Appl Phys B
DOI 10.1007/s00340-009-3365-7
The airborne multi-wavelength water vapor differential
absorption lidar WALES: system design and performance
M. Wirth ·A. Fix ·P. Mahnke ·H. Schwarzer ·
F. Schrandt ·G. Ehret
Received: 30 October 2008 / Revised version: 17 December 2008
© Springer-Verlag 2009
Abstract A high-performance airborne water vapor differ-
ential absorption lidar has been developed during the past
years. This system uses a four-wavelength/three-absorption
line measurement scheme in the 935 nm H
2
O absorption
band to cover the whole troposphere and lower stratosphere
simultaneously. Additional high spectral resolution aerosol
and depolarization channels allow precise aerosol charac-
terization. This system is intended to demonstrate a future
space-borne instrument. For the first time, it realizes an out-
put power of up to 12 W at a high wall-plug efficiency us-
ing diode-pumped solid-state lasers and nonlinear conver-
sion techniques. Special attention was given to a rugged op-
tical layout. This paper describes the system layout and tech-
nical realization. Key performance parameters are given for
the different subsystems.
PACS 42.65.Yj · 42.68.Wt · 92.60.Jq
1 Introduction
The primary objective of the project WALES (derived
from WAter vapor Lidar Experiment in Space) of the DLR
M. Wirth (
) · A. Fix · G. Ehret
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut
für Physik der Atmosphäre, Oberpfaffenhofen, Münchner Str. 20,
82234 Wessling, Germany
e-mail: martin.wirth@dlr.de
P. Mahnke
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut
für Technische Physik, Stuttgart, Germany
H. Schwarzer · F. Schrandt
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut
für Robotik und Mechatronik, Berlin, Germany
(Deutsches Zentrum für Luft- und Raumfahrt) was the
preparation of a space-borne mission to overcome the short-
comings of radio-sondes and passive satellite sensors in
mapping the global water vapor distribution. While the for-
mer do not cover the globe uniformly and do not provide
reliable water vapor observations in the upper troposphere
and lower stratosphere, the latter suffer from insufficient
vertical resolution and accuracy [1]. In contrast, a space-
borne multi-wavelength H
2
O-DIfferential Absorption Lidar
(DIAL) could provide global water vapor observations suit-
able for a reliable assessment of its temporal and spatial evo-
lution. These data would lead to an improved description of
climate processes in general circulation models (GCMs) and
to benefits in numerical weather prediction (NWP) [2].
The methodology of DIAL has been developed during
the late 1960s and 1970s, and a large number of studies ap-
peared in the following years, see [3, 4]forareviewofthe
principle of measurement and [5, 6] for an overview of ex-
isting systems. First proposals for a space-borne H
2
O-DIAL
dating back to the 1980s and 1990s [6, 7] suffered from a
high power-aperture product, driving the system costs and
lacking coverage of the upper troposphere. To overcome
these problems, a measurement scheme was developed at
DLR that uses four wavelengths in the 935 nm absorption
band of H
2
O, each one especially adapted to a restricted alti-
tude range of the atmosphere. Figure 1 shows the absorption
cross section of water vapor in this wavelength region near
ground and at 10 km altitude and indicates possible lines for
DIAL measurements. In this way, relatively large absorption
coefficients can be chosen, which allow for short averaging
times even at high noise levels, thus lowering the system’s
power-aperture product considerably [2, 9–11].
One major step undertaken during recent years to vali-
date the four-wavelength concept was the realization of an
airborne demonstrator. This instrument not only implements
M. Wirth et al.
the basic multi-wavelength concept envisaged for a space
WALES but is also based on laser technologies which are
suited for in-space operation by using high-efficiency solid-
state lasers and nonlinear conversion techniques. Special at-
Fig. 1 H
2
O-absorption lines used for the WALES demonstrator. Ab-
sorption cross section data calculated from HITRAN 2006 [8] for sea
level conditions (solid line) and at 10 km altitude (dashed line)using
the US-standard atmosphere. Possible wavelengths of operation are in-
dicated by arrows, where the current system is able to use four at a
time
tention has also been given to a rugged optical layout avoid-
ing large resonators and long beam paths. The assembly of
the new instrument was finished in Summer 2007, and first
measurements were carried out on board the DLR research
aircraft Falcon F20 during the field experiments COPS (July
2007) [12], SAMUM II (Jan./Feb. 2008) [13], IPY-Thorpex
(Feb./Mar. 2008) [14], EUCAARI (May 2008), and T-PARC
(Aug./Sept. 2008). Results of these activities will be pre-
sented in separate publications.
2 Transmitter layout
The basic requirement for the transmitter system was to gen-
erate nanosecond single-frequency light pulses at four wave-
lengths between 935 and 936 nm having a total average
power in the range of 10 W. Based on the experience with its
precursor system [15, 16], it was decided to use an Nd:YAG
laser in master oscillator/power amplifier configuration fol-
lowed by two nonlinear conversion stages (see Fig. 2).
First, the radiation of the pump laser is frequency dou-
bled by a second-harmonic generator (SHG) and then con-
verted to a wavelength of 935 nm by an optical parametric
oscillator (OPO), both stages using potassium titanyl phos-
phate (KTP) as the nonlinear material. The output of the
OPO is repetitively switched between two wavelengths at a
Fig. 2 Block diagram and pulse
timing scheme of the
transmitter. Two identical chains
of lasers and non-linear
conversion stages are used to
generate pulsed radiation on
four selectable wavelengths
between 935 and 936 nm. Each
transmitter module is alternated
between two wavelengths at a
rate of 50 Hz. The time
difference t between output
pulses of the transmitters may
be set to an arbitrary offset.
Currently a value of 5 ms is
used, resulting in an equidistant
pulse-train. Additionally, the
unused pump radiation at 1064
and 532 nm is transmitted to the
atmosphere for aerosol lidar
measurements
The airborne multi-wavelength water vapor differential absorption lidar WALES: system design
Table 1 Transmitter performance parameters
Parameter Value
Repetition rate per laser 100 Hz
Pump laser pulse energy @ 1064 nm 400 mJ
Pump laser pulse energy @ 532 nm 220 mJ
System output @ 1064 nm 120 mJ (160 mJ)
a
System output @ 935 nm 45 mJ (60 mJ)
a
System output @ 532 nm 75 mJ (100 mJ)
a
Pulse length @ 1064 nm 8 ns (FWHM)
Pulse length @ 935 nm 5.5 ns (FWHM)
Pulse length @ 532 nm 7.5 ns (FWHM)
Beam quality M
2
@ 1064 nm 1.5
Beam quality M
2
@ 935 nm 7.6
Beam quality M
2
@ 532 nm 1.8
Beam divergence (all) 1 mrad
Line width @ 1064 nm 54 MHz (FWHM)
Line width @ 935 nm 150 MHz (FWHM)
Frequency stability @ 1064 nm ≤1MHz
Frequency stability @ 935 nm ≤30 MHz
Spectral purity @ 935 nm ≥99.9%
Spectral purity @ 532 nm ≥99.995%
a
Energy values in brackets are valid for a reduced spectral purity of
99%
rate of 50 Hz. Two identical laser systems are operated tem-
porally interleaved, resulting in a total pulse rate of 200 Hz
and a repetition rate for the four-wavelength pulse train of
50 Hz (see Fig. 2). The residual pump radiation at 1064
and 532 nm, which is not converted by SHG and/or OPO,
is also transmitted into the atmosphere and used for aerosol
measurements. Table 1 summarizes measured values for the
most important performance parameters of the transmitter.
The individual sub-systems are described in more detail in
the following sections.
2.1 Pump laser
To guarantee stable single-mode operation, the master oscil-
lator is implemented as a passively Q-switched monolithic
Nd:YAG ring laser [17] with inherent longitudinal mode se-
lection (MephistoQ, Innolight GmbH). For highest stability
under pressure/temperature changes and vibrations, the op-
tical components of this laser were integrated into a custom-
made pressure-tight, monolithic housing. This master laser
operates at a repetition rate of 4 kHz with an output energy
of 40 µJ per pulse at 1064 nm. The pulse length is 7.7 ns full
width at half maximum (FWHM).
The frequency of this laser is temperature-tunable with a
coefficient of 3.86 GHz/K. Each 4.5 GHz a mode hop oc-
curs. The free spectral ranges overlap by nearly half of their
Fig. 3 Time series and histogram of the pulse-to-pulse frequency jit-
ter of the Nd:YAG master oscillator, measured against a Lightwave
124-1064-100 CW laser using a heterodyne technique. Measurements
were taken at the pump repetition rate of the amplifier chain (100 Hz).
Since the reference laser was not absolutely stabilized, variations on
time scales larger than one minute were removed by high pass filtering
width, so that for each given wavelength, it is always pos-
sible to find a temperature where the laser runs stably in a
single longitudinal mode. In addition it is possible to modu-
late the output wavelength of the master oscillator by induc-
ing a small amount of mechanical stress on the laser crystal.
This is done by a piezo transducer and allows for a full wave
modulation of up to 30 MHz. Both, temperature and stress
tuning schemes are used to stabilize the laser to a iodine (I
2
)
line for High Spectral Resolution Lidar (HSRL) measure-
ments of aerosol extinction (see [18] for an explanation of
this technique). To accomplish the locking to the I
2
-line, the
laser is stress-modulated with a 100 Hz sine wave in phase
with the pump pulse for the power amplifier stages (see be-
low). Part of the output radiation is frequency doubled in
a 10-mm-long KTP crystal and transmitted through a 50-
mm-long absorption cell, filled with about 50 Pa of iodine.
A standard lock-in technique is used to stabilize the laser to
the line center by controlling the oscillators crystal temper-
ature. Typically the I
2
line at 18787.8098 cm
−1
(line 1109
according to [19]) is used for HSRL measurements. The ab-
solute stability is estimated to be better than 1 MHz, and the
short-term (<60 s) pulse-to-pulse frequency jitter was mea-
sured to be better than 300 kHz RMS (see Fig. 3)usinga
heterodyne technique [20–22].
Since the master oscillator is passively Q-switched, the
timing of the output pulses is not controllable on a shot-to-
shot basis. To synchronize the timing to the power amplifier
stages, the current of the pump laser diode of the master is
controlled by a phase-locked loop. The residual timing jitter
is below ±1 µs as can be seen from Fig. 4. This is sufficient
M. Wirth et al.
to synchronize the master laser to the pump pulse of the am-
plifier chain. The timing of the lidar data acquisition, where
a lower jitter is mandatory, is triggered by the outgoing light
pulses.
A three-stage amplifier chain, side-pumped by diodes, is
used to achieve the desired output pulse energy. The am-
plifiers are pumped at a repetition rate of 100 Hz and syn-
Fig. 4 Time series and histogram of the pulse-to-pulse timing jitter
of the Nd:YAG master oscillator relative to an external clock source.
Measurements were taken at the pump repetition rate of the amplifier
chain (100 Hz)
chronized with the master oscillator using a phase-locked
loop (see Fig. 5), i.e., every 40th master oscillator pulse is
amplified. Within each amplifier pump chamber, 60 quasi-
continuous-wave (QCW) laser diodes emitting at 808 nm
are arranged in a five-fold symmetry around the 110 mm
long Nd:YAG rod. These diodes are operated at a current of
85 A resulting in a total optical power of 5 kW per chamber.
The pump chambers were designed by Rofin Sinar AG and
use diodes from DILAS GmbH.
The preamplifier stage uses an Nd:YAG rod with 3 mm
diameter and 0.9% Nd doping. For a pump duration of
140 µs, the resulting single-pass small-signal gain is approx-
imately 100. The preamplifier is built up in a polarization-
coupled double pass configuration. For better compensation
of the thermally induced birefringence within the laser rod,
a Faraday rotator is used for polarization rotation after the
first pass [23]. The total saturated double-pass gain is about
120 resulting in an output pulse energy of 50 mJ. Since the
preamplifier has a very high double-pass small-signal gain
of more than 10
4
, it is very sensitive to back-reflections.
Therefore Faraday isolators are put into the entrance and
exit paths of this stage. A half-wave plate mounted into a
motorized rotation stage allows control of the input energy
to the main amplifier stages.
The main amplifier uses two identical pump chambers,
each equipped with rods of 6 mm diameter, 110 mm length
and 0.5% Nd doping. Operated at 220 µs pump-pulse du-
ration, these stages give a total saturated gain of more than
Fig. 5 Schematic optical layout
of the pump laser
The airborne multi-wavelength water vapor differential absorption lidar WALES: system design
Fig. 6 Pump laser beam propagation calculations. Shown are the nominal second-moment radii for both axes of the beam. The master laser is
located at distance zero
Fig. 7 Measurement of beam propagation parameter M
2
at the fun-
damental wavelength of the pump laser at 405 mJ pulse energy. Data
shown for the principal axis a (blue)andb (red) and the equivalent
radius r =
√
r
a
r
b
(green)
8 resulting in a total output pulse energy >400 mJ. A 90°
polarization rotator in between the two chambers is used for
birefringence compensation [23].
The thermal lens induced by a radial thermal gradient
within the rods is compensated by means of a convex mirror
(R = 750 mm) for the preamplifier and two adjustable tele-
scopes before and after the main amplifier stages. To guide
a proper layout and adjustment of these compensation el-
ements, Gaussian beam propagation calculations were per-
formed. Figure 6 shows the nominal beam radius along the
optical path of the laser. The beam propagation parameter
M
2
was adjusted within each rod to measured values to ac-
count for distortions caused by spatially nonuniform ampli-
fication, beam truncation, and higher-order thermal aberra-
tions. The M
2
values measured after the oscillator, pream-
plifier first pass, preamplifier second pass, first main ampli-
fier, and second main amplifier (whole laser) are in this or-
der: 1.05, 1.1, 1.2, 1.35, and 1.5. Figure 7 shows a beam
parameter measurement for the total laser at 405 mJ out-
put pulse energy. The beam propagation parameter was mea-
sured according to the second-moment method as described
Fig. 8 Far-field beam profile of the Nd:YAG pump laser at the fun-
damental wavelength and a pulse energy of 405 mJ. The solid drawn
ellipse shows the second-moment radius determined according to the
ISO11146 standard. The background is estimated from the rectangular
regions shown near the corners
by the ISO11146 [24] standard using an f = 500 mm lens
and a CCD camera with 6.75 µm spatial resolution (see
Fig. 8 for a single-beam profile measurement). Since the
beam of the master laser is slightly elliptical, calculations
are shown for both principal axes. These calculations were
verified by beam parameter measurements at various points
along the optical axis. The beam propagation through the
main amplifiers was chosen symmetrically with respect to
the plane of the 90° polarization rotator to optimize the ef-
fect of birefringence compensation (see [23, 25, 26] and ref-
erences therein).
![Fig. 1 H2O-absorption lines used for the WALES demonstrator. Absorption cross section data calculated from HITRAN 2006 [8] for sea level conditions (solid line) and at 10 km altitude (dashed line) using the US-standard atmosphere. Possible wavelengths of operation are indicated by arrows, where the current system is able to use four at a time](/figures/fig-1-h2o-absorption-lines-used-for-the-wales-demonstrator-wdq3fcea.webp)