases
the responsiveness of the materials. The shearing result of a submout with increased
bonding temperature and argon treatment is shown in figure 5.14b. The wettability
of the solder could be increased to approximately 94%. Diodes that were bonded to
submounts after argon treatment showed an improved thermal conductivity between
chip and submount and fulfilled the requirement of optical output power.
For characterizing the amplifier, the mount temperature of the structure holding the
submount was stabilized to the temperature named in the measurement conditions. For
some amplifiers, the actual submount temperature was recorded during characteriza-
tion as well, allowing conclusions of the thermal behaviour of the amplifier indepen-
dent of the thermal conductivity of the measurement setup. In figure 5.15 the optical
output power at ITA =2000 mA is plotted against the temperature of the submount
5.1. LASER DIODE PRE-CHARACTERIZATION 75
Figure 5.15: Output power dependent on submount temperature for various PA chips
of the PAs. The blue line marks the output power qualification criterion of 1W. The
evaluation includes diodes with both RW-section lengths, 1 mm and 2 mm, and both
mounting options, p-up and p-down. According to figure 5.15 p-down mounted am-
plifiers with 2mm RW section (illustrated as blue and green dots) reach higher output
power values for comparable submount temperatures than p-up mounted amplifiers.
The p-down mounted diodes show a linear decrease of output power for increasing
submount temperatures with -26mW/K (black line), independent of the wavelength
they amplify. At 40°C 1250mW can be achieved, with a deviation of 4.3 mW/K. The
black dotted line shows the linear fit of p-down diodes with an offset of +30mW. The
dotted line captures all but two measurement points of the p-down values that show
up underneath the linear fit. Probably the wettability of the solder to the submount
of these two diodes is not as high as for the other p-down diodes in this sample. As
described above decreased wettability results in decreased thermal coupling and thus
increased thermal resistance and decrease in optical output power.
The p-up mounted diodes of the 780nm 2mm long RW-section chips are illustrated
with orange squares. All of them, with one exception, are below the power-temperature
line estimated for the p-down mounting, indicating that p-down mounted amplifiers are
able to emit more power than p-up-mounted diodes for the same temperatures. This
can be explained by the limited thermal capacity of the heat spreader on top of the
diodes. However, the distribution of the p-up mounted diodes is scattered, a meaningful
prediction of the power-temperature behaviour could not be found based on this data.
The data points of the p-up mounted 767nm amplifiers with 1mm RW-section (in
pink), however, decrease with 20.5mW/K (±2.8mW/K) and show an offset of ap-
proximately -400mW compared to the p-down mounted PAs. The offset can both be
explained by the mounting and by the increased length of the TA-section. A longer ac-
tive area results in decreased carrier density for identical injection currents, resulting
in decreased current-to-optical power conversion efficiency. The p-up mounting, again
76 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
increases the thermal resistance of the device, resulting in lower output power.
All characterized diodes that emitted more than 500 mW optical output power ful-
filled the spectral qualification criteria of ASE suppression and distribution of the seed-
ing beam spectrum. For the other diodes, having problems with the thermal conduc-
tion, the spectral behaviour was not evaluated because the output power criteria could
not be fulfilled.
5.1.5.3 Summary Pre-Integration PA Characterization
Figure 5.16: Total yield of the PA selection process
34% of the 100 available amplifier diodes passed the qualification procedure. The
main reason for rejection was defects detected during facet inspection with an error
rate of 30% (compared to 12% for DFBs). This can be explained with the different
mounting concept for PAs in order to increase the thermal conduction in the diode
at the cost of higher defect rates of the facet. The rejection rate of the PAs after EO
characterization, on the other hand, is only 14 % of all diodes, and with that small
compared to 36% for the DFB diodes. This is due to the fact that PAs are not expected
to change spectral behaviour and therefore the qualification criteria for EO qualification
of the PAs are less tight than for MO diodes.
The 22% of chips with incomplete data are chips that passed the facet inspection
but were not characterized electro-optically because the required number of amplifiers
for integration into laser modes was reached. Due to this, the total yield of the PA
qualification cannot be directly compared to the DFB diode qualification yield. If the
distribution of qualification of the 22 not-characterized diode equals the qualification
rate of the other 48 diodes that run through the EO characterization, than a total yield
of 50% for PAs could have been reached. If so, the yield of both the DFB and the PA
qualification process would in the range of 50%.
The PA qualification process revealed problems in the diode assembly that led to
an adaptation of the assembly process. After the adaptation of the assembly process
the yield of qualification improved. With that the qualification process could satisfy a
controlling function of the production processes.
5.2. PASSIVE COMPONENT PRE-INTEGRATION CHARACTERIZATION 77
5.2 Passive Component Pre-Integration Characteriza-
tion
Additionally to the semiconductor chips, the optical micro-isolators, lenses, the MIOBs
themselves and the CCP are qualified before integration.
Micro-Isolators The qualification of the isolators includes an examination of the
facets for dirt or defects, as well as an optical characterization of the isolation, the
polarization extinction ratio (PER), the insertion loss and the rotation angle of the out-
put beam compared to the input beam. The purpose of the optical characterization is
to confirm the optical performance as it is defined in the data sheet. The characteriza-
tion became necessary after it was discovered that some isolators came with a wrong
waveplate, resulting in a poor PER and a rotated output polarization. Single stage and
semi-double stage isolators have been tested, both for 767 nm and 780 nm. A semi-
double optical isolator provides a better isolation on the cost of a higher insertion loss,
compared to a single stage optical isolator.
For the isolator characterization, a specific measurement setup was designed with
defined linear polarized light passing the DUT isolator. The DUT is placed on a
V-grove between a polariser and analyser device, formed by two polarizing beam split-
ters mounted on motorized rotational stages. The optical power of the beam, passing
the setup, is detected with a photodiode. Depending on the alignment of the polariser
and analyser, and the orientation of the DUT, the isolation, PER, and rotation angle are
determined.
An isolator passes the visual test, when the facets are free of dirt or scratches. To
pass the optical characterization, single stage isolators have to provide isolation larger
than 27dB with an insertion loss less than 5dB. The isolation of a semi-double isolator
has to be larger than 50dB with the insertion loss not exceeding 8dB. The PER has to
be larger than 27dB, and the rotation angle of the output beam shout not exceed 10 °.
The qualification criteria are based on the data sheets and the experience obtained with
the micro-isolators.
In total, 198 micro-isolators were characterized. Visual checks show that 31% of the
DUTs facets were affected by dirt. In some cases, however, the dirt could be removed
without touching the facets by turning the isolator upside-down or by blowing it off
with a stream of pure nitrogen. The optical characterization is summarized in table
5.4.
The average isolation of a single stage isolator and semi-double isolator was mea-
sured to be 38.7±6.5dB, and 57.1±5.5dB, respectively. With that, the isolation of the
single stage isolators were better than expected (30 dB). As expected semi-double stage
isolators show a significantly higher isolation than single stage isolators, the measure-
ment of the isolation of the semi-double isolators where limited by scattered photons in
the measurement setup. The isolation performance of 767nm isolators does not differ
significantly from 780nm. However, one batch of single stage isolators for 767 nm had
to be rejected due to significantly less isolation (23.9dB compared to 41.9 ±2.6 dB for
qualified batches of 767nm single stage isolators). The cause of this failure could not
78 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
isolator type isolation [dB]ins. loss [dB]PER [dB]pol. rotation [°]
SiSt 780nm 39.7±5.2 3.5±0.4 32.7±4.5 3.1±4.0
SDSt 780nm 56.8±4.7 6.3±0.3 34.1±6.7 3.1±4.4
SiSt 767nm 38.1±7.7 4.0±0.5 33.0±8.8 4.1±3.4
SDSt 767nm 54.9±3.2 7.7±0.6 32.6±4.5 3.5±2.6
Table 5.4: List average performance of micro-isolators, classified into wavelength:
767nm and 780nm, as well as in design: Single stage (SiSt) and Semi-double stage
(SDSt)
be determined. Still, identifying this batch as faulty proves the necessity of the optical
qualification of the isolators.
In terms of insertion loss, the 767 nm isolators show higher absorptions of the in-
coming beam than the 780 nm isolators. This can be explained with the material prop-
erties of the thin film plates inside the isolators. They were once optimized for higher
wavelength (>1000nm), resulting in the effect of increasing absorption for decreasing
wavelength of the input beam. In contrast, isolators tested with 1070 nm input beam
showed an insertion loss of 2.4±0.2dB.
The average PER value of the tested isolators, independent of wavelength and de-
sign, is consistently larger than 32dB, as shown in table 5.4. However, there were
in total only 7 isolators with a PER less than 27dB. 5 of those were characterized
with a different wavelength than specified. When using a wavelength other than the
specified wavelength of the device, e.g. 767 nm for 780 nm isolators, the PER drops
to approximately 17dB. This behaviour is expected due to the wavelength dependent
characteristics of the optical materials of the isolator.
In average, the absolute value of the rotation angle of the polarization was
3.1±3.8°. 767nm isolators have a slightly higher excursion of the rotation angle,
compared to 780nm isolators. 7% of all isolators rotated the polarization more than
5.5°. The uncertainty of this measurement was <1°. Again, the cause of the different
beam rotation angles might be due to varying alignments during production.
In summary it can be stated that the performance of 767nm isolators is constricted,
compared to the 780nm isolators. This can be explained with the wavelength depended
performance of the materials used in the isolator design. However, the isolator’s isola-
tion is not affected and they are generally suitable for integration into micro-integrated
laser modules emitting at 767nm. The results of the single isolators allowed to iden-
tify a faulty batch that was returned to the manufacturer. 54 % of the revised micro-
isolators passed the qualification. The majority of isolators that did not apply to the
qualification criteria in the optical test were either in the faulty batch or were char-
acterized with a wavelength other than the specified wavelength. The visual check
allowed the improvement of the facets of the isolator before integration. Still, 30% of
the isolators failed the visual qualification of the facet.
5.3. CONCLUSION OF PRE-INTEGRATION QUALIFICATION 79
Micro-Lenses The lenses are only examined visually for dirt, scratches or flaking at
the edges. None of the lenses failed this test before first integration, but some lenses had
to be replaced after their facets were contaminated with adhesive during the integration
process.
Micro-optical Bench The MIOB itself has to be checked electrically and visually be-
fore integrating the submounts. The benches are cleaned either with pure nitrogen or
with vacuum tweezers to avoid damage on the bonding or gluing areas. The electrical
interfaces are tested for their input resistance and the connection to the bond pads.
All 30 MIOBs were tested. One MIOB was rejected for not providing electrical contact
of the pre-amplifier section. This MIOB was replaced by the manufacturer. All other
benches passed the test.
Mechanical Interface The planarity of the top and bottom surface of the mechanical
adapter CCP is fundamental to avoid mechanical stress and misalignment of the optical
components. The fail criterion was therefore defined as a maximal bow of 10 µm over
the entire length of the CCP. All CCPs are characterized with a non-contact, optical
profilometer Zygo NewView 6300. After adjusting the manufacturing process the CCPs
showed an average bow of 3µm of the bottom site and 1.4µm bow on the side carrying
the MIOB. All CCPs could be qualified for usage with MOPA modules.
5.3 Conclusion of Pre-Integration Qualification
This chapter shows that a pre-integration qualification process is highly recommended
to increase the yield of module production and decrease costs and working hours. With-
out the qualification, the bonding problems for the PAs would have been undetected
until after the first chips had been integrated onto the MIOB. In that case the MIOBs
hosting a PA with poor thermal contact would have turned into rejections, negating the
work of MO collimation and increasing the demand on the customized MIOBs. With a
yield of around 50% each for DFBs and PAs, the number of modules that would have
to be built in order to fulfil the delivery demands, would have quadrupled, since some
of the defects would have affected the performance of the module only after some time
of operation. Possibly, some modules could have degraded over time or failed spon-
taneously in the customers experiment, increasing the assembly time or stopping the
experiment.
However, the measurement setup, described in section 5.1, revealed problems with
the handling of the heat dissipated by the amplifiers into the submounts. Hence, the
temporary mount should be redesigned in order to guarantee a larger contact surface
with the submount. The mechanism, holding the submount in place, has to be strong
enough to withstand the resilience of the bond wires but may not touch or puncture
the gold (Au) wire bond surfaces. With a submount size of 2 mm width and 150 µm
distance of bond Au to the edge, this is a challenge. Adjustment pins can be used to
guarantee a reproducible positioning of the mount in the setup for each measurement.
A positioning robot can help to reproduce the position of the in-coupling lens of the
80 CHAPTER 5. PRE-INTEGRATION COMPONENT CHARACTERIZATION
seed laser, an automatized robot may optimize the seeding for each measuring point,
based on the output power of the amplifier.
Although the chip setup has potential for optimization, as shown above, it served
its purpose for the MAIUS mission.
Also the qualification of the passive components revealed variations in the perfor-
mance due to production uncertainties and failures, as seen for the micro-isolators. The
qualification of all components is therefore recommended.
Chapter 6
Characterization and Qualification
In the course of this thesis, 20 modules were assembled, designed and integrated as
described in chapter 3 and 4. Out of these, 18 modules, of which 5 modules host a
master oscillator only, have been delivered to project partners to be used in various
experiments such as the missions described in chapter 2. This chapter deals with the
individual performance and behaviour of these laser modules. The electro-optical char-
acteristics describe CW behaviour at working point operation parameters, as optical
output power, spectrum, linewidth, and modulation capabilities. The environmental
characterization compares the performance of the laser modules after an intense me-
chanical and thermal stress test, relevant for the operation of the laser modules in the
field and in space.
6.1 Electro-optical Characterization
The electro-optical characterization results describe the behaviour of the laser modules
for certain operation temperatures and injection currents. With that the suitability of
the laser modules according to the experiment’s requirements can be verified and the
performance of the laser modules can be predicted.
6.1.1 Optical Output Power
The output power of a MOPA is dependent on the injection current applied to the am-
plifier section, the temperature of the device and the seeding power of the master oscil-
lator. Figure 6.1 plots the output power against the injection current into the amplifier
section for various operating temperatures. The MO input power corresponds to 25 mW
behind the isolator. For all temperatures, the optical output power increases exponen-
tially for an input current <750mA, then develops linearly with 0.95W/A (35°C) up to
1.16W/A (15°C). For currents >2500 mA an increasing thermal resistance (see equa-
tion 3.15) results in a minimized increase of the optical output power. However, no
thermal roll-over could be observed for amplifier injection currents <3500 mA. Even
for 30°C an optical output power of >3W can be achieved.
82 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
Figure 6.1: Relation of output power and MIOB temperature against amplifier input
current (left axis), and conversion efficiency for 25°C (right axis)
The thermal dependency of the output power can also be derived from figure 6.1.
At 2000mA the power decreases linearly with 20mW/K.
(a) (b)
Figure 6.2: Optical output power of MOPA module (a) and amplifier gain (b) against
MO optical output power.
The green graph belonging to the right axis in figure 6.1 shows the conversion
efficiency of the MOPA system at 25°C for various TA injection currents. The maximal
efficiency is reached at 2000 mA with 29% and decreases when thermal resistance in
the amplifier increases, affecting the power performance. The power consumption of
the MO and RW section is included in this calculation. The power consumption of a
MOPA can be estimated by the injection current and voltage over the semiconductor
section, including the electrical interface. The MO voltage typically is in the range of
2.3V, the RW voltage typically is around 2.0 V and the TA voltage is around 2.4V. With
6.1. ELECTRO-OPTICAL CHARACTERIZATION 83
a maximum of 200mA for the MO and RW section and 2000mA at the TA section, the
power consumption of a MOPA typically will be around 5.7 W.
The dependency of the output power and the seeding power is shown in figure 6.2a.
The blue curve depicts the output power of the MOPA system versus the output power
of the DFB behind the optical isolator, corresponding to the seed power of the amplifier,
the red curve is the injection current of MO diode. The power emitted for IMO ≤55mA
is the ASE output of the amplifier. 55mA is the threshold current of the DFB in this
module. The maximal gain of the MOPA is 24.1dB. The gain saturation is reached
for an input power of 5.2mW with a saturation power of 1577mW. In figure 6.2b,
depicting the gain of the amplifier, the ASE background of 850mW was subtracted in
order to show the gain caused by the input power of the master laser. However, this
is just an approximation, because the ASE background also decreases with increasing
seed power.
The required optical free-space output power, as defined in table 2.2, can be
achieved also for WP temperatures >35°C and for PA injection currents <2000 mA
within the defined power consumption boundaries.
6.1.2 Optical Spectrum
The optical spectrum of a MOPA is dependent on the MO current and the MO tem-
perature. Since the laser modules are designed to operate at a certain wavelength,
the working point, the combination of IMO and TMO which allows for operation at the
aimed wavelength, has to be defined.
(a) (b)
Figure 6.3: Wavelength of a MOPA, hosting a 3 mm long master laser versus (a) MO
current and (b) MO temperature. The spectra are normalized to the global peak of all
emission spectra in the figure.
Figure 6.3a shows a false colour plot of the emission spectrum versus the MO injec-
tion current of a 3mm long DFB diode. It was recorded with an Advantest Q8384 OSA
with a resolution bandwidth (RBW) of 10pm (FWHM) and calibrated with a High Fi-
nesse WS-7 wavelength meter. The spectra in the figures are normalized to the global
84 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
peak of all emission spectra. Red colour indicates high power spectral intensity, val-
ues below –50dB are shown in white. There were no values taken in the range of
zero to 30mA since the threshold current of the MO in this laser module is >30 mA.
This protects the fully pumped amplifier (with IRW =200mA and ITA=2000 mA) from
overheating, since the carriers in the PA cannot be cleared without an optical input.
The dashed line indicates the desired emission wavelength, in this case 780.241 nm
as defined by the Rb D2 line. This emission wavelength can be provided by the laser
module. The laser module shows stable single mode behaviour, no mode hop can be
observed. The wavelength shifts with an average of 1.2 nm/A. However, a DFB laser
with a 1.5mm long resonator shows a current tuning of 2.7nm/A due to the higher
carrier density in the shorter active region.
Figure 6.3b shows the dependency of the wavelength on the temperature of the
optical master oscillator on the same laser module. MO and PA currents are kept sta-
ble at the designated values. A linear tuning of 0.053 nm/K can be observed for this
3mm DFB laser. A 1.5mm DFB laser provides with 0.056 nm/K a temperature tuning
coefficient that is in the same range as the temperature tuning coefficient of a 3mm
diode. With that, the length of the diode has only a minor influence on the temperature
tuning.
Since the operating temperatures of the MOPA are limited by the dew-point of the
environment (typically 15° C) and by the thermal roll-over of the output power, a
typical temperature range of 15°C – 40°C allows for approximately 1.4nm thermal
wavelength tuning. The current tuning range is limited in order to decelerate the ageing
process of the laser chip, and contributes to a longer life time of the laser. Hence, the
MO current tuning range in the MOPA module typically spans 150mA for a 1.5mm DFB,
which corresponds to 0.4nm current tuning range, and 300 mA for a 3 mm DFB, which
corresponds to 0.36nm current tuning range. With that, the total spectral tuning range
of a DFB laser module is approximately 1.8 nm or 880 GHz for 1.5 mm short laser chips
and approximately 1.7nm or 740GHz for 3mm long chips. These values do not differ
much, since the smaller current tuning coefficient of 3 mm DFBs is compensated with
a larger current range that can be applied to the chip due to the longer active area in
order to reach the same energy density in the active area. The requirement of a tuning
range >100GHz could be achieved with current tuning alone. However, the enhanced
tuning range due to thermal tuning also guarantees that the desired wavelength can
be provided in spite of production variances of the DFB laser wafers and stimulation of
competing optical modes in the laser’s resonator.
The amplifier currents have no significant influence on the optical spectrum as long
as the amplifier is saturated. Without an optical input, the PA emits non-coherent ASE
with a peak emission at its gain maximum. When saturated, the ASE background of the
PA is reduced and the MO signal is amplified. Figure 6.4a shows the single spectrum
for IMO =0mA and IMO =270mA. In the blue curve, the broad ASE output is clearly
visible but is significantly reduced when the MO signal is injected. However, the red
curve also reveals that the gain maximum of the amplifier at 787nm is red-shifted
compared to the carrier signal at 780.241nm. This causes an increased ASE emission
because carriers at the gain maximum wavelength are not entirely cleared. The ASE
power will not exceed -40dB suppression compared to the carrier emission. In the
6.1. ELECTRO-OPTICAL CHARACTERIZATION 85
(a) (b)
Figure 6.4: (a) Comparison of spectral output with (red) and without (blue) input
from the master oscillator. (b) MOPA spectrum calibrated, optical power in mW vs.
wavelength.
range of ±1nm around the MO carrier emission, the SMSR is even at 50dB. Emission
outside this range is not of interest due to the nature of quantum experiments as the
atoms will not be influenced by these wavelengths. The single-spectrum measurement
in figure 6.4 allows for the estimation of the total ASE power when the total emission
power is known, which in this case was 945 mW. Figure 6.4b shows the same curve but
calibrated in mW. It has to be mentioned that the maximum peak power shown in this
plot does not reflect the actual beam power because the linewidth of the MO is much
smaller than the resolution of the OSA, see section 6.1.3. However, the calibration is
based on the assumption that the integrated power of the spectrum matches the power
measured with the power meter. With the calibrated power in mW, the ratio between
the total ASE emission power in the wavelength range between 770 nm and 800 nm
and the peak emission power can be estimated. The ASE emission adds up to be in
total 16mW, the ratio between peak power and ASE power is 17.5 dB. With that, the
ASE is clearly not dominating the laser module’s emission and can be neglected.
6.1.3 Linewidth
The laser linewidth, introduced in section 3.1.3.4, indicates the spectral stability of a
laser. A detailed analysis of the spectral stability can be done by measuring the fre-
quency noise power spectral density (PSD) of the laser and is described in [111,112].
Based on the PSD of the frequency noise, the FWHM of the linewidth and the white
noise floor of the linewidth can be determined. From that the technical linewidth,
based on the FWHM value, and intrinsic linewidth, based on the white noise level, can
be calculated.
6.1.3.1 Measurement Setup
Various methods exist to measure the linewidth of lasers. Fabry-Perot interferome-
ters [113]and beat note measurements [114]are the most common techniques. How-
86 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
ever, both suffer from the problem of frequency drifts that are typically larger than the
linewidth of the lasers on the time scales that are typical for linewidth measurements.
In optical frequency discriminators the frequency resolution is dependent on the long-
term frequency stability of the laser, requiring active frequency stabilization in order
to be able to measure the linewidth. However, frequency stabilization affects the PSD
for frequencies smaller and equal the servo bandwidth of the servo loop, distorting the
frequency noise analysis for low Fourier frequencies.
In the heterodyne measurement method, the laser beams of two comparable lasers,
the device under test (DUT) and the optical local oscillator (OLO), are mixed on a high
speed photo diode, the photo diode signal then is processed by a RF spectrum analyser,
as shown in figure 6.5a. However, RF spectrum analysers cannot cope with frequency
drifts of the carrier frequencies of both lasers. The spectrum at various RF frequencies
is recorder sequentially, resulting in the problem that different parts of the RF spectrum
belong to different beat note carrier frequencies. With that, the RF spectrum analysis
shows systematic deviations. In addition, amplitude noise can be misinterpreted as
frequency noise.
(a)
(b)
Figure 6.5: Measurement setup for linewidth measurement: Beat note with frequency
difference capable for photo diode (PD) detection is created either with (a) lasers DUT
and OLO or (b) self-delayed with a DUT, modulated with an AOM, first order output of
AOM is delayed with a long fibre. PD signal is analysed by signal analyser (SA) which
delivers IQ-data.
For beat note measurements, the self-delayed heterodyne (SDH) technique [115]
is an approach to overcome frequency drifts. With the SDH technique only the DUT
laser is required, as depicted in figure 6.5b. An acusto-optical modulator (AOM)1splits
the DUT laser beam, shifting the frequency of the first diffraction order beam output
1We use the AOM Intra Action ATM-804DA2B
6.1. ELECTRO-OPTICAL CHARACTERIZATION 87
by several MHz 2. One of the beams is delayed by a long fibre in order to provide
an incoherent reference beam for the analysis, both beams are mixed again on a high
speed photo diode. The measured frequency noise PSD values have to be divided by
2×(1−cos(2πf·Tdela y )) in order to compensate for the superposition of two electrical
fields on the photo diode [116], and to obtain the laser’s PSD. That requires the exact
knowledge of Tdela y and excludes frequencies from the analysis which are integer mul-
tiples of 1/Tdela y due to division by zero. This, again, leads to an incomplete frequency
noise PSD analysis. With a L=2 km long polarization maintaining fibre as delay line
and a refractive index nof 1.5 for the fibre material the Tdela y =nL/cwould be 10 µs.
A resolution limit for SDH measurements is the inverse delay time, resulting in a limit
of 100kHz. However, for the determination of the technical and intrinsic linewidth
values, these limitations are not relevant.
Instead for recording a PSD with a spectrum analyser to estimate the white noise
floor and to record the beat note spectrum separately in order to estimate its FWHM,
we chose the I-Q data approach described by Schiemangk in 2014 [112]. Only a single
measurement is required to acquire a data set from which both the phase noise and the
beat note spectrum for various integration times can be reconstructed. Also the misin-
terpretation of amplitude noise can be avoided. A heterodyne measurement setup, as
depicted in figure 6.5, can be used without frequency stabilization being necessary. In
addition, the I-Q method allows the linewidth characterization over all frequencies of
the aimed bandwidth of free-running laser modules when using a heterodyne setup, as
shown in figure 6.5a.
In our measurements a beat note signal was generated by superimposing the beams
in an x-coupler either of an OLO and the DUT, or self-delayed, which then are sent
to a fast photo detector (New Focus NFO-1554-B). The voltage of the photo diode
corresponds to
Vbeat (t) = V0(t)·cos(∆ωt+ϕ(t)) (6.1)
with voltage amplitude V0, the differences of the angular frequency of the two laser
beams ∆ωand phase ϕ(t).
The second output of the X-coupler can be used for power or spectrum monitoring.
The beat note signal is then recorded by a RF spectrum analyser (Rhode & Schwarz
FSW). The spectrum analyser down-converts the signal to an intermediate frequency,
digitizes this signal, and generates the in-phase (I) and quadrature (Q) components:
I(t) = ˜
V0(t)·cos(∆˜
ωt+ϕ(t)) (6.2)
Q(t) = ˜
V0(t)·sin(∆˜
ωt+ϕ(t)) (6.3)
where ˜
V0is an adapted amplitude and ∆˜
ωt+ϕ(t)corresponds to the instantaneous
signal phase. These I-Q components are recorded and can be accessed for further data
analysis.
The instantaneous signal phase can be extracted from the I-Q data with arctan(Q/I).
The linear component ∆˜
ωtcan be subtracted from the signal phase after obtaining a
linear fit to the remaining signal. The phase noise PSD Sϕ(f)is calculated from the
278MHz in our setup
88 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
remaining ϕ(t)with a fast Fourier transformation and finally can be transformed into
frequency noise PSD Sν(f)by multiplying with f2, since
Sν(f) = Sϕ(f)·f2(6.4)
The intrinsic linewidth of the laser module corresponds to the white noise floor F
of Sν(f)multiplied by π/2.
∆νintrinsic =F·π
2(6.5)
The beat note spectrum can be reconstructed with a Fourier transform of the I-Q
signal. With the beat note spectrum the FWHM linewidth can be determined. This is
done by fitting a Voigt profile to the beat note spectrum. According to Mercer [116],
the relationship between the Voigt linewidth ∆νV, the Gaussian linewidth ∆νG, and
the Lorentzian linewidth ∆νLcan be described with
∆νV=1.0692∆νL+Æ0.866639∆ν2
L+4∆ν2
G
2(6.6)
≈∆νG(6.7)
whereas the Lorentzian linewidth corresponds to the intrinsic linewidth,
∆νL=∆νintrinsic ∆νG.
In order to achieve the FWHM linewidth of a single laser, the FWHM value of ∆νG
has to be divided by p2, since the FWHM of the two superimposed Gauss profiles is
p2 times the FWHM value of a single Gauss profile.
∆νFW H M,Voigt =1
p2∆νV(6.8)
6.1.3.2 Measurement Results
Figure 6.6a shows a frequency noise PSD of a MOPA (blue graph) and the reconstructed
beat note spectrum of the self-delayed heterodyne measurement (red graph) with I-Q
analysis. The analysis is limited to 100kHz due to the limitations of the delay line,
as explained earlier. The DFB laser in this laser module has a length of 3mm and
was driven with 270mA at 34.5° C. The PSD shows a 1/f characteristic, as predicted
by [116], for frequencies <1 MHz, the white noise characteristic is observed for fre-
quencies >10MHz.
The influence of the chip length on the spectral behaviour can also be found in
the linewidth measurements. Figure 6.6b shows the FWHM and intrinsic linewidth for
1.5mm diodes (marked with red lines) and 3mm diodes (marked with black lines).
The x-axis shows the DFB output power, measured before micro-isolator integration.
Both depicted intrinsic linewidths show a 1/P behaviour, as predicted in equation 3.25.
The 1.5mm DFB has a minimum intrinsic linewidth of 106kHz at an output power
of 120mW. The linewidths of the 3mm chip are, expectedly according to [88, 89],
even narrower due to the larger resonator length. The intrinsic linewidth is 14 kHz
at a comparable output power. The FWHM linewidth remains constant at a level of
1.16MHz for the 1.5mm DFB module and 386kHz for the 3mm DFB module.
6.1. ELECTRO-OPTICAL CHARACTERIZATION 89
(a) (b)
Figure 6.6: Linewidth of MOPA: frequency noise PSD and beat note spectrum (a) and
linewidth distribution against MO output power (b)
6.1.4 Modulation Capability of the MIOB
As described in section 4.1.2.2, the MIOB provides an electrical interface that allows
the modulation of the MO injection current on-board the MIOB through a transistor. By
applying a signal at the transistor modulation interface (MMOD) and through that to
the transistor’s gate, the transistor branches off a fraction of the DC injection current,
supplied by the MDC interface.
In order to test the effect of the transistor, a MIOB with an integrated 3mm long DFB
laser as MO but without an amplifier was chosen. A DC voltage in the range of +0.5V to
-2.5V was applied at the MMOD interface additionally to the current input of 270 mA
at the MDC interface. The output power and wavelength of the MO were recorded for
varying transistor gate voltages, as depicted in figure 6.7. As expected for a normally-
on n-channel JFET transistor, the optical output power is at its minimum when the gate
voltage is maximized. If the gate voltage is smaller than the source voltage, the source-
drain-current will be reduced. The transistor "closes", depleting less current from the
laser diode supply. Since the source is connected to ground potential, the gate voltage
has to be negative. Less injection current results in less optical output power, but also,
as shown in section 6.1.2, in a tuning of the laser emission frequency. Both can be
observed in figure 6.7, confirming the depletion of the injection current. Between -1V
and +0.5V a linear tuning can be approximated. A linear fit of the emission frequency
reveals a frequency deviation of 8.9GHz/V for a 3mm long DFB laser for modulation
intensities in the range of -1V and +0.5V, depicted in red in figure 6.7. Since the
spectral current tuning range is given in section 6.1.2 with 1.2nm/A, corresponding to
591GHz/A, the current depletion can be calculated to be 15 mA/V. With that and the
known current tuning coefficient for 1.5mm long DFB diodes the effect of the MMOD
based current depletion can be estimated to be 18.6GHz/V for shorter laser diodes.
For gate voltages smaller than - 2 V hardly any changes in either the output power
or the emission frequency can be detected. It can be concluded that the transistor is
closed completely at this gate voltage, the entire injection current is supplied to the
90 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
laser diode.
This test reveals that the common-source transistor circuit operates as expected.
The recommended signal intensity for the modulation signal is in the range of 0V to
-1V.
Figure 6.7: DC Influence on MMOD voltage on wavelength and output power of a 3 mm
long DFB. The output power is measured behind a 60dB µ-isolator.
6.2 Environmental Tests
Environmental stress tests are important to prove the suitability of a component for its
operational environment. For operation in space very often the requirements are based
on the environment in the transport vehicle, like the rocket, or the spacecraft, such as
described in section 2.1.3. In addition to these mission requirements, there are general
qualification specifications, e.g. the specifications of the European Space Components
Coordination (ESCC) [117]3, which specify the test methods that should be used for
qualification of the hardware. The mechanical stress tests with random vibration and
shock are of particular importance for sounding rocket missions. Complex modules,
using different materials and assembly technologies come with the risk of falling apart
or simply breaking when being subjected to mechanical stress. Thermal loads become
relevant during transportation and storage for sounding rocket campaigns. Different
3Not to be confused with the ECSS, the European Cooperation for Space Standardization which aims
at "a coherent, single set of user-friendly standards for use in all European space activities" [118]. The
ESCC, however, provides procedures and standards for "the production and use of electrical, electronic
and electro-mechanical components suitable for use in space" [119]
6.2. ENVIRONMENTAL TESTS 91
materials come with different values for the coefficient of thermal expansion (CTE). A
differential thermal expansion adds stress to the bond areas between parts, potentially
causing damage to bonds or at least misalignment of the optical components. Both
thermal and mechanical tests were performed and are described in this section. These
tests exceed the requirements for the TEXUS [47]and MAIUS sounding rocket mission
in order to find limitations and prepare for future space missions.
Besides mechanical stress tests and temperature stress tests, radiation stress tests
are important in the qualification process towards qualification of a component [117].
Semiconductor elements that operate high above the surface where the Earth’s mag-
netic field is weaker compared to the surface are threatened by ionizing particles,
causing single event effects. For microprocessors this can result in bit-flips or latch-
ups [120]. In semiconductor laser diodes the impact of radiation causes an increasing
threshold current and a decreasing output power [121], hence a degradation of the
electro-optical performance, but typically no fatal damage. However, radiation hard-
ness tests are not relevant for sounding rocket experiments because the exposure time
in space is, with approximately 15 minutes, very short compared to satellite missions
with several years of operation times. With an apogee of 260 km the sounding rocket
also does not reach the Van Allen belts. With a short operation time and the close
vicinity to Earth the dose of radiation impact can be neglected.
6.2.1 Definition of Test Parameters
The random vibration test and temperature cycling test were performed with varying
intensities in order to first ensure the performance according to the load specifications
defined by the sounding rocket environment [47], and in a second step, to find the
limitations of the technology. Additionally to the random vibration tests and temper-
ature cycles, a mechanical shock test was performed. The specifications of the stress
test campaign in the second step are based on the ESCC "Evaluation Test Programme
Guidelines for Laser Diode Modules" [117]. These guidelines point out designated MIL
specification methods and test parameters appropriate for qualifying diode laser mod-
ules. Table 6.1 gives an overview of the environmental tests performed in this work.
The tests took place in the order given in the table, starting with an 8 gRMS vibration
test and finishing with the –55 ° C ... +85°C temperature cycles. The electro-optical
performance of the DUTs was recorded before and after each stress test.
6.2.1.1 Random Vibration Tests
Two random vibration tests have been performed. The one is carried out at 8 gRMS load
level, which corresponds to the qualification level for the TEXUS sounding rockets. The
load profile of this test was determined experimentally in a TEXUS mission. The profile
of the second test, at 30gRMS, was defined by a more general test procedure, covering
a broad band of vibrating frequencies. The profiles are given in table 6.2.
4+15°C corresponds to the dew point of water on cold surfaces when the surrounding has a humidity
of 50% a room temperature. +70 °C is the maximal temperature defined by the ESCC [117].
92 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
Test Duration Repetition Specification
random vibration
8gRMS 2 min /axis 1 DLR, TEXUS mission [47]
30 gRMS 3 min /axis 1 MIL-STD-202 (Method 214;
Condition I-H) [122]
half sine shock 0.5ms/axis 3 MIL-STD-883 (Method 2002;
Condition B) [123]
temperature cycles
+15°C ... +70°C ≤12K/min 10 laboratory storage4
–30°C ... +45°C 2K/min 10 DLR, TEXUS mission [47]
–55°C ... +85°C 2K/min 10 MIL-STD-883 (Method 1010;
Condition A) [123]
Table 6.1: Overview of environmental tests performed in this chapter
In addition to the tests on system level, all modules integrated into the rocket pay-
loads of FOKUS and MAIUS run through the same qualification test at 8 gRMS on subsys-
tem and unit level. Although it is not required by the TEXUS stress test specifications,
one MOPA module was fully operated when performing the first vibration test. The
output power was recorded before, during, and after the vibration test.
The second random vibration test specifications are based on the ESCC document
[117], referring to the MIL-STD-202, Method 214 [122]. However, since the aim of
this test was not the qualification according to ESCC regulations but to show the ability
for future qualification, some details of the specification were modified in order to
increase the chance of survival of the modules. The duration of the test phase was
reduced from 7.5minutes, as defined by the ESCC [117], to 3minutes, as suggested
in the MIL-STD-202 method 214-H test definition to which the ESCC test refers. In
addition, the modules were not operated during this test.
Both tests were performed with a 35 kN TIRA shaker. Before and after the random
vibration test, the vibration frequency was swept in order to identify resonances in the
modules and verify that these resonances have not changed after the random vibration
test. Each test was performed in the x, y and z-axis of the module.
Test Frequency Range PSD Component
8gRMS
20 – 400Hz 0.0045g2/Hz
400 – 600Hz 0.0675g2/Hz
600 – 1300Hz 0.0045g2/Hz
1300 – 2000Hz 0.0675g2/Hz
30gRMS
50 – 100Hz +6dB/oct
100 – 1000Hz 0.6g2/Hz
1000 – 2000Hz -6dB/oct
Table 6.2: PSD loads per frequency range of random vibration test
6.2. ENVIRONMENTAL TESTS 93
During the 8gRMS test, the output power of one MOPA was recorded. To do so a
macroscopic collecting lens with a diameter of 2 inch compensated the movement of the
vibration table by focussing the optical output beam to a photo detector mounted to an
Ulbricht sphere. The Ulbricht sphere prevents saturation of the detector by distributing
the insert beam equally to its inner surface. The data was recorded with a sample rate
of 300Hz, limited by the available recorder. The temperature of the MOPA was not
stabilized during the test.
6.2.1.2 Shock Tests
Additionally to the random vibration tests, a half-sine shock test according to ESCC and
MIL-STD-883; Method 2002; Condition B [117,123]was performed. A pivot-mounted
hammer applies a half sine shock with an amplitude of 1500g and a duration of 0.5ms
to block that transfers the shock to the DUT mounted in a pre-defined direction. The
shock is applied three times to each axis in each direction, resulting in 18 shocks per
module.
6.2.1.3 Thermo Cylce Tests
The temperature cycles were performed with a CTS T-65/50 test chamber. The first test,
testing for laboratory storage environments between +15° C ... +70° C, was performed
with a dwell time of 40min, but with the maximal slope of up to 12K/min.
The second test parameters are given by the experimental environment of a TEXUS
mission. –30° C might occur under transportation and recovery of the payload. The
upper temperature limit of +45° C is the maximal operation temperature of the mod-
ules in the experiment. The dwell time was chosen to be 70 min, resulting in a slope of
1.9K/min. The number of cycles was not defined in the TEXUS environmental speci-
fications. However, a number of 10 cycles should exceed the number of temperature
cycles the experiment is exposed to during the mission and therefore covers the stress
limits.
The third test, again, is based on general qualification specifications, namely the
MIL-STD-883; Method 1010; Condition A, giving a temperature range of –55 ° C ...
+85° C. The slope, again, was 2 K/min, with a dwell time of 80min for the rising edge
and 100min for the falling edge. The hold time at the minimum temperature was
extended to make sure that the test chamber reaches the minimum temperature. This
test is also suggested in the ESCC document, but with a maximal temperature of 70 °C.
Since the modules have passed the 70°C in the first test, we decided to increase the
temperature range even further to investigate the limitations of the DUT modules.
6.2.2 Results of Environmental Tests
All laser modules, integrated in the FOKUS and MAIUS experiment, passed through the
qualification tests on system level as defined in the TEXUS specification [47].
Two additional laser modules were tested under more serve test conditions, such as
higher random vibration loads in the vibration test, shock test, and temperature cycles.
94 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
Before and after each test, the output power and spectral behaviour were recorded.
6.2.2.1 Results of the Mechanical Stress Test
The visual inspection of the MOPAs did not reveal any changes on the module, the
lenses and isolators as well as the cables attached to the electrical interface remained
in place.
After the random vibration tests and shock tests no significant modifications of the
performance could be observed. The MOPA output power remained stable at the WP,
it rather increased slightly from 1.4W before vibration and shock tests to 1.48W after
the tests.
(a) (b)
Figure 6.8: Spectral performance of DUT before and after vibration and shock tests.
(a) spectral map, (b) single spectrum at WP. Performance remained stable.
Figure 6.8a shows the spectral behaviour of the MOPA recorded after each random
vibration test and the shock test. No changes, such as shifting of the mode hops, ap-
pearing of additional mode hops, and shifted threshold current, in the spectrum map
could be observed. Figure 6.8b shows the single spectrum at the WP. It can be observed
that the noise level for emitted light >0.5nm next to the peak is stable below -50dB.
The side-mode is suppressed by >45 dB in all measurements. However, it also can be
observed that the wavelength of the peak maximum is scattered by 26pm, without a
6.2. ENVIRONMENTAL TESTS 95
tendency during the test series. The resolution bandwidth of the OSA is 10pm, and
the measurement step size of 2 mA, corresponds to approximately 6 pm according to
section 6.1.2. This is not sufficient to explain this wavelength scattering alone. The
main contribution to the variance of emission wavelength is based on different tem-
peratures of the submount of the DUT. In module MOPA11-13 the thermal resistance
of the MO submount to the MIOB appeared to be higher than in the other modules,
resulting in a poor thermal dissipation of the heat generated by the laser diode, and in
problems when stabilizing the temperature of the module to the measures of the sub-
mount based temperature sensor. Therefore the module was stabilized to the sensor
at the MIOB, resulting in varying temperatures at the MO submount. The temperature
difference between the measurements amount to 0.38K, corresponding to 20pm wave-
length tuning, as described in section 6.1.2. With that, the module MOPA11.13 is not
qualified for high precision measurement applications, still, together with the measure-
ment uncertainties named above the emission wavelength variation can be explained.
The FWHM and intrinsic linewidth also stayed stable after the tests.
Since the spectral deviations can be explained with the uncertainties in the mea-
surement setup and the optical output power was stable after all mechanical stress
tests, it can be stated that the mechanical stress tests did not influence the behaviour
of the laser module.
As described in section 6.2.1.1, the optical output power of one MOPA module was
recorded during the TEXUS launch qualification test at 8 gRMS. The aim of this test was
to show that the MOPA modules can be operated during the rocket launch.
(a) (b)
Figure 6.9: (a) Power recording (above) and RMS deviation (below) of a MOPA emit-
ting during lateral (along the x-axis of the shaker) vibration test of 8gRMS, (b) RMS
values of these power recordings before, during and after the vibration tests. pre: be-
fore the test, no external movement, vib: random vibration run, post: after the test, no
external movement. X,Y, and Z are the directions of movement. The longitudinal (Y)
vibration run was performed first.
In its upper panel, figure 6.9a shows in blue the raw data of the power measurement
for the x-axis (lateral) vibration run, including the sine-sweep resonant search before
96 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
and after the random vibration run. In black, the root mean square (RMS) power fluc-
tuation of this data is presented in the lower panel of this figure. The resonance search
was carried out starting at second 100 and 820 and lasted for 180 s each. The ran-
dom vibration test was performed between t=460s and t=580s. During the times
marked with the green arrows, no external force was applied to the modules. Before,
during and after the vibration test the MOPA emitted with a median of 1202mW with a
maximal peak-to-peak variation of 59mW. It can be seen that the optical output power
stayed stable also during the random vibration test, the coupling of the MO-beam into
the power amplifier was hardly influenced by the external vibration of the environ-
ment. However, both the median and maximum RMS fluctuation of the power values
increased slightly during the test, from a median of 3.5mW before the random vibration
tests to 4mW during the random vibrations. The maximum RMS deviation of 22 mW
before the vibration was overtaken with 25 mW during the test. However, the max-
imum RMS deviation of 31mW appeared only after the random vibration run when
the shaker was stood still. This shows that externally applied vibrations are not the
main source of power fluctuations in this setup. The temperature of the MOPA module
was not stabilized, and varying temperature has an influence on the output power, as
described in 3.1.3.3. After the random vibration run, the DUTs were inspected visually.
Draught arising from open doors and moving people can be the cause of the increased
power fluctuation after the random vibration test of the lateral (X) axis as well as the
internal cooling mechanisms of the shaker apparatus.
Figure 6.9b shows the evolution of the power fluctuation in RMS values during the
8gRMS test campaign. Also for the Y (longitudinal) and Z (vertical) axis the maximum
deviation of the median power value did not exceed 45mW, stating that there also was
no interruption of the power emission during the random vibration runs of these axes.
However, the power fluctuation increases during the test campaign, from 2.75 mW to
5.25mW. This can be explained by displacements in the nm range of the optical compo-
nents of the MOPA module due to the vibrational loads, influencing the beam coupling
and the clearing of photons in the amplifier.
Unfortunately it was not possible to record the optical output power with a sampling
rate higher that 300Hz. Since the maximum vibration frequency is 2 kHz, a sampling
rate of at least 4kHz would have allowed a detailed analysis of the applied vibration
frequencies on the power fluctuation.
6.2.2.2 Thermal Stress Test Results
Two MOPA modules were tested in a thermal test chamber according test parameters
given in table 6.1. It was expected that thermal cycling causes misalignment between
the optical and electro-optical components of the laser module due to CTE mismatch
of the different materials and adhesives used. Both modules still fulfilled the spec-
ifications after the test campaign. However, a degradation in the in-coupling of the
beam from the master laser into the amplifier could be observed for both modules,
as shown in figure 6.10a. This figure illustrates the optical output power for various
MO injection currents and with that the saturation behaviour of the amplifier. After
the first thermal cycling the saturation behaviour, marked with red, follows the satu-
6.3. CONCLUSION OF CHARACTERIZATION AND STRESS TESTS 97
ration behaviour previous to the thermal cycling campaign, marked in black. But after
the second thermal cycling the graph flattens, in green and blue, indicating that the
amplifier saturation requires a higher MO injection current, reducing the in-coupling
efficiency. Despite the reduction of the in-coupling efficiency the total optical output
power remained stable, as depicted in figure 6.10b. The degradation of the in-coupling
efficiency is likely caused by misalignment of the lens positions. The adhesive fixating
the lenses to the MIOB is specified for a temperature range of - 15 ° C to +60 ° C before
ageing [124]. The DUTs were not aged before test. This may have caused a softening
of the adhesive, allowing the lenses to change their position slightly.
(a) (b)
Figure 6.10: Output power development after several thermal cyclings. (a) over TA
injection current, (b) over DFB injection current
The thermal stress did not affect the spectral behaviour of the DFB lasers as the
misalignment of the lenses did not affect the DFB operation. The saturation of the
amplifier suppresses the ASE background on the same level as in prior tests at the WP
settings. The linewidth also remained stable during and after the tests.
With the relatively small number of 10 thermal cycles per temperature range, the
laser modules remained within the specifications and proved that they can withstand
temperatures in the range of -55°C ... +85°C. This is more than enough for sounding
rocket missions and would also be enough for short term space missions such as the
(no longer existing) Space Shuttle experiments [125]or operation on the International
Space Station. For long-term satellite missions however, the need for temperature sta-
bilization of the laser module is inevitable also when not operated. Alternatives for
the adhesive that keeps the optics in place, such as soldering the lenses [126], could
be evaluated to avoid the thermal degradation of the laser modules performance that
show an effect in these test.
6.3 Conclusion of Characterization and Stress Tests
The characterization process shall verify the performance of the laser modules accord-
ing to the requirements of the application. Table 6.3 lists the characterization results
98 CHAPTER 6. CHARACTERIZATION AND QUALIFICATION
of the laser modules that correspond to the requirements listed in tables 2.1 and 2.2.
Measurement conditions and additional analysis are described earlier in this chapter.
As listed in table 6.3, all requirements can be fulfilled. This does not only apply to the
laser modules presented in this chapter but to all laser modules produced in the course
of this thesis.
parameter required value achieved value comment
electro-optical
wavelength (λ) 780.241nm 780.241nm with ±0.2nm current
tuning range at const. TMO
FWHM linewidth ≤6 MHz 1.16MHz for 1.5mm long DFBs
386kHz for 3mm long DFBs
tunability around WP 100GHz 880 GHz for 1.5 mm long DFBs
740GHz for 3mm long DFBs
optical power per MOPA >1W (CW) 1.5W at ITA =2 A, TMO=35° C
optical power of MO >10 mW (CW) >20 mW at IMO <200 mA, TMO=36° C
max. power consumption 7.2W typ. 5.7W at WP operation
environmental
vibration loads 8.1gRMS passed also 29 gRMS passed
shock loads ≥250g 1500g passed
WP temperature ≥35° C achieved
storage temperature –30° C ... +45 °C passed –55° C ... +85 °C passed
Table 6.3: Results of laser module characterization and stress tests compared to re-
quirements
For the environmental stress tests, the test parameters were enhanced in order to
apply to recommended test specifications for future applications and to find the bound-
aries of environmental conditions of the laser modules. All laser modules that went
through the mechanical and thermal stress tests kept their specified performance after
each test. The qualification tests for the sounding rocket missions were passed without
affecting the performance at all. The extensive mechanical stress tests also triggered
no change in behaviour. However, the thermal stress tests for temperatures below zero
revealed a degradation in the in-coupling of the MO beam into the amplifier. This
degradation does not affect the performance at the WP and therefore is not critical.
However, for storage under these conditions preventive measures should be taken to
guarantee long-term performance of the laser modules.
Chapter 7
Evaluation of Sounding Rocket
Missions
On 23rd of April 2015, the TEXUS 51 rocket was launched. One of the four payloads on
board this rocket was the FOKUS experiment, including the first semiconductor laser
frequency locked on an alkali transition line operated in space. The Rb spectroscopy
was driven by one of the DFB-MO laser modules, developed in this thesis.
On 22nd of January 2017 the MAIUS rocket was launched and created the first BEC
in space. This rocket hosted 5 fully integrated MOPA modules and one DFB-MO module
that were developed and characterized in this thesis.
With FOKUS and MAIUS, our laser modules have proven their functionality in space
operation and obtained technology readiness level (TRL) 9 for sounding rockets. The
experimental setup and flight environment were already described in chapter 2.1. This
chapter presents the performance data of the laser modules and results of the missions.
7.1 FOKUS Mission
7.1.1 Qualification and Ground Operation
The rubidium reference system consists of a DFB laser diode integrated in a MIOB and a
Zerodur based spectroscopy module, hosting a Rb gas cell and photodiodes as described
in section 2.1. After integration the system was subjected to random vibration tests with
8.1gRMS, thermal cycling tests between 5°C and 40°C. These tests were simulating
launch conditions of the rocket and storage conditions of the assembled payload. The
system passed all tests [42].
In order to perform a LPI experiment, precision measurements of the path differ-
ences of the two clocks are necessary. Long-term drifts of the clock frequencies limit the
accuracy of the measurement. Main contribution to frequency drifts are the fluctuation
of the locking point of the Rb laser, but also drifts of the reference clock (CSAC) and
uncertainties due to the light shift and Zeeman shift [42]. In order to verify the stability
and reliability of the overall system, a long-term measurement of the clock differences
was performed prior to the mission launch. To do so, the DFB laser frequency was
stabilized on the Rb transition frequency and compared to the CSAC by means of the
100 CHAPTER 7. EVALUATION OF SOUNDING ROCKET MISSIONS
frequency comb. The frequency difference of the Rb laser and the CSAC, characterizing
the overall stability of the FOKUS system, was recorded over a time span of 200,000s
(65h). Figure 7.1 shows the absolute frequency shift in the time domain, and relative
stability of this measurement.
Figure 7.1: Frequency deviation of beat note signal from median (upper panel) and
relative stability σ(lower panel) of long-term measurement of Rb reference system
compared to the CSAC
In total, the frequency shifts by ±150kHz compared to the median frequency. The
stability is estimated by calculating the Allan deviation. The Allan deviation is defined
as the square root of the Allan variance, a widely accepted measure for frequency sta-
bility [127]. The Allan variance σ2is defined as
σ2
y(τ) = 1
2Σ[(¯
yn+1−¯
yn)2](7.1)
with ¯
ynas the nth average normalized frequency deviation from nominal frequency
over a finite time interval with the length of τ[128]. With τ=1sthe systems instability
corresponds to 6 ×10−11, integrating with 1/pτ. However for the rocket flight, time
scales in the range of τ=1000sare relevant. The system provides a relative frequency
instability of 5 ×10−11.
The CSAC contributes to the instability with a specified Allan deviation of only 8 ×
10−12 at 1000s, the experiments stability therefore must be limited by the optical clock.
The increasing Allan deviation for integration times >60s is caused by drifts in the
analogue control electronics triggered by external temperature fluctuations [60]. Apart
from this, the relative frequency instability corresponds to the frequency instability
reported for other DFB laser system, as described in [129].
7.1. FOKUS MISSION 101
In order to improve the stability, the linewidth of the optical oscillator has to be
improved. An approach could be to use an ECDL as optical oscillator as described
in [130]. However, the improvement of the linewidth by factor 10 comes with the cost
of spectral mode hops with distances of one free spectral range. This limits the spectral
tuning range and demands a more complex locking scheme. The linewidth of the DFB
laser diode matches the natural linewidth of the reference point for stabilization. To
improve the stabilization, it is also suggested to choose another atom species, such as
Al, transition 1S0,F=5/2→3P0,F=5/2[131], with a narrower natural linewidth
and smaller sensitivity factor to minimize frequency shifts of the reference point. The
stabilization can also be improved by optimizing the modulation bandwidth of the elec-
tronics and making use of the modulation interface provided by the MIOB. In FOKUS
the modulation signal for frequency correction is added to the injection current exter-
nally. This limits the modulation bandwidth on a hardware level e.g. due to losses in
the transmission cable.
The limited stability of the optical clock assembly also limits the accuracy of the
experiment. However, the stability is sufficient to demonstrate the potentials of this
technology.
7.1.2 The Flight Campaign
The FOKUS experiment is both a technology demonstrator and a proof of concept of a
LPI experiment based on an optical clock in space. This section describes the mission
process and the results of the flight campaign.
7.1.2.1 Mission Schedule and Performed Operations
The mission was launched at t0=9:35:00 April 23rd 2015 CET from Kiruna Space
Station. Table 7.1 gives a detailed overview of the flight events.
7.1.2.2 Results of the Flight Campaign
As described in table 7.1, the system was switched on and the DFB lasers frequency
was stabilized on the Rb reference before launch. The frequency stayed stable until
330s after launch, when frequency stabilization was deactivated on purpose. The peak
accelerations during launch were 8.1g and 12.6 g, through the first and second stage
flight of the rocket. At t0+330s, a scan over the DFBs injection current was initiated
to force the system out of the lock in order to test the auto-lock procedure during flight.
The graphs in figure 7.2 illustrate the PD readout signal (in blue) and the FMS error
signal (in red). The constancy of the error signal and the photo diode signal together
indicate that the frequency of the DFB laser remained stable and in lock over the launch
period until the intended auto-lock test 330s after launch. The dips in the error signal
after 330s are caused by 87Rb F=2⇒F’ transition. The frequency lock acquisition was
achieved right after the scan was terminated and stayed in lock until shut down of the
system at t0+600s. The Rb system performed as expected and without any problems,
102 CHAPTER 7. EVALUATION OF SOUNDING ROCKET MISSIONS
mission time event
t0- 5:30h switching on the system to thermalize
locking of comb and Rb system, collection of ground reference data
t0- 10s system in secure state: Rb system locked, comb system unlocked
t0launch at 9:35:00 April 23rd, 2015 CET
t0+13.4s separation of first rocket stage
t0+59s separation of second rocket stage
switching on the locking of comb system
t0+72s µg regime reached, <10−4g
t0+84s comb frequency was locked
t0+261s apogee reached
t0+330s remotely controlled Rb spectroscopy scan
t0+360s spectroscopy scan of Rb system finished
t0+442s µg phase ended
t0+600s system shut down for landing
t0+884s touch down
Table 7.1: Time table of FOKUS mission flight
the technology demonstration of alkali references optical oscillators in space was hence
successful.
In order to demonstrate a LPI experiment the fractional frequency difference
(f1−f2)/fin a varying gravitational field ∆Uhas to be recorded. The frequency
difference is related to ∆Uthrough
(β1−β2) = f1−f2
f·c2
∆U(7.2)
where βiis a dimensionless parameter reflecting the sensitivity of each clock to a
potential violation of the LPI.
(f1−f2)/fis the clock frequency comparison data which was collected until
t0+450s, as long as both, the Rb system and the Er comb system were locked. The
collected data is shown in blue in figure 7.3. A linear correction of a Rb lock drift
of 122Hz/s, probably caused by the increasing temperature of the payload, was sub-
tracted from the data before plotting. The gap of data between 0 – 84 s is caused by
the deactivated stabilization of the frequency comb system during launch. Between
330s and 440s after launch the Rb system was not stabilized due to the forced scan of
the injection current of the DFB laser. At 478s the frequency comp system was out of
lock due to the landing procedure. The black line illustrates the gravitational poten-
tial ∆U/c2the FOKUS experiment was travelling through, calculated from the altitude
of the rocket and a spherical Earth model, using Gas gravitational constant, MEarth
as Earth’s mass, rEar th as the radius of a spherical Earth, and hrocket as altitude of the
payload:
U(t) = G·mEarth
rEarth +hrocket (t)(7.3)
7.1. FOKUS MISSION 103
Figure 7.2: Photodiode read out (blue) and error signal (red) of the Rb reference system
before and during flight. The laser frequency was stabilized before launch and stayed
stable the entire mission except for the planned scan over the injection current.
The slow slope between 500 and 600s is caused by the re-entry into the atmosphere.
At 600s the system was shut down for landing.
With the available data in figure 7.3, equation 7.2 and the least square method,
the LPI violation indicator was estimated to be (β1−β2) = 0.186 ±0.260. Although
this estimation is orders of magnitude worse than state of the art experiments 1, the
measurement principle, using optical clocks as references is not the limiting factor. The
absolute clock evaluation is limited by long-term drifts of the reference frequencies, as
discussed in section 7.1.1. The measurement accuracy can be improved by using state
of the art optical clocks, and by enlarging the gravitational potential, e.g. with a space
mission on a trajectory bringing the instrument close to the sun.
However, this was the first demonstration of an LPI test between two different types
of clocks in space. The micro-integrated diode laser module performed as expected,
providing the oscillator of the compact optical clock. This compact and robust opti-
cal clock is a pioneer in referencing frequencies in future spaceborne precision experi-
ments.
7.1.3 Post-Flight Characterization of Laser Module
The DFB laser module, integrated into the Rb reference system in the FOKUS exper-
iment, was characterized electro-optically before delivery (20th August 2012), after
each of both 8.1gRMS vibration tests (4th and 10th October 2012) and after the flight
1∆β=β1−β2<10−6, shown by Fortier in 2007 [132]
104 CHAPTER 7. EVALUATION OF SOUNDING ROCKET MISSIONS
Figure 7.3: Relative changes of clock frequency comparison (blue squares) and gravita-
tional potential (black dotted line) before and during flight. A linear drift of 122Hz/s
was subtracted from the optical frequency reading.
campaign (6th October 2015). The electro-optical characterization corresponds to the
characterization described in section 6.1, including the optical output power, the emis-
sion spectrum and the linewidth characteristics (both FWHM and intrinsic) for various
injection currents. However, after the flight campaign, it was not possible to measure
either the free-space output power of the laser module or the total power in the fibre,
because the laser module was integrated into the Rb module, the fibre coupled output
of the laser module was spliced to a fibre optic splitter. The power values therefore are
not directly comparable to each other.
(a) (b)
Figure 7.4: Optical output (a) power and linewidth (b) of FOKUS laser module before
and after mechanical stress tests.
7.1. FOKUS MISSION 105
(a) (b)
Figure 7.5: Spectral performance of FOKUS diode laser module before and after vi-
bration tests and after the flight campaign, (a) spectral map, (b) single spectrum at
WP.
The optical output power is shown in figure 7.4a. In the pre-flight measurements,
the output power did increase after the first vibration test. This behaviour can be ex-
plained with the burn-in process of a laser diode and was observed before, also by
Spießberger in 2012 [111]. The slight lost in power after the second vibration test
could be explained with different absorption behaviour of the optical isolator that had
to be exchanged between the vibration tests. The free-space output power was approx-
imately 19mW at a WP of 150mA and 36°C before flight. The coupling efficiency of
the free-space light into the optical fibre was roughly 60 %, estimated before system in-
tegration. The splitter between laser and spectroscopy module guides nominally 90%
of the beam to the comb system. With that, 10.3mW at the port for beat detection was
expected. Post-flight, 6.8 mW could be measured at the designated WP at the optical
output of the Rb reference module, featuring a fibre coupler, fibre splices, and the fibre
splitter in the optical propagation path. The measured value does not fulfil the expec-
tations. The reason probably is a combination of either: the laser diode degraded over
106 CHAPTER 7. EVALUATION OF SOUNDING ROCKET MISSIONS
time, the coupling efficiency is less than the expected 60 %, there are in-fibre losses,
e.g. at the splices, or the fibre splitter propagates less than 90% of the light in this fibre
output e.g. due to the mismatch of the design wavelength of the splitter. Since a refer-
ence value after integration into the FOKUS module is missing, it cannot be determined
whether the power degraded in the fibre during the flight campaign.
If the laser diode degraded over time, this did not have an impact on the spectral
behaviour. The FWHM and intrinsic linewidth remained constant within the measure-
ment tolerance in all measurements, as can be seen in figure 7.4b. The FWHM linewidth
was 0.996MHz in 10µs with a standard deviation of 75 kHz, the intrinsic linewidth,
based on the white noise floor, was 64kHz, with a standard deviation of 3kHz. These
values lie well within the expected performance as described in section 6.1.3 and the
experimental requirements, given in table 2.1.
The emission spectrum remained constant at first view, as depicted in figure 7.5a.
Comparing the emission wavelength at a single working point, as done in figure 7.5b,
reveals that the SMSR remains constant below - 50 dB. The peak wavelength, however,
seems to decrease after each mechanical stress with an overall drift of 44pm. As de-
scribed in section 6.2.2.1, the accuracy of the OSA measurement instrument is limited,
with a RBW of 10pm and an absolute accuracy in the range of ±200pm. The drift
therefore could be explained by the inaccuracy of the OSA. Nevertheless, the current
tunability remains stable at 3.2pm/mA, noting that the spectrum seems to be shifted
by an offset. The current tuning range of the laser module also remains stable above
450pm2. Even if the emission spectrum drifted due to degradation of the laser diode,
the drift can be compensated by the current tuning range of the laser.
The characterizations reveal that the spectral laser module behaviour remained sta-
ble all through the stress tests and the FOKUS flight campaign. Despite the uncertainty
of the degradation of the power in the fibre output there is no evidence of degradation
of the performance of the laser module itself. The laser module itself proofs suitable for
sounding rocked based optical sensor operations and can be used for further missions.
2Since the minimal optical output power is defined to be 10 mW, the injection current range is 85mA
to 250mA
7.2. MAIUS MISSION 107
7.2 MAIUS Mission
The MAIUS-1 mission was successfully launched at t0=3:30:18 January 22nd 2017
CET from Kiruna Space Station. Telemetry data gathered during the mission and trans-
mitted to the operation centre revealed that during the mission flight time it was possi-
ble to generate a BEC with Rb atoms, the first BEC produced in space [133]. Table 7.2
gives the main events of the flight.
mission time event
t0launch at 3:30:18 January 22nd, 2017 CET
t0+64s µg regime reached
t0+421s µg phase ended
t0+873s touch down
Table 7.2: Time table of FOKUS mission flight
The experiment data, stored on the on-board computer still has to be evaluated.
Further details about the physics experiments therefore cannot be presented in this
thesis. The remote telemetry data, however, contained the reading of the MO injection
currents, the free-space photodiodes at the rear of the MOPAs, and the in-fibre photo-
diodes monitoring the optical power emitted by the laser modules into the fibres. To
reduce the telemetry link budget the data sample rate of the housekeeping data was
reduced to 0.1Hz. The in-fibre PDs were monitored with 0.2Hz.
Figure 7.6: Telemetry data of MO injection current (top), the rear free-space optical
power of the MO (centre), and optical power in fibre (bottom) of MAIUS laser modules
108 CHAPTER 7. EVALUATION OF SOUNDING ROCKET MISSIONS
Two of the 5 MOPA modules were cold redundancy and therefore not operative dur-
ing flight. Figure 7.6 shows the injection current and PD readings of the three science-
MOPAs and the DFB frequency reference laser. The recorded PD voltage, corresponding
to the optical power at the rear output and in the fibre, is normalized to the reading
in stable operation. The values are expected to be stable because switching and prepa-
ration for the experiment is done in the switching and distribution module behind the
optical power monitoring. The data reveals that the lasers have been operating before
and during launch. There is an interruption of the injection current 20s after launch.
This time corresponds to the separation of the first rocket stage. Since this data point
is not reflected in the output power values, it might be caused by a reading or trans-
mission error. The injection current and optical power remains stable during the entire
µg phase. After the µg phase, during descent, the current driver of the 2D-MOT laser
module increased the current and with that increasing the optical power of the MO, as
shown in black in the top and middle graph in figure 7.6. As expected, the MO power
variances were not propagated through the amplifier of the 2D-MOT laser. Also after
the µg phase, the current of the re-pump laser shows multiple interruptions, resulting
in power fluctuations both for the MO and in the fibre. These interruptions were most
likely cause by a lose SMA cable connection of the re-pumpers MO injection current. It
can be seen, that for an injection current close to zero the MO output power drops to
zero and the amplified in-fibre power is reduced to ASE emission. The 3D-MOT laser
and the reference laser remained stable during the entire flight.
Figure 7.7: RMS power fluctuations of the in-fibre power readings of the laser modules
Although the power of all modules remained stable during ascent and µg phase,
there are deviations in power when comparing the in-fibre data with the free-space
data. The RMS fluctuations in figure 7.7 reveal that the deviation of the mean value of
the MOPA modules is not larger than 5 %. There are short-term ripples and long-term
drifts observable. The ascent phase after launch until t0+60 s is clearly visible by fluc-
7.2. MAIUS MISSION 109
tuation peaks in all data sets. However, during the mission sine-like fluctuations, lasting
several minutes, dominate. This long-term drift is very clearly visible for the reference
laser with up to 6.5%, although, due to the smaller optical power in the fibre thermal
dependencies might show a stronger effect. Since this drift can only be observed in the
fibre path reading, but not in the free-space data at the rear port, the drifts are most
likely introduced by the fibre components of the laser system. They are probably caused
by temperature differences in the laser system and different temperature dependencies
of the components in the optical path, such as the fibres, fibre couplers, and the in-fibre
PDs themselves, and not by the amplifier and collimation optics. This assumption also
is supported by the data of the reference laser that does not host an amplifier.
However, all interruptions of optical power can be traced back to fluctuations of
the injection current. It also has to be stated that all lasers performed stable with no
interruption from launch and through the entire µg phase when the experiments took
place. Power fluctuations in fibre were rather small and not caused by the laser modules
themselves.
After recovery of the payload and minor repairs in the optical distribution system,
the experiment apparatus still could be operated and was able to create a BEC, under-
lining the robust design of the entire MAIUS experiment apparatus. As shown by these
preliminary results of the MAIUS mission the MOPA modules proved their suitability
for quantum optical experiments in space.
110 CHAPTER 7. EVALUATION OF SOUNDING ROCKET MISSIONS
Chapter 8
Summary and Outlook
The main objective in this work was the development and production of high-power,
narrow-linewidth semiconductor laser modules with improved frequency stabilization
capabilities for applications on board of sounding rockets. This chapter gives a brief
summary of the results achieved in this work and identifies some aspects that could
be improved in future design. Future prospects of hybrid integrated laser modules for
space missions are named in the outlook section.
8.1 Summary
In this thesis a semiconductor based laser module for high precision mobile quantum
optical sensors was developed that is capable for operation in harsh environments. To
do so, a hybrid micro-integration approach was chosen, allowing the integration of ac-
tive semiconductor devices as well as passive optical components in order to enhance
the spectral performance of the laser modules and form the beam output. DFB laser
diodes, optimized for narrow linewidth emission at the demanded emission frequen-
cies, as well as TPA diodes with sufficient gain at the demanded emission frequencies
were chosen to provide the spectral stability and sufficient optical power as required
for quantum optical sensor applications. Optical feedback on the DFB diodes is pre-
vented by the integration of a miniaturised optical isolator, maintaining the spectral
performance of the MO. The optical output beam was collimated with miniaturized
lenses in order to meet beam parameters for fibre coupling.
The laser diodes and micro-optics are integrated on a micro-optical bench based
on a 4mm thick AlN ceramic with a footprint as small as 25mm ×80mm. The ce-
ramic base plate offers a high mechanical stability and good thermal conduction. The
latter is important for thermal stabilization of the laser diodes in order to stabilize the
laser’s emission frequency. The thermal stabilization of the laser module is supported
by thermal sensors placed close to the laser diodes. These temperature sensors allow
for a more explicit regulation of the temperature of the laser diodes since the thermal
environment of the diode can be measured more precisely. The MIOB also offers an
electrical interface for coaxial cable connectors and a RF interface for modulation of
the injection current of the diodes in short distance to the semiconductor. With that
112 CHAPTER 8. SUMMARY AND OUTLOOK
the frequency stabilization capabilities of the laser module are improved. The laser
module omits movable parts. The micro-optics are adhesively attached to the MIOB
after active alignment. This decreases the probability of misalignment and increases
the mechanical stability of the laser module.
Methods and procedures for producing a small series of laser modules were estab-
lished, including pre-integration qualification of the components, burn-in of laser diode
and amplifier chips, and active alignment procedures for the laser module integration.
With help of the qualification of the diodes only the ones with the best performance
were selected for integration, increasing the yield of laser module production. A statisti-
cal analysis of the pre-integration characterization of the semiconductor chips revealed
a yield of less than 50%, both for DFB diodes and for TPA amplifiers. The failure
causes, though, were different for both kinds of chips. For DFB diode lasers a poor
spectral performance, such as instability or mismatch of the wavelength, dominated.
Amplifiers most often failed the qualification due to poor facet quality. Because of small
production numbers, the semiconductor processing and coating processes at the FBH
are not fully automated and can be seen as experimental. The processes therefore
cannot be fully optimized to increase the yield. The resulting yield of less than 50 %
therefore was expected and underlines the necessity of a qualification process prior to
integration of the semiconductor diodes into a complex laser system.
The laser modules were characterized electro-optically, showing an optical output
power of >1W and an electrical-to-optical efficiency up to 29%. The optical behaviour
of DFB diode lasers depends on the resonator length of the DFB laser. Lasers with a
resonator length of 1.5mm provide a higher efficiency and larger current tuning. 3 mm
long lasers allow for a larger current range, and provide a much narrower linewidth.
The overall single-mode spectral tuning range of 1.8nm (corresponding to 886GHz)
can be reached, when tuning both the injection current and the operating tempera-
ture. The linewidth improves for longer chips: the 3mm DFB chip provides an intrinsic
linewidth less than 50kHz and a short-term FWHM linewidth of less than 400kHz
(10µs) for an optical output power of the DFB of 120mA, whereas a 1.5mm DFB chip
only provides an intrinsic linewidth of in the range of 100kHz and a short-term FWHM
linewidth in the range of 1MHz (10µs) for a the same optical output power.
The effect of environmental stress on the laser modules was tested as well. Random
vibration up to 29gRMS and shock tests with 1500 g did not influence the performance
of the laser modules. The thermal cycling, however, caused degradation in the ampli-
fier saturation, caused by misaligned in-coupling into the amplifier. Since the optical
power of the MO is significantly larger than the optical power required to saturate the
amplifier, no changes in the performance at WP operation could be observed after the
thermal stress tests. For long-term missions, however, provisions, such as temperature
control and thermal shielding, should be implemented for storage and operation mode
to guarantee stable performance over a long time of the mission.
A micro-integrated DFB laser on a MIOB was employed in the FOKUS sounding
rocked experiment where it was frequency-stabilized to a Rb transition in a Doppler-
free spectroscopy setup. The optical beam then was compared to a CSAC, operating in
the microwave range. This experiment served as technology demonstrator for an opti-
cal atomic clock based LPI experiment in space, aiming to prove or disprove the EEP.
8.2. OUTLOOK 113
The test was successful, all components operated as expected, although the sensitivity
of the measurement was limited by long-term drifts of the reference frequency. The
laser modules were still functional after the mission with comparable performance. In
January 2017 the MAIUS apparatus, hosting five MOPA modules and one DFB module,
was launched and generated the first BEC in space. The apparatus was, after minor
repairs, fully operational after recovery. The laser modules were not harmed and oper-
ated as expected during the whole mission. The data collected in this mission still has
to be analysed.
The successful execution of the FOKUS and MAIUS missions proves not only that
it is possible to miniaturize laser sources with high precision frequency emission, but
also that laser modules developed in this thesis fulfil all requirements demanded by
quantum optical high precision sensor experiments in micro-gravity environments. To
our current knowledge, the laser modules flown in MAIUS are the smallest and lightest
laser modules providing more than 1W optical output power ever operated in space.
Besides the sounding rocket experiments, laser modules developed in this thesis
are also used in drop tower experiments as described by Kulas in [134]. These laser
modules are not only suitable for spaceborne applications but serve also in rough en-
vironments on Earth and can save space in ordinary laboratories as well.
8.2 Outlook
The next logical step towards a laser module for field applications is the integration of
a fibre coupler and a protective housing in order to provide a plug and play system for
the users of quantum optical sensors.
Figure 8.1: A laser module of the next generation, hosting a fibre coupling. Beam ad-
justment is done by mirrors on spherical mounts, the fibre ferrule contains a collimating
lens.
The housing would prevent contamination of the optics by dirt particles. Further-
more, a hermetic housing would contain an atmosphere around the semiconductor
chips even in a vacuum operation environment. This would enable the use of laser
114 CHAPTER 8. SUMMARY AND OUTLOOK
modules in long-term operation in space such as on satellites. As mentioned before, an
ECDL setup as MO would increase the frequency stability of the emission. However,
existing ECDL modules provide an output power not larger than 100 mW [130]and
have to be amplified in a separate module.
The MIOB of the next generation of laser modules is depicted in figure 8.1. It can
be equipped modularly with different laser types, allowing for combinations such as
ECDL with amplifier, combining a very narrow linewidth with high output power, or
two laser chips pointing in opposite directions. The fibre couple concept makes use of
a collimated fibre ferrule and mirrors on polished spherical joints. Since the MIOB pro-
vides fibre coupling for the front and for the rear output, besides the classical front and
rear output configuration, there also is the option of two fibre coupled master oscillators
on one module, minimizing space and weight in case the optical power requirements
allow the absence of an amplifier.
The next generation of laser modules also considers aspects that were learned dur-
ing the production of the laser modules described in this work. The new MIOB provides
a wider channel for optics integration. The channel in the current version has a width of
2.2 mm. A width of >4 mm is recommended for various reasons. First, the submount
can be wider, allowing for additional bond-pads that omit multiple wire-bonding on
the semiconductor chips during pre-characterization and integration. And second, ad-
hesive, applied to hold the SACs during integration, will not flow in the space between
the SAC lens and the MIOB side wall because the gap will be too large. Lateral pulling
of the SAC, and with that displacement, during UV curing will be avoided. The beam
propagation properties, adjusted before the curing of the adhesive, can be maintained.
With these changes and additional features, the successful laser module, described
in this thesis, will be improved and the range of applications can be increased.
Appendix A
List of diodes used in this work
The following table A.1 identifies the laser diodes and amplifier chips integrated in the
laser modules described in this thesis. Further, the appearance of measurement results
obtained with these laser modules is given. The diode ID, MO ID, and PA ID consist of
the number of the wafers test field (TF), the row in the test field in two digits (RR),
and the diode number of the row in two digits (DD). They are arranged as TF-RR-DD.
module ID wafer MO MO ID wafer PA PA ID appearance
in figure
MOPA11-01 C1152-6-1 02-09-03 C2358-3 06-03-xx 5.2
MOPA11-07 C2212-6-1 00-08-21 no PA 7.4,7.5
MOPA11-09 C1152-6-1 02-09-04 C2358-3 08-09-21 6.2,6.10
MOPA11-10 C2212-6-1 01-01&02-10 C2358-3 08-07-14 6.3,6.4,6.6,6.7
MOPA11-13 C1152-6-1 02-09-07 C2358-3 08-07-15 6.8,6.9,
MOPA12-16 C2212-6-1 00-09-16 C2358-3 08-07-13 6.1
MOPA12-18 C2610-6-1 01-04-10 C2850-3 01-01-05 4.13,4.15
Table A.1: List of laser modules used in this work. The diode number of the PA in
MOPA11-01 is unknown.
The following table A.2 gives the remaining diodes of which measurement results
are presented in this thesis, but which are not integrated into a laser module.
diode type diode wafer diode ID appearance
in figure
DFB C2212-6-1 01-01&02-17 5.1a
TPA C2358-3 08-02-05 5.1b
Table A.2: List of single diodes used in this work
116 APPENDIX A. LIST OF DIODES USED IN THIS WORK
119
part value description
MO interface PCB
A 10dB attenuator of MRF signal
C3 1nF bias-tee MRF
L1 2.2nH
L2 2.75µH bias-tee MRF
T n.a. JFET transistor
R1 100Ωpull-down resistor for transistor gate
PA interface PCB
L1 1.35µH filter
T n.a. JFET transistor
R1 100Ωpull-down resistor for transistor gate
temperature sensor interface PCB
C1 100nH filter, in parallel to C2
C2 1nH filter, in parallel to C1
NTC 10kΩtemperature sensor
Table B.1: Components list of the electrical interface of the MIOB
120 APPENDIX B. SCHEMATICS OF MIOB PCBS
Appendix C
Examples of Facet Inspection Ratings
This appendix gives an overview of facet issues, as rated in chapter 5.
C.1 Qualified Facets
C.1.1 Blameless Facets
(a) passed DFB front facet (b) passed DFB rear facet
(c) passed TPA rear facet, RW (d) passed TPA front facet
Figure C.1: Facets of DFB diode C2212-6-1: 00-09-22 (a,b) and TPA diode C2358-3:
08-02-13 (c,d)
122 APPENDIX C. EXAMPLES OF FACET INSPECTION RATINGS
C.1.2 Qualified Facets with Non-critical Defects
(a) particle close to active region (b) shine above the ridge
(c) particles at bulk (d) blopp
(e) bulk demolition, p-down (f) particle outside active area, rear facet
Figure C.2: Questionable facets of DFB diode C2610-6-2: 02-0304-21 (a), (b), 01-04-
13 (d) and C2212-6-1: 00-09-16 (c), as well as TPA diodes C2358-3: 08-06-08 (e) and
08-09-11 (f)
124 APPENDIX C. EXAMPLES OF FACET INSPECTION RATINGS
(a) lose bulk material (b) flaked off coating at rear facet
(c) terraces at front facet
Figure C.4: facets failed in inspection because of bulk defects. DFBs: C2610-6-2 01-
04-17 (a) and 01-01-15 (b); TPA: C2358-3 08-07-16 (c)
(a) before burn-in (b) after burn-in
Figure C.5: Facets of diode C2610-6-2: 02-0304-10, destroyed front facet after qualifi-
cation process
Appendix D
Terms and Acronyms
2D-MOT two-dimensional magneto-optical trap .... . .... . .... . .... . .... . .... . .... 16
3D-MOT three-dimensional magneto-optical trap. ..... ..... ..... ..... . .... ..... ..16
Al aluminium....................................................................35
AlN aluminiumnitride............................................................7
AOM acousto-opticmodulators...................................................16
AR anti-reflection................................................................28
facet coating option
ASE amplifiedspontaneousemission.............................................29
Au gold..........................................................................79
AuSn goldtin....................................................................39
BEC Bose-Einsteincondensate.....................................................3
BPSK binaryphase-shiftkeying....................................................2
CCD charge-coupleddevice......................................................14
CCP conductivelycooledpackage................................................43
CEO carrier-envelope-offset.......................................................8
COD catastrophicopticaldamage................................................30
CSAC chipscaleatomicclock......................................................6
CTE coefficientofthermalexpansion.............................................91
Cu copper.......................................................................38
CuW copper-tungsten............................................................71
125
126 APPENDIX D. TERMS AND ACRONYMS
CW continuouswave.............................................................16
DC directcurrent................................................................41
DFB distributedfeedback..........................................................7
laser diode
DLR "Deutsche Zentrum für Luft- und Raumfahrt e. V." ..... ..... .... . .... . .... . . 129
DLR-MORABA "Mobile Raketenbasis des Deutschen Zentrums für Luft- und Raum-
fahrt"........................................................................8
DOF degreeoffreedom..........................................................36
DUT deviceundertest...........................................................61
ECDL extendedcavitydiodelaser................................................27
laser module configuration
EDM electricaldischargemachining..............................................43
EEP Einsteinequivalenceprinciple.................................................6
EO electro-optical................................................................66
Er erbium.........................................................................8
ESA EuropeanSpaceAgency......................................................2
ESCC European Space Components Coordination.... . .... . .... . .... . .... . .... . ...90
FAC fastaxiscollimator..........................................................37
FBH "Ferdinand-Braun-Institut, Leibniz-Institut für Höchstfrequenztechnik" ..... .. 35
FMS frequencymodulationspectroscopy...........................................7
FSR freespectralrange..........................................................27
FOKUS "Faserlaserbasierter Optischer Kammgenerator unter Schwerelosigkeit" .. ... 5
sounding rocked mission
FSR freespectralrange..........................................................27
FWHM fullwidthathalfmaximum...............................................10
GaAs galliumarsenide...........................................................21
GaN galliumnitride..............................................................21
GSE groundsupportequipment...................................................9
HR highreflection...............................................................22
127
IAE "Instituto de Aeronáutica e Espaço", Brazil... ..... ..... ..... ..... ..... ..... .... 8
InP indiumphosphide............................................................21
ISS InternationalSpaceStation...................................................12
JFET junctiongatefield-effecttransistor..........................................41
laser light amplification by stimulated emission of radiation .... . .... . .... . .... . .. 19
LED lightemittingdiode.........................................................24
LF lowfrequency................................................................40
typically in the range of 300Hz to 30kHz
LIDAR lightdetectionandranging.................................................1
LPI localpositioninvariance.......................................................6
LUH "Leibniz-Universität Hannover", Institut für Quantenoptik.... . .... . .... . .... .11
MASER microwave amplification by stimulated emission of radiation ...... ..... .. 19
MAIUS "Materiewelleninterferometrie unter Schwerelosigkeit" ... ..... .... . .... . ... 6
sounding rocked mission
MDC DCportformodulationoperation..........................................40
MO interface on MIOB
MIOB micro-opticalbench.........................................................7
MMCX micro-miniaturecoaxial..................................................40
cable interface type
MML Radiallmicro-miniaturecoaxial............................................40
cable interface type
MMOD transistormodulationport...............................................40
MO interface on MIOB
MO opticalmasteroscillator......................................................16
MOPA masteroscillatorpoweramplifier..........................................16
laser module configuration
MOT magneto-opticaltrap.......................................................13
MRF RFmodulationport.........................................................40
MO interface on MIOB
µgmicrogravity...................................................................5
NASA "National Aeronautics and Space Administration".... ..... ..... ..... ..... ....1
NF nearfield.....................................................................52
128 APPENDIX D. TERMS AND ACRONYMS
Ni nickel.........................................................................43
OLO opticallocaloscillator.......................................................86
OSA opticalspectrumanalyser...................................................49
PA opticalpoweramplifier.......................................................34
PBS polarizingbeamsplitter.......................................................7
PCB printedcircuitboard........................................................39
PD photodiode....................................................................7
PER polarizationextinctionratio.................................................77
PSD powerspectraldensity......................................................85
QAM quadratureamplitudemodulation...........................................2
QUANTUS "Quantengase unter Schwerelosigkeit". . .... . .... ..... ..... ..... ..... . .11
project association featuring drop tower experiments
Rb rubidium......................................................................6
RBW resolutionbandwidth.......................................................65
RF radiofrequency...............................................................13
typically in the range of 3kHz to 300GHz
RW ridgewaveguide.............................................................30
RWDC DC port for the RW pre-amplifier section .... . .... . .... . .... . .... . .... . .... 42
RW interface on MIOB
RMS rootmeansquare...........................................................96
SAC slowaxiscollimator.........................................................37
SDH self-delayedheterodyne.....................................................86
SHG secondharmonicgeneration.................................................6
frequency doubling technique
SMA SubMiniatureversionA.....................................................43
cable interface type
SMSR sidemodesuppressionratio...............................................65
SNR signaltonoiseratio.........................................................14
SOA semiconductoropticalamplifier.............................................28
129
TA taperedsectionofanopticalamplifier.........................................35
TADC DCportforthetaperedsection.............................................42
PA interface on MIOB
TAMOD modulation port for the PA gain section modulation ... ..... ..... ..... .... 42
PA interface on MIOB
TEXUS "Technologische Experimente unter Schwerelosigkeit" . . .... . .... . .... . .... . 8
rocket mission program of "Deutsche Zentrum für Luft- und Raumfahrt e. V."
(DLR)
TPA taperedpoweramplifier.....................................................72
TRL technologyreadinesslevel...................................................99
UHH "Universität Hamburg", Institut für Laserphysik .. .... . .... . .... . .... . .... . .. 45
UV ultraviolet...................................................................48
radiation with wavelengths in the range of 100nm –380nm
VSB-30 "Veículo de Sondagem Booster – 30" (engl.: Booster Sounding Vehicle)... ..8
Brazilian sounding rocket, used in FOKUS and MAIUS
WP workingpoint................................................................10
parameter set a diode laser operates with the aimed characteristics
130 APPENDIX D. TERMS AND ACRONYMS
Nomenclature
αlinewidth enhancement factor, also: Henry factor
αmcavity losses
¯
ynnth average normalized frequency deviation from nominal frequency over a finite
time interval
βisensitivity to potential violation of the LPI
βsp spontaneous emission factor
∆νsp linewidth enhancement due to spontaneous emission
∆ωthe difference of the angular frequency of two laser beams
∆Uvarying gravitational field
εconversion efficiency of electrical power into optical power.
ηiinjection efficiency, fraction of Ithat generates carriers in the active region
ηconversion efficiency, η=Pin/Po
ηddifferential quantum efficiency, defined by the number of photons emitted per
injected electron
Γconfinement factor, ratio between the volume of the active area and the volume
of the optical mode
νthe frequency of the photon
ρth thermal conductivity
σ2Allan variance
τtime interval, carrier lifetime
τmequivalent lifetime of photon in resonator before leaving through mirror
τpaverage lifetime of a photon in a resonator
τst lifetime of stimulated carriers
131
132 APPENDIX D. TERMS AND ACRONYMS
ϕ(t)the phase of the beat note signal
ALD footprint of the laser diode
cspeed of light, c=299792458 m/s
Emexcited energy state of an electron
Enground energy state of an electron
Emode stored optical energy in the cavity
ffrequency
fmmodulation frequency
Ggravitational constant, G=6.674 ×10−4m3/kgs2
ggain
gmmaterial gain coefficient
gamp amplifier gain coefficient
gthr gain above threshold
gusat unsaturated small signal gain
hthe Planck constant, h=6.626070 ×10−34 Js
hνenergy per photon
hrocket altitude of the experiment above ground
Hvert vertical height of the heat sink
Iinjection current
Imdifference of injection current, induced by modulation
Isstatic injection current, without modulation
Ithr threshold current
Kspontaneous emission enhancement factor, also: Petermann factor
Lresonator length
Lmchange of resonator length, induced by modulation
MEarth mass of Earth, M=5.974 ×1024 kg
Nnumber of carriers
133
nccarrier density
nr,mchange in the refraction index in the resonator due to carrier modulation
nrrefraction index
Pddissipation power
Pooptical output power
Pi,sat saturation input power
Po,sat saturation output power
qelementary electrical charge, q=1.602177 ×10−19 C
Rddifferential electrical resistance of the laser diode
Rlleakage rate
Rsserial resistance of laser diode
rEarth radius of Earth, r=6378 km
Riinjection rate of electrons
Rnr non-radiative recombination rate
Rrec electron recombination rate
Rsp spontaneous recombination rate
Rst stimulated recombination rate
Rth thermal resistance of laser diode
Snumber of photons
Ttemperature
T0characteristic temperature, given by the material system
t0time of rocket launch
Tmtemperature difference, induced by modulation
Tsstatic temperature, without modulation
TW P external applied temperature, e.g. working point temperature
Usstatic bias voltage
Vvolume
134 APPENDIX D. TERMS AND ACRONYMS
Vddiode voltage, quasi-Fermi level separation
vgvelocity of a photon in a gain medium
Vsseries voltage over laser diode, current-independent
Vbeat beat signal in heterodyne measurement, voltage of the photo diode
yfractional frequency
Ztthermal impedance of the laser diode
Appendix E
Publications
The following scientific publications have been prepared in connection with this thesis:
E.1 Print
Compact narrow linewidth diode laser modules for precision quantum optics
experiments on board of sounding rockets
A. Kohfeldt, Ch. Kürbis, E. Luvsandamdin, M. Schiemangk, A. Wicht, A. Peters, G.
Erbert, G. Tränkle
Proc. SPIE, Quantum Optics, volume 9900, no. 99001G, 2016.
Miniaturized lab system for cold atom experiments in microgravity
S. Kulas, Ch. Vogt, A. Resch, J. Hartwig, S. Ganske, J. Matthias, D. Schlippert, T. Wen-
drich, W. Ertmer, E.M. Rasel, M. Damjanic, P. Weßels, A. Kohfeldt, E. Luvsandamdin,
M. Schiemangk, Ch. Grzeschik, M. Krutzik, A. Wicht, A. Peters, S. Herrmann, C.
Lämmerzahl
Microgravity Science and Technology, volume 29, no. 1:37, 2017.
A compact and robust diode laser system for atom interferometry on a sounding
rocket
V. Schkolnik, O. Hellmig, A. Wenzlawski, J. Grosse, A. Kohfeldt, K. Döringshoff, A.
Wicht, P. Windpassinger, K. Sengstock, C. Braxmaier, M. Krutzik, A. Peters
Applied Physics B, volume 122, no. 8:217, 2016.
Space-born Frequency Comb Metrology
M. Lezius, T. Wilken, Ch. Deutsch, M. Giunta, O. Mandel, A. Thaller, V. Schkolnik, M.
Schiemangk, A. Dinkelaker, M. Krutzik, A. Kohfeldt, A. Wicht, A. Peters, O. Hellmig,
H. Duncker, K. Sengstock, P. Windpassinger, K. Lampmann, T. Hülsing, T. Hänsch, R.
Holzwarth
Optica, volume 3, no. 12:1381, 2016.
136 APPENDIX E. PUBLICATIONS
High power, micro-integrated diode laser modules at 767 and 780 nm for
portable quantum gas experiments
M. Schiemangk, K. Lampmann, A. Dinkelaker, A. Kohfeldt, M. Krutzik, Ch. Kürbis A.
Sahm, St. Spießberger, A. Wicht, G. Erbert, G. Tränkle, A Peters
Applied Optics, volume 54, no. 17:5332, 2015.
High power, narrow linewidth, micro-integrated semiconductor laser mod-
ules designed for quantum sensors in space
A. Kohfeldt, A. Bawamia, Ch. Kürbis, E. Luvsandamdin, M. Schiemangk, A. Wicht, G.
Erbert, A. Peters, G. Tränkle
CLEO: 2014 OSA Technical Digest, paper JTh2A.36, 2014.
A frequency comb and precision spectroscopy experiment in space
T. Wilken, M. Lezius, T. W. Hänsch, A. Kohfeldt, A. Wicht, V. Schkolnik, M. Krutzik, H.
Duncker, O. Hellmig, P. Windpassinger, K. Sengstock, A. Peters, R. Holzwarth
CLEO: 2013, OSA Technical Digest, paper AF2H.5, 2013.
Micro-integrated, high power, narrow linewidth master oscillator power amplifier
for precision quantum optics experiments in space
A. Kohfeldt, M. Schiemangk, St. Spießberger, A. Wicht, A. Peters, G. Erbert, G. Tränkle
CLEO 2012: OSA Technical Digest, paper JW3C.2, 2012.
E.2 Oral
This following list only contains presentations, where the author of this thesis was first
author:
Micro-integrated, high power, narrow linewidth master oscillator power amplifier
for precision quantum optics experiments in space
Conference on Lasers and Electro-Optics (CLEO): Science and Innovations 2012, San Jose
(USA), 6-11 May 2012.
Micro-integrated diode laser modules for high precision quantum sensors
in space
Deutsche Physikalische Gesellschaft (DPG), AMOP Annual Meeting, Heidelberg (Ger-
many), 23-27 March 2015.
Micro-integrated semiconductor laser modules designed for quantum sen-
sors in space
Laser Optics Berlin, Berlin (Germany), 18-20 March 2014.
Lichtquellen für quantenoptische Sensoren im Weltraum
Kolloquium des FBH, Berlin (Germany), 6 June 2014.
E.3. POSTER 137
E.3 Poster
This following list only contains presentations, where the author of this thesis was first
author:
Compact narrow linewidth diode laser modules for precision quantum op-
tics experiments on board of sounding rockets
SPIE Photonics Europe, Brussels (Belgia), 4-7 April 2016.
Micro-Integrated, Narrow Linewidth Master Oscillator Power Amplifier De-
signed for Quantum Sensors in Space
Conference on Lasers and Electro-Optics/European Quantum Electronics Conference
(CLEO Europe), Munich (Germany), 21-25 June 2015.
DFB-Master-Oscillator-Power-Amplifier system for high precision optical sensors
Deutsche Physikalische Gesellschaft (DPG), SKM Annual Meeting, Berlin (Germany),
15-20 March 2015.
High power, very narrow linewidth, micro-integrated diode laser modules
designed for quantum sensors in space
International Conference on Atomic Physics (ICAP), Washington DC (USA), 3-8 August
2014.
High power, narrow linewidth micro-integrated semiconductor laser mod-
ules designed for quantum sensors in space
Conference on Lasers and Electro-Optics (CLEO), San Jose (USA), 8-13 June 2014.
Dynamic properties of diode laser modules for space application
WIAS Workshop: Nonlinear Dynamics in Semiconductor Lasers, Berlin (Germany),
12-15 May 2014.
Micro-integrated semiconductor laser modules for precision quantum sen-
sors in space
Deutsche Physikalische Gesellschaft (DPG), AMOP Annual Meeting, Berlin (Germany),
17-21 March 2014.
Micro-integrated, high power, narrow linewidth diode lasers for precision
quantum optics experiments in space
European Frequency and Time Forum (EFTF), Gothenburg (Sweden), 23-27 April 2012.
138 APPENDIX E. PUBLICATIONS
Bibliography
[1]Maiman, T.H., “Stimulated Optical Radiation in Ruby”, Nature, volume 187, no.
4736: 493, 1960.
[2]Spetz, J., “Physicians and Physicists: The Interdisciplinary Introduction of the
Laser to Medicine”, in N. Rosenberg, A.C. Gelijns, and H. Dawkins (editors),
“Sources of Medical Technology: Universities and Industry”, chapter 3, pp.
41–66, Institute of Medicine (US) Committee on Technological Innovation in
Medicine, Washington, D.C., 5th edition, 1995.
[3]Steen, W., Watkins, K., and Mazumder, J., Laser Material Processing, Springer
London, London, 4 edition, 2010.
[4]Gowar, J., Optical Communication Systems, Prentice Hall, 2nd edition, 1993.
[5]Lambert, S.G. and Casey, W.L., Laser Communications in Space, Artech House,
1995.
[6]Donges, A. and Noll, R., Laser Measurement Technology: Fundamentals and Ap-
plications, Springer, Heidelberg, 1st edition, 2014.
[7]Browell, E., Ismail, S., McCormick, M., et al., “Spaceborne lidar system for mea-
surements of atmospheric water vapor and aerosols”, in “Earth Observing Sys-
tems Conference”, pp. AIAA–85–2091, 1985.
[8]Winker, D.M., Couch, R.H., and McCormick, M.P., “An overview of LITE: NASA’s
Lidar In-space Technology Experiment”, Proceedings of the IEEE, volume 84,
no. 2: 164, 1996.
[9]ESA, “Earth Explorers – Technical & Programmatic Annex to EarthCARE mis-
sion”, Technical Report, ESA, 2004.
[10]Chan, V.W.S., “Optical Satellite Networks”, J. Lightwave Technol., volume 21,
no. 11: 2811, 2003.
[11]Sodnik, Z., Furch, B., and Lutz, H., “Free-space laser communication activities
in Europe: SILEX and beyond”, in “Conference Proceedings - Lasers and Electro-
Optics Society Annual Meeting-LEOS”, pp. 78–79, 2006.
[12]Gaudel, I., “Operational Use of SILEX Optical Laser Link between SPOT4 and
ARTEMIS Satellites”, in “Space OPS 2004 Conference”, pp. 1–12, 2004.
139
140 BIBLIOGRAPHY
[13]Norimatsu, S. and Iwashita, K., “Linewidth requirements for optical synchronous
detection systems with nonnegligible loop delay time”, Journal of Lightwave
Technology, volume 10, no. 3: 341, 1992.
[14]Ip, E. and Kahn, J.M., “Carrier Synchronization for 3- and 4-bit-per-Symbol Op-
tical Transmission”, Journal of Lightwave Technology, volume 23, no. 12: 4110,
2005.
[15]Steck, D.A., “Rubidium 87 D Line Data”, Technical Report, Oregon Center for
Optics and Department of Physics, Univers ity of Oregon, Oregon, 2010.
[16]Tiecke, T.G., “Properties of Potassium”, Technical Report v1.02, Harvard Univer-
sity, 2011.
[17]Kasevich, M. and Chu, S., “Measurement of the gravitational acceleration of
an atom with a light-pulse atom interferometer”, Applied Physics B, volume 54,
no. 5: 321, 1992.
[18]Planck, M., “Zur Theorie des Gesetzes der Energieverteilung im Normalspek-
trum”, in “Verhandlungen der Deutschen Physikalischen Gesellschaft”, volume 2,
p. 237, Berlin, 1900.
[19]Einstein, A., “Über einen die Erzeugung und Verwandlung des Lichtes betref-
fenden heuristischen Gesichtspunkt”, Annalen der Physik, volume 17: 132, 1905.
[20]Einstein, A., “Zur Quantentheorie der Strahlung”, Mitteilungen der Physikalische
Gesellschaft Zürich, volume 18: 47, 1916.
[21]Fedak, W.A. and Prentis, J.J., “The 1925 Born and Jordan paper ”On quantum
mechanics””, American Journal of Physics, volume 77, no. 2: 128, 2009.
[22]De Broglie, L., “Waves and Quanta”, Nature, volume 112, no. 2815: 540, 1923.
[23]Abbott, B.P., Abbott, R., Abbott, T.D., et al., “Observation of Gravitational Waves
from a Binary Black Hole Merger”, Phys. Rev. Lett., volume 116, no. 6: 16, 2016.
[24]Weiss, D.S., Young, B.C., and Chu, S., “Precision measurement of hbar/m_{Cs}
based on photon recoil using laser-cooled atoms and atomic interferometry”,
Applied Physics B, volume 59, no. 3: 217, 1994.
[25]Wicht, A., Hensley, J.M., Sarajlic, E., et al., “A Preliminary Measurement of the
Fine Structure Constant Based on Atom Interferometry”, Physica Scripta, volume
2002, no. T102: 82, 2002.
[26]Schlippert, D., Hartwig, J., Albers, H., et al., “Quantum Test of the Universality
of Free Fall”, Physical Review Letters, volume 112, no. 20: 203002, 2014.
[27]Peters, A., Chung, K.K.Y., and Chu, S., “Measurement of gravitational accelera-
tion by dropping atoms”, Nature, volume 400, no. 6747: 849, 1999.
BIBLIOGRAPHY 141
[28]Tino, G.M., Cacciapuoti, L., Bongs, K., et al., “Atom interferometers and optical
atomic clocks: New quantum sensors for fundamental physics experiments in
space”, Nuclear Physics B - Proceedings Supplements, volume 166: 159, 2007.
[29]Poli, N., Oates, C.W., Gill, P., et al., “Optical atomic clocks”, arXiv: 1401.2378
[physics], 2014.
[30]Cornell, E.A. and Wieman, C.E., “Nobel lecture: Bose-Einstein condensation in
a dilute gas, the first 70 years and some recent experiments”, Reviews of Modern
Physics, volume 74, no. 3: 875, 2002.
[31]Ketterle, W., “Nobel lecture: When atoms behave as waves: Bose-Einstein con-
densation and the atom laser”, Reviews of Modern Physics, volume 74, no. Octo-
ber: 1131, 2002.
[32]Torii, Y., Suzuki, Y., Kozuma, M., et al., “Mach-Zehnder Bragg interferometer for
a Bose-Einstein condensate”, Physical Review A, volume 61, no. 4: 41602, 2000.
[33]Kasevich, M. and Chu, S., “Atomic interferometry using stimulated Raman tran-
sitions”, Physical Review Letters, volume 67, no. 2: 181, 1991.
[34]de Angelis, M., Bertoldi, A., Cacciapuoti, L., et al., “Precision gravimetry with
atomic sensors”, Measurement Science and Technology, volume 20, no. 2: 22001,
2009.
[35]Dickerson, S.M., Hogan, J.M., Sugarbaker, A., et al., “Multiaxis Inertial Sens-
ing with Long-Time Point Source Atom Interferometry”, Physical Review Letters,
volume 111, no. 8: 83001, 2013.
[36]Müntinga, H., Ahlers, H., Krutzik, M., et al., “Interferometry with Bose-Einstein
Condensates in Microgravity”, Physical Review Letters, volume 110: 93602,
2013.
[37]Peters, A., Chung, K.Y., and Chu, S., “High-precision gravity measurements using
atom interferometry”, Metrologia, volume 38: 25, 2001.
[38]Kessler, T., Hagemann, C., Grebing, C., et al., “A sub-40-mHz-linewidth laser
based on a silicon single-crystal optical cavity”, Nature Photonics, volume 6,
no. 10: 687, 2012.
[39]Nicholson, T.L., Campbell, S.L., Hutson, R.B., et al., “Systematic evaluation of
an atomic clock at 2 x 10e(-18) total uncertainty”, Nature communications, vol-
ume 6: 6896, 2015.
[40]Bresson, A., Bidel, Y., Bouyer, P., et al., “Quantum mechanics for space applica-
tions”, Applied Physics B: Lasers and Optics, volume 84, no. 4: 545, 2006.
[41]Udem, T., Holzwarth, R., and Hänsch, T.W., “Optical frequency metrology”, Na-
ture, volume 416, no. 6877: 233, 2002.
142 BIBLIOGRAPHY
[42]Lezius, M., Wilken, T., Deutsch, C., et al., “Space-born Frequency Comb Metrol-
ogy”, Optica, volume 3, no. 12: 1381, 2016.
[43]Duncker, H., Hellmig, O., Wenzlawski, A., et al., “Ultrastable, Zerodur-based op-
tical benches for quantum gas experiments.”, Applied optics, volume 53, no. 20:
4468, 2014.
[44]Baumann, E., Giorgetta, F.R., Nicholson, J.W., et al., “High-performance,
vibration-immune, fiber-laser frequency comb”, Optics Letters, volume 34, no. 5:
638, 2009.
[45]Hänsel, W., Holzwarth, R., Doubek, R., et al., “Laser with non-linear optical loop
mirror”, patent no. US 8873601 B2, 2013.
[46]Wilken, T., Lezius, M., Hänsch, T.W., et al., “A frequency comb and precision
spectroscopy experiment in space”, in “CLEO: 2013”, p. AF2H.5, OSA, 2013.
[47]TEXUS Team, “TX-RIBRE-ICD-TBD TEXUS Technical Report”, Technical Report,
EADS Astrium.
[48]Seidel, S.T., Eine Quelle für die Interferometrie mit Bose-Einstein-Kondensaten
auf Höhenforschungsraketen, Ph.D. thesis, Gottfried Wilhelm Leibniz Universität
Hannover, 2014.
[49]Deutsches Zentrum für Luft- und Raumfahrt (DLR), TEXUS - Forschen in Schw-
erelosigkeit, Missions-Broschüre, DLR, 2007.
[50]Cronin, A.D., Schmiedmayer, J., and Pritchard, D.E., “Optics and interferometry
with atoms and molecules”, Reviews of Modern Physics, volume 81, no. 3: 1051,
2009.
[51]van Zoest, T., Gaaloul, N., Singh, Y., et al., “Bose-Einstein Condensation in Mi-
crogravity”, Science, volume 328, no. 5985: 1540, 2010.
[52]Dittus, H., Lämmerzahl, C., and Turyshev, S.G., Lasers, Clocks, and Drag-Free:
Exploration of Relativistic Gravity in Space, Springer, Berlin, 2009.
[53]Phillips, W. and Metcalf, H., “Cooling and Trapping Atoms”, Scientific American,
volume 256, no. 3: 50, 1987.
[54]Chu, S., “Laser Trapping of Neutral Particles”, Scientific American, volume 266,
no. 2: 70, 1992.
[55]Chu, S., Hollberg, L., Bjorkholm, J.E., et al., “Three-Dimensional Viscous Con-
finement and Cooling of Atoms by Resonance Radiation Pressure”, Physical Re-
view Letters, volume 55, no. 1: 48, 1985.
[56]Metcalf, H.J. and Straten, P.V.D., “Laser cooling and trapping of atoms”, Journal
of Optical Society of America, volume 20, no. 5: 887, 2003.
BIBLIOGRAPHY 143
[57]Anderson, M.H., Ensher, J.R., Matthews, M.R., et al., “Observation of Bose-
Einstein condensation in a dilute atomic vapor”, Science (New York, N.Y.), volume
269, no. 5221: 198, 1995.
[58]Hess, H.F., “Evaporative cooling of magnetically trapped and compressed spin-
polarized hydrogen”, Physical Review B, volume 34, no. 5: 3476, 1986.
[59]Grosse, J., Thermal and Mechanical Design and Simulation for the first High Preci-
sion Quantum Optics Experiment on a Sounding Rocket, Ph.D. thesis, Universität
Bremen, 2016.
[60]Schkolnik, V., Probing gravity with quantum sensors - on ground and in space -,
Ph.D. thesis, Humboldt-Universität zu Berlin, 2016.
[61]Wells, H.G., The War of the Worlds, William Heinemann, 1898.
[62]Schawlow, A.L. and Townes, C.H., “Infrared and Optical Masers”, Physical Re-
view, volume 112, no. 6: 1940, 1958.
[63]von Neumann, J., Collected Works 5: Design of computers, theory of automation
and numerical analysis, Pergamon Press, 1963.
[64]von Neumann, J., “Notes on the photon-disequilibrium-amplification scheme
(JvN), September 16, 1953”, IEEE Journal of Quantum Electronics, volume 23,
no. 6: 659, 1987.
[65]Hall, R.N., Fenner, G.E., Kingsley, J.D., et al., “Coherent Light Emission From
GaAs Junctions”, Physical Review Letters, volume 9, no. 9: 366, 1962.
[66]Alferov, Z.I. and Kazarinov, R.F., “Semiconductor laser with electric pumping”,
patent no. 181737 [in Russian], 1963.
[67]Kroemer, H., “A proposed class of hetero-junction injection lasers”, Proceedings
of the IEEE, volume 51, no. 12: 1782, 1963.
[68]Alferov, Z., Andreev, V., Garbuzov, D., et al., “Investigation of the influence of the
AlAs - GaAs heterostructure parameters on the laser threshold current and the
realization of continuous emission at room temperature”, Sov. Phys.-Semicond.,
volume 4: 1573, 1971.
[69]Siegman, A., Lasers, University Science Books, 1st edition, 1986.
[70]Sarzała, R.P., Piskorski, Ł., Szczerbiak, P., et al., “An attempt to design long-
wavelength (>2µm) InP-based GaInNAs diode lasers”, Applied Physics A: Mate-
rials Science and Processing, volume 108, no. 3: 521, 2012.
[71]Diehl, R., High-power diode lasers, Springer, Berlin, 2000.
[72]Spießberger, S., Schiemangk, M., Wicht, A., et al., “Narrow linewidth DBR-RW
lasers emitting near 1064 nm”, Journal of Lightwave Technology, volume 28,
no. 17: 2611, 2010.
144 BIBLIOGRAPHY
[73]Lewoczko-Adamczyk, W., Pyrlik, C., Häger, J., et al., “Ultra-narrow linewidth
DFB-laser with optical feedback from a monolithic confocal Fabry-Perot cavity”,
Optics Express, volume 23, no. 8: 9705, 2015.
[74]DeMaria, A. and Hennessey, T., “The CO2 Laser: The Workhorse of the Laser
Material Processing Industry”, SPIE Professional Magazine, pp. 1–27, 2010.
[75]Webster, S.A., Oxborrow, M., and Gill, P., “Subhertz-linewidth Nd:YAG laser.”,
Optics letters, volume 29, no. 13: 1497, 2004.
[76]Lévèque, T., Faure, B., Esnault, F.X.X., et al., “PHARAO Laser Source Flight
Model: Design and Performances”, arXiv:1502.06376 [physics], pp. 1–12, 2015.
[77]Theron, F., Carraz, O., Renon, G., et al., “Narrow linewidth single laser source
system for onboard atom interferometry”, arXiv:1407.4684 [physics], 2014.
[78]Chow, W.W. and Koch, S.W., Semiconductor - Laser Fundamentals, Springer, 1999.
[79]Servagent, N., “Specific applications of sensors for photonic
and imaging - Laser Diode Principle”, http://www.optique-
ingenieur.org/en/courses/OPI_ang_M05_C04/co/Contenu.html, accessed
on 2015-12-22, 2007.
[80]Bartl, J., Fíra, R., and Jacko, V., “Tuning of the laser diode”, Measurement Science
Review, volume 2, no. 3: 9, 2002.
[81]Petermann, K., Laser Diode Modulation and Noise, Kluwer Academic Publishers,
1988.
[82]Coldren, L.A., Corzine, S.W., and Masanovic, M.L., Diode Lasers and Photonic
Integrated Circuits, Wiley, 2nd edition, 2012.
[83]Paoli, T.L., “Saturation Behavior of the Spontaneous Emission from Double-
Heterostructure Junction Lasers Operating High Above Threshold”, IEEE Journal
of Quantum Electronics, volume 9, no. 2: 267, 1973.
[84]Coenning, W. and Bürkle, H.P., “Phasendiagramme zur Beschreibung des Hal-
bleiterlasers bei Frequenzmodulation”, Archiv für Elektrotechnik, volume 75,
no. 5: 369, 1992.
[85]Kobayashi, S., Yamamoto, Y., Ito, M., et al., “Direct Frequency Modulation In Al-
GaAs Semiconductor Lasers”, IEEE Transactions on Microwave Theory and Tech-
niques, volume 30, no. 4: 428, 1982.
[86]Petermann, K., “Calculated Spontaneous Emission Factor for Double-
Heterostructure Injection Lasers with Gain-Induced Waveguiding”, IEEE Journal
of Quantum Electronics, volume 15, no. 7: 566, 1979.
[87]Henry, C.H., “Theory of the linewidth of semiconductor lasers”, IEEE Journal of
Quantum Electronics, volume 18, no. 2: 259, 1982.
BIBLIOGRAPHY 145
[88]Kojima, K., Kyuma, K., and Nakayama, T., “Analysis of the spectral linewidth of
distributed feedback laser diodes”, Journal of Lightwave Technology, volume 3,
no. 5: 1048, 1985.
[89]Nguyen, T.P., Schiemangk, M., Spießberger, S., et al., “Optimization of 780 nm
DFB diode lasers for high-power narrow linewidth emission”, Applied Physics B,
volume 108, no. 4: 767, 2012.
[90]Luvsandamdin, E., Spießberger, S., Schiemangk, M., et al., “Development of
narrow linewidth, micro-integrated extended cavity diode lasers for quantum
optics experiments in space”, Applied Physics B: Lasers and Optics, volume 111,
no. 2: 255, 2013.
[91]Kogelnik, H., “Coupled-Wave Theory of Distributed Feedback Lasers”, Journal of
Applied Physics, volume 43, no. 5: 2327, 1972.
[92]Brox, O., Bugge, F., Mogilatenko, A., et al., “Distributed feedback lasers in the
760 to 810 nm range and epitaxial grating design”, Semiconductor Science and
Technology, volume 29, no. 9: 095018, 2014.
[93]Connelly, M., Semiconductor Optical Amplifiers, Kluwer Academic Publisher,
Boston, 2002.
[94]Moser, A., “Thermodynamics of facet damage in cleaved AlGaAs lasers”, Applied
Physics Letters, volume 59, no. 1991: 522, 1991.
[95]Ressel, P., Erbert, G., Zeimer, U., et al., “Novel passivation process for the mir-
ror facets of Al-free active-region high-power semiconductor diode lasers”, IEEE
Photonics Technology Letters, volume 17, no. 5: 962, 2005.
[96]Zah, C., Bhat, R., Menocal, S., et al., “1.5 um GaInAsP angled-facet flared-
waveguide traveling-wave laser amplifiers”, IEEE Photonics Technology Letters,
volume 2, no. 1: 46, 1990.
[97]Sochava, S., “Semiconductor optical amplifier for an external cavity diode laser”,
patent no. US20080304826 A1, 2008.
[98]Bendelli, G., Komori, K., Arai, S., et al., “A new structure for high-power TW-
SLA (Travelling wave semiconductor laser amplifier)”, IEEE Photonics Technology
Letters, volume 3, no. 1: 42, 1991.
[99]Fiebig, C., Blume, G., Uebernickel, M., et al., “High-power DBR-tapered laser at
980 nm for single-path second harmonic generation”, IEEE Journal on Selected
Topics in Quantum Electronics, volume 15, no. 3: 978, 2009.
[100]Fiebig, C., Diodenlaser mit Trapezstruktur und hoher Brillanz für die Realisierung
einer Frequenzkonversion auf einer mikro-optischen Bank Elektrotechnik und In-
formatik, Ph.D. thesis, Technische Universität Berlin, 2013.
146 BIBLIOGRAPHY
[101]Paschke, K., Spießberger, S., Kaspari, C., et al., “High-power distributed Bragg
reflector ridge-waveguide diode laser with very small spectral linewidth”, Optics
letters, volume 35, no. 3: 402, 2010.
[102]Kohfeldt, A., Kürbis, C., Luvsandamdin, E., et al., “Compact narrow linewidth
diode laser modules for precision quantum optics experiments on board of
sounding rockets”, Proc. of SPIE, volume 9900, no. 0: 99001G, 2016.
[103]Eppich, B., “Optical Design of Beam Delivery and Beam Forming Systems”, Optik
and Photonik, volume 2, no. 2: 48, 2008.
[104]Schkolnik, V., Hellmig, O., Wenzlawski, A., et al., “A compact and robust diode
laser system for atom interferometry on a sounding rocket”, Applied Physics B,
volume 122, no. 8: 217, 2016.
[105]Schiemangk, M., Lampmann, K., Dinkelaker, A., et al., “High power, micro-
integrated diode laser modules at 767 nm and 780 nm for portable quantum
gas experiments”, Applied Optics, volume 54, no. 17: 5332, 2015.
[106]Meschede, D., Optik, Licht und Laser, Vieweg+Teubner, Wiesbaden, 3rd edition,
2008.
[107]Chik, K.C.D., Devenyi, T.F., and Dyment, J.C., “Method for screening laser
diodes”, patent no. US4489477 A, 1985.
[108]Koenig, M., Hartmayer, R., Thiagarajan, P., et al., “Automated tests improve laser
diode manufacturing”, Laser Focus World, volume 34, no. 8, 1998.
[109]Johnson, L.A., “Laser Diode Burn-in and Reliability Testing”, IEEE Communia-
tions, volume 44, no. 2: 4, 2006.
[110]Huang, J.s., “Burn-In Aging Behavior and Analytical Modeling of Wavelength-
Division Multiplexing Semiconductor Lasers : Is the Swift Burn-In Feasible for
Long-Term Reliability Assurance ?”, Advances in OptoElectronics, volume 2013,
no. 568945, 2013.
[111]Spießberger, S., Compact Semiconductor-Based Laser Sources with Narrow
Linewidth and High Output Power, Ph.D. thesis, Technische Universität Berlin,
2012.
[112]Schiemangk, M., Spießberger, S., Wicht, A., et al., “Accurate frequency noise
measurement of free-running lasers”, Applied Optics, volume 53, no. 30: 7138,
2014.
[113]Kourogi, M. and Ohtsu, M., “Novel optical frequency discriminator for FM noise
reduction of semiconductor lasers”, Optics communications, volume 81, no. 3-4:
204, 1991.
BIBLIOGRAPHY 147
[114]Spießberger, S., Schiemangk, M., Wicht, A., et al., “DBR laser diodes emitting
near 1064 nm with a narrow intrinsic linewidth of 2 kHz”, Applied Physics B:
Lasers and Optics, volume 104, no. 4: 813, 2011.
[115]Okoshi, T., Kikuchi, K., and Nakayama, A., “Novel method for high resolution
measurement of laser output spectrum”, Electronics Letters, volume 16, no. 16:
630, 1980.
[116]Mercer, L.B.L.B., “1/f frequency noise effects on self-heterodyne linewidth mea-
surements”, Journal of Lightwave Technology, volume 9, no. 4: 485, 1991.
[117]European Space Components Coordination, “Evaluation Test Programme Guide-
lines for Laser Diode Modules, ESCC Basic Specification No . 23201”, Technical
Report 23201, ESA, 2014.
[118]European Cooperation for Space Standardization, “Web page of the ECSS”,
http://ecss.nl/, accessed on 2016-12-16, 2016.
[119]European Space Components Coordination, “Web page of the ESCC”,
https://spacecomponents.org, accessed on 2016-12-16, 2016.
[120]Sexton, F.W., “Destructive single-event effects in semiconductor devices and ICs”,
IEEE Transactions on Nuclear Science, volume 50 III, no. 3: 603, 2003.
[121]Garden, C., “Effects of Proton Radiation on InGaAs Photodiodes and Laser-
diodes”, Technical Report, ESA, 2000.
[122]Department of Defense, “MIL-STD-202G: Test Method Standard Electronic and
Electrical Component Parts”, Technical Report G, US Department of Defense,
Columbus, 2002.
[123]Department of Defense, “MIL-STD-883J: Test Method Standard Microcircuits”,
Technical Report J, US Department of Defense, Columbus, 2012.
[124]Nordland, “Norland Optical Adhesive 88”, Technical Report, Norland Products,
Cranbury, 2016.
[125]NASA, “Space Shuttle Specification Environmental Acceptance Testing”, Techni-
cal Report SP-T-0023, NASA, 2001.
[126]Burkhardt, T., Hornaff, M., Kamm, A., et al., “Low-strain laser-based solder join-
ing of mounted lenses”, in “Proc. SPIE 9574, Material Technologies and Ap-
plications to Optics, Structures, Components, and Sub-Systems II”, San Diego,
California, 2015.
[127]Robins, W.P., Phase Noise in Signal Sources: Theory and Applications, IET, 1984.
[128]Barnes, J. and Allan, D.W., “Variances based on data with dead time between
the measurement”, NIST Tech. Note, volume 1318, 1990.
148 BIBLIOGRAPHY
[129]Affolderbach, C. and Mileti, G., “A compact laser head with high-frequency sta-
bility for Rb atomic clocks”, Review of Scientific Instruments, volume 76, no. 7:
1, 2005.
[130]Luvsandamdin, E., Kürbis, C., Schiemangk, M., et al., “Micro-integrated ex-
tended cavity diode lasers for precision potassium spectroscopy in space”, Optics
Express, volume 22, no. 7: 7790, 2014.
[131]Ludlow, A.D., Boyd, M.M., Ye, J., et al., “Optical atomic clocks”, Reviews of Mod-
ern Physics, volume 87: 637, 2015.
[132]Fortier, T.M., Ashby, N., Bergquist, J.C., et al., “Improved limits on variation
of the fine structure constant and violation of Local Position Invariance”, Pro-
ceedings of the IEEE International Frequency Control Symposium and Exposition,
volume 070801, no. 2: 663, 2007.
[133]Pro-Physik.de, “Bose-Einstein-Kondensat im All”, http://www.pro-
physik.de/details/news/10417601/Bose-Einstein-Kondensat_im_All.html,
accessed on 2017-03-16, 2017.
[134]Kulas, S., Vogt, C., Resch, A., et al., “Miniaturized lab system for cold atom
experiments in microgravity”, Microgravity Science and Technology, pp. 1–12,
2016.
List of Figures
2.1 Schematic of FOKUS experiment: the RF atomic CSAC is compared to
the optical atomic clock, formed by the Rb reference system, via the
frequency comb system. Optical signals are illustrated in red, electrical
signals in blue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 FOKUS Rb reference system, hosting a spectroscopy board (right) with
master laser (left, underneath semi-transparent lid) in flight model hous-
ing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Flight configuration of a VSB-30 sounding rocket . . . . . . . . . . . . . 8
2.4 Acceleration load (a) and vibration level (b) of a typical launch. Burnout
of 1st stage at 13.5s, maximum load is reached before burnout of 2nd
stage at 36s. Data from TEXUS 45 mission [47,48]. . . . . . . . . . . . 9
2.5 Official logo of MAIUS-1 mission . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6 Hyperfine structure of 87Rb D2 transition line and frequencies differ-
ences. Data from [15]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.7 Sketch of VSB-30 rocket with MAIUS payload, modified version in [59]15
2.8 Schematic of laser system with four experiment lasers, spectroscopy
module, and switching and distribution module, as presented in [60]. 15
3.1 Definition of the spatial orientation of an edge emitting laser diode in
this thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2 Emitting Junctions a) Homojunction. b) Double-heterojunction. In the
case of a heterojunction, a better confinement of the photon carriers can
be obtained. [79], licence: CC-PD . . . . . . . . . . . . . . . . . . . . . . . 22
3.3 Example of amplifier gain vs. output power . . . . . . . . . . . . . . . . . 29
3.4 Concept of monolithic and hybrid semiconductor laser MOPA . . . . . . 31
4.1 Hybrid integrated MOPA module on AlN MIOB with integrated electrical
interface. A modified version published in [102]. . . . . . . . . . . . . . 34
4.2 Lateral and vertical view of optical configuration on MOPA . . . . . . . 35
4.3 Detailed picture of a MOPA: DFB on submount with temperature sensor,
coax connector for temperature sensor, collimating lenses and optical
micro-isolator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.4 MIOB with integrated DFB and fibre coupling on a Cu mount adapter
(Image courtesy of V. Schkolnik [104]) . . . . . . . . . . . . . . . . . . . . 38
4.5 Concept of p-up and p-down mounting . . . . . . . . . . . . . . . . . . . 40
149
150 LIST OF FIGURES
4.6 Electrical interface provided by a MIOB. TADC: DC injection current of
PA gain section, TAMOD: modulation of PA gain section, RWDC: DC in-
jection current of PA pre-amplifier section, DC: DC injection current of
MO (bypassing modulation electronics), MDC: DC injection current of
MO (to be modulated by modulation electronics), MRF: RF modulation
port, MMOD: LF modulation port, TSMO: MO temperature sensor port,
TSMIOB: MIOB temperature sensor port, TSPA: PA temperature sensor
port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.7 Electrical interface on MO-MIOB: the DC port supply can either be mod-
ulated with a common source circuit (MMOD with T1) or with a bias-tee
path (MRF). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.8 CCP with serial number (marked red) and imprints of the safety spring
washers (marked blue) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.9 Housing package with integrated electrical adaptor (a) on left side, and
(b) on the right side of the optical axis . . . . . . . . . . . . . . . . . . . . 44
4.10 Overview of laser module integration process flow . . . . . . . . . . . . . 47
4.11 Schematic of assembly station and monitoring . . . . . . . . . . . . . . . 48
4.12 Overview of master oscillator optics integration . . . . . . . . . . . . . . 50
4.13 Variation of near field beam diameter by varying the position of the
(a) FAC and (b) SAC lens. The beam is collimated at the Zero-position
in this plot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.14 Overview of power amplifier optics integration . . . . . . . . . . . . . . 54
4.15 Variation of amplifier output power by varying the position of the (a)
FAC and (B) SAC coupling lens. The MO-beam is coupled best into the
the amplifier at zero-position in this plot. . . . . . . . . . . . . . . . . . . 55
5.1 Identified in component qualification: multi-mode spectrum of a DFB
laser (a), and thermal roll-over of a PA(b) . . . . . . . . . . . . . . . . . . 58
5.2 Degradation in the spectrum of a DFB diode over time . . . . . . . . . . 60
5.3 Overview of diode qualification process . . . . . . . . . . . . . . . . . . . 61
5.4 Submount holder with electrical interface and a 3 mm DFB mounted (a),
and an open burn-in box for four submount holders (b) . . . . . . . . . 62
5.5 Schematic of pre-characterization setup in PA characterization configu-
ration. (i) seed from PM fibre, (ii) DUT with current driver and temper-
ature control, (iii) voltmeter for RW photo current measurement, (iv)
power meter with Ulbricht sphere, and (v) fibre coupling for OSA. . . . 63
5.6 Total yield of facet inspection and failure causes . . . . . . . . . . . . . . 68
5.7 Yield of facet inspection for each wafer and length . . . . . . . . . . . . . 68
5.8 Causes of failure in EO characterization . . . . . . . . . . . . . . . . . . . 69
5.9 Yield of EO characterization for each wafer and length . . . . . . . . . . 69
5.10 Total yield of pre-characterization of DFB diodes. . . . . . . . . . . . . . 70
5.11 Mounting of the PAs: either p-up with heat spreader (left) or p-down
with the submount as heat sink (right) . . . . . . . . . . . . . . . . . . . . 71
5.12 Total yield of facet inspection of PA diodes. . . . . . . . . . . . . . . . . . 73
5.13 Distribution of facet defects due to mounting configuration. . . . . . . . 73
LIST OF FIGURES 151
5.14 PA submounts after shearing of the diode. Areas where the solder did not
wet the diode surface are marked in red. On the left: Bonding without
Ar-cleaning process, wettability: 40%. On the right: submount was
cleaned with Ar before bonding, wettability: 95%. . . . . . . . . . . . . 74
5.15 Output power dependent on submount temperature for various PA chips 75
5.16 Total yield of the PA selection process . . . . . . . . . . . . . . . . . . . . . 76
6.1 Relation of output power and MIOB temperature against amplifier input
current (left axis), and conversion efficiency for 25°C (right axis) . . . 82
6.2 Optical output power of MOPA module (a) and amplifier gain (b) against
MO optical output power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.3 Wavelength of a MOPA, hosting a 3mm long master laser versus (a) MO
current and (b) MO temperature. The spectra are normalized to the
global peak of all emission spectra in the figure. . . . . . . . . . . . . . . 83
6.4 (a) Comparison of spectral output with (red) and without (blue) input
from the master oscillator. (b) MOPA spectrum calibrated, optical power
in mW vs. wavelength. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.5 Measurement setup for linewidth measurement: Beat note with fre-
quency difference capable for photo diode (PD) detection is created ei-
ther with (a) lasers DUT and OLO or (b) self-delayed with a DUT, mod-
ulated with an AOM, first order output of AOM is delayed with a long
fibre. PD signal is analysed by signal analyser (SA) which delivers IQ-
data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.6 Linewidth of MOPA: frequency noise PSD and beat note spectrum (a)
and linewidth distribution against MO output power (b) . . . . . . . . . 89
6.7 DC Influence on MMOD voltage on wavelength and output power of a
3mm long DFB. The output power is measured behind a 60 dB µ-isolator. 90
6.8 Spectral performance of DUT before and after vibration and shock tests.
(a) spectral map, (b) single spectrum at WP. Performance remained
stable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.9 (a) Power recording (above) and RMS deviation (below) of a MOPA
emitting during lateral (along the x-axis of the shaker) vibration test
of 8gRMS, (b) RMS values of these power recordings before, during and
after the vibration tests. pre: before the test, no external movement, vib:
random vibration run, post: after the test, no external movement. X,Y,
and Z are the directions of movement. The longitudinal (Y) vibration
run was performed first. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.10 Output power development after several thermal cyclings. (a) over TA
injection current, (b) over DFB injection current . . . . . . . . . . . . . . 97
7.1 Frequency deviation of beat note signal from median (upper panel) and
relative stability σ(lower panel) of long-term measurement of Rb refer-
ence system compared to the CSAC . . . . . . . . . . . . . . . . . . . . . . 100
152 LIST OF FIGURES
7.2 Photodiode read out (blue) and error signal (red) of the Rb reference
system before and during flight. The laser frequency was stabilized be-
fore launch and stayed stable the entire mission except for the planned
scan over the injection current. . . . . . . . . . . . . . . . . . . . . . . . . 103
7.3 Relative changes of clock frequency comparison (blue squares) and grav-
itational potential (black dotted line) before and during flight. A linear
drift of 122Hz/s was subtracted from the optical frequency reading. . 104
7.4 Optical output (a) power and linewidth (b) of FOKUS laser module be-
fore and after mechanical stress tests. . . . . . . . . . . . . . . . . . . . . . 104
7.5 Spectral performance of FOKUS diode laser module before and after vi-
bration tests and after the flight campaign, (a) spectral map, (b) single
spectrum at WP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.6 Telemetry data of MO injection current (top), the rear free-space optical
power of the MO (centre), and optical power in fibre (bottom) of MAIUS
laser modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
7.7 RMS power fluctuations of the in-fibre power readings of the laser mod-
ules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8.1 A laser module of the next generation, hosting a fibre coupling. Beam
adjustment is done by mirrors on spherical mounts, the fibre ferrule con-
tains a collimating lens. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
B.1 Functional schematic of MO electrical interface . . . . . . . . . . . . . . . 117
B.2 Functional schematic of PA electrical interface . . . . . . . . . . . . . . . 118
C.1 Facets of DFB diode C2212-6-1: 00-09-22 (a,b) and TPA diode C2358-3:
08-02-13 (c,d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
C.2 Questionable facets of DFB diode C2610-6-2: 02-0304-21 (a), (b), 01-
04-13 (d) and C2212-6-1: 00-09-16 (c), as well as TPA diodes C2358-3:
08-06-08 (e) and 08-09-11 (f) . . . . . . . . . . . . . . . . . . . . . . . . . 122
C.3 facets failed in inspection because of active area defects. DFBs: C2212-
6-1: 00-07-07 (a) and 00-10-05 (b), TPA C2358-3: 01-01-03 (c) . . . 123
C.4 facets failed in inspection because of bulk defects. DFBs: C2610-6-2
01-04-17 (a) and 01-01-15 (b); TPA: C2358-3 08-07-16 (c) . . . . . . . 124
C.5 Facets of diode C2610-6-2: 02-0304-10, destroyed front facet after qual-
ification process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
List of Tables
2.1 Requirements on laser module of the Rb reference system due to exper-
imental needs and microgravity platform environment . . . . . . . . . . 10
2.2 Requirements on laser modules in the MAIUS laser system according to
the scientific and environmental needs . . . . . . . . . . . . . . . . . . . . 17
4.1 List of beam propagation properties of DFB lasers and amplifier diodes
used in FOKUS and MAIUS modules. Values behind "/" refer to rear facet
properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.1 List of components that are qualified before integration, and their qual-
ification property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.2 Overview of DFB integration candidates. Denotation in brackets: (in-
complete facet data, incomplete electro-optical data) . . . . . . . . . . . 67
5.3 Overview of PA DUTs. All testes chips are 4 mm long. Denotation in
brackets: (incomplete facet data, incomplete electro-optical data) . . . 72
5.4 List average performance of micro-isolators, classified into wavelength:
767nm and 780nm, as well as in design: Single stage (SiSt) and Semi-
double stage (SDSt) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.1 Overview of environmental tests performed in this chapter . . . . . . . 92
6.2 PSD loads per frequency range of random vibration test . . . . . . . . . 92
6.3 Results of laser module characterization and stress tests compared to
requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.1 Time table of FOKUS mission flight . . . . . . . . . . . . . . . . . . . . . . 102
7.2 Time table of FOKUS mission flight . . . . . . . . . . . . . . . . . . . . . . 107
A.1 List of laser modules used in this work. The diode number of the PA in
MOPA11-01 is unknown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
A.2 List of single diodes used in this work . . . . . . . . . . . . . . . . . . . . . 115
B.1 Components list of the electrical interface of the MIOB . . . . . . . . . . 119
153
