LYRA, a solar UV radiometer on Proba2

Abstract

LYRA is the solar UV radiometer that will embark in 2006 onboard Proba2, a technologically oriented ESA micro-mission. LYRA is designed and manufactured by a Belgian–Swiss–German consortium (ROB, PMOD/WRC, IMOMEC, CSL, MPS and BISA) with additional international collaborations. It will monitor the solar irradiance in four UV passbands. They have been chosen for their relevance to Solar Physics, Aeronomy and Space Weather: (1) the 115–125 nm Lyman-α channel, (2) the 200–220 nm Herzberg continuum range, (3) the Aluminium filter channel (17–70 nm) including He II at 30.4 nm and (4) the Zirconium filter channel (1–20 nm). The radiometric calibration will be traceable to synchrotron source standards (PTB and NIST). The stability will be monitored by onboard calibration sources (LEDs), which allow to distinguish between potential degradations of the detectors and filters. Additionally, a redundancy strategy maximizes the accuracy and the stability of the measurements. LYRA will benefit from wide bandgap detectors based on diamond: it will be the first space assessment of a pioneering UV detectors program. Diamond sensors make the instruments radiation-hard and solar-blind: their high bandgap energy makes them insensitive to visible light and, therefore, make dispensable visible light blocking filters, which seriously attenuate the desired ultraviolet signal. Their elimination augments the effective area and hence the signal-to-noise, therefore increasing the precision and the cadence. The SWAP EUV imaging telescope will operate next to LYRA on Proba2. Together, they will establish a high performance solar monitor for operational space weather nowcasting and research. LYRA demonstrates technologies important for future missions such as the ESA Solar Orbiter.

Full text

Page 1

LYRA: the Solar UV radiometer aboard the
ESA Proba–2?
J.-F. Hochedeza,∗, W. Schmutzb, M.Nesladekc,j, Y. Stockmand,
U. Sch¨ uhlee, A. BenMoussaa, S. Kollerb, K. Haenenc,j,
D. Berghmansa, J.-M. Defised, J.-P Halaind, A. Theissena,
V. Delouillea, V. Slemzing, D. Gillotayf, D. Fussenf,
M. Dominiquef, F. Vanhellemontf, D. McMullinh,
M. Kretzschmari, A. Mitrofanovg, B. Niculaa, L. Wautersa,
H. Rothb, E. Rozanovb, I. R¨ uedib, C. Wehrlib, H. Amanom
R. VanderLindena, A. Zhukova, F. Clettea, S. Koizumi?,
V. Mortetc, Z. Remesc, R. Petersenj, M. D’Olieslaegerc,j,
J. Roggenk, P. Rochusd,
aRoyal Observatory of Belgium (ROB), Circular Avenue 3, B-1180 Brussels,
Belgium
bPhysikalisch-Meteorologisches Observatorium Davos (PMOD) and World
Radiation Center (WRC), Dorfstrasse 33, CH-7260 Davos Dorf, Switzerland
cInstitute for Materials Research, Limburgs Universitair Centrum,
Wetenschapspark 1, B-3590 Diepenbeek, Belgium
dCentre Spatial de Li` ege (CSL) - Avenue du Pr´ e Aily B-4031 Angleur - Belgium
eMax-Planck-Institut f¨ ur Sonnensystemforschung (MPS) - D-37191
Katlenburg-Lindau - Germany
fBelgian Institute for Space Aeronomy (BISA), Circular Avenue 3, B-1180
Brussels, Belgium
gLebedev Physical Institute (LPI), 53 Leninsky Prospect, Moscow, 119991, Russia
hNaval Research Laboratory (NRL), 4555 Overlook Avenue, S.W., Washington,
DC 20375, USA
iIstituto Fisica dello Spazio Interplanetario, Consiglio Nazionale delle Ricerche,
Via del Fosso del Cavaliere, 100, I-00133 Roma, Italy
jDivision IMOMEC, IMEC vzw, Wetenschapspark 1, B-3590 Diepenbeek, Belgium
kIMEC, Kapeldreef 75, B-3001 Leuven, Belgium
?Advanced Materials Laboratory, National Institute for Materials Science (NIMS),
1-1 Namiki, Tsukuba 305-0044, Japan
mDepartment of Materials Science and Engineering, Meijo University, 1-501
Shiogamaguchi,Tempaku-ku, Nagoya 468-8502, Japan
Preprint submitted to Elsevier Science 22 October 2004

Page 2

Abstract
LYRA is the solar UV radiometer that will embark in 2006 aboard Proba–2, a tech-
nologically oriented ESA micro-mission. LYRA is designed and manufactured by a
Belgian–Swiss–German consortium (ROB, PMOD/WRC, IMOMEC, CSL, MPS &
BISA) with additional international collaborations. It will monitor the solar irra-
diance in 4 UV passbands. The channels have been chosen for their relevance to
Solar Physics, Aeronomy, and Space Weather: 1/ 115-125 nm (Lyman–α), 2/ the
200–220 nm Herzberg continuum range, 3/ Aluminium filter channel (17–30 nm)
including He II at 30.4 nm, and 4/ Zirconium filter channel (1–20 nm). The ra-
diometric calibration will be traceable to synchrotron source standards (PTB &
NIST), and the stability will be monitored by on-board calibration sources (VIS
& NUV LEDs). These allow to distinguish between possible degradations of the
detectors and filters. Additionally, a redundancy strategy maximizes the accuracy
and the stability of the measurements. LYRA will benefit from wide bandgap de-
tectors based on diamond: it will be the first space assessment of a pioneering
UV detectors program. Diamond sensors make the instruments radiation-hard and
solar-blind: their high bandgap energy makes them insensitive to visible light and,
thus, make dispensable visible light blocking filters, which seriously attenuate the
desired ultraviolet signal. Their elimination augments the effective area, and hence
the signal-to-noise, therefore increasing the precision and the cadence. The SWAP
EUV imaging telescope will operate next to LYRA on Proba–2. Together, they will
establish a high performance solar monitor for operational space weather nowcast-
ing and research. LYRA demonstrates technologies important for future missions
such as the ESA Solar Orbiter.
Key words: Sun: Irradiance, Sun: UV radiation, Sun: flares, solar-terrestrial
relations, Aeronomy, Instrumentation: detectors, Diamond, Techniques:
Radiometry
PACS: 78.20.-e, 78.40.Fy, 78.66.Db, 85.60.Dw, 94.10.-s, 94.20.-y, 94.80.+g,
07.87.+v, 07.89.+b, 95.55.Ev, 95.55.Qf, 95.75.Wx, 95.85.Mt, 96.60.Rd, 96.60.Tf
1Introduction
The knowledge of the solar spectral irradiance is of large interest to Solar
Physics, Aeronomy, and to other fields of heliospheric or planetary research.
The solar ultraviolet (UV) irradiance below 300 nm is the main source of the
energy converted in the Earth’s atmosphere, controlling its thermal structure,
?http://lyra.oma.be/
∗Corresponding author.
Email address: hochedez@oma.be (J.-F. Hochedez).
2

Page 3

dynamics and chemistry through photo-dissociation and photo-ionization. Of
special interest for life and mankind is the balance of ozone formed by radiation
below 240 nm in the stratosphere and mesosphere, but photodissociated above
200 nm in the stratosphere (e.g. Egorova et al., 2004; Rozanov et al., 2004).
The VUV irradiance variability has significant effects on human technologies
too, currently addressed in the frame of Space Weather studies.
The solar spectral irradiance is a function of both time and wavelength, and
one would ideally like to sample it with maximal temporal and spectral resolu-
tions, together with the highest absolute accuracy, precision and time coverage.
These quests however encounter physical limits, making them conflicting and
imposing trade-offs. The strategy is thus to bridge the most reliable observa-
tions, with the best possible models (e.g. Warren et al., 2001; Woods, 2002;
Lean et al., 2003).
Numerous experiments are currently monitoring the solar full-disk UV– and
X–Ray irradiance. All of them differ in spectral coverage, time coverage, time
cadence, and nature of the instrument (spectrograph, photometer, or im-
ager). Some data characterizing these missions are summarized in Table 1.
Full-disk spectrographs are used in UARS–SOLSTICE (Solar Stellar Irradi-
ance Comparison Experiment; Rottman et al. (1993); Woods et al. (1993)),
in UARS–SUSIM (Solar Ultraviolet Spectral Irradiance Monitor, Vanhoosier
et al. (1981); Brueckner et al. (1993)), in SoHO–SEM (Solar EUV Monitor,
Judge et al. (1998)), in SORCE–SOLSTICE II, in SORCE–SIM (Spectral
Irradiance Monitor), and in TIMED–SEE (Solar EUV Experiment, Woods
et al. (1998)). SEE includes XPS, a photometer system, for the short wave-
length range, whereas photometer systems are exclusively used in GOES–XRS
(X–Ray Sensor) and in SNOE–SXP (Solar X–Ray Photometer, Bailey et al.
(2000)).
In principle, more information can be extracted from spectrographs than from
photometers such as LYRA, however, under the sacrifice of cadence. Also, not
all missions are designed for continuous data aquisition. For instance, mea-
surements with TIMED–SEE are made for only 3 minutes per 97 minute orbit
during which the Sun moves per-chance into the field of view. While this is not
ideal for the study of phenomena that occur unexpectedly and vary in time,
flares have been observed during such short observation periods(Woods et al.,
2004). Generally, the required integration times are higher for spectrographs,
and time has to be spent on spectral scanning. It all unfavourably affects their
time cadence. The SORCE–SOLSTICE experiment allows spectral scans to be
restricted to Lyα and Mg II (280 nm), thereby achieving its highest cadence
of ≈1 minute. Lyα profile variations during flares have been detected that
way (Wang et al., 2000; Woods et al., 2004). While this is beyond its scope,
LYRA has the advantage of continuous monitoring in the day phase of its
orbit. Moreover, it offers the novelty of very high cadence observations down
3

Page 4

Table 1
Comparison of LYRA with other current and future solar irradiance experiments
Satellite Experiment Spectral range
[nm]
typeNominal freq.
of aquisition
time of
operation
Proba–2LYRA1–20, 17–70
Lyα, 200-220
30.4, 17.5
19.5, 28.4
0.1–0.8, 0.05–0.4
0.6–6.0
17-220
180-3200
310nm–100µm
2–7, 6–19, 17–20
pcadence ≥ 0.01 s 2006–2008–
CoronasSPIRITi cadence: 5”–15’
day: 45’, orbit: 93.5’
cadence 0.5 s
cadence 1 min
15 spectra per day
15 spectra per day
cadence ≥ 1 min
cadence 62.8 ms,
not continuous
cadence 15 sec
few per day
Aug. 2001–
Goes XRS
SXI
SolACES
SOLSPEC
SOVIM
SXP
p
i
s
s
p
p
since 1974
since 2001
2006–2009–
2006–2009–
2006–2009–
03/11/1998
–12/13/2003
1996–
1996–
Iss
Snoe
Soho SEM
EIT
0.1–50
30.4, 17.1,
19.5, 28.4
115–310
200–3000
1–35, Lyα
0.1–195
s
i
SorceSOLSTICE
SIM
XPS
SEE
s
s
p
s+p
at least every 6 h
4 times per day
06/03/2003–
06/03/2003–
06/03/2003–
22/01/2002– Timedcadence 10 s
not continuous
UarsSOLSTICE
SUSIM
119–420
115–410
s
s
Oct. 1991–
Oct. 1991–
type: p = photometer, s = spectrograph, i = imager.
to 10 milliseconds, which is of interest for the study of solar flares (Woods
et al., 2003; Snow et al., 2004), and for the limb occultation technique (See
section 3.3). The maximum cadence of LYRA is higher than the one of SNOE–
SXP, an instrument widely used for atmospheric studies and similar to LYRA
except that it lacks longer wavelength channels at Lyα and at the Herzberg
continuum. Also, SXP does not monitor the Sun in a continuous fashion. Con-
tinuous long-term time series of the EUV solar irradiance can bring insights
into fundamental questions such as coronal heating (Greenhough et al., 2003),
but here also, the higher the sampling rate, the less the bias of the statistics.
Full-disk imagers such as EIT (EUV Imaging Telescope) on SOHO, SPIRIT on
CORONAS, SXT (Soft X–Ray Telescope) on YOHKOH, or SXI (Solar X–Ray
Imager) on the GOES series of satellites enable irradiance measurements with
the additional (and actually primary) benefit of spatial resolution (Newmark
4

Page 5

et al., 2001). Sub-second cadence, however, is not yet achieved.
Further future instruments measuring solar UV– and X–Ray irradiances are
SolACES (Solar Auto-Calibrating EUV/UV Spectrophotometers), SOVIM (So-
lar Variability and Irradiance Monitor), SOLSPEC (Solar Spectrum Measure-
ment Instrument) onboard the International Space Station (ISS), and EVE
(EUV Variability Experiment) aboard SDO. There is no guarantee though
that there will be no gap in the future regarding the time–wavelength cover-
age.
Of maximum benefit for astrophysical and atmospherical studies is the com-
bination of data of complementary instruments. Spectrographs, imagers and
photometers are all designed for their specific purposes, and LYRA will add
sub-second cadence capabilities to the currently available ensemble of solar
irradiance experiments.
This paper describes LYRA (the LYman–α RAdiometer), a solar VUV pho-
tometer, and the preparation to exploitation of its observations. One purpose
of the instrument is to demonstrate several technologies able to enhance vac-
uum ultraviolet measurements by increasing the overall effective area, and the
ability to maintain calibration. The former feature permits a better precision
versus cadence trade-off, the latter, a higher accuracy. LYRA will benefit from
diamond detectors: it will be the first space assessment of a pioneering UV
detectors program (Hochedez et al., 2000, 2001, 2002, 2003a,b; Sch¨ uhle et al.,
2004).
2Instrument description
LYRA is part of the Proba–2 (Project for On-board Autonomy) space mission
of the European Space Agency (ESA), which aims at demonstrating technolo-
gies embedded in its technical or scientific payload. Proba–2 is a follow-up of
the successful Proba–1 program in orbit since October 2001 (Teston et al.,
1999). It includes major Belgian contributions. It is developed under an ESA
General Support Technology Program (GSTP) contract by a consortium led
by Verhaert Design & Development (Belgium). It will be launched as a piggy
back payload, to reach a heliosynchronous polar orbit stabilized at 10:30 for
a 2-year mission.
Beyond LYRA, the Proba–2 Science payload contains the Thermal Plasma
Measurement Unit and the Dual Segmented Langmuir Probe for Space envi-
ronment, and SWAP (Sun Watcher using an Active Pixel Sensor and image
Processing; Defise et al. (2004); Berghmans (2005)). LYRA and SWAP teams
emphasize their synergies for fundamental scientific research, and operational
5

Page 6

Fig. 1. Exploded view of one of the three identical LYRA units. The collimator, the
filters, the LEDs and the head can be seen. The detectors are hidden by the head.
Space Weather nowcast/forecast.
2.1Radiometer design
LYRA is a compact solar VUV radiometer, designed, manufactured and cali-
brated by a Belgian–Swiss–German consortium with additional international
collaborations. It encompasses three redundant units including four spectral
channels each: 1–20 nm (Zr), 17–30 nm (Al), 115-125 nm (Lyα) and 200–
220 nm (Herzberg). Each channel consists in a collimator and a head with a
precision aperture, a spectral filter, a detector and two LED light sources (See
Fig. 1). The Physikalisch Meteorologisches Observatorium Davos (PMOD,
Swizterland) provides the optical–, electronical– and mechanical design. The
solar-blind diamond detectors have been designed and fabricated at IMO-
MEC, Belgium with the collaboration of the National Institute for Materials
Science (NIMS), Tsukuba, Japan. The LYRA development takes into account
cleanliness and thermal issues. Despite a non-optimal orbit, limited platform
resources, and only 16 months of development, LYRA will normally be deliv-
ered to the platform in time (spring 2005).
The dimensions of LYRA are 315 mm × 92.5 mm × 222 mm, and its weight is
5.0 kg. Given the geometry of the collimator, view-limiting apertures of 8 mm
diameter, precision apertures of 3 mm diameter, and detector sensitive area
6

Page 7

Table 2
Expected signal and spectral purity for generic detectors (PiN for Herzberg, MSM
otherwise)
LYRA Channelpurity definition Active SunPuritySolar Min Purity
Herzberg
Lyman–α
Al
Zr
200–220 nm
121.6 ± 2.5 nm
HeII (30.5 ± 1.25 nm)
0–20 nm
12.2 nA
55.4 pA
1.4 nA
671.3 pA
87.0 %
73.6 %
53.5 %
99.4 %
12.2 nA
37.0 pA
355 pA
135 pA
87.0 %
63.3 %
56.2 %
99.4 %
of 4 mm, the unobstructed FOV is 2.06oand the full opening angle is 4.69o.
Provisions have been taken for mounting tolerances, off-points, and spacecraft
jitter.
2.2Filters, definition of channels, and radiometric model
The two detector types built for LYRA are a photoresistive device (metal-
semiconductor-metal junction, MSM) and a pin Schottky type junction (PiN).
A radiometric model based on the solar spectral irradiance, transmittance of
the LYRA filters, and detector responsivity (PiN and MSM) is used to deter-
mine the anticipated photocurrents and their spectral purity. The accuracy
of the current draft calculations is not well determined; some approximations
such as the extrapolation towards the infrared of the absolute transmittance
of interference filters were made. A generic detector was used until all flight
model devices were calibrated across the full necessary range. Updating the
radiometric model with the most reliable data is a permanent process within
the LYRA project. To assess the signal currents and spectral purity, we have
summarized both detector types from median responsivity graphs. Table 2
shows the results based on the generic MSM and PiN devices. There is no
special difficulty with the Herzberg channel thanks to the relatively large sig-
nal expected (> nA). However this must be moderated by the small variability
(<2 %), and hence the need for high precision measurement. For Lyman–α,
Al and Zr channels, the results based on the generic MSM anticipate a lack
of signal with these channels. For the Al and Zr channels the above is alle-
viated by the relatively large fluctuation expected. As to Lyα, the diamond
detectors perform almost as well as an AXUV diode (commercially available
reference detector from IRD, based on Si technolgy), which would have im-
plied an additional filter, and its associated concerns, to reach 72.9 pA with
84.8 % purity.
Porous filters specially designed and manufactured by the Lebedev Physical
Institut (Russia) are available for LYRA. They have a very high porosity
(>20%) making them transmittive in the XUV-EUV. Details of the principle
7

Page 8

and fabrication procedure can be found in Mitrofanov and Apel (1989), and
following papers. A special effort was made to decrease the size of the pore in
order to better reject the VIS and the NUV. A positive decision on a porous
filter channel for LYRA is pending a measurement of the actual transmittance
in the 1–50 nm range.
2.3 Diamond Detectors and LEDs
Diamond, a wide bandgap material, makes the sensors radiation-hard and
‘solar-blind’, which makes dispensable filters that block the unwanted visible,
but attenuate seriously the desired UV radiation. Two types of diamond detec-
tors are investigated with LYRA: PiN (photodiode) and MSM (photoresistor)
structures. PiN devices are intended for the Herzberg (Zr is TBC) channel(s)
because they promise maximal solar-blindness, that improves the spectral pu-
rity feature. For a channel using thin metallic filters, maximal solar-blindness
helps if few pinholes appear. Although MSM devices are less solar-blind and
less linear than PiN diodes, they are foreseen for the Lyman–α, Al, and per-
haps Zr channels, in order to maximise their effective area.
2.3.1PiN detectors
P-i-N detector structures were developed in a joint collaboration between
IMOMEC and NIMS (Koizumi et al., 2001). The homoepitaxial layers of
the detector were deposited in two different ULVAC stainless steel chamber
plasma-enhanced microwave deposition reactors. Each apparatus is solely used
for one type of doping, i.e. p-type (B) or n-type (P), in order to avoid un-
wanted contaminations of the CVD diamond layers. After a careful selection
and chemical oxidation of {111}-oriented, 5 mm in diameter, 0.5 mm thick,
HPHT Ib substrates from Element Six, an epitaxial p-type B-doped layer is
deposited. Boron-doping is accomplished by adding trimethylboron (TMB)
to the typical low-concentration methane-to-hydrogen mixture of 0.05%. The
TMB-to-methane ratio was varied between 100 and 4000 ppm. Other typi-
cal process parameters are a pressure of 100 Torr, a substrate temperature
around 900–950oC and a microwave power of about 400 W. The intrinsic and
P-doped n-type layer are grown in the same deposition run making use of the
second growth apparatus. Starting with a normal H2/CH4 mixture for the
intrinsic layer, phosphine (PH3) is added after a certain amount of time to
get an n-type layer. Compared to the first process, these layers are deposited
at a slightly higher microwave power (700 W) and lower substrate tempera-
ture (870–900oC), while the PH3/CH4ratio is about 5000 ppm for the n-type
layer. In order to make an electrical contact to the B-doped layer, part of the
intrinsic/P-doped layer needs to be removed using reactive ion etching (RIE).
8

Page 9

Fig. 2. Photograph of a LYRA MSM detector
Fig. 3. Responsivity in A/W of the generic MSM detector
Therefore a circular Al film (4.5 mm in diameter) is deposited on top of the
structure. This layer, acting as a mask during the RIE, is chemically removed
afterwards. As a last step, two ohmic contacts are evaporated. For the p-type
layer this is Ti/Al (50/50 nm) ring, while a thin Al layer with a transmittance
of typically 20-50 % acts as top contact.
2.3.2 MSM detector
Fabrication of a new circular MSM photodetector using a Ti/Pt/Au multilayer
structure for the interdigitated electrodes was carried out at IMO–IMOMEC.
A 0.8 µ thick diamond layer was grown epitaxially on single IIa (natural) dia-
mond substrates (500 µm thickness), also from Element Six, using microwave–
9

Page 10

plasma enhanced chemical vapor deposition (MECVD) technology. Details of
the growth procedure can be found in Remes et al. (2005). The optimisation of
the circular contact geometry was done by Garching Analytics GmbH in Ger-
many. A MSM detector is shown in Fig. 2 with its rectangular ceramic (Al203)
package. This packaging (not commercially available) has the same dimensions
as the Hamamatsu S1337–66BQ. A metal ring contact on the perimeter of the
mesa defines the active area of the MSM photodetector, which has an inner
diameter of 4.2 mm, corresponding to an optical detection area of 13.9 mm2.
2.4 Calibrations
2.4.1Traceability to radiometric standards
It is a scientific goal of LYRA to improve the absolute accuracy of the mea-
surement (goal 5%); hence the need for sub-systems and system calibrations,
on-ground and in-flight. The radiometric responsivity of each LYRA channel
has to be determined over a wavelength range that is extremely large: from the
soft X–Rays (1 nm) to the near infrared. First, subsystems (filters, MSM de-
tectors, PIN detectors) are characterized for their UV responsivity, visible light
blocking and suppression, background noise, dark current, temporal stability
within different wavelength ranges. Second, the LYRA instrument will be cal-
ibrated, each channel separately, consisting of a precision aperture, filter, and
detector. The measurements are being carried out by several teams at radio-
metric calibration facilities who are in collaboration with the LYRA project:
CSL, PTB/BESSY, NIST, IMOMEC... To cover such a large spectral range,
synchrotron beamline facilities are required. They also provide the traceabil-
ity to a primary source standard. The calibration results obtained with the
different detector types and filters will be reported in a dedicated publica-
tion. The global calibration of the LYRA instrument is foreseen in 2005 at the
calibration beamline for detector radiometry of the Physikalisch-Technische
Bundesanstalt at the electron storage ring BESSY II in Berlin.
Started in November 2003, the pre-delivery test and calibration activities, are
expected to finish in Spring 2005 (covering a period of 15 months). The LYRA
calibration plan consists of the following calibration programs:
(1) First detector campaign at IMOMEC and PTB/BESSY. This campaign
was used to characterize and compare the spectral responsivity of MSM
and PiN structures over the spectral range from 40 nm to 240nm (Ben
Moussa et al., 2004).
(2) Precision aperture area measurement at PMOD. Only the precision aper-
ture are critical, thus calibrated. The open aperture diameter is 3.0 mm.
The manufacturing tolerance of the precision apertures is H7, or +10/-
10

Page 11

0 µm in diameter. After manufacturing, the apertures will be sent to the
Swiss Federal Office of Metrology and Accreditation (METAS) for cali-
bration. The inspection of roundness allows area calibration by diameter.
The overall accuracy must be considered as not better than 1.2 µm in
diameter, that is 8.10−4in relative terms.
(3) LEDs calibration campaign at the Amano Laboratory and at IMO. The
emissivity of the diodes are tested by calibrated detectors. Reliability
tests are also carried out versus prolonged duration, and temperature
stress. Finally, the LEDS are tested with the LYRA calibrated detectors.
(4) Filters calibration campaign at CSL, MPS and PTB/BESSY. Measure-
ments of the light transmission is performed. The flight filters are doc-
umented with a complete package of certifications. After calibrations,
the filters will be mounted on the Flight Model (FM) instrument for
vibration, thermal and possible irradiation tests. After these tests, a re-
calibration of the filters will be done by ARC (Acton Research Corp.,
USA) for the Lyα and the Herzberg filters, and by PTB/BESSY for the
Aluminium and Zirconium filters.
(5) Second detector campaign at IMOMEC, CSL, MPS, NIST and PTB/BESSY.
This calibration program of FM–detectors is designed to know their XUV-
to-VIS response, and the stability of their performance with time (Ben
Moussa et al, 2005). Part of this campaign addresses other characteriza-
tions (linearity, flat-field).
(6) Global instrument calibration campaign at PTB/BESSY. This program
will assess the radiometric performance of the whole LYRA in order to
provide the most accurate knowledge of its spectral response with the
flight electronics configuration.
After delivery and during flight operations, further ground calibrations will
occur on the LYRA ’calibration head’ that is intended to be identical to the
FM ones.
2.4.2In–flight calibration
The redundancy strategy requires that all three units are made as similar as
possible. One is used continuously, the second on a weekly basis, while the
other remains closed most of the time, and is only used a few times during the
mission. In this way, the radiometric evolution of the sensors and filters will
be assessed. Furthermore, the LEDs will help disentangle filters and detec-
tors aging, and in-flight flat-field campaigns will look for unexpected burn-in
effects.
11

Page 12

2.5Onboard software and operations
The Proba–2 mission is about autonomous operations. LYRA’s onboard soft-
ware will compliantly accomplish the needed tasks in a most optimal manner
along an orbit. The Science programs will be uploaded daily from the Opera-
tion Center at ROB. Some pre-defined operating modes are implemented in a
so-called ’instrument manager’: a sun–centered acquisition state with redun-
dancy or not, a calibration state with either LED type on, and a standby state
for the night part of the orbit. For flat-field evaluation, the platform will reg-
ularly perform a special sequence of off-pointings. Few non-routine activities
are intended with the intervention of the ground operators.
Its high cadence is a scientific asset of LYRA, but it generates some non-
negligeable amount of data. To reduce the telemetry, the on-board ’data man-
ager’ reorganizes the acquired data and compresses them losslessly. To make
the scheme robust the time-lines are interleaved in three groups, so as to dis-
tribute contiguous records in separate packets.
3Science Preparation
The span of LYRA Science is large; its reviewing is out of the scope of this
paper. We present in this section some dedicated studies meant to feedback
on the development decisions, and to prepare the timely exploitation of our
instrument in view of its design.
3.1 Time series analysis and flares
The time series that will be recorded by LYRA can be modelled as the sum of
a pure signal perturbed by some noise. Numerous tools have been developed
in the statistical literature to analyse univariate (one-channel) or multivariate
(multi-channel) time series. In our case, we aim at objectively providing a
level of confidence to the detection of flares, whether small or not. We want
to also establish correspondances between blips appearing in the four different
wavebands.
Flares appear as ‘peaks’ in the time series. To detect these peaks, we use the
multiscale approach of Bigot (2003a,b) that relies on the maxima of a contin-
uous wavelet transform. This algorithm also produces the structural intensity
function, that gives the importance of the flare detected. This quantity is con-
tinuous, as opposed to the threshold that are used nowadays to determine
12

Page 13

0,2000,4000,6000,8000, 10000, 12000,14000,16000, 18000,
0,
2,
4,
6,
8,
10,
12,
14,
Estimation of the wavelet maxima
0, 2000,4000, 6000,8000,10000,12000, 14000, 16000,18000,
0,
20,
40,
60,
80,
100,
Estimation of the intensity function
0, 2000, 4000,6000, 8000,10000,12000, 14000,16000, 18000,
0,
0,5
1,
1,5
2,
2,5
3,
3,5x 10
−7
Original data and location of bumps
Fig. 4. Detection of peaks in the GOES data of February 9, 2001, in the 0.5-4˙A
waveband. (a) Maxima lines in the CWT (top). Structural intensity function (b)
Stars ‘*’ indicate the location of events detected.
the category of the flares (C-flare, M-flare, X-flare). As an illustration, Fig. 4
shows the analysis of one GOES X-ray time series using the algorithm devel-
oped by Bigot. After having detected the important events, it will be useful
to link the four times series. One possibility is to align the curves using a
registration method. Bigot (2003a) has proposed such a registration method
based on the alignement of landmarks —characteristic features of the signal.
Work in this direction is in progress.
3.2Irradiance modelling and interplay with SWAP
A set of LYRA data may be used to compute a differential emission measure
(DEM), which would enable the calculation of irradiances in individual opti-
cally thin spectral lines, also outside the spectral range of LYRA, using the
equation
Ik=
∞
?
0
Gk(T)ζ(T)dT, (1)
where ζ(T) denotes the DEM and the function Gk(T), dependent on temper-
ature T, comprises the elemental abundance and the contribution function
for the transition identified by the index k for the respective ion. G(T) is in
principle known for any optically thin line from theoretical calculations. Thus
if the DEM is known, line irradiances for unobserved lines can be predicted.
Direct inversion of Eq. 1 is troublesome, since the problem is ill–contrained
and the solution not unique. The situation gets worse if only a few lines at few
formation temperatures can be measured. Often the DEM is then expressed
in some functional form, mostly as a sum of Chebychev polynomials, ensuring
a solution which resembles known well–measured DEMs. In case of LYRA,
13

Page 14

Fig. 5. Sum of line contribution functions, weighted with instrumental throughput
(see Eq. 2). All lines from the CHIANTI database (Dere et al., 1997) enter the
calculation, using solar coronal abundances from Feldman et al. (1992) and ioniza-
tion equilibria from Arnaud and Rothenflug (1985); Arnaud and Raymond (1992);
Landini and Fossi (1991). Note the units at the ordinate: prior to any calibration,
the LYRA signal is given in Amperes per Watt, thus the units of the function g(T)
are Amperes × m3.
the situation is even more complicated, not only because there are only four
measurements at hand, but also because LYRA does not measure individual
spectral lines but the sum of many lines of various formation temperatures
and wavelengths within the wavelength range of each of the four bandpasses.
Moreover, contributions from contiuum radiation and from optically thick lines
for which Eq. 1 does not apply have to be corrected for.
Eq. 1 will apply for a LYRA measurement Ik=1,2,3,4after replacing the functions
Gk? with
gk(T) =
?
k?
Fk?kGk?(T),k = 1,2,3,4. (2)
The index k now denotes the four LYRA bandpasses, whereas the index for
line transitions is replaced by k?. The summation runs over all optically thin
lines k?within the LYRA bandpasses. Fk?kdenotes the response of the LYRA
channel k at the wavelength of the transition k?. Fk= Fk(λ) results from the
known filter transmittances and detector responses. Ideally, the four LYRA
channels could be designed in such a way that the four functions g peak at
well distinct formation temperatures over the range of interest to ensure a
well defined DEM. For instance, a comb–filter could pick out all relevant lines
in some narrow range of formation temperature. In practise, however, one
is restricted to the possibilities of filter and detector technology, and a best
compromise has to be found. Fig. 5 shows the four functions gk(T), suggesting
that a sufficiently distinct temperature coverage is given in the range between
4 ? log(T) ? 6.
14

Page 15

Fig. 6. Reconstruction of a DEM (dotted line) based on LYRA data constructed
from an input DEM of the quiet Sun (solid line). The dots mark the temperature
values which maximize the contribution functions to the four LYRA channels.
We found the best method to calculate the DEM to be an iteration scheme
originally proposed by (Withbroe, 1975) and later refined by (Sylwester et al.,
1980). Fig. 6 shows the result after 100 iteration steps of a DEM reconstruction
to artificial LYRA data calculated on the basis of a DEM for the quiet Sun.
While the reconstructed DEM reproduces the LYRA data within less than 1%,
the DEM itself does not satisfactorily resemble the input DEM. This is due
to the non-uniqueness of the solutions for Eq 1. No reproduction of the input
DEM could be expected for temperatures above ≈ 106K, as the functions g
in Eq. 5 are too ambiguous in that domain. The reconstruction around the
two higher temperature values is also not as desired, due to the fact that the
g functions do not peak very clearly at these temperatures. As a result, the
reconstructed DEM seems to be averaged out. Only in the domain of cooler
transitions does the fit approach the original DEM. It should be possible to
enforce a better agreement by demanding a certain functional form resembling
known DEMs, and iteratively fit such a form to agree with the input data.
An approach similar to the one used for the NRLEUV model (Warren et al.,
2001) will take advantage of the availability of the SWAP telescope next to
LYRA. In this way we hope as well to automatically recognize in SWAP the
temporal features identified in LYRA. Work in these directions is in progress.
3.3Aeronomy
Direct Sun observation from space allows the implementation of the limb oc-
cultation technique, a major tool in atmospheric remote sounding. The main
advantage of the occultation mode lies in the derivation of an absolute quan-
tity (the slant path optical thickness τ) from a relative measurement. Indeed,
the determination of the atmospheric transmittance along the optical path is
obtained as the ratio of the measured spectrum through the atmosphere by the
15