Full Text 01

The Neutral Particle Detector
on the Mars and Venus Express missions
Alexander Grigoriev
Swedish Institute of Space Physics

Kiruna
September 2007
c
Copyright 2007 Grigoriev Alexander
Doctoral thesis at the Swedish Institute of Space Physics,
Kiruna, 2007
The Neutral Particle Detector on the Mars and Venus Express missions
Typeset by the author in LATEX
IRF Scientific Report 290

ISSN 0284-1703
ISBN 978-91-7264-349-9
Printed at the Swedish Institute of Space Physics
Box 812, SE-981 28 Kiruna, Sweden
September 2007
Abstract
The Neutral Particle Detector (NPD) is a new type of instrumentation for energetic neutral
atom (ENA) diagnostics. This thesis deals with development of the NPD sensor designed as
a part of the plasma and neutral particle packages ASPERA-3 and ASPERA-4 on board Mars
Express and Venus Express, the European Space Agency (ESA) satellites to Mars and Venus,
respectively. It describes how the NPD sensors were designed, developed, tested and calibrated.
It also presents the first scientific results obtained with NPD during its operation at Mars.
The NPD package consists of two identical detectors, NPD1 and NPD2. Each detector
has a 9◦ × 90◦ intrinsic field-of-view divided into three sectors. The ENA detection principle
is based on the surface interaction technique. NPD detects ENA differential fluxes within the
energy range of 100 eV to 10 keV and is capable of resolving hydrogen and oxygen atoms by
time-of-flight (TOF) measurements or pulse height analysis.
During the calibration process the detailed response of the sensor was defined, including
properties such as an angular response function and energy dependent efficiency of each of the
sensor sectors for different ENA species.
Based on the NPD measurements at Mars the main scientific results reported so far are:
- observation of the Martian H-ENA jet / cone and its dynamics,
- observations of ENA emissions from the Martian upper atmosphere,
- measurements of the hydrogen exosphere density profile at Mars,
- observations of the response of the Martian plasma environment to an interplanetary shock,
- observations of the H-ENA fluxes in the interplanetary medium.
Keywords: ENA imaging, exosphere, magnetosphere, Mars, Venus, solar wind interaction
Sammanfattning
Den Neutrala Partikel Detektorn (NPD) är en ny typ av instrumentering som används för analys
av energirika neutrala atomer (ENA). Denna avhandling omfattar utvecklingen av NPD-sensorn
som ingår i plasma- och neutralapartikelinstrumenten ASPERA-3 och ASPERA-4 ombord på
Mars Express och Venus Express, vilka är den europeiska rymdstyrelsens (ESA) satteliter till
Mars och Venus. Avhandlingen beskriver hur NPD-sensorerna konstruerades, utvecklades, tes-
tades och kalibrerades. De första vetenskapliga resultaten från NPD-sensorns verksamhetstid
vid Mars presenteras också.
NPD-sensorn består av två identiska detektorer, NPD1 och NPD2. Varje detektor har ett
synfält på 9◦ × 90◦ som är indelat i tre sektorer. Principen för ENA-detektering baserar sig på
tekniken för växelverkan mellan ytor. NPD-sensorn detekterar flödet av ENA inom energiin-
tervallet 100 eV till 10 keV och är kapabel att urskilda väte- och syreatomer genom "time-
of-flight"-mätningar (TOF) eller pulshöjdsanalys. Under kalibreringsprocessen identifierades
NPD-sensorns egenskaper som inkluderar vinkelsvarsfunktionen och den energiberoende ef-
fektiviteten för varje sensors sektor för olika ENA-typer. De huvudsakliga vetenskapliga resul-
taten, baserade på NPD-mätningar vid Mars och som hittills har rapporterats är:
- observationer av H-ENA strålar/koner och dess dynamik,
- observationer av ENA-emissioner från den övre delen av atmosfären,
- mätningar av vätedensitetens profil i exosfären,
- observationer av plasmamiljöns svar vid en interplanetär chock,
- observationer av H-ENA flöden i det interplanetära mediet.
Nyckelord: ENA-avbildning, exosfär, magnetosfär, Mars, Venus, solvindens växelverkan

Contents
Introduction 1
1 Energetic neutral atoms in space 3

1.1 Production mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 Charge exchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.2 Back-scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1.3 Sputtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 ENAs at non-magnetized planets . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.1 ENA environment of Mars . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.2 ENA environment of Venus . . . . . . . . . . . . . . . . . . . . . . . 11
2 Energetic neutral atoms detection 15

2.1 ENA imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Principle functions of ENA instruments . . . . . . . . . . . . . . . . . . . . . 17
2.3 Deflection systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 UV rejection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5 ENA detection and analysis: Instrument examples . . . . . . . . . . . . . . . . 22
2.5.1 Foils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.5.2 Surface interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.5.3 High frequency shutters . . . . . . . . . . . . . . . . . . . . . . . . . 29
3 The ASPERA-3 and ASPERA-4 experiments 31

3.1 Scientific objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.1.1 ASPERA-3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.1.2 ASPERA-4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 Instrument overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.2.1 The Ion Mass Analyzer (IMA) . . . . . . . . . . . . . . . . . . . . . . 36
3.2.2 The Electron Spectrometer (ELS) . . . . . . . . . . . . . . . . . . . . 38
3.2.3 The Neutral Particle Imager (NPI) . . . . . . . . . . . . . . . . . . . . 39
3.2.4 The Digital Processing Unit (DPU) and the scanner . . . . . . . . . . . 40
4 The Neutral Particle Detector (NPD) 41

4.1 The measurement technique . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2 NPD mechanical design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2.1 Deflector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2.2 Start unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.2.3 Stop unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.2.4 Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.3 Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.4 MCP assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.5 Data formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
I
4.6 Instrument level qualification . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.7 NPD response to high energy particles . . . . . . . . . . . . . . . . . . . . . . 60
5 The NPD calibrations 63

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.1.1 Calibration facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.1.2 Calibration setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.2 Theoretical principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.2.1 MCP characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.2.2 Beam characterization . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.2.3 Geometrical Factor calculation . . . . . . . . . . . . . . . . . . . . . . 67
5.2.4 Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.2.5 Energy resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.2.6 Mass resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.3 Measurement principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.3.2 Angular response measurements . . . . . . . . . . . . . . . . . . . . . 72
5.3.3 Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.3.5 Mass resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.4 ASPERA-3 / NPD calibration results . . . . . . . . . . . . . . . . . . . . . . . 74
5.4.1 Calibration objectives . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.4.3 Efficiency measurements . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.4.4 Angular response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.4.5 Geometrical factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.4.7 Mass resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.4.8 Heater and temperature sensor characterization . . . . . . . . . . . . . 93
5.4.9 Dark noise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.5 ASPERA-4 / NPD calibration results . . . . . . . . . . . . . . . . . . . . . . . 96
5.5.1 Calibration objectives . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.5.3 Efficiency measurements . . . . . . . . . . . . . . . . . . . . . . . . . 97
5.5.4 Angular response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
5.5.5 Geometrical factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.5.7 Mass resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.5.8 Heater and temperature sensor characterization . . . . . . . . . . . . . 116
5.5.9 Dark noise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
6 Scientific results. The NPD measurements at Mars. 119

6.1 Subsolar ENA jet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.1.2 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
6.1.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.1.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.2 Observations of the Martian subsolar ENA jet oscillations . . . . . . . . . . . . 128
6.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.2.2 Observation geometry . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.2.3 ENA jet fluctuation observation . . . . . . . . . . . . . . . . . . . . . 130
6.2.4 Statistics on the intensity variations . . . . . . . . . . . . . . . . . . . 131
6.2.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
6.2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.3 Other results by the ASPERA-3 / NPD . . . . . . . . . . . . . . . . . . . . . . 136
6.3.1 Global response of Martian plasma environment to an interplanetary
structure: From ENA and plasma observations at Mars . . . . . . . . . 136
6.3.2 The Hydrogen Exospheric Density Profile Measured with
ASPERA-3 / NPD . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.3.3 Energetic Hydrogen and Oxygen Atoms Observed on
the Nightside of Mars . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.3.4 First ENA observations at Mars: ENA emissions from the Martian up-
per atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
6.3.5 Direct Measurements of Energetic Neutral Hydrogen in the Interplan-
etary Medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
6.3.6 Energetic Neutral Atoms from the Heliosheath . . . . . . . . . . . . . 144
7 Summary and future prospects 145
A NPD data processing 147

A.1 In addition to the NPD operation modes . . . . . . . . . . . . . . . . . . . . . 147
A.2 Log-compression algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
A.3 NPD data display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
Bibliography 153
Glossary of Acronyms 161
Acknowledgments 163
Introduction
The solar wind is a supersonic flow of tenuous solar plasma which interacts with all bodies
in our solar system. It possesses a magnetic field which is considered to be frozen in to the
flowing plasma. In general every celestial body possesses a neutral gas environment of varying
thickness. The interaction of the solar wind with them can be roughly divided into three types:
interaction with magnetized bodies, interaction with unmagnetized bodies possessing an atmo-
sphere, and interaction with those having negligible or no atmosphere at all. This PhD thesis
is related to the solar wind interaction with unmagnetized bodies possessing an atmosphere,
namely, the planets Mars and Venus. These planets no longer have a global magnetic field to
deflect the solar wind, which causes atmospheric erosion through the interaction with the upper
part of the planetary atmospheres.
The interaction between charged and neutral particles is a common phenomenon in space
plasmas. An energetic neutral atom (ENA) is born whenever an energetic ion undergoes a
charge exchange process in a collision with a neutral background atom. An ENA can also
appear as a result of atmospheric and surface sputtering processes. The newly born ENA
becomes independent from the surrounding plasma and the influences of magnetic and electric
fields, and its trajectory is defined solely by the initial momentum and gravitational forces.
With the exception of the case of very low energy (< 10 eV) atoms, gravitational effects can be
disregarded. Considering the case of charge exchange collision, one can assume that creation
of an ENA preserves both the direction and magnitude of the energetic ion velocity before the
collision. The movement of ENAs along the ballistic trajectory resembles the movement of
photons in space. Hence, principles of the imaging technique used in optics can be applied to
image ENAs. ’Imaging’ is used to refer to the detection of the direction and wavelength of
photons originating from some source of light. The term ’ENA imaging’ is used for recording
ENA fluxes as a function of observational direction. A global image of the object of interest
can be reconstructed from a set of ENA images. In ENA imaging it is not only the angular
distribution that is measured, but also the energy and mass of ENAs originating from an ENA
source region. By determining ENA flux angular distribution as well as ENA energies and
masses it is possible to establish plasma ion composition and distribution function remotely.
This makes it possible to probe inaccessible regions in space from afar, as well as to obtain
instantaneous information about the object.
ENA imaging can be used to: diagnose plasma processes on the global scale; reveal plasma
boundaries resulting from the interaction of the solar wind with magnetized planets (e.g.,
Williams et al., 1992); and characterize solar wind interaction processes with unmagnetized
planets (Lichtenegger et al., 2002; Barabash et al., 2002; Holmström et al., 2002). While elec-
tron and ion distributions in the planetary environment can only be measured locally, remote
ENA imaging can give the whole picture of the interaction processes between different plasma
populations and neutral background gas that result in ENA generation. Furthermore, in situ
plasma measurements have the drawback that it is not possible to use them to resolve temporal
1
2
and spatial variations unambiguously. Remote ENA imaging, on the other hand, can reveal
spatial variations. Therefore the global ENA imaging technique is an important complement
to local measurements of electrons and ions. Modern planetary missions now include ENA
detectors together with plasma packages.
The European Space Agency (ESA) missions towards Mars and Venus, namely Mars Ex-
press and Venus Express, carry the plasma and neutral particle packages, Analyzer of Space
Plasma and Energetic Atoms (ASPERA-3 and ASPERA-4), among the scientific payload.
Mars Express is Europe’s first spacecraft to the Red Planet. Launched from the Baikonur
launch site in Kazakhstan on board a Russian Soyuz-Fregat launcher, it travelled to Mars in
seven months, going into orbit on December 25, 2003. Mars Express was inserted into a
6.5 hour elliptical near-polar orbit with apogee ∼3 R M (where R M stands for the radius of Mars)
and perigee as low as ∼265 km. Venus Express is ESA’s first mission to Earth’s nearest plan-
etary neighbour, Venus. After 5 months cruise to Venus the Venus Express spacecraft entered
orbit round the planet on April 11, 2006. It was inserted into a 24 hour elliptical orbit with
apogee ∼11 RV (where RV stands for the radius of Venus) and perigee ∼300 km.
The general scientific objective of both the ASPERA-3 and ASPERA-4 experiments is
to study the solar wind – atmosphere interaction and to characterize the plasma and neutral
gas environment in the vicinity of Mars and Venus through the use of ENA imaging and by
measuring local ion and electron plasma populations. The ASPERA packages comprise 4
instruments, namely an electron spectrometer; an ion spectrometer; and two ENA sensors, the
Neutral Particle Imager (NPI) and the Neutral Particle Detector (NPD) (Barabash et al., 2004,
2006). NPD is an ENA detector, designed to perform mass and energy analysis of incoming
ENAs, with a moderate angular resolution. This dissertation is focused on NPD development
and calibration.
The thesis is organized as follows: Chapter 1 introduces the basics of ENAs along with a
short description of the ENA environment of Mars and Venus. Chapter 2 provides the princi-
ples of ENA imaging followed by a review of ENA measurement techniques and ENA instru-
mentation examples. Chapter 3 contains a comprehensive description of the ASPERA-3 and
ASPERA-4 packages that are providing plasma and ENA measurements at Mars and Venus,
respectively. Chapter 4 presents a detailed design description of the NPD, followed by calibra-
tion results of both ASPERA-3 / NPD and ASPERA-4 / NPD in chapter 5. Chapter 6 reviews
selected papers based on data obtained by ASPERA-3 / NPD during its operation at Mars. Fi-
nally, chapter 7 sums up the thesis and outlines future prospects. Appendices contain details
on NPD operation modes, and a description of quick look NPD data display.
Chapter 1
Energetic neutral atoms in space
ENAs are neutral particles, possessing energy exceeding the thermal energy (i.e., several eV).
1.1 Production mechanisms

ENAs in space are produced by various processes of ion/atom – atom collision, mainly
- charge exchange of energetic ions with exospheric gasses in the near-planet environment
or interplanetary background neutral gas,
- sputtering of atmospheric or surface materials by precipitating energetic ions or neutrals,
- back-scattering of energetic particles precipitating on the planetary upper atmosphere or
surfaces.
1.1.1 Charge exchange

Charge exchange of singly-ionized plasma ions to produce ENAs is fundamental to many ENA
sources. ENAs are formed in charge exchange collisions between energetic plasma ions and
cold neutral gas atoms. The charge exchange process
A+energetic + Mcold → Aenergetic + Mcold

+
produces ENAs when an energetic (compared to thermal energies) ion A+energetic , collides with a
+ . Species M and A may be identical
cold neutral Mcold , resulting in an ENA and a cold ion Mcold
(i.e., H + + H → H + H + , resonance charge exchange) or not (i.e., H + + O → H + O+ ). Due to
the large internuclear distances during charge exchange, only negligible energy and momentum
are transferred in these interactions. Hence the initial velocity of an energetic particle is only
slightly changed in a charge exchange collision (Bransden and McDowell, 1992). Figure 1.1
illustrates the charge exchange (also known as electron pick-up) process between a fast ion and
a slow atom.
The probability that a given charge exchange process will occur in a collision is expressed
as a reaction cross-section. Figure 1.2 shows the charge exchange cross-sections for singly
charged hydrogen and oxygen ions with cold neutral gas. In general, at low ion energies the
cross-sections for charge exchange are within a range of 10−15 cm2 . The H + cross-section
begins to fall off for proton energies above 10 keV and drops off steeply above 50 keV. This is
a very important constraint on ENA production, and it assures that ENA hydrogen spectra will
be concentrated below ∼200 keV.
3
4 CHAPTER 1. ENERGETIC NEUTRAL ATOMS IN SPACE
ENA
0
1
00
11
fast ion
1
0 0
1
0
1
0
1
0
1
1
0 0
1
0
1 0
1
0
1
0
1 0
1
0
1 slow ion
3
slow atom 2
1
Figure 1.1: Charge exchange mechanism.
Figure 1.2: Charge exchange cross-sections of energetic H + and O+ ions as a function of incident ion
energy for electron pick-up from cold neutral hydrogen and oxygen atoms. From Wurz (2000).
1.1.2 Back-scattering
ENAs can be born during scattering, in a process of elastic and inelastic collision of energetic
charged- or neutral particles with slow neutral background atoms. The ENA back-scattering
production mechanism with reference to Mars is as follows:
The neutral solar wind (see Section 1.3.1) can enter the Martian upper atmosphere and
reach the exobase, where it experiences elastic and inelastic collisions (Kallio and Barabash,
2000). It possesses the energy of the solar wind bulk flow. A fraction of the neutral solar wind
1.1. PRODUCTION MECHANISMS 5
incident flux can be scattered back in the form of ENAs due to momentum transfer in collisions
and cascade of charge exchange and electron stripping processes.
1.1.3 Sputtering
Two types of sputtering are present in space, namely surface and atmospheric sputtering. Sur-
face sputtering occurs in general on the celestial bodies without atmosphere, while the atmo-
spheric sputtering occurs on the ones possessing an atmosphere.
Atmospheric sputtering. The atmospheric sputtering mechanism with reference to Mars is

shown schematically in Figure 1.3. Because the modern Mars lacks an intrinsic magnetic field,
the exosphere is directly exposed to the solar wind. Hence, the solar wind convection electric
field can accelerate O+ ions, originating from ionization of the exospheric atoms, to produce
’pick-up ions’. A fraction of these O+ pick-up ions can re-enter the Martian upper atmosphere
and reach the exobase due to a large gyro-radius. After the precipitation, O+ ions exchange
charges and the resulting fast O atoms undergo elastic collisions with cold O atoms in the
background gas (Luhmann and Kozyra, 1991). The large energy is imparted to surrounding
particles through further collisions, causing atmosphere sputtering. A certain fraction of the
particles which gained energy in these collisions is scattered back out of the atmosphere.
Figure 1.3: Atmospheric sputtering occurs when ionized O+ in the upper atmosphere is accelerated by
the solar wind convection electric field and strikes the exobase (step 1), causing a cascade of collisions
that results in the ejection of particles in the form of ENAs (step 2). Adapted from Kass (1999).
Surface sputtering. In the case of surface sputtering on celestial bodies without an atmo-
sphere, such as the Moon or Mercury, energetic ions coming directly from the solar wind as
well as energized planetary ions may precipitate onto the surface, resulting in extensive sput-
tering (Grande, 1997; Lukyanov et al., 2004). ENAs originating from the sputtering process
possess energy of a few tens of eV (Massetti et al., 2003).
1.2 Classification
NPD
VLENA
Figure 1.4: ENA classification by energy and sources. Adapted from Wurz (2000).
The energy range of ENAs is generally considered to cover four sub-ranges: very low-
energy neutral atoms (VLENA) ∼0.1 – 10 eV; low-energy neutral atoms (LENA) ∼10 – 1000
eV; medium-energy neutral atoms (MENA) ∼0.5 – 30 keV; and high-energy neutral atoms
(HENA) ∼10 – 200 keV.
This arbitrary division (with overlapping ranges) derives from the necessity of employing
different experimental techniques in different energy ranges (since no single analyzer can cover
the entire concerned range), rather than from the different physical natures of these ENAs.
Figure 1.4 gives an overview of the different sources of energetic neutral particles that can be
observed in space, together with their approximate energy range. The upper limit for HENAs
is a consequence of the inherent energy and species-dependent cut-off values for ENA charge
exchange cross-sections. The NPD detection range is indicated.
ENA fluxes come from different ion populations, with different compositions, flux levels
and energy, and spatial, and temporal dependencies. Table 1.1 gives an overview of the typi-
cal parameters characterizing ENAs, generated near various celestial objects within the Solar
system.
1.3 ENAs at non-magnetized planets

1.3.1 ENA environment of Mars
The solar wind interaction with Mars is complex and results in the production of ENAs in
a wide energy range. In order to get a better understanding of such interaction, the general
picture of the Martian plasma boundaries is shown in Figure 1.5.
1.3. ENAS AT NON-MAGNETIZED PLANETS 7
ENA source Energy, Flux, cm−2 sr−1 s−1 keV−1 Species

keV
Intestellar medium 0.013 5 × 105 H
0.050
Heliospheric shock 0.2 - 1 (1-4)×102 cm−2 sr−1 s−1 H
@ 1AU
Interplanetary shocks 10 4 · 101 − 102 H
100 10−2 − 10−1
Coronal mass ejection 2-7 104 @ 1AU H
Martian magnetosphere 1-8 105 − 106 H, O
20-80 10−1 − 101 @ 1AU
Terrestrial magneto- 10-20 103 − 104 H, O, He
sphere 20-30 101 − 102 @ 5 RE
Jovian magnetosphere 15-65 10−3 -10−2 cm−2 s−1 keV−1 H, O, S
@ 100 R J
Saturnian magnetoshere >40 10−1 cm−2 s−1 keV−1 H, O
@ 45 RS
Outgassing asteroids 1-5 104 − 106 H
(Phobos)
Table 1.1: Overview of different ENA population parameters (AU - astronomical unit, RS - Saturn
radius, RE - Earth radius, R J - Jupiter radius).
Induced magnetosphere boundary
Figure 1.5: Structure of the Martian plasma environment in the plane of the interplanetary magnetic
field. From Fedorov et al. (2006).
Mars has a very thin atmosphere with a pressure of ∼7 mbar at the surface, consisting
mostly of CO2 . It extends a long way out due to low gravity at Mars. The upper part of the
atmosphere is partly ionized by extreme ultra-violet (EUV) / ultra-violet (UV) solar radiation.
The absence of an intrinsic magnetic field (Acuña et al., 1998) leads to a direct interaction of
the solar wind with the upper part of the extended neutral atmosphere of Mars. Planetary ions
become picked-up by the solar wind convection electric field, which results in a significant
mass-loading of the frozen-in interplanetary magnetic field (IMF). The boundary upstream and
around the ionosphere, where the IMF mass-loading and draping around this obstacle occur,
is called the magnetic pile-up boundary (MPB) (Vignes et al., 2000) or the induced magneto-
sphere boundary (IMB) (Lundin et al., 2004). Upstream of the IMB, the solar wind protons are
thought to be the dominant ion species, while below the IMB, heavy ions of planetary origin
(mostly O+ and O+2 ) prevail. A bow shock (BS) appears upstream of the IMB, where the so-
lar wind plasma flow is slowed from supersonic to subsonic, thermalizes and begins to divert
around the obstacle. The region in-between the BS and IMB, containing both shocked solar
wind ions and picked-up planetary ions, is called a magnetosheath. Behind the planet, in the
planetary shadow, the plasma sheet region is located (Rosenbauer et al., 1989), where the dense
flow of the heavy ions can be encountered.
The neutral gas density in the solar wind interaction region where the BS and IMB are
located, can reach 104 - 106 cm−3 due to the low gravity on Mars. The solar wind plasma
can, therefore, interact strongly with the exospheric gases, mainly H, through the collisional
interactions (Section 1.1), resulting in strong ENA production.
Nowadays the Martian plasma and ENA environment is well understood. The measure-
ments carried out on a number of missions, such as Phobos-2 (Lundin et al., 1989), MGS (Vi-
gnes et al., 2000), and with the ASPERA-3 experiment on Mars Express (Barabash et al.,
2004, 2006) as well as numerous numerical simulations performed by Brecht (1997a); Holm-
ström et al. (2002); Kallio and Janhunen (2002); Ma et al. (2002); Lichtenegger et al. (2002)
complement the general picture of the Martian ENA environment.
The main sources of ENAs in the Martian environment are:
• Upstream solar wind ENAs – neutral solar wind (NSW)
Some part of the supersonic solar wind flux is neutralized due to charge exchange with
the interplanetary neutral gas. Such solar wind ENAs have been detected by the LENA
instrument on board the terrestrial IMAGE mission, which is capable of looking directly
towards the Sun (Collier et al., 2001). Also, the undisturbed solar wind flow upstream of
the Martian BS can experience charge exchange with the Martian hydrogen exosphere,
extended over very long distances (exceeding 4 Martian radii, R M ).
The resulting narrow (∼10◦ ) anti-sunward beam of solar wind ENAs, called the NSW,
has the same energy as the bulk solar wind flow (∼1 keV). NSW flux was likely de-
tected with the sunward pointed ASPERA-3 / NPI sensor, right after entering Martian
eclipse (Brinkfeldt et al., 2006b). The planetary disk blocks solar UV photons in such a
configuration, allowing NPI to avoid being solar blinded and thus detect the NSW. The
NSW can reach the umbra region mostly due to thermal spreading and scattering in the
upper atmosphere (Kallio et al., 2006).
• Shocked solar wind
The shocked thermalized solar wind is a strongest source of H-ENAs with an energy
of a few hundreds eV. The shocked solar wind flow moves around the Martian obstacle
through the comparatively dense hydrogen exosphere. Therefore charge exchange in-
teractions between protons and cold planetary neutral species in the magnetosheath are
very probable. The ENA fluxes, generated from the shocked solar wind, are sensitive
to the neutral hydrogen distribution, which is controlled by the exobase temperature and
density (Holmström et al., 2002). Detailed modeling of the shocked solar wind ENA
production was performed by Kallio et al. (1997); Holmström et al. (2002); Gunell et al.
(2006). Figure 1.6 shows images of H-ENA emissions near Mars simulated by Holm-
ström et al. (2002) for several vantage points at different solar zenith angles.
Intense fluxes of H-ENAs emitted from the subsolar exosphere of Mars (so-called ENA
jets (cones), see Section 6.1) were detected by the ASPERA-3 / NPD sensor on board
Mars Express (Futaana et al., 2006a). The differential flux was estimated to be 4 – 7
×105 cm−2 sr−1 s−1 in the energy range of 0.3-3 keV/amu. These ENAs are likely to
be generated through charge exchange between the shocked solar wind protons and the
Martian exosphere in the subsolar region, where the solar wind plasma penetrates to its
lowest altitude and where the neutral gas density is high.
Figure 1.6: Images of ENA emissions near Mars. The look direction is toward the center of Mars. The
view position is at a distance of 3 R M . The angle of the view position to the Mars-Sun line is, from
left to right: top to bottom, 80◦ , 100◦, 120◦ , 140◦, 160◦, and 180◦. The images have a field-of-view of
180◦ and show the intensity (cm−2 sr−1 s−1 ) as a function of direction (q, j) in a polar format, with the q
coordinate as the polar angle and j in the radial direction. The axes show the angle to the look direction,
j (deg). The circle is the obstacle boundary, of radius 1.05 R M . The up direction is perpendicular to the
ecliptic plane, along the z axis. From Holmström et al. (2002).
• Accelerated planetary ions

Because the Martian upper atmosphere is directly exposed to the solar wind, the cold
planetary atomic and molecular species, once ionized, are being picked up and accele-
rated by the solar wind convection electric field and can subsequently escape the planet.
A fraction of these ions can experience charge exchange reactions resulting in a spe-
cific ENA signal. Lichtenegger et al. (2002) and Barabash et al. (2002) investigated the
details of such ENA fluxes associated with hydrogen pick-up and oxygen pick-up.
Using the empirical model of the solar wind plasma flow near Mars developed by Kallio
(1996), Barabash et al. (2002) solved the kinetic equation numerically to obtain the
global distribution of oxygen ions. This distribution was then converted to the cor-
responding ENA fluxes. The differential fluxes of oxygen ENAs were estimated for solar
minimum conditions to reach 105 cm−2 sr−1 s−1 eV−1 in the energy range 0.1-1.7 keV. It
was found that the majority of oxygen ENAs have energies below 600 eV. For these en-
ergies the integral fluxes of O-ENAs could reach 104 cm−2 s−1 eV−1 . Figure 1.7 shows
simulated O-ENA images in a fish-eye projection, obtained for different vantage points
in the noon-midnight meridian plane, and the corresponding vantage points.
Figure 1.7: O-ENA images simulated for vantage points with different solar zenith angles (SZA). The
energy range is 0.1-1.65 keV to cover the main oxygen ion population. The projection is a polar one,
with the radius being the angle to the axis pointing toward the planetary center and the polar angle is
the angle to the solar direction in the plane perpendicular to the planetary center direction. The position
of the vantage points are shown in the inserts as well as electrical and magnetic field vectors. All points
are in the OXZ plane. From Barabash et al. (2002).
The total ENA production rates were estimated by Galli et al. (2006a) from observa-
tions for both H-ENA and O-ENA to be 2-3 ×1024 s−1 and < 1022 s−1 , respectively. This
corresponds to a total escape of both H and O as <1g/s.
• Back-scattered hydrogen (ENA albedo)

The generation mechanism of back-scattered ENAs (H – ENA albedo) is described in
Section 1.1.2. Kallio and Barabash (2000) used a three-dimensional Monte Carlo model
to investigate the back-scattered ENAs. The ratio of the particle flux of the back-scattered
ENAs to the impinging ENAs was found to be 0.58. The average energy of the back-
scattered ENAs was 60% of that of the impinging ENAs.
Yet, according to the models, some of the solar wind ions directly impact the Martian up-
per atmosphere near its exobase (∼180 km altitude) because their gyro-radii are too large
to behave as a deflected ’fluid’ in the subsolar magnetosheath (Brecht, 1997a; Kallio and
Janhunen, 2001) and/or because they are partially thermalized in the BS (Kallio et al.,
1997). These protons, reaching the exobase, experience the similar elastic and inelastic
collision processes and a portion of them is scattered back as hydrogen atoms, resulting
in the ENA albedo. The ENA flux generated by this proton – ENA albedo process was
estimated by Holmström et al. (2002) to be 103 - 104 cm−2 sr−1 s−1 , i.e., negligible in
comparison with the ENA albedo flux produced by precipitating hydrogen atoms.
The H-ENA albedo on the dayside of Mars was detected by the ASPERA-3 / NPD sen-
sor (Futaana et al., 2006b). The back-scattered ENAs have energies of 0.2-2 keV with
an average energy of ∼1.1 keV. The differential flux of back-scattered H-ENAs was esti-
mated to be 1.5-2.0 ×106 cm−2 sr−1 s−1 .
• Sputtered O-ENAs
The solar wind motional electric field accelerates ionized cold exospheric components
producing pick-up ions (mostly H + and O+ ). All ions are accelerated up to twice the
speed of the solar wind. Thus, because of higher mass, O+ ions reach the highest energy.
Atomic O ionization to produce O+ ions can be due to various processes, e.g., photo-
ionization by UV photons (O + hv → O+ + e− ), photo-chemical processes in the upper
atmosphere, charge exchange reactions in a collision with solar wind high energy H +
(O + H +∗ → O+ + H ∗ ) and electron impact ionization.
Because of a gyro-radius comparable with the Mars size, some of these pick-up ions can
re-enter the Martian upper atmosphere and reach the exobase. After the precipitation O+
ions charge exchange and the resulting fast O atoms undergo elastic collisions with cold
O atoms in the background gas (Luhmann and Kozyra, 1991). The sputtering mechanism
is shown schematically in Figure 1.3. The large energy is imparted to surrounding parti-
cles through further collisions, causing atmosphere sputtering. A certain fraction of the
particles, gained energy in these collisions, is scattered back out of the atmosphere. This
process occurs on mainly the dayside. The simulated energy spectrum of the sputtered
oxygen is shown in Figure 1.8.
1.3.2 ENA environment of Venus

The ENA environment of Venus is very similar to that of Mars. This is because both planets
are non-magnetized and the solar wind can directly interact with the upper atmospheres. ENAs
are produced in charge exchange collisions between solar wind protons and neutral atoms in
Figure 1.8: Calculated low energy flux from the pick-up O+ ion precipitation. From Luhmann and
Kozyra (1991).
the upper part of the atmospheres of the planets. Of course, the planetary environments differ
considerably. The Venusian atmospheric pressure is about 90 bar at the surface, while the
Martian one is about 7 mbar at the surface. But as the gravity of Venus is about 3 times larger
than that of Mars, the Venusian barometric scale height is lower than the Martian one.
The ENA flux and production rates at Venus are lower than at Mars even though the solar
wind flux is greater at Venus. The reason for this is that the neutral gas density at relevant
heights is lower in the exosphere of Venus that at Mars. The neutral density falls off more
rapidly with altitude at Venus, due to its stronger gravity field. The dominant contribution
to the neutral density at high altitudes at Mars during solar minimum conditions is the large
hydrogen corona (Krasnopolsky and Gladstone, 1996). The hydrogen density at Mars is greater
than that at Venus everywhere above the exobase, and hydrogen is by far the most important
specie for ENA production at Mars (Holmström et al., 2002). The ENA production rate at Mars
at solar maximum conditions is about the same as that at Venus.
Gunell et al. (2005) compared ENA production rate for Mars and Venus (Table 1.2). At
solar minimum a lower ENA production rate is expected for Venus. At solar maximum the
ENA production rate for both planets is expected to be comparable, as the neutral density at
Mars decreases at high altitudes.
The ionopause altitude at Venus is not well known for solar minimum conditions (Luh-
mann, 1992). It is thought to vary with the solar cycle, but since all in situ measurements were
made during solar maximum conditions this variation is still unconfirmed.
Gunell et al. (2005) have investigated the ENA emissions as a function of ionopause dis-
tance by scaling the ionopause altitude in the plasma model. The ENA flux from the local
emission maximum near the planet decreases with increasing ionopause altitude, since with a
higher ionopause altitude the protons pass through a region with lower neutral density. This
also affects the ENA production and escape rates. The ionopause is thought to be close to the
lower end of that range at solar minimum because of lower ionospheric pressure (Luhmann,
1992).
Venus Express has arrived at Venus during solar minimum conditions. The ASPERA-4
instrument provides ENA images of the solar wind – Venus interaction region. Such images
Table 1.2: A comparison of ENA fluxes at Venus and Mars. Values for Venus are given for
ionopause (IP) altitudes 250 km and 400 km respectively. Venusian upper atmosphere is approximately
the same independent of the solar cycle. The values for Mars from Holmström et al. (2002) are all for
solar minimum conditions. Values from the MHD simulation of Mars were taken from Gunell et al.
(2006). "Max. flux" refers to the maximum flux in an ENA image of the interaction region downstream
of the BS. Solar minimum and maximum conditions are denoted by "min" and "max" respectively.
From Gunell et al. (2005).
have been simulated (Figure 1.9) through the integration of the ENA production along lines-
of-sight (LOS) to a virtual ENA instrument (Fok et al., 2004; Gunell et al., 2005). The ENA
images are generated by evaluating LOS integrals in the same way as has previously been
done to simulate ENA images of the Martian environment (Holmström et al., 2002; Gunell
et al., 2006). Gunell et al. (2005) have used a semi-analytical magnetohydrodynamics (MHD)
model (Biernat et al., 1999, 2001) to describe the plasma flow around Venus, and a neutral
gas density model based on published data from measurements. The maximum flux observed
at 3 RV (RV denotes the Venus radius), coming from the interaction region on the dayside of
Venus is 5.8 ×1010 sr−1 m−2 s−1 , which occurs for the lowest ionopause altitude, i.e., 250 km
at the subsolar point. The ENAs that are produced in the solar wind upstream of the BS are
not included in this number. For higher ionopause altitudes (400 km) the ENA flux decreases
and is below 3.8 ×1010 sr−1 m−2 s−1 . The corresponding number for Mars at solar minimum
conditions, computed by Holmström et al. (2002), is about 3 ×1011 sr−1 m−2 s−1 , which is five
times larger than the value obtained for Venus with an ionopause altitude of 250 km.
Figure 1.9: ENA images of Venus from vantage points 3 RV from Venus (planetocentric distance) and
solar zenith angles θ = 80◦ , 100◦, 120◦, 140◦, 160◦, and 180◦. The ENA flux is shown in units of
sr−1 m−2 s−1 , and the axes show the polar angle in degrees. The altitude of the ionopause is 250 km at the
subsolar point. The dominant contribution to the ENA flux comes from a region between the ionopause
and the BS on the day-side of Venus, except in the lower right panel where θ = 180◦, and this region is
occulted by Venus. The second maximum toward the right side of the images with 100◦ < θ < 140◦, is
produced upstream of the BS in the solar wind. Each image has its own colour scale. From Gunell et al.
(2005).
Chapter 2
Energetic neutral atoms detection
2.1 ENA imaging

The capability to detect ENAs with high mass, energy and angular resolutions constitutes the
basis of ENA imaging (Gruntman, 1997). By recording ENA fluxes as a function of observa-
tional direction, one can reconstruct a global image of a remote object of interest. An ENA
image is a two-dimensional (2D) map of the ENA fluxes given by LOS integrals of the ion
distribution convolved with the neutral density through the whole plasma volume. An ENA’s
ability to fly along straight lines resembles the photon’s movement in space. Thus the ENA
imaging concept resembles that used in optical instruments.
All ENA imaging instruments can be classified within three groups, namely non-imaging
detectors, one-dimensional (1D) imaging instruments and 2D imaging instruments. Non-
imaging detectors have a well defined but narrow field-of-view (FOV) and no intrinsic imaging
capability (like telescopes). To obtain a 1D "image" with a non-imaging detector, the latter can
be installed on a scanning platform or make use of a spin of a spacecraft to point the detector in
the desired direction. Combining both the spacecraft spin and the scanning platform allows us
to obtain a 2D image with a non-imaging detector. But as the image is being obtained sequen-
tially, pixel by pixel, total image accumulation time can be very long, as it equals the number of
pixels composing the image, multiplied by a pixel accumulation time. In a 1D imaging scheme,
an ENA imager records the arrival direction of particles only in one dimension, with the second
dimension being narrowly collimated. To scan over an object of interest and obtain a 2D image,
the 1D imager can either be placed on a scanning platform or make use of the spacecraft spin.
A 2D ENA imager records the arrival direction of a particle in two dimensions simultaneously,
while pointing at the object of interest. It can have a sufficiently large FOV to cover the entire
object and is typically located on a three-axis stabilized spacecraft. Time resolution of a 2D
imager can be very high (if counting statistics allows), as it obtains a 2D image momentarily.
However the angular resolution is not high. Charge particle rejection becomes a challenging
problem.
The 1D imager is a compromise between a non-imaging ENA detector with a narrow FOV
and long image accumulation time, and the 2D imager with a large FOV but complex instru-
mentation. 1D imaging has therefore become most common.
ENA imaging (only 1D and 2D imaging schemes) can be divided into two groups, one
obtaining non-inverted images and another one obtaining inverted images. Both concepts are
shown schematically in Figure 2.1. The object plane on the plot is a remote object of interest.
The image plane is the plane onto which a non-inverted or inverted image is mapped. The
15
16 CHAPTER 2. ENERGETIC NEUTRAL ATOMS DETECTION
Collimator
Collimator
pinhole
Image plane Image plane
Object plane Object plane

a) b)
Figure 2.1: ENA imaging concepts: non-inverted (a) and inverted (b) image.
image plane can be a position sensitive detector. The aperture of a non-inverting instrument
(Figure 2.1a) is a long narrow slit. Imaging is performed in one dimension, while the second
dimension is collimated by charge particle rejection plates to a small angle, corresponding to a
width of 1 pixel. This configuration resembles a set of telescopes placed parallel to each other.
Wrapping the aperture slit around 360◦ allows us to obtain a 360◦ viewing plane (as in the NPI
imager, Section 3.2.3).
Another way is to use the pinhole camera concept (Figure 2.1b), where imaging is achieved
by letting an incoming particle pass through a small aperture (pinhole) and impinge on a 1D
or 2D imaging detector, located at a certain distance apart from the pinhole. From an impact
location on the image plane of the detector the arrival direction of an incoming particle can be
estimated. The advantage of this concept is that we can obtain a large FOV with sufficiently
good angular resolution. However, this technique requires longer accumulation times due to
the small aperture size. A larger pinhole size allows us to decrease the accumulation time, but
compromises an angular resolution of the image (blurring). The typical area of the pinhole may
range from 1 mm2 to 1 cm2 . A pinhole camera itself does not provide any information on the
mass or energy of the registered particles. Additional information is obtained by combining
the pinhole imaging concept with different ENA detection techniques (see Section 2.5).
ENA imaging provides images of the plasma region under investigation and gives the ob-
server spectral, compositional and spatial information. ENA imaging is a powerful technique
for obtaining the temporal and spatial evolution of space plasmas on a global scale and is com-
plementary to local plasma measurements.
The ENA flux, originating from charge exchange, that reaches the observer from a given
direction, is a LOS integral of the ion distribution convolved with the neutral density through
the whole plasma volume. Other mechanisms of ENA production are reviewed in Chapter 1.
Considering only the prevalent mechanism, namely charge exchange, the uni-directional dif-
ferential flux fi (E) (in units of cm−2 sr−1 s−1 keV−1 ) of charge exchange neutrals is then given
by an integral along the LOS in equation 2.1, assuming no loss.
X Z
fi (E) = σik (E) ji (E, l) nk (l) dl (2.1)
k l
where σik (E) denotes the energy-dependent charge exchange cross sections for involved vari-
ous ion i and neutral k species. ji (E, l) is the directional singly charged ion flux along the LOS
at each point l for species i within the source volume, nk (l) is the density of the component
k of the neutral gas. The sum extends over all constituents of the neutral gas contributing to
2.2. PRINCIPLE FUNCTIONS OF ENA INSTRUMENTS 17
the charge exchange. The extinction of ENAs due to re-ionization by EUV/UV photons or
collisions with electrons, ions and neutral particles along the way is not included. This ap-
proximation is valid for ENAs traveling through a so-called ’ENA thin’ or ’ENA transparent’
medium, where interaction of ENAs with the medium is negligible.
As soon as an ENA image is accumulated, extraction of the quantitative information from
the image, which contains an admixture of information on energetic ion and cold neutral dis-
tributions, requires deconvolution of the sum of ji (E, l) nk (l) products from the integral 2.1.
Parameterized models of the ion distribution of the observed plasma volume as well as the
density distribution of the neutral gas are needed. Different image processing techniques to
recover plasma distribution parameters of the remote object can be used.
One approach to interpreting an ENA image is a forward modeling technique. In one model
a set of initial parameters is chosen and is used together with another model to simulate an ENA
image, which is compared with the observed one. The model parameters are changed until a
simulated ENA image which matches the recorded ENA image is obtained (Chase and Roelof ,
1995). The forward modeling technique can be rather time consuming if the models contain
large numbers of parameters to change.
Another approach is extraction of the original ion distribution information from ENA
images using an ENA image deconvolution technique (Roelof and Skinner, 2000; Perez et al.,
2000). This is a method which allows deconvolution of the sum of ji (E, l) nk (l) products. This
technique is usually very fast, although it is much more complicated mathematically than the
forward modeling, as it uses inverted equations of the models. Roelof and Skinner (2000)
developed (for investigating the terrestrial ring current) several algorithms to extract the para-
meters of the model ion distribution by minimizing the differences between a simulated image
and an observed image or set of images. Even if the information in the image is insufficient to
determine all details of the ion distribution, the developed algorithms can still often provide a
quantitative estimate of the range of ion intensities or densities on a time scale comparable to
the exposure time required to acquire the images themselves.
The ENA imaging technique can be used for remote sensing of planetary magnetospheres,
and it has been successfully applied to the terrestrial magnetosphere (Mitchell et al., 2000;
Pollock et al., 2000; Moore et al., 2000). Apart from the Earth, it has been considered for
Saturn (Curtis and Hsieh, 1989), for the Martian plasma environment (Holmström et al., 2002;
Barabash et al., 2002) and for that of Venus (Gunell et al., 2005). ENA imaging technique was
also used to image ENA production resulting from the interaction of Titan’s exosphere with
Saturn’s magnetosphere (Amsif et al., 1997; Dandouras and Amsif , 1999).
2.2 Principle functions of ENA instruments

In order to detect ENAs and obtain a sufficiently good signal-to-noise ratio, an ENA sensor has
to be able to perform the following basic functions:
- detection of incoming ENAs with angular, and preferably energy and mass, resolution,
- rejection of charged particles,
- suppression of EUV/UV photon background.
LENA / MENA detection is usually performed by ionization of neutral particles by means
of interaction with foils or conversion surfaces, followed by detection of ionized components.
Depending on which particular ENA sensor detection technique is used, certain angular, energy
and mass resolutions can be achieved. Typically, one can achieve an angular resolution of 5◦ –
40◦ in one direction and 1◦ – 10◦ in another one (in the case of a 1D imager), energy resolution
∆E/E of 30% – 50% for H-ENA and the ability to distinguish between H-ENA and O-ENA.
Charged particle background should be removed from the incident flux prior to detection,
because incident energetic charged particles and ENAs of the same energy and species are
indistinguishable to a particle detector (McEntire and Mitchell, 1989). Often, the local ion and
electron fluxes in the ENA detector environment can exceed the ENA flux to be diagnosed. A
deflector of the ENA sensor prevents ions and electrons from entering the instrument. Charged
particle rejection factor of >103 is typically required.
Suppression of EUV/UV photon fluxes is vital for every ENA instrument because EUV/UV
photons cause photo-electron emission from every lit surface. Emitted photo-electrons can
consequently be registered by particle detectors (e.g., MCP detectors) used for incident ENA
trajectory and energy determination. Moreover, EUV/UV photons can trigger particle detectors
directly. Therefore, an intensive UV flux can induce an unacceptably high noise background in
the ENA measurements and therefore has to be rejected. A typical UV photon rejection factor
of >105 is required.
Other ENA instrument requirements are:
- large geometrical factor, >10−3 cm2 sr / pixel,
- sufficiently wide dynamic range. The ENA sensor should ideally be capable of detecting
ENA fluxes of up to 107 cm−2 sr−1 s−1 in a covered energy range. Thus, particle counting
rate would range up to ∼104 counts/s.
stage 1
stage 2 stage 3 stage 4 stage 5
Charged
particle
rejection Detect ENA
UV / EUV Energy / mass Detect
attenuation Ionize ENA analysis of +
f f f
+
Collimate f f
ENA flux
Stop event
Start event
Detect SE Detect SE
Flux
Direction
Mass Analyze Energy Analyze
SEY Mass SEY
Figure 2.2: Principle functions of a generic ENA instrument are shown. For details see text.
The schematic structure of a conventional ENA instrument is shown in Figure 2.2. The
instrument performs several basic functions which are shown as a sequence of stages. The first
stage is a deflection system, which rejects charged particle fluxes and collimates the incoming
ENA flux.
At the second stage EUV/UV photon flux attenuation is performed, followed by either an
ENA detection or an ENA ionization, release of secondary electrons, and/or generation of a
start event for the TOF measurement, at the third stage. Mapping secondary electrons to a
position sensitive particle detector can also give the direction of an incoming ENA. Analysis
of secondary electron yield (SEY) can also provide a crude mass resolution. Very often UV
photon flux attenuation is performed at the same time with an ENA ionization, i.e., stages 2
2.3. DEFLECTION SYSTEMS 19
and 3 are combined (e.g., thin foils provide both UV rejection and ENA ionization).
At the fourth stage the ionized fraction of neutrals, i.e. positively and/or negatively charged
particles (f+ and f− , respectively), can be analyzed for energy and/or mass. If this stage is
present, then either ion energy or mass or both can be obtained. Otherwise, a particle mass can
be derived by means of secondary electron yield analysis, and particle energy can be calculated
from the TOF measurement and estimated mass information.
At the fifth stage the ionized components (and in some cases neutral fraction as well) are de-
tected at a low-energy particle detector (e.g., MCP) directly, striking the detector, or indirectly,
striking a surface optimized for high secondary electron yield, with consecutive detection of
released secondary electrons. Analysis of secondary electron yield would give also a crude
mass resolution. A stop timing signal is obtained at this stage for the TOF measurements,
yielding incident ENA velocity. Combining this information with that of a particle mass, the
ENA energy can be derived.
The first three stages are present in the vast majority of LENA/MENA detecting sensors.
Any of the other stages or a combination of these can also be present, depending on the practical
design of the sensor. In order to increase an EUV/UV photon rejection factor, UV attenuation
can be done during any other stage, if feasible. There is a detailed description of stages 1, 2, 3
in the following sections.
2.3 Deflection systems

For most missions on which ENA instruments are flown, there are likely to be significant local
charged particle fluxes at the location of the spacecraft. In fact, for magnetospheric missions
one has to assume the local energetic charged particle fluxes in the ENA imager environment
to be orders of magnitude higher than the ENA flux to be measured. Charged particles have
to be prevented from entering the ENA instrument or greatly reduced prior to any interaction
with a detector in an ENA imager. This is because for most particle detectors the detection
of charged particles is similar to that of ENAs of the same energy. Charged particles up to a
certain energy can be prevented from reaching a detector by deflecting them out of the path of
the ENAs in the entrance system of an ENA imager and absorbing them in the structure. This
can be realized with either electric or magnetic fields, or both.
Electric fields are normally used for charged particle rejection. The first element of an
ENA instrument is typically a mechanical collimator, which defines the overall FOV of the
instrument. The collimator can be built in a such way as to consist of two or more closely
spaced parallel metallic plates of length L and separation D, and serve also as an electrostatic
deflector. If these plates are biased with high voltages of alternate polarity, transverse electric
field is created in-between the plates, and charged particles can be deflected to the collimator
plates and be absorbed, while the ENAs are not affected. Charged particles with energy below
the rejection energy Ee will be deflected to the collimator plates. Ee is defined by the applied
voltage V between two adjacent plates and the geometrical dimensions of those, according to
equation 2.2 (McEntire and Mitchell, 1989). Fringe electric fields can be ignored, if L ≫ D.
L 2
" #
Ee = qV 1 + (2.2)
4D
where q stands for the elementary charge. For example, for V = 10 kV, L = 0.12 m and D =
0.004 m, the propagation of ions and electrons of energy Ee /q < 570 kV beyond the deflection
plates is prevented.
Permanent magnets can also be included in the collimator. With a magnetic field strength
B, and magnet length Lm , particles with mass m and energy below the rejection energy Eb will
be deflected into the collimator structure, according to equation 2.3 (Wurz, 2000).
!2
qB2 D2 + L2m
Eb /q = (2.3)
2m 2D
For example, for B = 0.01 T, Lm = 0.02 m and D = 0.004 m, electrons with energy 6 24 keV
are prevented from moving beyond the deflection system. Magnets have an advantage of zero
power consumption. However, in order to deflect ions or high energy electrons higher magnetic
field strengths and, hence, larger/heavier magnets are required as well as a magnetic shield that
can be unacceptably heavy for space-borne instrumentation. Therefore permanent magnets are
used primarily to deflect incident electrons away from the detector.
A charged particle rejection factor of >103 is achievable by means of the deflection tech-
niques described. In practice, it is not enough to deflect incoming ions to hit a plate in the
collimating system, since the deflected ions and, in particular, electrons may scatter further
into the instrument and cause an increase of background signal. Thus, roughening of the de-
flection plates or anti-scatter serration is an important part of a deflection system design.
2.4 UV rejection
UV photon fluxes in the interplanetary environment can be large (especially at distances less
than 2 astronomical units (AU) to the Sun), with a dominant line being H Lyman-α at 1216
Å. This radiation is resonantly scattered on neutral hydrogen atoms in the planet’s exosphere
to create widespread Lyman-α emissions from the planetary corona, with intensities, e.g. at
Earth, ranging from ∼5×107 to 109 photons·cm−2 sr−1 s−1 depending on the vantage point and
observation directions (McEntire and Mitchell, 1989). EUV/UV photons have sufficient energy
to stimulate low-energy particle detectors, such as MCP or channel electron multiplier (CEM).
Since there is a direct optical path from the exterior into an ENA detector, UV fluxes can
cause unacceptable background count rates and thus severely disturb ENA measurements or
even, at most, damage an ENA sensor. Therefore, incident UV photon flux attenuation down
to acceptable levels is a very important aspect of an ENA imager design. There are different
techniques to counter intensive UV flux, depending on the energy range of the ENA sensor, as
shown in Figure 2.3 (Funsten et al., 1998).
UV rejection
Technique
Figure 2.3: Applicability of UV rejection techniques for different ENA energy ranges. Adapted
from Funsten et al. (1998).
Thick UV blocking foils (∼10 -15 µg/cm2 ) in front of an imaging ENA detector screen
out a large fraction of UV flux, though allowing only high energetic neutrals to pass through
2.4. UV REJECTION 21
(Mitchell et al., 1998). Such foils have to be thick enough to block the UV photons efficiently,
but still be thin enough to allow the passage of ENAs without blurring the image. These
conflicting demands result in blocking foils with only moderate photon suppression factors in
the range of 10 to 103 , depending on foil thickness (Hsieh et al., 1991).
Thin (<10 µg/cm2 ) and ultra-thin conversion foils (∼1 µg/cm2 ) have low UV attenuation
factors, and allow LENA/MENA to penetrate. ENAs, converted to ions after passing the foils,
can be separated from UV photons afterwards, by means of electrostatic deflection.
Low energy ENAs cannot penetrate foils. Hence, for lower energy ranges, surface interac-
tion techniques are utilized. A solution to separate LENAs from the incoming UV photon flux
is to convert LENAs to negative ions on low work function surfaces and subsequently deflect
them electrostatically to separate newly created ions from the ambient UV (Gruntman, 1993).
Another way to separate UV photon flux and ENA flux is to mount free-standing transmis-
sion gratings over an ENA imager aperture. Transmission gratings consist of a set of parallel
bars supported by a large mesh grid. The gratings with a period of 200 nm, 100-140 nm bar
width and, therefore, 100-60 nm slit width were developed and investigated (Funsten et al.,
1995). Such gratings act as a diffraction filter to suppress incident UV radiation, while permit-
ting ENA transmission. Figure 2.4 shows UV transmission through gold gratings as a function
of wavelength. A Lyman-α flux transmission factor of 4 ×10−5 is achieved (McComas et al.,
1998). More efficient UV suppression (by a factor of up to 107 ) is achieved by two sequentially
located and perpendicularly oriented gratings, as the UV suppression is lower in the dimension
along the slit. The total geometric transparency of the grating, i.e, transmission to ENA fluxes,
ranges from 0.08 to 0.15 (Funsten et al., 1995).
Figure 2.4: Plots of measured (solid symbols) and calculated (open symbols and lines) UV transmission
through gold gratings as functions of wavelength. From McComas et al. (1998).
This technique differs importantly from previously suggested thin foil/surface interaction in
that the large EUV/UV flux is rejected prior to ENA detection, and therefore, it is not necessary
to extract the small signal from very large background counting rates.
For even lower energy range a high frequency shutter technique can be used. A very high
suppression factor for incident EUV/UV photon flux is achieved, as the shutters perform veloci-
ty filtering of the incoming particles and are basically closed during the detection time window
(see Section 2.5.3).
Blackening of ENA detector internal structure surfaces as well as deflection plates allows
the scattered EUV/UV photon flux to be attenuated. Furthermore, utilizing TOF coincidence
techniques in an ENA imager allows further reduction of background noise caused by UV
photons.
2.5 ENA detection and analysis: Instrument examples

Various ENA ionization techniques are utilized in ENA instrumentation for different energy
ranges. Detectors for MENA or the high energy part of the LENA range (> 600 eV) use
thin/ultra-thin foils in order to convert an incoming ENA to an ion in a foil passage. Different
detection techniques are needed for the low energy part of the LENA range. This is because
such low energy (< 600 eV) is insufficient for a particle to penetrate through an ultra-thin foil.
Several approaches are now possible. One approach is to use a particle interaction with a
conversion surface upon scattering to initiate detection. Different implementations of particle–
surface interaction processes have led to development of a number of techniques. Some of
them are discussed below, namely ENA detection based on ENA-to-negative ion conversion,
particle reflection, secondary ion emission, and kinetic emission of secondary electrons. An-
other approach is to use high frequency shutters. These ENA detection concepts are reviewed
in the following sections, complemented by examples of ENA instrumentation utilizing the
respective ENA detection techniques.
2.5.1 Foils
The ionization of incoming ENAs while passing through one or several thin foils is the most
well-developed technique with heritage from ion spectroscopy. In ENA instrumentation, foils
play a three-fold role: they serve for ENA ionization, for incident EUV/UV radiation attenu-
ation and for producing secondary electrons to generate a detection event. In order to allow
passage of ENAs through a foil, it must be freestanding without a solid substrate support and,
at the same time, be mechanically robust to withstand vibrations and acoustic shocks during the
rocket launch. A high-transparency (90% - 95%) metal grid usually provides the required foil
support with slight reduction of the effective area of the foil. Materials such as carbon or sili-
con are often used for thin foil production, since low atomic number materials reduce scattering
and energy losses of penetrating particles. Ultra-thin foils of carbon are characterized by high
mechanical strength and technological simplicity. Foil composition is usually optimized to
maximize UV suppression while providing high secondary electron yield (Hsieh et al., 1991).
Different units are used for foil thickness, µg/cm2 and Å. Foils of different thickness are used
in space-borne ENA detecting instrumentation: thick (>10 µg/cm2 ), thin (1-10 µg/cm2 ) and
ultra-thin (61 µg/cm2 ). Foils as thin as ∼20 Å (∼0.1 µg/cm2 ) have been reported by McComas
et al. (1991).
ENA penetration of a thin foil results in particle energy loss, scattering, possible change
of the initial charge state, and emission of electrons from the foil surface (Gruntman, 1997).
Figure 2.5 gives an overview of processes which occur during an ENA passage through a thin
foil. ENA scattering (i.e., deviation of the final direction of flight from the initial one by the
angle φ) and energy loss occur due to collisions and interactions with solid body electrons
2.5. ENA DETECTION AND ANALYSIS: INSTRUMENT EXAMPLES 23
Figure 2.5: ENA penetration through a thin foil (TF) results in particle energy loss, scattering, possible
change of particle charge state (fractions f+ , f− , f◦ ), and secondary electron emission (backward and
forward). From Gruntman (1997).
and lattice ions. These interactions are a statistical process, and an incident monoenergetic
and collimated particle beam would be characterized by energy and angular distributions after
leaving the foil.
An incident particle loses its initial charge after passing through the first 2-3 atomic layers
in the foil, and the exit charge state does not therefore depend on the initial charge. An incident
ENA can either be stripped to a positive ion (f+ ) or emerge as a negative ion (f− ), or stay neutral
(f◦ ). ENA ionization efficiency as well as the charge state of a particle exiting the foil, are a
function of particle species and energy (McComas et al., 1998), as shown in Figure 2.6.
Figure 2.6: Ionization efficiency for H, He, and O as a function of energy. From McComas et al. (1998).
Secondary electron emission occurs from both sides of the foil, forward and backward,
while the particle penetrates it (Gruntman, 1997). Electron emission can be used to provide
a Start event in a conventional TOF spectrometer designed for MENA detection and using
thick (10-15 µg/cm2 ) foil, allowing efficient blocking of background EUV/UV photon flux.
Meanwhile, in LENA instruments utilizing ultra-thin foils, the photon flux is only partially
attenuated. As the foils are directly exposed to the incoming radiation, photo-electrons emitted
from foils used for coincidence and TOF measurements, that are able to trigger MCP detectors,
can increase the background count rate during the ENA measurement. This is because photo-
electrons are indistinguishable from the secondary electrons emitted from a thin foil by ENA
passage.
A concept of a foil-based ENA imager (Funsten et al., 1995) for a spinning spacecraft,
designed to detect ENAs in the energy range 0.9 – 30 keV, is shown in Figure 2.7. ENAs
Figure 2.7: Schematic of an ENA imager, based on the ENA stripping in an ultra-thin foil. LENA
– ENA in the energy range of 0.9 – 30 keV, ESA – electrostatic analyzer, SE – secondary electron.
From Gruntman (1997).
pass a collimator that sets an arrival angle, and then transit an ultra-thin (∼1.1 µg/cm2 ) carbon
foil. Newly formed ions enter a hemispheric electrostatic analyzer and are subsequently ana-
lyzed for energy per charge (E/q). The electrostatic analyzer also effectively rejects EUV/UV
photons. Ions that pass through the electrostatic analyzer enter an ion detector, located at the
exit focal plane of the analyzer. The ion detector (Figure 2.7 right panel) is a combination of
another ultra-thin foil and two MCP position-sensitive detectors. An ion transits this second
foil, yielding secondary electrons, detected by one of the MCP detectors, and is detected at
the second MCP detector. Measurement of the coordinates of both electron and particle im-
pact positions allows reconstruction of the trajectory of the incident ENA. The time interval
between electron and particle detection establishes the velocity of the ion and can be employed
for effective suppression of noise counts. Since the ion energy is selected by the electrostatic
analyzer, the ENA mass can be determined. The instrument instantaneous FOV is 120◦ × 2◦
with 120◦ × 360◦ coverage during one spacecraft spin and a nominal pixel resolution of 2◦ × 2◦ .
Estimated energy resolution ∆E/E is ∼10% for 5 keV H-ENA.
2.5.2 Surface interaction
Currently, several methods are employed for particle detection based on surface interaction.
Ionization
Direct detection of ENAs with energies below a few hundred eV is not feasible. Since secon-
dary electron emission, on which most of the particle detectors rely, gives useful secondary
electron yield only at particle energies in excess of a few hundred eV, conversion of atoms into
charged particles is necessary. One approach to ionize LENAs is based on an atom-to-negative
ion surface conversion (Gruntman, 1993). The concept is shown schematically in Figure 2.8.
Ion ( f )
ENA
Conversion surface
Figure 2.8: Schematic view of a neutral-to-negative ion conversion process.
The first instrument to use an atom-to-negative ion conversion technique is the IMAGE /
LENA instrument (Moore et al., 2000), designed to detect ENA in the energy range of ∼10 -750
eV. The instrument concept is shown in Figure 2.9.
Figure 2.9: The IMAGE / LENA principle schematics and ray-tracing. Adapted from Moore et al.
(2000).
The ENA trajectories are depicted by color lines, corresponding to different energies. All
incoming particles enter a collimator and charge particle rejection system, designed to reject
charged particles with an energy of 6100 keV/q. Incoming neutrals pass through a ∼1 cm2
pinhole, strike a near-conical tungsten conversion surface at a shallow angle of 15◦ , where a
considerable fraction of nearly specularly reflected particles becomes negatively charged.
The ionized fraction of the neutrals is accelerated and collected by an ion extraction lens
that focuses them spatially and disperses in energy. A broom magnet in front of an electrostatic
analyzer removes electrons with energies of below 200 keV. Accelerated negatively-charged
ions, mapped onto the electrostatic analyzer entrance, are delivered to a TOF sensor and 2D
imaging system, that sorts the particles by polar angle of arrival and energy. The imaging TOF
subsystem contains ∼2 µg/cm2 carbon foils at the entrance aperture, placed in the focal plane
of the electrostatic analyzer. Upon striking the carbon foil, the negative ions produce secondary
electrons, which provide a start pulse as well as the azimuth and radial position information.
Particles (both ions and neutrals) exit the carbon foil and proceed to a Stop MCP.
As a result, a direct correlation between the azimuth direction and the position on the plane
of the carbon foil is achieved, allowing the original arrival direction of an ENA to be deduced.
In a similar manner, energy information is extracted from the radial impact position. EUV/UV
photon rejection is done by means of an electrostatic analyzer. LENA provides a 90◦ × 8◦ FOV
with a pixels size of 8◦ ×8◦ . Energy resolution ∆E/E ∼100% and mass resolution ∆M/M ∼ 0.25
are achieved.
Reflection
The basic concept for detecting ENA by means of ENA reflection on a reflection surface with
following detection at a MCP detector is shown in Figure 2.10.
UV Stop surface
ENA
Stop surface
Deflector e−
e−
Stop MCP
Figure 2.10: ENA deflection on a reflection surface.
The LENA / MENA strikes a reflection surface at a shallow angle (< 30◦ ) and leaves it as
a negatively (f− ) or positively (f+ ) charged particle or neutral (f◦ ). The emerging particles pro-
ceed to another reflection on a second reflection surface. Finally, a certain fraction of emerging
particles (f+ , f◦ ) strikes a MCP position-sensitive detector. Negatively charged particles as well
as photo-electrons are repelled from the MCP detector by means of a retarding grid in front of
the MCP, or negative potential of the MCP front surface.
This approach was used in the Prelude In Planetary Particle Imaging (PIPPI)–MCP in-
strument on ASTRID, the first dedicated ENA instrument to detect magnetospheric ENAs
(Barabash, 1995; Barabash et al., 1998) and in Neutral Particle Imager (NPI) of the ASPERA-
3 / ASPERA-4 packages on Mars Express and Venus Express missions (Barabash et al., 2004,
2006).
The PIPPI imager comprises two sensor heads, namely PIPPI–SSD, a high energy detector
and PIPPI–MCP, a low energy detector (Figure 2.11). The latter is dedicated to measuring
0.1-70 keV LENA/MENA by means of ENA conversion into secondary particles, followed by
detection at an imaging MCP detector.
SSD sensors Deflectors / collimators
SSD sensor card

0V
e− +4 kV
ENA
ions 0V
+4 kV e−
ENA
_
ions
SSD sensor card
+
MCP sensor card
Target
Retarding grid Imaging MCP
Figure 2.11: The PIPPI sensor head schematics. From Barabash (1995).
The PIPPI–MCP deflection system was designed to reject charged particles with energy
6 70 keV/q. The space between collimation disks is divided into 32 sectors each with 9◦ × 18◦
FOV. Neutrals, passing through the deflection system, strike a 32-sided cone target at ∼20◦
angle of incidence, producing secondary particles, that are further detected by a MCP detector.
UV suppression is achieved by blackening the collimator plates and by selection of the target
block coating. Secondary and photo-electrons are retarded by the potential of the front surface
of the MCP, operating in the ion detection mode. The sensor provides a FOV of 9◦ × 360◦ and
covers almost 4π in a half satellite spin period (with a spacecraft rotation plane perpendicular
to the 360◦ viewing plane).
Advantages of such instrumentation are simplicity and the possibility of obtaining a high
angular resolution in one dimension. The application field is, however, rather constrained
as such instruments normally have low sensitivity and do not have sufficient UV rejection
efficiency, or capability for energy- or mass analysis.
Secondary ion emission
Figure 2.12 shows an approach for detecting ENA in very low energy ranges making use of
secondary ion sputtering from a sensitive surface by incident ENA.
The first instrument utilizing secondary ion emission for detection of ENAs in a very low
energy range is the GAS experiment on board Ulysses, the interstellar neutral helium detec-
tor (Witte et al., 1992). It is designed to measure interstellar neutral helium flux directly in
the energy range of 30-100 eV. Schematic cross-sections of the GAS sensor head are shown in
Figure 2.13.
Incoming particles first pass electrostatic deflection systems, which serve as filters against
charged particles. Incident ENA are converted into secondary ions and electrons upon hitting
the lithium fluoride (LiF) surface. When bombarded with neutral helium atoms, the yield of Li+
from the surface is much higher than that of F− because of the better momentum transfer from
He to Li. As LiF is an ionic crystal, the efficiency for secondary ion production is comparatively
high and is about 10−2 for incident ENA energy around 80 eV. Towards lower energies the
ENA
UV Li
LiF − layer
Figure 2.12: Schematic view of a secondary ion sputtering from a LiF layer. LiF is transparent to UV
radiation, thus UV is suppressed.
Figure 2.13: Cross-sections of the Ulysses / GAS sensor head (schematic): 1) conversion plate with
heater, evaporated with lithium-fluoride (LiF), 2) quartz crystal for monitoring the LiF evaporation pro-
cess, 3) furnace with LiF supply, 4) CEM, 5) CEM electronics, 6) tungsten filaments to stimulate the
CEMs, 7) vacuum-tight cover in closed (dashed lines) and open position, 8) electrostatic deflection sys-
tem, 9,10,11, and 12) circular apertures defining the field of view of channel I and II, 13) light baffle.
From Witte et al. (1992).
efficiency falls off rapidly. Therefore, an instrument based on LiF as a conversion surface is
only useful for neutral helium detection if the incident particle energy is above ∼30 eV and if
there are no much heavier atoms (e.g., O) to be detected. It is ideal for detection of interstellar
neutrals as interstellar He population is next most abundant after H and its energy range is well
defined.
Since LiF is transparent to UV radiation, Lyman-α photons have a low photo-electron yield
from the LiF surface. To keep the surface transparency to UV radiation high, the LiF cover is
refurbished periodically.
The yielded Li+ - ions are accelerated by the electric field to a particle detector, and at the
same time photo-electrons are suppressed. Therefore, a detector system based on the release
of Li+ - ions from a LiF surface is very insensitive to UV photons. An integrated turn-table
together with the rotation of a spacecraft allow coverage of the whole celestial sphere. The
advantages of the secondary ion emission instrumentation are an ability to detect He ENAs in a
very low energy range, and a sufficiently high UV photon rejection factor. However this type of
instrumentation allows particle detection with no capability for mass analysis and very limited
energy analysis using the sputtering threshold energy and the change in the apparent energy of
incident ENAs due to the orbital motion of the spacecraft.
Secondary electron emission

A newly-developed method for detecting ENA in low and medium energy ranges is based on
the kinetic emission of secondary electrons upon a particle scattering from a dedicated surface.
The basic principle of this approach is shown schematically in Figure 2.14. ENAs strike a
Start surface at a shallow angle of incidence of ∼15◦ . Secondary electrons, yielded from the
surface and detected at a Start MCP detector, give a start signal for a TOF measurement. The
reflected particles of any charge state proceed towards the second, Stop surface, and strike it
nearly normally to reduce TOF path length variation. Secondary electrons released from the
Stop surface, detected at a Stop MCP detector, give a stop signal for the TOF measurement.
The energy dependence of the secondary electron yield limits this technique to low energies
of >100 eV.
Stop MCP
Stop surface
Start MCP
e−
UV e−
UV attenuation
ENA f, f ,f
Conversion surface
Figure 2.14: Schematic view of a TOF spectrometer, based on kinetic emission of secondary electrons.
The ASPERA-3 / ASPERA-4 NPD sensors, described in this thesis in Chapter 4, utilize
this concept of ENA detection (Barabash et al., 2004, 2006). The main advantage of this
technique is the high efficiency obtained, as no particle separation by charge is required after
ENA reflection on the Start surface. Comparatively simple design and the light weight of the
instrumentation allow it wide applicability.
2.5.3 High frequency shutters

If one replaces a conventional carbon foil and conversion surface (used in TOF mass spectro-
meters) with high frequency shutters (Funsten et al., 1995) one can build ENA sensors without
intrinsic energy range limitations, and ones that are more robust against EUV/UV. Figure 2.15
illustrates two different implementations of a shutter-based ENA direct detection technique.
In Figure 2.15(a) two shutters, separated by a distance d, are involved. Each shutter is
a set of plates, one plate is stationary and the other is movable. Each plate has an aperture,
the width of which is less than half the travel distance of the moving plate. The oscillating
Figure 2.15: Schematic views of using high frequency shutters in ENA detection. From Funsten et al.
(1995).
plates of both shutters are 180◦ out of phase. UV trapping proceeds as shown in panel (a).
The apertures of the first shutter plates are aligned (i.e., shutter is opened), so that both ENAs
and UV photons can enter the region between the shutters. The apertures of the second shutter
plates are misaligned (i.e., shutter is closed), so that UV photons cannot enter the detector
region. Next, both shutters are closed, so the region between the shutters acts as a light trap
in which UV photons are absorbed. Finally, the second shutter is opened, so that ENAs can
enter the detector region. This cycle is repeated. The typical time period when the shutters are
closed is set to be much longer than that with opened shutters; UV photon flux is absorbed and
photo-electrons are removed by means of electrostatic deflection during this period.
The technique illustrated in Figure 2.15(b) is similar, except the second shutter is replaced
with a detector that is electrically gated to operate only when the shutter is closed and to shut
off when the shutter is opened. An incoming ENA is identified by having it pass through the
opened shutter (at that moment a start signal is generated) and then by measuring the elapsed
time until the particle hits a particle detector at a given distance. Basic shutter ’open’ state
time is much shorter than that of the ’closed’ state and the free-field path d is sufficiently long.
Hence, the UV-induced signal in the detector can be clearly separated from the one caused by
ENAs, as the particles arrive a while after the shutter closes. This results in ENA detection
while UV photon flux is rejected.
The first space-borne instrumentation utilizing the concept based on Micro-Electro-Mecha-
nical System (MEMS) technology, is under development (Brinkfeldt et al., 2006a). It is shown
in Figure 2.15(b). The verification of the technique will be first performed using MEMS shut-
ters in an ion spectrometer. The expected ion energy range is ∼5 -100 eV/q, energy resolution
∼15%. The sensor typical oscillating frequency is in the range of 100 - 300 kHz. Duty cycle of
the sensor times MEMS transparency is designed to be >10−5 , which will place constraints on
a lower detection limit of the ENA flux. Flight verification of the instrument is due to be done
during the PRISMA mission in 2008 (Wieser et al., 2006).
Chapter 3
The ASPERA-3 and ASPERA-4

experiments
The Analyzer of Space Plasmas and Energetic Atoms (ASPERA)-3 experiment (Figure 3.1)
on board the European Space Agency (ESA) Mars Express mission is a comprehensive plasma
diagnostics package to measure ENAs, electrons and ions with a wide angular coverage from a
three-axis stabilized platform (Barabash et al., 2004). The ASPERA-4 experiment (Figure 3.2)
on board the Venus Express mission is a replica of the ASPERA-3 package.
NPD2
NPD1
Figure 3.1: Main unit of the ASPERA-3 flight instrument, covered with black multi-layer insulation
(MLI). The NPD with GN2 purging tubes connected is identified. All sensors have red protective covers.
31
32 CHAPTER 3. THE ASPERA-3 AND ASPERA-4 EXPERIMENTS
NPD2
NPD1
Figure 3.2: Main unit of the ASPERA-4 flight instrument, covered with white MLI. The NPD with
GN2 purging tubes connected is identified. All sensors have red protective covers.
3.1 Scientific objectives

3.1.1 ASPERA-3
The studies of the solar wind – atmosphere interaction address the fundamental question: How
strongly do the interplanetary plasma and electromagnetic fields affect the Martian atmosphere?
This question is already coupled with the problem of what happened to the Martian water that
once flowed in numerous channels? As we know from the experience of our Earth, together
with an inventory of organic compounds and external energy sources, liquid water is a funda-
mental requirement for life as we know it. Therefore, a clear understanding of the fate of the
Martian water is a crucial issue in resolving the problem of whether or not life existed on Mars
in the past.
In order to study the processes related to the impact of the solar wind – Mars interaction
on the atmosphere, the ASPERA-3 experiment was designed to measure electrons and ions in
the hot plasma energy range, as well as to provide remote sensing (diagnostics) of the plasma–
neutral gas interaction via ENA measurements.
The general scientific objective of the ASPERA-3 experiment is to study the solar wind–
atmosphere interaction and to characterize the plasma and neutral gas environment within the
space near Mars through the use of ENA imaging and by measuring local ion and electron
plasma populations.
The general scientific task can be subdivided into specific scientific objectives, listed in
3.1. SCIENTIFIC OBJECTIVES 33
Scientific objectives Associated measure- Measurement requirements

ments
Determine the instanta- ENAs originating Measure the ENA flux in the energy range tens
neous global distributions from the shocked of eV - few keV with 4π coverage.
of plasma and neutral gas solar wind ENA flux > 104 cm−2 s−1 keV−1
near the planet Measure the upstream solar wind parameters
Study plasma induced at- ENAs originating Mass resolving (H / O) ENA measurements in
mospheric escape from the inside of the energy range up to tens of keV.
the magnetosphere ENA flux > 103 cm−2 s−1 keV−1
Investigate the modifica- ENA albedo Mass resolving (H / O) ENA measurements in
tion of the atmosphere the energy range down to tens of eV from
through ion bombard- the nadir direction
ment ENA flux > 106 cm−2 s−1 keV−1 (100 eV)
Investigate the energy de- Precipitating ENAs ENA measurements in the energy range tens
position from the solar of eV - few keV.
wind to the ionosphere ENA flux > 104 cm−2 s−1 keV−1
Search for the solar wind ENAs originating ENA measurements in the energy range tens
– Phobos interactions from Phobos of eV - few keV with 4π coverage
ENA flux 104 cm−2 s−1 keV−1
Define the local char- Ions and electron Ion and electron measurements in the energy
acteristics of the main measurements of range few eV - tens of keV with 4π coverage
plasma regions hot plasma
Table 3.1: The ASPERA-3 scientific objectives
Table 3.1, together with the corresponding instrument requirements.

The key objectives of the ASPERA-3 experiment are:
1. to determine as precisely as possible the total ion escape (particles/s) for the major ion
species (O+ , O+2 , CO+2 ),
2. to study momentum, energy and mass deposition from the solar wind into the upper
atmosphere / ionosphere and its response (sputtering),
3. to investigate the morphological structure of the Martian interaction region and define its
local plasma characteristics.
3.1.2 ASPERA-4
The ASPERA-4 experiment is designed to study ENAs, ions and electrons from Venus orbit.
The scientific objectives of the ASPERA-4 experiment are similar to those of the ASPERA-3
experiment. The specific scientific questions are:
- What is the structure of the solar wind interaction region?
- How is the Venus atmosphere coupled with the solar wind? How is mass added to and
removed from the atmosphere due to this coupling?
- The previous question is connected to the issue of "where did the water go?" The early
atmosphere must have contained water equivalent to a global ocean a few meters deep as
follows from the H/D ratio consideration (Donahue and Hartle, 1992). Could the solar
wind interaction have contributed to the water escape (mainly increasing H escape)? Is
the process the same as for the water escape from Mars?
- What is the mass composition of the escaping plasma?
- What is the neutrals – plasma interaction on Venus? How does the presence of the neutral
gas affect plasma dynamics?
- What are similarities and differences in the solar wind interaction with Mars?
The ASPERA-4 experiment is the first at Venus to cover such a wide range of scientific
objectives in the area of the solar wind interaction. The other unique feature of the ASPERA-4
experiment is that it is a replica of the ASPERA-3 experiment orbiting Mars. The combi-
nation of the observations made by two identical instruments at two non-magnetized planets
exhibiting a similar type of solar wind interaction (but differing significantly in terms of inter-
planetary conditions, atmospheric characteristics, and size and mass) opens up completely new
perspectives for comparative magnetospheric studies.
3.2 Instrument overview

As the ASPERA-4 package (Figure 3.2) is a replica of the ASPERA-3 one, this section de-
scribes both the ASPERA-3 and the ASPERA-4 instruments. The packages are designed to
work in different environments, hence there are certain differences between them, listed below.
Mechanically, ASPERA-3 / ASPERA-4 consists of two units, the Main unit and the Ion
Mass Analyzer (IMA). The Main unit comprises three sensors (the Neutral Particle Imager,
the Neutral Particle Detector, the Electron Spectrometer (ELS)) and a digital processing unit
(DPU), shown in Figure 3.3. The Main unit is located on a turnable platform of a mechanical
scanner. The scanner sweeps the three sensors, mounted on it, through 180◦ to give 4π steradian
coverage (ideally) when the spacecraft is 3-axis stabilized. In practice, part of the FOV is
blocked by the spacecraft body. All electrical interfaces of the instrument with the spacecraft
are made through the scanner.
IMA is an improved version of the ion mass spectrographs Freja/TICS, Mars-96/ASPE-
RA-C/IMIS and Planet-B/IMI (Norberg et al., 1998). IMA is a stand alone instrument with
its own DPU and high voltage (HV) power supplies. Since IMA is not accommodated on the
scanner, electrostatic sweeping is used to achieve ±45◦ elevation coverage. Electrically, IMA
interfaces only the Main unit.
Total mass of the ASPERA-4 instrument is 9.00 kg; the Main unit flight model mass with-
out thermal hardware is 6.63 kg, while the mass of the IMA flight model is 2.37 kg without
thermal hardware. The maximum power consumption is 18 W. The Main unit envelope is
350 × 263 × 288 mm; for IMA it is 287 × 187 × 165 mm.
Total mass of the ASPERA-3 instrument is 8.2 kg, the power consumption is 13.5 W, and
the envelopes are 359 × 234 × 393 mm and 255 × 150 × 150 mm for the Main unit and IMA,
respectively.
The ENA sensors (NPI, NPD) of the ASPERA-3 / ASPERA-4 instruments complement
each other. The NPI is designed to provide ENA measurements with relatively high angular
resolution, but no mass and energy discrimination, while the NPD performs mass and energy
analysis of the incoming ENAs, but the angular resolution is crude. This approach also gives
the necessary redundancy as well as the independent cross-checking that is necessary for such
measurements in a new environment. The charged particle sensors not only provide charac-
terisation of the local plasma environment, but also support ENA measurements in terms of
charged particle background and inter-calibrations. The ELS sensor represents a new genera-
3.2. INSTRUMENT OVERVIEW 35
NPD2
NPD1
NPI ELS
scanner DPU
Figure 3.3: ASPERA-3 main unit. The NPD and ELS apertures are covered with protective covers
(red). The NPI protective cover has been removed to show the NPI and ELS sensors.
tion of ultra-light, low-power electron sensors. It is a standard top-hat electrostatic analyzer in

a very compact design and with high energy resolution.
The ASPERA-3 / ASPERA-4 instrument design, while based on a modular structure,
demonstrates a high degree of packaging and sharing of the internal resources. The instrument
DC/DC converters are shared between all 5 units, including the two identical NPD sensors.
The DPU mechanical structure also serves as a carrying support for mounting the NPD sensors
and the NPI, which in turn is carrying the ELS. The internal walls that separate the DPU, NPI
and both NPD sensors have been replaced by conductive kapton foils to minimise mass, while
maintaining sufficient electromagnetic shielding.
Table 3.2 summarizes the ASPERA-3 / ASPERA-4 instrument performance.
The differences between the ASPERA-3 and the ASPERA-4 instruments are mainly due to
the environments they are designed to be immersed in. As Mars is located more than twice as
far from the Sun as Venus, the solar thermal flux delivered to the Mars Express spacecraft orbit-
ing Mars as well as all on board instruments is much lower than that for Venus Express orbiting
Venus. The thermal design is therefore different: the ASPERA-3 multi-layer insulation (MLI)
outermost layer is black kapton, while for ASPERA-4 it is white kapton. ASPERA-4 is also
equipped with a radiator covered by optical secondary reflection (OSR) mirrors (Figure 3.2).
Parameter NPI NPD ELS IMA

Particles to be measured ENA ENA Electrons Ions
Energy range, keV/q ∼0.1 - 60 0.1 - 10 0.01 - 20 0.01 - 30
Energy resolution, ∆E/E – 0.5 0.08 0.07
Resolved masses, amu – 1, 16 – 1, 2, 4, 8, 16, >32
Intrinsic field-of-view 9 × 344◦
◦ 9◦ × 180◦ 10 × 360◦
◦ 90◦ × 360◦
Angular resolution (FWHM) 4.6◦ × 11.5◦ 5◦ × 30◦ 10◦ × 22.5◦ 4.5◦ × 22.5◦
G-factor / pixel, cm2 sr 2.5 × 10−3 6.2 × 10−3 7 × 10−5 1.6 × 10−6
(ε not incl.) (ε not incl.)
Efficiency, ε, % ∼1 1-15 Inc. in G Inc. in G
Time resolution (full 3D), s 32 32 32 192
Massa , kg 0.7 1.3b 0.3 2.37
Power, W 0.8 1.5 0.6 3.5
a ASPERA-4
b2 sensors
Table 3.2: The baseline performance of the NPI, NPD, ELS and IMA sensors. FWHM stands for full
width at half maximum, G-factor a geometrical factor.
Open surface treatment is also different. Those ASPERA-4 surfaces, not covered by MLI,
are covered with PCBE white paint. The radiation requirements were also different for the two
missions, 5 kRad for Mars Express and 30 kRad for Venus Express. In order not to change the
component selection, the point shielding (tantalum) was used to protect ASPERA-4 compo-
nents which were not sufficiently radiation hard. To protect the ASPERA-4 / ELS electronics
an aluminum protecting skirt was introduced around the entire sensor.
a) b)
Figure 3.4: The ASPERA-4 overall configuration with the Main unit (left) and IMA (right). The
NPI protective cover has been removed to show both the NPI and ELS sensors. From Barabash et al.
(2007b).
3.2.1 The Ion Mass Analyzer (IMA)

IMA is a top-hat electrostatic analyzer (Figure 3.5a) resolving H + , He++ , He+ , O+ and mole-
cular ions in the energy range 10 eV/q - 30 keV/q. Being located separately from the scanner, it
a) b)
Figure 3.5: Photograph (a) and a cross-section (b) of the IMA sensor. Adapted from Barabash et al.
(2007b).
is equipped with an electrostatic deflector to scan over elevation. The principal diagram of the
instrument is shown in Figure 3.5b. Ions first pass through a grounded grid and enter the deflec-
tion system comprising two curved electrodes that deflect ions arriving in the instrument over
elevation range ±45◦ (shown as trajectories 1, 2 and 3) and over 360◦ azimuthal FOV into the
entrance of the top-hat analyzer. The latter selects only ions with a given energy according to
the interplate voltage of the analyzer’s hemispheres. The ions, passed through the electrostatic
analyzer, exit the annular space separating the hemispheres and travel through the magnetic
velocity analyzer section (magnetic separator). The cut of this section in the azimuthal plane
is shown on the right side of Figure 3.5b. The magnets are shown as blue wedges. Sixteen
gaps between the magnets correspond to 16 azimuthal sectors of the instrument of 22.5◦ each.
In the magnetic mass analyser, the ions pass through a static, cylindrical magnetic field, which
deflects them outward, away from the central axis of the analyzer system. The lighter the ions,
the stronger its trajectory bending. As the ions leave the magnetic mass analyzer, they are reg-
istered at the MCP detector with a position sensitive anode. A system of 32 concentric rings
measures the radial impact position, which corresponds to the ion mass, and 16 sector anodes
measure azimuthal impact position, which corresponds to the ion azimuth entrance angle. To
measure light ions, such as H + , at low energies, which have too small gyro-radius to reach the
MCP, the ions can be post-accelerated in the gap between the electrostatic analyzer exit and the
magnet assembly entrance. Varying the post-acceleration voltage allows one to select the mass
range and to adjust the mass resolution.
The IMA sensor also includes an IMA DPU and a HV unit, providing the MCP bias and
sweep voltages for all electrodes. Energy range sweeps from 30 keV down to 10 eV over 96
logarithmically equidistant steps. The polar angle is scanned from −45◦ up to +45◦ over 16
steps. The time to complete a full 3D spectrum is 192 s.
a)
b)
Figure 3.6: Photograph (a) and cut-off view (b) of the ELS sensor. Adapted from Barabash et al.
(2006).
3.2.2 The Electron Spectrometer (ELS)

ELS is a top-hat electrostatic analyzer to measure electrons in the energy range 10 eV - 20 keV
entering over 360◦ FOV (Figure 3.6a). It comprises a collimator system and a spherical top-hat
electrostatic analyzer with the radii of the inner and outer hemispheres equal to 14.9 mm and
15.9 mm respectively (Figure 3.6b).
Entering electrons are deflected into the spectrometer by a positive potential on the inner
hemisphere. A spectral measurement is achieved by stepping the plate voltage. The time of one
energy sweep, consisting of 128 steps, is four seconds. The ELS sweep is fully programmable
within the constraint of the maximum decay rate of 32 steps/s. The electrons, filtered by energy,
are detected by an MCP detector. There are 16 anodes behind the MCP defining 16 azimuthal
sectors of 22.5◦ each.
ELS was designed to be sun blind, so that it may operate during exposure to direct sunlight.
EUV/UV photon flux is minimized through the use of a series of light baffles in the collimator
and a series of light traps at the entrance to the spherical deflection plates. Photo-electrons
are reduced by use of a special coating, based on a modified Ebanol-C process, applied to the
deflection surface, light trap and the collimator system (Johnstone et al., 1997). To eliminate
the photo-electron flux, a retarding grid is introduced between the MCP and the analyzer exit.
3.2.3 The Neutral Particle Imager (NPI)
NPI (Figure 3.7a) is designed to measure ENAs with a relatively high angular resolution. All
incoming particles, entering over 360◦ FOV, pass between two 150 mm diameter disks sepa-
rated by a 3 mm gap. The disks and radial spokes in-between them collimate the incoming
beam over the elevation and azimuth angle to give a pixel of 9◦ × 18◦ FOV and 4.5◦ × 11.5◦ full
width at half maximum (FWHM). The 5 kV potential between the grounded and biased disks
results in a strong electric field, which sweeps away charged particles with energy less than 60
keV.
a)
b)
Figure 3.7: Photograph (a) and cut-off view (b) of the NPI sensor. Adapted from Barabash et al. (2006).
Neutrals that pass through the deflector system hit a 32-sided conical target at a grazing
angle of incidence of ∼30◦ . On impact with the target block the incident neutral can either
be reflected in positive, negative or neutral charge state, or produce secondary particles. The
secondaries can be both ions and electrons. The particles, leaving the target, are then detected
by an MCP detector followed by a 32 sector anode. The MCP operates in an ion mode with
a negative bias of -2.3 kV applied to the front side and thus detects both positive and neutral
particles reflected as well as sputtered off the target, while electrons are repelled. The internal
view of the NPI sensor is shown in Figure 3.7b. The UV suppression is based on the target
material selection, blackening of the internal surfaces, and the MCP mode operation. The NPI
is located on the scanner platform. Ideally it can cover 4π sr in a 180◦ scan, whereas approxi-
mately a half of its FOV is blocked by the spacecraft body. Also two sectors are mechanically
blocked to accommodate the ELS cable.
3.2.4 The Digital Processing Unit (DPU) and the scanner

The DPU includes two boards: the DPU board itself and a Housekeeping (HK) board. These
are connected together with the sensor control electronics and a power supply via a common
bus system with 8 address- and 16 data lines. The bus also includes control-, analog- and
power supply lines. The software is built around a real-time system with a scheduler and an
interrupt handler. All executable routines are defined inside a routing table, which resides in
an Electrically Erasable Programmable Read Only Memory (EEPROM) and can be modified
during the flight. In this way new or modified software routines can be stored inside a free
area of the EEPROM, verified and added to the operating software by including their start
address into this routing table. A macro feature of the telecommand (TC) handler offers the
possibility to generate sequences of standard TCs automatically according to a pre-defined list,
reducing the need for complex TC groups to be up-linked over and over again. Besides detector
activation and parameter control, compression and averaging of measurement data allow the
reduction of the amount of telemetry generated.
The scanner platform (Figure 3.8) was originally designed for the ASPERA-C experiment
on the Russian Mars-96 mission. Some modifications were made for ASPERA-3 on Mars
Express as well as for ASPERA-4 on Venus Express, most of them concerned with optimization
of the performance during long-term operations and reduction of its mass.
scanning platform
Figure 3.8: The ASPERA-3 scanner. Adapted from Barabash et al. (2006).
The scanner serves as a bearing structure for the ELS, NPI, NPD sensors and the DPU,
as well as providing all electrical interfaces with the spacecraft. It constitutes the 0◦ to 180◦
rotating platform on which the sensors as well as the DPU, HV power supply (HV PS), HK and
DC/DC boards are situated. Rotation is accomplished by means of a worm gear mechanism.
The scan platform is made as a plug-in unit for the sensor assembly.
Chapter 4
The Neutral Particle Detector (NPD)
The Neutral Particle Detector (Figure 4.1) developed for the ESA Mars Express and Venus Ex-
press missions is a compact, low weight (∼650 g), high efficiency TOF sensor to image low and
medium energy ENAs that result from the solar wind interaction with the Martian / Venusian
exospheres. The ASPERA-3/ASPERA-4 packages each contain two identical NPD detectors.
NPD is a pinhole camera with a 9◦ × 90◦ intrinsic field-of-view. The measurement principle is
based on a surface interaction technique. The sensor is capable of resolving hydrogen (H) and
oxygen (O) atoms and provides measurements of an ENA differential flux over the energy range
of 100 eV - 10 keV with a coarse angular resolution of 5◦ × 30◦ and total efficiency of 1-16%.
Figure 4.1: The NPD flight model. The cover has been removed.
41
42 CHAPTER 4. THE NEUTRAL PARTICLE DETECTOR (NPD)
4.1 The measurement technique

Figure 4.2 provides a conceptual view of the detector. Incoming ENA particles pass between
two parallel deflector plates (shown in red), separated by a 3.0 mm gap with an inter-plate
electric potential difference of up to 10 kV, and enter a pin-hole of 3.0 × 4.5 mm2 size. Charged
particles (ions [f+ ] and electrons [f− ]) with an energy of 670 keV/q are swept away by the
transverse electric field created between the plates. The collimated ENA beam emerging from
the pinhole hits a Start surface (shown in grey) at a 15◦ grazing angle, causing secondary
electron emission. The angle between an ENA trajectory and the Start surface is chosen to
provide a maximal ENA reflection from the surface as well as a high secondary electron yield.
Secondary electrons are transported by a system of collecting grids (shown as green boxes) to
one of two Start MCP assemblies (inside the green boxes), giving a start event. The incident
ENAs, reflected from the Start surface nearly specularly, hit a Stop surface. The emerging beam
contains both neutral and ionized components since the charge state equilibrium is established
during the interaction with the surface. No further separation by charge is made in order to
increase the total efficiency. Therefore particles of any charge state–negative, neutral, positive–
a)
TOF
Stop MCP
PH
Start MCP
ENA Pinhole
+5kV secondary
+ electrons
f
f
−5kV
Start Surface Stop Surface
b) Stop MCP secondary

electrons
Start MCP
ENA
Collimator
Start Surface
Start MCP
Stop Surface
Figure 4.2: The principle design of the NPD sensor. a) view from the side, b) view from the top. For
details see the text.
4.2. NPD MECHANICAL DESIGN 43
impact the Stop surface. A black line depicts an ENA trajectory. Secondary electrons yielded
from the Stop surface accelerate towards one of three Stop MCP assemblies, mapping the
position of the impact, and produce a stop event. The start and stop signals are then used for
TOF measurements. The particle velocity is defined by the TOF over a fixed distance between
the Start and Stop surfaces of on average about 8 cm. Three Stop MCPs give an angular
resolution of roughly 30◦ within the 90◦ azimuth acceptance angle. The secondary electron
yield from the Stop surface depends on the atomic mass number of a neutral and, therefore,
is different for ENAs of different mass moving with the same velocity. The pulse height (PH)
distribution analysis of the stop signals can be used to estimate the mass (H or O) utilizing this
phenomenon. Moreover, the difference in the TOF between H and O atoms is sufficient for
direct mass identification within the particular energy range. UV suppression is provided by
properties of the Start and the Stop surface coatings as well as by the coincidence between start
and stop signals within a defined TOF window. Figure 4.3 shows a 3D model of NPD with the
main components identified.
4.2 NPD mechanical design

NPD consists of several functional units, namely, a deflector, Start unit, Stop unit, Stop surface,
housing, and electronics (Figure 4.4).
The sensor is enclosed in a rectangular chassis with a size of 200 × 140 × 60 mm (Figu-
re 4.4e). The top and bottom sides are closed by an aluminum plate and/or kapton membranes
with a conductive coating. The deflector housing (Figure 4.4a) is a triangular box, located in
Figure 4.3: The 3D model of the NPD sensor.

f)
b)
a)
e)
d)
c)
Figure 4.4: NPD functional units. a - Deflector, b - Start unit, c - Stop unit, d - Stop surface with a
holder, e - housing (chassi), f - electronics.
the upper left corner of the chassis. The opening is covered by a highly transmitting (>90%)
stainless grid. The collimator plates are located parallel to each other inside the deflector. A
strong transverse electric field is created across the incoming beam path when plates are biased.
The central deflector plane is inclined 15◦ away from the Start surface plane.
The Start unit (Figure 4.4b) is located next to the deflector. The entire unit is electrically
insulated from the chassis. The Start surface, two MCP assemblies, electron optics system and
the pinhole plate are placed on the Start unit frame. High voltage- and signal cables, and the
electron optics system bias cable, are attached to the corresponding connectors on one side of
the unit. The connectors and parts of the MCP electronics boards, exposed to the ion beam
emerging from the Start unit, are closed by grounded blackened (Ebanol-C) shields.
The Stop unit (Figure 4.4c) is located next to the Start unit. It comprises a frame holding
three MCP detectors. The frame is electrically coupled to the chassis. High voltage- and signal
cables are attached to the connector on the back side of the frame.
The Stop surface (Figure 4.4d) is located next to the Stop unit. The Stop surface is part of
a sphere of 108 mm radius to cover all possible particle trajectories from the Start unit.
All sensor elements are fixed to the outer shell of the mechanical structure (chassis). This
configuration provides easy access to all elements, mechanical stiffness, and minimizes the
structural mass.
Four signal cables are guided from the Start and Stop units through the sensor and con-
nected to the electronics package (Figure 4.4f) located in the opposite corner to the deflector.
All high voltage cables as well as the electron optics bias cable bypass the sensor and connect
to the Main unit HV board. The Start and Stop MCP detectors are biased separately. The digital
electronics connection with the Main unit DPU is performed via a MDM-37 connector.
The sensor has a quite large empty space between the Start unit and the Stop surface as well
as in the electronics compartment, and is therefore very light. The high voltage is generated
externally, in the Main unit HV PS.
Pointing of the NPD sectors FOVs (Dir0, Dir1, Dir2) is shown in Figure 4.5, depicted by
black arrows.
Dir0
Dir1
Dir2
Figure 4.5: Engineering drawing of NPD. The black arrows depict the pointing of the NPD sectors.
The NPD package is comprised of two identical sensors, NPD1 and NPD2. It covers a
hemisphere (half a hemisphere per sensor) when placed on a scanning platform, which per-
forms 180◦ sweeps (see Figure 3.3). A conductive kapton film provides an electrical screening
between the units mounted together.
Main unit HV PS. The ASPERA-3 / ASPERA-4 Main unit HV PS provides high voltages
for the NPD1 and NPD2 sensors as well as for NPI. Each NPD sensor is connected to an
individual HV PS which is built around two base supplies. The base supply uses a conventional
coil transformer followed by a custom-made doubler. The first single polarity HV PS provides
base voltage from 0 to +3 kV regulated further by two AMPTEK HV601B optocouplers to bias
the Start and Stop MCP assemblies individually. The second double polarity HV PS provides
two voltages from 0 to ±5 kV for the NPD deflector electrodes. The regulation accuracy is 256
steps for each range, which is sufficient for this application.
Assembled NPD sensor. Figure 4.6 shows a cross-section of a 3D model of NPD with the
main units depicted by arrows, and an incoming ENA trajectory (as a pink line). Figure 4.7
shows the NPD flight model with covers removed to exhibit the main elements.
4.2.1 Deflector
The NPD deflection system is an electrostatic deflector, consisting of a pair of parallel plates
located inside the grounded deflector housing. Commandable potentials of up to ±5 kV are
ENA
Stop surface
Stop unit
Start unit
Deflector
Figure 4.6: Cross-section of a 3D model of NPD. Main units are depicted by arrows.
applied to the plates to sweep charged particles with energies 6 70 keV/q into the plates (ex-
cluding them from the detector). Each deflection electrode is a blackened aluminum plate of a
trapezoidal shape with serrations to inhibit forward scattering of incident particles. Figure 4.8
shows a cross-section of a 3D model of the NPD deflector with the lower half cut away. The
deflector electrodes are visible inside the housing. The entrance slit has dimensions of 3.0 mm
× 70.0 mm. It is covered by a stainless steel grid in order to prevent electric field leakage
outside the deflector housing.
Charged particles with energy below the cut-off energy Ee will be rejected. Ee of an electro-
static deflector, consisting of two parallel plates of the length L and a gap D in-between, can be
estimated roughly by Equation 2.2. The lengths of the deflector electrodes (along the incoming
beam path) are 30.0 and 31.0 mm. The gap between the electrodes is 3.0 mm. The maximum
potential supplied by the Main unit HV PS on the deflection plates is ±5 kV. Therefore, all
charged particles with energy up to ∼70 keV/q, will be swept away, while passing through the
deflector. Energetic charged particles with energy above Ee can pass through the deflector, but
they are not registered as valid events due to their very short TOF. Detailed investigation of the
sensor response to high energy particles is given in Section 4.7.
4.2.2 Start unit

The Start unit consists of the pinhole plate, the Start surface (see Section 4.2.4), the electron
optics system to collect and guide secondary electrons toward the Start MCPs, and two Start
MCP detectors. The Start unit is shown in Figure 4.9. The left panel is a 3D model of the
Start unit. All key elements are depicted by arrows. The Start surface is not shown. By a
system of collecting grids (shown in blue), secondary electrons emitted from the Start surface
are transported to one of two MCP assemblies (inside blue boxes), giving a start event. The
right panel shows a photo of a completely integrated Start unit. Figure 4.10 shows the key
elements of the Start unit during integration. Panel (a) shows the Start unit from above. Two
Digital electronics
Stop MCPs
Deflector
Start MCPs
Front−end−electronics
Stop surface
Figure 4.7: NPD flight model with the main elements shown.
MCP assemblies are located aside the trapezoidal Start surface. The pinhole is attached in
front of the latter. Panel (b) shows parts of the electron optics system. Panel (c) shows a view
of the completed Start unit from below. The electronics of the MCP detectors is depicted by
arrows. In the center of the Start unit frame can be seen a trapezoidal aluminum plate with
a temperature sensor and a heater attached (see details in the following section). The plate is
attached to the bottom side of the Start surface. Panel (d) shows this plate on its own.
High transmission grid
Deflector housing
Deflector plates
Figure 4.8: Cross-section of a 3D model of the deflector.
a) b)
ceil plate
pinhole MCP detector

frame Start surface
HV connector
electron optics system signal connector
Figure 4.9: The Start unit assembly. (a) 3D model of the Start unit with key elements depicted by
arrows. The Start surface is not shown. (b) a photo of the integrated Start unit.
Electron optics system
The main purpose of the Start unit electron optics (Figure 4.9a, colored blue) is to provide an
effective transport of secondary electrons yielded from the Start surface to Start MCPs located
inside the cages, preventing direct exposure of the MCP detectors to UV photons coming into
the sensor. Figure 4.11 shows an electrical model of the electron optics system. The cages,
main elements of the optics, have openings toward the Start surface closed by highly transparent
(91%) golden grids. The grids are biased with +50 V. All inner sides of the cages have the
same potential as the Start unit frame (+12 V). Small electrodes, biased differently (+20 V),
are embedded into the inner back sides of the cages. The top horizontal plate of the electron
optics, located above the Start surface, has a negative potential of -12 V to repel electrons
towards the cages. All potentials are set with respect to that of the Start unit frame. The entire
Start unit is biased positively (+12 V), to prevent secondary and photo-electrons yielded from
the Start surface from reaching the Stop unit. The front side of the MCP detectors is biased by
+300 V.
a) b)
c) d)
MCP detectors electronics

Figure 4.10: (a) the Start unit from above; (b) electron optics system cages; (c) the Start unit from
below; (d) an aluminum plate with a heater and a temperature sensor attached.
As proven by ion ray-tracing, the Start electron optics introduces minor disturbance to the
reflected ions leaving the Start surface with an energy above 80 eV. Figure 4.12 shows the
results of the electron and ion ray-tracing in the Start electron optics assembly. The electron
collection efficiency of the optics varies from 80% to 95% depending on the azimuth angle of
incident ENAs.
The heater and the temperature sensor
The heater is intended for Start surface pre-conditioning in space. The surface performance was
investigated during thermal vacuum testing performed at IKI, Moscow. The thermal cycling
included five heating cycles from +15◦ to +90◦ and back to +15◦ and five cooling cycles
from +15◦ to -60◦ and back to +15◦ with a temperature gradient of 1◦ C/min. The surface
performance was measured before and after the tests. He+ beams of different incident energy
were used. Figure 4.13 show TOF spectra before (panel a) and after (panel b) the thermal
cycling. The black, yellow, and green curves show the TOF spectra obtained during the tests
using ion beams of energies of 500 eV, 1500 eV, and 2500 eV, respectively. The vertical red
lines show the position of the corresponding peaks obtained before the test. The energy loss at
the surface was found to reduce by several per cent after the thermal vacuum test. A hydrogen
recoil peak (a small bump on the 500 eV He+ TOF spectrum at ∼210 ns) decreased as well
as the electron event peak (at ∼50 ns, it is caused by secondary electrons yielded from the
Start surface reaching the Stop unit). These improvements are related to the cleaning during
- 12V
+20V Slit2
Z, mm
Start Surface
+300V +50V
Y, mm
Start Unit
Figure 4.11: An electrical model of the Start unit electron optics system.
a) b)
Figure 4.12: The ray-tracing of the trajectories of secondary electrons (panel a) and 80 eV ions (panel
b) in the Start unit electron optics system.
heating cycles.
TM
The heater is a 0.4 W flexible Termofoil MINCO heater (see Figure 4.10d) without tem-
perature regulation to heat the Start surface when powered with +5V. It allows an increase of the
Start surface temperature by ∼50◦ above the ambient. The heater operation is commandable.
The temperature sensor AD590J is used to monitor the Start surface temperature continu-
ously. It reflects the Start surface temperature and can be used to monitor the NPD temperature
when the heater, located close to it, is switched off. Both the heater and the temperature sensor
are attached to a thin aluminum plate (Figure 4.10d), which is located in a pocket inside the
Start surface and is pressed tightly to a bottom side of the latter. Using the intermediate alu-
minum plate between the Start surface and the heater helps avoid hot spot heating as well as
making the heater easy to replace.
4.2.3 Stop unit

The Stop unit registers secondary electrons yielded from the Stop surface (Section 4.2.4), pro-
duces stop signals and provides a coarse analysis of the direction of an incoming ENA. It
comprises an aluminum frame (Figure 4.14) to hold MCP detectors and three identical MCP
detectors with some passive HV components.
TOF of He of varying Enegry
2500eV 1500eV 500eV

E [eV] ∆Ε/Ε E [eV] ∆Ε/Ε
a) 500 0.92
b) 500 0.8
1500 0.9 1500 0.86
2500 1.12 2500 0.82
Elevation = 12 deg
Azimuth = −26 deg
Figure 4.13: TOF analysis of different energy beams made a) before and b) after the thermal vacuum
test. The vertical red lines show the position of the corresponding peaks obtained before the test.
The frame is fixed right above the Stop surface in such a way as to achieve the highest
secondary electron collection efficiency. Every MCP detector is shielded by a grounded gold-
plated cover to inhibit electric field leakage outward from the biased MCPs through the MCP
assembly holders. The electric potential of all of the MCP front surfaces is kept the same and is
about +300 V in order to effectively collect secondary electrons released from the Stop surface.
Each MCP detector produces a stop pulse with a magnitude proportional to the amount of
secondary electrons reaching the MCP detector simultaneously. The MCP bias is commandable
and is nominally in the range of 2750 to 3100 V.
Figure 4.15 shows different views of the assembled Stop unit. The MCP detector supporting
electronics and connectors are located at the back of the Stop unit (left panel). The Stop unit
MCP detectors are shown in the right panel. The frame is blackened to inhibit UV reflections
and photo-electron yield.
Figure 4.14: The Stop unit frame.

Supporting electronics
a) b)
HV connector signal connectors MCP assemblies
Figure 4.15: The Stop unit assembly: a) back side with the MCP detector electronics and connectors
depicted by arrows, b) front side of the Stop unit with facing MCP detectors.
4.2.4 Surfaces
The selection of the Start and Stop surfaces was the most challenging part of the NPD develop-
ment. Extensive studies have been performed at University of Bern, Switzerland (Jans, 2000),
and Brigham Young University, USA, to optimize the performance of the surfaces, which must
satisfy a number of requirements, namely, high secondary electron yield, high UV absorption
even at grazing incident angles, high particle reflection coefficient (Start surface), low angular
scattering (Start surface) and low photo-electron yield.
A multi-layer coating composed of a thin layer of Cr2 O3 , covered by a thicker layer of
MgF and topped with a thin layer of WO2 , was used for the Start surface (Figure 4.16). The
coating was optimized for the absorption of the 121.6 nm Lyman-α line at an incident angle of
∼15◦ . It was applied on a titanium substrate polished down to 100 Å roughness. The reflection
coefficient was about 30%. The Start surface is of a trapezoidal shape. A pocket in the surface
substrate is used to accommodate the temperature sensor and the heater.
a) b)
1 cm
Figure 4.16: The NPD Start surface, view from the top (a) and bottom (b).
The Stop surface (Figure 4.17) is part of a graphite sphere (roughness around 100 nm)
covered by a MgO layer of ∼500 nm thickness. This combination has a very high secondary
electron yield, low photo-electron yield, and high UV absorption. Much effort has been ex-
pended into increasing the stability of the MgO coating against moisture absorption. It was
established that polishing the graphite down to the roughness of about 100 nm improves the
surface stability, so no problems related to surface performance were encountered during as-
sembly, integration, and verification (AIV) activities (i.e., due to absorption of air humidity).
4.3. ELECTRONICS 53
Thus, the Stop surfaces did not require special maintenance to retain their stability and preserve
the surface properties unchanged from their calibrated state.
Figure 4.17: The Stop surface in a NPD chassi (right panel). The left panel shows a batch of Stop
surfaces in a transport container.
4.3 Electronics
Each NPD sensor is an "intelligent" device and its electronics perform a substantial portion
of the data pre-processing. The NPD electronics consist of two boards: Front-end electronics
(FEE, Figure 4.18a) and digital TOF electronics (DigTOF, Figure 4.18b). All high voltages are
provided and controlled externally.
The FEE block diagram and typical signal shapes are shown in Figure 4.19. Pulses from
four MCP anodes (one Start and three Stops) are fed to discrete-charge sensitive preamplifiers
based on fast MOS FET (frequency cut-off at 6 GHz) and followed by operational amplifiers.
The wave shaper generates a fast logic pulse for the Time-to-Digital converter (TDC) of the
DigTOF board, and fast Texas TLV1562 video Analog-to-Digital converters (ADC) perform
a pulse height analysis. The signals are up-shifted by 0.8 V to reach the ADC working range
of 0.8 – 3.8 V. Four Digital-to-Analog converters (DAC) provide input threshold control. The
thresholds are commandable. The FEE provides a theoretical TOF resolution of 0.1 ns that is
well within the limits of what is required for this measurement technique (1 ns).
DigTOF electronics serve a number of tasks, including:
- TOF measurement from one start to one out of three stop signals,
- serving FEE and initiating analog-to-digital conversion of stop pulses,
- coincidence check and selection of valid TOF – pulse height pairs,
- events counting,
- three different data types buffering in Static Random Access Memory (SRAM),
- interfacing DigTOF to DPU.
DigTOF is realized in two Field Programmable Gate Arrays (FPGA) Actel 54SX32. Figu-
re 4.20 shows the DigTOF block diagram focused on the functionality of the diffirent com-
ponets and FPGA subroutines.
a) b)
Figure 4.18: Technological models of a) Front-end electronics (FEE) and b) digital TOF electronics
(DigTOF) are shown.
Figure 4.19: Front-end electronics block diagram.
The occurrence of a start signal followed by a stop signal leads to a TOF measurement in
the TDC and generation of a data item with uncorrected TOF information. The time obtained
is then corrected to compensate for a possible temperature drift. The TDC calibration is per-
formed continuously using a 6 MHz clock supplied by the DPU, and the calibration constants
are continuously updated at DigTOF. The TDC control and TOF correction are performed by
a TDC Management unit, a part of FPGA. The final time data item is a TOF value with 12-bit
binary time information. Parallel to this, a sampling process by FEE ADC follows the occur-
rence of a stop signal, thus generating a 10-bit data item with information of both direction and
4.4. MCP ASSEMBLY 55
Figure 4.20: Digital TOF electronics block diagram.
8-bit stop pulse height. The coincidence control checks then for a valid coincidence of these
two data items and additionally flags an occurrence of more then one start or stop signal during
the TOF measurements. NPD registers a ’valid event’ when a signal from one (and only one)
Stop detector corresponds to one start event within the given TOF window. This now leads
to a 25-bit raw data item formation that is used for the pending storing process. The memory
control has to handle three different memory areas in the SRAM. For the binning array, the raw
data is compressed into 10-bit data that represents the bin number inside the array. The respec-
tive bin is incremented by one up to the bin depth of 65535 (16 bit). For the raw data array,
incoming raw data items are successively stored until this array is filled completely (512 en-
tries). The PH array is filled in the same way as the binning array, but with the compressed stop
pulse height data together with the respective direction. All data arrays are filled in parallel.
The binning array is excluded, if the coincidence (1 Start – 1 Stop event) does not fit. Readout
and the following initialization of these arrays are performed by using burst read access from
the DPU. Several 16-bit event counters and two registers facing the preamplifier board are also
implemented in FPGAs. One of these registers is used to program FEE DACs, the other one
for directly commanding FEE. All control, counter and memory data are accessible from the
DPU over 16-bit registers; physically the connection to the DPU is a 16-bit bus.
4.4 MCP assembly

There are five MCP detectors in each NPD: two Start and three Stop MCP assemblies. Fi-
gure 4.21 shows the Start (panel a) and the Stop (panel b) MCP assemblies. Micro-channel
plates are known as high gain, low noise, two-dimensional detectors of charged particles and
electromagnetic radiation. MCP is an array of solid state electron multipliers (Wiza, 1979)
consisting of millions of microscopic electron multipliers (105 - 107 channels/cm2 ) all fused
together in a solid array. Typical channel diameter is 6 - 25 µm. The operational principle of
the MCP detector is shown in Figure 4.22 (panel a). Panel (b) shows a cross-section of a 3D
model of a Stop MCP assembly.
Each NPD MCP assembly comprises a pair of standard ’Photonis’ custom cut rectangu-
a) b)
Figure 4.21: (a) Start and (b) Stop MCP assemblies.
shield
b)
a)
cover
secondary electron
+300V
electrode
MCP
+2700V
+2800V electrode
anod
MCP holder
FEE
HV PCB
Figure 4.22: (a) MCP detector operation principle. Arrows depict major elements. (b) Cross-section of
a 3D model of a Stop MCP assembly.
lar MCP plates of 1.0 mm thickness. MCPs operate in a chevron configuration to avoid an ion
feedback. Secondary electrons reaching the MCP detector are attracted by the +300 V potential
of the front surface of the first MCP. As soon as an electrons enter the MCP channels, it collides
with the channel wall resulting in secondary electrons. The resultant secondary electrons are
accelerated down the channel by the electric field of the bias high voltage applied across the
plates. Further collisions with the channel walls produce more secondary electrons. Finally, an
electron cloud is formed and extracted from the output side of the MCP and accumulates by an
anode, separated from the rear side of MCP by a gap of 1 mm. The anode potential is ∼100 V
higher than that of the back of the MCP and is about +2800 V. The collected charge is trans-
ferred to FEE through a decoupling capacitor. The typical gain of the MCP detectors is in the
range of (6 - 8) ×106 .
The size of the Start and Stop MCP detectors is different. The Start MCP is 8 × 12 mm, and
the Stop MCP is 16 × 12 mm. Also, Start and Stop MCP detectors are connected differently.
Both Start MCP detectors possess a common anode to efficiently collect secondary electrons
yielded from different regions of the Start surface. Stop MCP detectors possess equal potential
on the front MCP surfaces providing an evenly distributed electric field over the Stop surface
in order to allow an angular response. MCP resistance differs between Stop MCP assemblies.
Therefore, a certain resistor R was chosen for each Stop MCP assembly to provide the same
MCP voltage across the plates in all three Stop MCP assemblies.
4.5. DATA FORMATS 57
Start MCP detectors (two assemblies) and Stop MCP detectors (three assemblies) are con-
trolled by two tele-commands independently.
4.5 Data formats

The sensor hardware mode is always the same but DigTOF generates in parallel two data flows
(formats), namely, raw event data (RAW) and binned matrix data (BIN). Depending on the
settings, the DPU reads out one of the flows. Thus, the sensor operates either in the RAW
mode or in the BIN mode.
The RAW mode. The data buffer contains 512 data entries. Each 32-bit depth item corre-
sponds to one recorded event. Incoming raw data items are stored sequentially until the data
buffer is filled completely. Each data entry contains uncompressed TOF values with a 0.5 ns
resolution as well as information about the direction, magnitude (pulse height) of a stop signal,
and a coincidence flag (see Table A.1). The TOF spectral resolution is reduced by the DPU
from 0.5 ns to 1 ns. The TOF window of 1 ns to 2048 ns is defined. PH data values are scaled
to 255 levels.
The BIN mode. Both TOF and PH data are accumulated for each of three directions. TOF
of valid events (one Start – one Stop) is stored in 16 logarithmically divided TOF bins. The
size of the TOF bins and their location in the TOF space are shown in Table A.1. A stop
signal magnitude value is stored in 16 linearly divided PH bins. The data buffer is an array
of 16 × 16 × 3 16-bit counters. TOF is measured in the range of 50 to 1900 ns. The lower
TOF limit allows for a cut off of electron events (’Start’ electrons reaching the Stop unit). The
higher TOF limit defines the upper energy of the range in question, i.e., no significant fluxes of
O ENAs of an energy above 10 keV are expected in the region where ASPERA-3 / ASPERA-4
performs measurements. The matrix of 16 × 16 × 3 can be reduced to 16 × 2 × 3 comprising 16
TOF steps and two PH steps for three directions (see Appendix A.1).
The TOF mode. The RAW event data is post-processed by the DPU in TOF mode in order to
reduce the bit-rate. The NPD sampling is fixed to 0.5 s. The direction and TOF of the correlated
signals is extracted from the RAW data 32-bit words by DPU. The TOF spectrum resolution is
downgraded to 8 ns, resulting in 256 linearly divided 8-bit TOF bins for every direction. Each
of 256 × 3 16-bit counters is increased by one when addressed by 8-bit TOF and 2-bit direction.
After accumulation, 16-bit counter values are log-compressed to 8 bits (Appendix A.2).
The PH mode. There is also an engineering data flow. The data buffer in PH mode contains
16 × 3 16-bit counters addressed by pulse height values.
Each of the NPD sensors (NPD1 and/or NPD2) can be set independently to any operation
mode or be disabled (masked out). Data is not received by the DPU from the masked-out
sensor. The NPD sampling time is commandable and equals 2n × 62.5 ms, where n=1, 2, . . . 7;
but nominally is 1 sec. The basic sampling periods are 0.5, 1.0, 2.0 sec (8, 16, 32 sampling
periods of 62.5 ms), which correspond to three scanner rotation periods, 32, 64, and 128 s.
The NPD electronics also include linear counters to monitor the performance. Every start
and stop event is counted by a corresponding 16-bit counter. There are four counters available
in every mode to count start events (Start counter) and stop events (Stop0, Stop1, Stop2 coun-
ters). In RAW mode two additional counters are used. The Tof counter is used to count valid
events. Again, a ’valid event’ is produced when a signal from one (and only one) Stop detector
corresponds to one start event within the given TOF window. The Raw counter is used to count
ADC conversion events.
4.6 Instrument level qualification

The NPD sensors were flight qualified on the instrument level as well as on the sensor level
(vibration test). The environment tests on the instrument level included a mechanical test,
electromagetic compatibility (EMC) test, and thermal vacuum (TVAC) test. The protoflight
approach was used, meaning the hardware was tested to the qualification levels. The quali-
fication level is usually 1.4 times the highest load the component should ever experience in
operation (for mechanical test).
Mechanical test. Mechanical test is called sometimes a vibration test. It was done in order
to simulate vibration caused by the rocket engine during the launch. Vibration was simulated
in three transverse directions. The standard testing procedure included: sine vibrations over a
specified frequency range to search for resonances, and random vibrations simulating the real
mechanical environment during the launch.
Figure 4.23 shows ASPERA-4 mounted on a vibration table, with accelerometers depicted
by arrows. Mechanical tests of ASPERA-4 were performed on an LDS964LS electro-dynamic
vibrator with LDS 1200 × 1200 sliding table. An electrical functional test of the instrument
was performed every time a given vibration test was over.
Figure 4.23: ASPERA-4 mounted on a vibration table. Arrows depict accelerometers. GN2 purging
tubes are connected between the test sessions.
4.6. INSTRUMENT LEVEL QUALIFICATION 59
EMC test. The purpose of the EMC test was to verify the electromagnetic susceptibility and
emissive properties of the ASPERA-3/ASPERA-4 instruments. It was done in order to prove
an electromagnetic compatibility of devices with the spacecraft systems as well as other sen-
sors on board the spacecraft. The testing sequence included measurements of electromagnetic
radiation of the instrument and identification of whether it is resistant to external electromag-
netic radiation in a certain frequency and magnitude range. During the EMC test the instrument
was running in different modes defined by the given EMC test.
TVAC test. The purpose of the TVAC test was to verify the capability of the ASPERA-3 /
ASPERA-4 instruments to withstand the thermal load to which the units were to be exposed
during their operational lifetime. This test mostly affects electrical components and movable
parts. The ambient temperature varied cyclically over the defined temperature interval. Fig-
ure 4.24 show the temperature cyclogram (panel a) and measured temperature profiles (panel
b). The instrument functionality was verified in vacuum over the required temperature range
(-30◦ to +50◦ centigrade) for a temperature variation of 1◦ /min (Figure 4.25).
a)
b)
Figure 4.24: TVAC test of ASPERA-4. The measured (a) and predefined (b) temperature profiles are
shown.
Figure 4.25: The ASPERA-4 instrument mounted in a vacuum chamber during the TVAC test. The
small ion guns to simulate particle flux are also visible.
4.7 NPD response to high energy particles

The NPD deflection system cut-off energy is ∼70/q keV, hence particles of higher energy
(>100 keV) can pass through the deflector and produce scattered particles of lower energy
that could be detected by the NPD. We performed simulations of the scattering of high energy
protons on a mono-crystal tungsten (W) surface covered with a thin layer of WO2 (simulating
the Start surface). The simulations were based on the ’The Stopping and Range of Ions in
Matter’ (SRIM) code (Ziegler et al. (1985); www.srim.org).
The NPD geometry was used in the simulation process. As many as 16730 projectile
particles of energy of 100 keV were injected at a grazing angle of 15◦ to the surface plane.
Scattered particles were collected and both angular- and energy distributions were analyzed.
Figure 4.26 shows the energy distribution of the particles for all scattering angles.
Figure 4.27 shows the energy distribution of particles which can reach the Stop surface.
Note that particles scattered approximately specularly (i.e., towards the Stop surface) lose
more than half of their initial 100 keV energy. A number of particles in the energy range
of 0.1 – 10 keV (the NPD energy band) that reached the Stop surface was ∼170. Hence, the
amount of high energy protons that can be detected by the NPD sensor is 61% of the incoming
flux.
If the typical differential flux of 100 keV H + ions is 103 cm−2 sr−1 s−1 keV−1 (such particles
can be born at a distant shock, travel along magnetic field lines and reach the sensor), then we
estimate that the detected integral flux number will be less than 10 – 102 cm−2 sr−1 s−1 . This
flux is several orders of magnitude lower than the expected ENA flux in the energy range of
0.1 – 10 keV. The ENA flux in the Martian environment is in the range of 104 −105 cm−2 sr−1 s−1
as reported by Futaana et al. (2006b); Galli et al. (2006b) (see Chapter 6). Moreover, the sim-
ulated energy spectrum of high energy H + differs from that of the detected ENA signals (Galli
et al., 2006a). Therefore, we conclude that high energy H + can be disregarded as a source of
intense ENA signal at Mars.
4.7. NPD RESPONSE TO HIGH ENERGY PARTICLES 61
Figure 4.26: Simulated energy distribution of H particles with an inital energy of 100 keV, scattered on
the Start surface in all directions.
Figure 4.27: Simulated energy distribution of 100 keV H, scattered on the Start surface, which can
reach the Stop surface.
This page is left in-

tentionally blank
Chapter 5
The NPD calibrations
5.1 Introduction
The calibration of NPD is performed in order to define the detailed characteristics of the sen-
sors, such as the angular response function, efficiency, geometric factor of each of the sensor’s
sectors. The NPD sensors were calibrated using ion beam facilities. The ion beam is easily
to operate and its absolute intensity is easier to monitor than an ENA beam. ENAs or ions
interacting with the Start surface, lose their initial charge state within a few Å of the material
and achieve a charge equilibrium which does not depend on the initial charge. Therefore, an
ENA instrument calibration can be performed using an incoming ion beam. The results of the
sensor calibrations using the ion beam can be directly applied to ENAs (except deflector system
calibrations, which in any case were out of scope of the calibrations).
5.1.1 Calibration facilities

The NPD sensors were calibrated at the Swedish Institute of Space Physics (IRF, Kiruna) using
its ion beam facilities. The calibration facility includes two ion sources with energies of 0.4 – 25
keV and 0.02 – 50 keV that provide large diameter (6 10 cm) parallel beams of selected energy
and mass and with variable intensity. The systems are also equipped with turn-tables having
four degrees of freedom. The system used for the NPD particle calibration is the PSX-2751 50
keV Ion Source from Peabody Scientific. It is capable of producing a parallel ion beam of a
flat profile with an energy of up to 50 keV/q. The system can be run in three different energy
ranges, namely 0.05 – 1.3 keV/q, 1.0 – 15 keV/q, and 15 – 50 keV/q. Ions are produced in a
duoplasmatron in discharge between an anode and a filament, coated with Ba2 CO4 to increase
electron emission. The ion species are defined through injection of the corresponding gas into
the duaplasmatron. To increase the ionization efficiency, a strong magnetic field is applied in
the axial direction of the duoplasmatron. Emitted electrons spiral along the magnetic field lines
towards the anode and produce long ionization paths. The duoplasmatron is kept under positive
HV potential defining the ion beam energy. Ions are extracted from the duoplasmatron / source
through an extraction electrode and pass through a three-element Einzel focusing lens, which
collimates the diverging ion beam into a parallel beam. The focal length of the lens can be
optimized by adjusting the voltage of the lens. The ions then enter an E × B Wien filter. The
Wien filter is based on transverse magnetic and electric fields. The electric field can be varied to
balance the magnetic force so that only ions of the desired velocity (mass) can pass through the
filter. Exiting the Wien filter, the ions are accelerated in an acceleration gap to achieve the final
63
64 CHAPTER 5. THE NPD CALIBRATIONS
beam energy. The beam is then guided through a 90◦ electrostatic analyzer, with a central radius
of 12.7 cm, in order to select ions within a narrow energy band and reject a neutralized fraction
of the beam. The electrostatic analyzer in the system has an energy resolution of 1%. To permit
broadening of the ion beam that emerges from the electrostatic analyzer (∼10 mm in diameter),
a defocus lens first diverges the beam and then a focus lens collects the diverging beam into a
parallel beam used for calibrations (Figure 5.1). Both lenses are three-element Einzel lenses of
different sizes. A retractable Faraday cup provides measurements of the ion beam intensity as
well as, indirectly, the beam location. It is mounted inside the vacuum tank. The Faraday cup
closes the beam outlet completely during the current measurements. A fraction of the outgoing
beam is neutralized by charge exchange between the ions and residual gas in the drift tube. The
neutral part of the beam does not exceed 3% of the total beam intensity.
The maximally achievable beam divergence is derived from the geometry of the ion source
(see Figure 5.1). The distance between the electrostatic analyzer exit and the aperture is ∼1.7 m,
the electrostatic analyzer exit diameter ∼10 mm, the aperture diameter 36 mm, and the distance
between the exit aperture and the turn-table ∼1.8 m, the maximal divergence of the ion beam
is thus estimated to be less than 2◦ . The ion beam spot at the center of the vacuum chamber,
where the instrument is located, can thus vary between 10 - 60 mm in diameter.
The system produces quite stable and homogeneous over profile (±10%) ion beam. During
the calibrations no direct beam profiling was performed.
5.1.2 Calibration setup

Each of the NPD units was calibrated on a sensor level. The calibration setup and coordinate
system are shown in Figure 5.1. The sensor, mounted on the turn-table inside the vacuum
chamber, was subjected to parallel ion beams of different energies and masses. The turn-table
allows the sensor to sweep in the azimuthal plane over [−45◦ , 45◦ ] and in the elevation plane
over [−7.5◦ , 7.5◦ ], with 1◦ steps. The turn-table can also be translated across the incident ion
beam in both vertical and horizontal directions.
Angles α, β denote the azimuth angle and elevation angle, respectively. The angular posi-
tion, α = α0 and β=β0 , shown in Figure 5.1, corresponds to the sector Dir1 pointing. In such
a way the deflector plates are placed horizontally, i.e., normally to the turn-table rotation axis.
Definition of the NPD sectors (Dir0, Dir1, and Dir2) pointing is shown in Figure 4.5. The
geometrical pin-hole area of the sensor is AG = 3.0 × 4.5 × cos 15◦ = 13.04 mm2 , where 15◦ is
the inclination of the pin-hole with respect to the entering beam (see Figure 5.2).
Figure 5.2 shows the cross-section of the sensor through the pin-hole and sector Dir1. All
key elements are identified. A thin black line shows an incoming beam trajectory through the
collimator, the pin-hole, towards the Start surface and then to the Stop surface. The azimuthal
rotation axis of the sensor is co-aligned with the rotation axis of the turn-table and passes
through the pin-hole geometrical center. It is normal to the deflector plane of NPD. In such a
configuration the NPD pin-hole is kept well inside the ion beam during the azimuthal scanning,
i.e., no translation of the sensor across the ion beam is required. The vertical shift of the sensor
aperture during elevation scanning is also small with regard to the ion beam spot size.
The sensor calibrations were performed with the following setup:
- Deflector electrodes were grounded because the sensor was calibrated using ions,
- Ion beam intensity was set to keep the Start MCP count rate 6 5 × 103 s−1 ,
- Pressure in the vacuum chamber was 6 4 × 10−7 mbar.
The calibration objectives together with the ion beam species and energy settings are given in
5.1. INTRODUCTION 65
elevation
rotation axis turn−table Faradey cup
focus lens defocus lens

− 45 deg
azimuth = 0
horizontal
translation NPD + 45 deg
1.7 m
3.5 m
pin−hole position
azimuthal aperture 36 mm
φ
rotation axis
− 7.5 deg
elevation = 0
ion gun
vertical + 7.5 deg electrostatic
translation analyzer exit
pin−hole
position
NPD calibration
mechanical setup
Figure 5.1: NPD calibration mechanical set up and coordinate system; view from the top (the upper
part) and from aside (the lower part).
Pinhole Azimuth rotation axis Stop MCP
e
ENAs
o
15
e
Deflector electrodes ENA trajectory

+/− 5 kV Start surface Stop surface
Figure 5.2: The instrument cross section, showing the azimuthal rotation axis.
Tables 5.1 and 5.20 respectively for the ASPERA-3 / NPD and ASPERA-4 / NPD.
5.2 Theoretical principles

5.2.1 MCP characterization
Before the calibration of the NPD sensor, MCP characterization was performed. MCP detector
characterization and tests included:
- HV test. MCP detector should withstand the bias voltage at least 10% above the nominal.
- Analysis of MCP count in dependency on the bias voltage and threshold level, and de-
termination of the working point, i.e., the nominal bias voltage at which MCP count
saturates.
- Dark noise determination under the nominal bias.
- Pulse height distribution analysis of MCP output signal in order to identify the sensitivity
threshold level for FEE. The nominal bias is applied.
5.2.2 Beam characterization

The ion beam characterization is performed in order to provide accurate calibrations of the
sensor efficiency. Beam monitoring in the system is provided by a retractable Faraday cup. It
is not possible to monitor the ion beam continuously during the measurements. Therefore, the
ion beam should be stable, and parallel. Its spatial distribution should be homogeneous.
The ion source provides a sufficiently stable beam (Section 5.1.1). The stability was verified
before each calibration run. The average value of the ion beam current measured before and
after the run was obtained and used in calculations.
To avoid the possible errors in the absolute calibrations of the sensor efficiency associated
with the beam intensity variations over the spot, a special transverse scan was implemented
5.2. THEORETICAL PRINCIPLES 67
for the efficiency calibration (see Section 5.3.3). Also errors due to measurements with a non-
parallel ion beam are taken into account by this scan procedure. Defocused ion beams can
otherwise lead to wrong estimation of the ion current density at the Faraday cup and the sensor.
5.2.3 Geometrical Factor calculation

The geometrical factor, (G), of a sensor is the integral of the effective area Ae f f (Ω̄) in the
direction Ω̄ of the sensor’s aperture over the sensor field-of-view
Z
G = Ae f f (Ω̄) dΩ
From simple spherical trigonometry the elementary solid angle dΩ will be given in the α, β
coordinates as dΩ = cos α dα dβ. We thus obtain the geometrical factor expression for NPD as
Z Z
G= Ae f f (α, β) cos α dα dβ, (5.1)
α β
where the integrals are over the sensor field-of-view. We assume that Ae f f can be presented by
the expression
Ae f f (α, β) = ε(E) · f (α, β) · Ao (5.2)
where ε is the efficiency and depends only on energy (see Section 5.2.4), f (α, β) is a function
which defines the shape of the angular response and max f (α, β)|α,β =1, and Ao is the pin-hole
area at the maximum f (α, β), i.e., Ao =A(αo , βo ). αo , βo are the azimuth and elevation at which
f (αo , βo )=1. Therefore,
Z Z
G= ε(E) · Ao · f (α, β) cos α dα dβ
α β
By defining a pure geometrical factor Go (with the efficiency excluded) as

Z Z
G o = Ao f (α, β) cos α dα dβ (5.3)
α β
we obtain the total geometrical factor as
G = ε(E) · Go (5.4)
The efficiency can thus be measured separately from the angular response, which avoids pos-
sible errors due to beam intensity variations over the spot. The calibration principle becomes
simple. We sweep the sensor over the beam over elevation and azimuth and normalize the
recorded count rate to the maximum value. Defining the azimuth and elevation of the maxi-
mum count rate, (αo , βo ), we also define the effective area at these angles. We assume
f (α, β)
Ae f f (α, β) = Ae f f (αo , βo )
f (αo , βo )
= Ae f f (αo , βo ) · f (α, β)
= ε(E) · Ao · f (α, β) (5.5)
C(α, β)
f (α, β) = (5.6)
C(αo , βo )
where C(α, β) is the count rate at the azimuth α and elevation β. G can be calculated finally as
Z Z
G = Ae f f (αo , βo ) f (α, β) cos α dα dβ
α β
Z Z
C(α, β)
= ε(E) · Ao cos α dα dβ (5.7)
α β C(αo , βo )
5.2.4 Efficiency
The sensor efficiency ε depends on a number of factors such as particle mass and energy, MCP
bias voltage, surface properties and conditions, etc. MCP characterization was performed and
the working point was defined prior to the tests. All tests were performed at this working point,
i.e., MCP bias was fixed. ε is measured for different incident ion species at different energies.
If the calibrating ion beam has the current density, j, which is constant over the effective
pin-hole area, the following expression can be written
j
C(α, β) = · Ae f f (α, β)
q
j
= · ε · Ao · f (α, β)
q
dI ε
= · · Ao · f (α, β) (5.8)
dA q
where q is the elementary charge, I is the beam current, dA the area element over the beam
dI
spot. Note that j = may vary over the beam spot (not homogenous beam). Rewriting the
dA
expression in the form
C(α, β)
ε · dI = q · · dA
Ao · f (α, β)
q
= · d C(α, β) (5.9)
f (α, β)
where d C(α, β) is the count rate over an elementary area over the beam spot, we can integrate
both sides and get the expression for the efficiency
CT otal (α, β) q
ε= · (5.10)
f (α, β) I
where CT otal (α, β) is the total count rate integrated over the whole beam spot Abeam .
Z
CT otal (α, β) = γ dC(α, β)
Abeam
XX
= γ Ci j (α, β) (5.11)
i j
where i, j - the indexes over coordinates in the plane perpendicular to the beam.
A pixel
γ= (5.12)
Ageom
where A pixel = 4.5 × 3.0 mm – a characteristics of the sweep, Ageom is a projection of the
pin-hole on the plane normal to the beam direction and depends on sensor direction. If the
measurement performed at α = αo and β = βo ,
CT otal (αo , βo )
ε = q· (5.13)
I
This expression does not depend on the area and thus can be applied even for non-homogenuos
beams. In practice, it implies that it is sufficient to cover the beam spot fully with the effective
pin-hole size pixels and record the total number of counts. In principle, it is not even necessary
to make the measurements at the maximum of the angular response function. If the efficiency
εm was measured at the azimuth αm and elevation βm , the sensor efficiency at αo , βo is given by
the obvious relation
εm
ε= (5.14)
f (αm , βm )
5.2.5 Energy resolution

Dependency of the particle time-of-flight T M on the initial particle energy E and mass M can
be expressed as
√
M
TM = L · √
E − ∆Eloss
√
M
= L· √ √
E 1−K
√
M(amu)
= 720 · L(cm) · √
E(eV) − ∆Eloss (eV)
√
M(amu)
= 720 · L(cm) · √ √ (5.15)
E(eV) 1 − K
where L is the TOF distance, ∆Eloss the energy loss when a particle interacts with the Start
surface, T M in ns, K = ∆Eloss /E the relative energy loss.
TOF measurements are performed for hydrogen and oxygen species in a range of particle
energies from 0.1 keV to 10 keV. TOF spectra are shown in Sections 5.4.6, 5.5.6. Using the
position of the maximum of the TOF spectra, we can obtain the TOF dependence on the particle
initial energy and define the energy loss.
5.2.6 Mass resolution

Hydrogen and oxygen atoms are assumed to be the most probable species to be detected in the
Martian and Venusian plasma environments.
Because of the large difference in velocity (for the same energy) and relatively narrow
TOF window (50 ns - 1900 ns), TOF measurements alone can be used for mass identification.
Indeed, TOF below 240 ns corresponds to either hydrogen with an energy above 580 eV or
oxygen above 10 keV. The oxygen ENA fluxes with such high energies are very low because
planetary ions do not have enough room to get accelerated in the small region with sufficiently
high exospheric densities. Therefore, TOFs below 240 ns can come only from hydrogen. TOF
above 650 ns can only come from oxygen because the instrument efficiency is very low for
100 eV hydrogen and such slow atoms cannot be detected by NPD. Inside the TOF range of
240 - 650 ns other mass resolving techniques are required, for example pulse height analysis.
PH analysis
PH analysis is based on the secondary electron yield dependency on the ENA atomic number
for different species with the same velocity. Particles of different mass and the same velocity
yield different amounts of secondary electrons from the Stop surface, which was optimized
for high secondary electron yield. The heavier the ions, the higher the secondary electron
yield. The more secondary electrons reach the Stop MCP simultaneously, the higher the output
signal, because electrons hit different MCP channels. The higher the secondary electron yield,
the higher the magnitude of the analog stop signal produced which is analyzed by a fast 8-bit
ADC. Hydrogen atoms yield fewer secondary electrons from the Stop surface than equally
fast oxygen atoms. Each 8-bit pulse height value is reduced to 4 bit resolution (16 levels or
channels) in the BIN mode.
Figure 5.3 shows pulse height distributions from H (the red curve) of an energy of 0.3 keV
and O (the black curve) of an energy of 5.0 keV (both species possess approximately the same
velocity) measured in the BIN mode. Both distributions are normalized to have the same area
of 1.
O H
H+O
Figure 5.3: Pulse height distributions from H (the red line) and O (the black line) with the same velocity
are shown. Both distributions are normalized to have the same area of 1. The green line shows a pulse
height distribution from a mixture of H and O normalized to have the area equal to 1 as well. The dashed
line is set to the channel N = 9. O gives a larger count in the range of channels 9 - 15.
As can be seen, areas of the tail parts of the pulse height distributions (integrated over the
channels from an arbitrary chosen channel N to 15) are different. H atoms produce signals
of lower amplitude and consequently give shorter tails in the pulse height distribution than
heavier O. Hence, the ratio of the tail to the total area of the distribution is larger for O than
for H. In the case of a mixture of H and O atoms, the pulse height distribution is a sum of the
distribution from H and O. The green line in Figure 5.3 shows a pulse height distribution from
a mixture of 50% H and 50% O. The distribution is normalized to have an area of 1. The tail
part of this distribution comprises partly H and partly O counts. If a measured pulse height
distribution has no counts in the tail part while the count rate is reasonably high (∼ 1 kHz), the
incoming particles are hydrogen. However, if the tail does contain some counts, the incoming
flux may contain a mixture of H and O. Let us show how one can deduce the mass composition
of the incoming flux during in-flight measurements from the pulse height distribution for H and
O obtained in the calibrations. The probability for hydrogen and oxygen to produce counts in
the tail part are, respectively,
15
X
CiH
i=N
PH = (5.16)
15
X
CiH
i=0
15
X
CiO
i=N
PO = (5.17)
15
X
CiO
i=0
where CiH and CiO are the respective amount of counts in the i-th bin of the pulse height distri-
bution for H and O. Numbers PH , PO , and N are defined in the calibrations. The total number
of counts Ntail from each specie in the tail region thus becomes
H
Ntail = PH · NTHotal (5.18)
O
Ntail = PO · NTOotal (5.19)
where NTHotal and NTOotal are the total amount of counts for H and O, respectively. In the case of
measuring a mixture of H and O, we know the number NTH+O otal and the relative amount of counts
in the tail region R,
15
X 15
X
CiH+O (CiH + CiO )
H O
i=N i=N Ntail + Ntail
R = = =
15
X NTH+O
otal NTH+O
otal
CiH+O
i=0
PH · NTHotal + PO · NTOotal
= (5.20)
NTH+O
otal
Keeping in mind NTHotal + NTOotal = NTH+O

otal , we arrive at
R − PH
NTOotal = · N H+O (5.21)
PO − PH T otal
PO − R
NTHotal = · N H+O (5.22)
PO − PH T otal
Channel N is defined in such a way as to minimize the value of PPHO . The values of N, PO , PH
are defined for every NPD sector during calibrations. Note, the method works only for good
statistics. This PH analysis is used in the BIN mode (due to good signal statistics achievable in
this mode) for the on board data reduction (see Appendix A.1).
Example. Let a total count of valid events NTH+O otal =1000 be detected by sector Dir1 of the
sensor during the sampling time ∆t. PH =0.12·10−2 and PO =1.25·10−2 as calculated during the
calibrations. Let the ratio R be calculated as 0.01.
Hence, the O fraction is NTOotal = (0.01−0.12·10−2 ) / (1.25·10−2 −0.12·10−2 ) ·1000 = ∼779.
The counts from H are: NTHotal = 1000 – 779 = 221.
5.3 Measurement principles

Figure 5.4 shows the characterization setup of the MCP detectors.
flange
HV
’electron gun’
_
e
MCA PC
HV
Figure 5.4: MCP characterization was done in the following configuration: MCA - multi-channel ana-
lyzer, PC - computer.
First the Start MCP assembly was inserted into a vacuum chamber, then it was changed
for the Stop MCP assembly. A small electron gun was used to produce an electron beam
to stimulate MCP detectors. MCP output signal was analyzed by means of a multi-channel
analyzer MCA8000. Bias of Start MCP U S tart and of Stop MCP U S top were changed in the
range of 2300 - 2900 V.
5.3.2 Angular response measurements

During the angular response measurements the sensor worked in the BIN mode, i.e., valid
events (1 Start and 1 Stop) were counted. NPD modes are described in Section 4.5. The
angular response was measured for different energies for both H and O species.
The sensor was translated to the center of the ion beam spot and set to a default position,
shown in Figure 5.1. An azimuthal scan was performed in the range of α = [−45◦ , +45◦ ], with
steps ∆α=5◦ . An elevation scan was performed in the range of β = [−7.5◦ , +7.5◦ ], with steps
∆β=1.5◦ for the ASPERA-3 / NPD, and in the range of β = [−6◦ , +6◦ ], with steps ∆β=1.0◦
for the ASPERA-4 / NPD. Accumulation time t was set to 1 s per pixel. The angles αo , βo
5.3. MEASUREMENT PRINCIPLES 73
were calculated afterwards by finding the angular position corresponding to the maximum of
the angular response function.
5.3.3 Efficiency
During the efficiency measurements, the sensor worked in the BIN mode. Absolute efficiency
measurements were performed for fixed angles αm , βm which may be different from αo , βo , for
various energies and species (see Tables 5.1 and 5.20). The instrument response was integrated
over the beam spot, performing a scan of an area larger than the ion beam spot at the NPD
location. No beam intensity variation over a small area of 3.0 × 4.5 mm2 (pin-hole size) is
assumed.
a)
b)
α=0 α=40 α=0 α=40
3.0
β=0 β=0 β=15 β=15

4.5
Figure 5.5: a) Scan sequence of the NPD sensor during efficiency test. Arrows depict direction of shift
of the sensor aperture; b) a pin-hole effective area (shown in grey) for different azimuthal and elevation
angles.
Scan procedure The ion beam spot location was found by scanning along the x- and y-axis.
Then the sensor was translated to the center of the spot. Then a scan sequence was initiated
(shown in Figure 5.5(a)). The very first and last measurements are made in the center of the
scanning matrix, i.e., in the center of the ion beam spot. The pinhole zigzagged as shown in
Figure 5.5(a) over the area A = 60 × 63 mm2 covering the ion beam spot entirely (its size is
estimated to be 660 mm2 in diameter at the position of the NPD). The turn-table translation
steps are chosen to correspond to the pin-hole size as 3.0 mm and 4.5 mm in vertical and
horizontal planes, respectively. The integrated count rate CT otal (αm , βm ) is calculated according
to equation 5.11. The pin-hole effective area A(α, β) dependency on the azimuth and elevation
angles is shown schematically in Figure 5.5(b).

Measurements are made in the RAW mode. TOF spectra are accumulated during the efficiency
calibration scans. The long exposure time allows us to obtain spectra with good statistics.

Pulse height spectra accumulated during the efficiency measurements are used. The NPD elec-
tronics worked in the BIN mode (Section 4.5) and produced both TOF and pulse height data.
A long exposure time allows us to obtain spectra with good statistics.
5.4 ASPERA-3 / NPD calibration results

5.4.1 Calibration objectives
The ASPERA-3 / NPD calibration objectives are given in Table 5.1.
N Calibrated Ion beam Comment

parameter/ function Energy, keV Species
1 MCP characterization 5.0 O Determine nominal bias of the Start /
Stop MCP detectors.
2 Efficiency measurements 5.0, 3.0, 1.3, H, O Energy / mass dependent efficiency ε for
0.7, 0.5, 0.3 sectors Dir0/Dir1/Dir2
3 Angular response 0.3, 1.3 H, O Azimuthal and elevation resolution
The geometrical constant Go estimation
4 Geometrical factor G 5.0, 3.0, 1.3, H, O Derived from:
0.7, 0.5, 0.3 1. the angular response,
2. the efficiency measurements
5 Energy and mass resolu- 5.0, 3.0, 1.3, H, O TOF and PH distribution functions
tion 0.7, 0.5, 0.3
6 Heater and temperature Start surface heating
sensor characterization
7 EMC related disturbances Functional test on the ASPERA-3 level
Table 5.1: The ASPERA-3 / NPD calibration objectives.

The MCP assemblies were tested up to 3100 V bias. The working point, i.e., a nominal MCP
bias, was defined. Figure 5.6 shows the Start, Stop0, Stop1, Stop2 MCP detectors count versus
MCP bias for the NPD1 (left panel) and NPD2 (right panel) sensors.
The working point of NPD1 is defined as 2750 V for both Start and Stop MCP detectors.
Calibration of NPD1 was performed at this MCP bias.
Though the saturation was not reached, the nominal bias was chosen at the point where the
curve derivative decreased. The lower bias was chosen to reduce a risk of discharge.
5.4.3 Efficiency measurements

Accumulation time t was set to 1.0 s. Efficiency measurements are performed for azimuthal
αm , and elevation βm angles. Tables 5.2 and 5.3 shows the angles αm and βm , the pin-hole
effective area, and the values of the angular response function at these angles for the NPD1 and
NPD2 sensors, respectively.
5.4. ASPERA-3 / NPD CALIBRATION RESULTS 75
working
point
Figure 5.6: The Start, Stop0, Stop1, Stop2 MCP detector count versus a MCP bias for the NPD1 (left
panel) and NPD2 (right panel) sensors. The dashed lines depict the working points.
Sensor Sector αm , deg βm , deg A(αm , βm ), mm2 f (αm , βm )

0.3 keV H 0.3 keV O
Dir0 40.0 0.0 10.0 0.43 0.69
NPD1 Dir1 0.0 0.0 13.0 0.77 0.73
Dir2 −40.0 0.0 10.0 0.39 0.72
Table 5.2: Azimuth and elevation angles αm , βm at which the efficiency was measured. A(αm , βm )
the pin-hole effective area at these angles. f (αm , βm ) is a value of the function defining the shape of
the angular response. It is calculated for the angles αm , βm from the angular response measurements
for 0.3 keV H and 0.3 keV O for NPD1 (Section 5.4.4).
Efficiency εm was calculated according to Equations 5.11 and 5.13:

CT otal (αm , βm )
εm = q ·
I
q XX
= ·γ· Ci j (αm , βm )
I i j
where i, j - the indexes over coordinates in the plane perpendicular to the beam, γ is defined
by Equation 5.12, Ageom = A(αm , βm ).
Then the efficiency was recomputed for angles αo , βo by Equation 5.14. Tables 5.4 and 5.5
show the efficiency ε (αo , βo ) of NPD1 and NPD2, respectively.
Figures 5.7 and 5.8 show the NPD sensors’ efficiency profiles together with the best fittings
by the function
ln (y) = a0 + a1 · x + a2 · ln (x) (5.23)
where x = ln (E), y = 104 · ε, and E an incident ion energy in units of eV.
The plots presented are valid for the energy range of 0.3 - 5.0 keV for both H and O.
Tables 5.6 and 5.7 show the coefficients for the efficiency fittings for NPD1 and NPD2, respec-
tively.

1.3 keV H
Dir0 40.0 0.0 10.0 0.67
NPD2 Dir1 0.0 0.0 13.0 0.84
Dir2 −40.0 0.0 10.0 0.60
Table 5.3: The format is the same as in Table 5.2. f (αm , βm ) is calculated from the angular response for
1.3 keV H of NPD2 (Section 5.4.4).
ASPERA-3 / NPD1
Efficiency ε (αo , βo ), ×10−2
E, keV 5.0 3.0 1.3 0.7 0.5 0.3
Dir0 − 8.86 3.17 0.67 − 0.02
O Dir1 − 7.89 2.12 0.73 − 0.03
Dir2 − 8.57 3.63 0.79 − 0.03
Dir0 − 9.07 7.67 2.70 − 0.44
H Dir1 − 5.26 3.69 1.70 − 0.40
Dir2 − 10.10 9.54 3.44 − 0.59
Table 5.4: Absolute efficiency of the ASPERA-3 / NPD1 versus ion mass and energy.
5.4.4 Angular response

Figures 5.9 and 5.11 shows the angular response of the NPD1 sensor for 0.3 keV H and
0.3 keV O, respectively. Each figure shows the angular response of the sectors Dir2, Dir1,
Dir0 (from left to right). The upper row is the measured azimuth-elevation response. The sec-
ond row from the top shows the fitting of the measured and smoothed data, normalized to have
the peak magnitude equal to 1. The contours show the polynomial fits based on the function
X
f (x, y) = k ji · xi · y j (5.24)
where x and y are azimuth and elevation angles scaled as x = α/5 + 10, y = β + 6. k ji are
coefficients.
Functions are valid in the range of α [-45, 45], and β [-4, 6] degrees. The third row shows the
azimuthal responses at β=βo and the bottom row shows the elevation responses at α=αo . Dia-
monds show the measured and smoothed data. The fittings are shown by thick lines. FWHM
ASPERA-3 / NPD2
E, keV 5.0 3.0 1.3 0.7 0.5 0.3
Dir0 13.25 8.27 2.75 0.19 0.19 0.06
O Dir1 9.10 6.75 2.42 0.64 0.29 0.02
Dir2 17.80 10.23 3.83 0.37 0.15 0.05
Dir0 7.00 6.78 4.06 2.36 1.25 0.36
H Dir1 4.88 4.58 2.92 1.69 0.98 0.35
Dir2 9.48 9.48 5.68 3.10 1.63 0.57
ASPERA-3 / NPD1
H O
a0 a1 a2 a0 a1 a2
Dir0 -38.96 -4.31 38.71 -76.60 -7.71 69.81
Dir1 -30.47 -3.37 30.67 -66.16 -6.48 59.96
Dir2 -36.62 -4.10 36.85 -73.10 -7.43 67.09
Table 5.6: Coefficients for the NPD1 efficiency fitting (Equation 5.23).
ASPERA-3 / NPD2
H O
a0 a1 a2 a0 a1 a2
Dir0 -33.57 -3.76 33.77 -57.66 -5.11 50.56
Dir1 -26.62 -2.87 26.80 -65.54 -6.62 60.13
Dir2 -32.24 -3.65 32.87 -72.71 -7.04 65.39
is given for each profile. The values of k ji for the fit in Figure 5.9 are given in Table 5.8. The
values of k ji for the fit in Figure 5.11 are given in Table 5.9.
Figures 5.10 and 5.12 show the azimuthal profiles for respectively 0.3 keV H and 0.3 keV
O of all three sectors of the NPD1 sensor plotted altogether. The dashed lines show the total
angular coverage of the NPD1 sensor in the azimuthal plane. All responses are normalized to
have the peak maximum equal to 1.
ASPERA−3 NPD1
Figure 5.7: The NPD1 absolute efficiency ε(αo , βo ) as a function of the ion energy and mass for different
sectors. Least-square fittings are shown by lines.
ASPERA−3 NPD2
Figure 5.8: The NPD2 absolute efficiency ε(αo , βo ) as a function of the ion energy and mass for different
sectors. Least-square fittings are shown by lines.
ASPERA−3 NPD1
Dir 2 Dir 1 Dir 0
α o , βo
α m , βm
Figure 5.9: Angular response of the NPD1 sensor for 0.3 keV H given for the sectors Dir2, Dir1, Dir0
(from left to right). The upper row shows the measured azimuth-elevation response. The second row
from the top shows the fitting of the measured and smoothed data. The third and fourth rows show
the cross-sections of the angular response function at βo and αo , respectively. FWHM is given for
each profile. Two red dots depict the angular positions corresponding to the position αm , βm at which
the efficiency measurements have been performed, and to the position αo , βo of the maximum of the
azimuth-elevation response function.
k ji i=0 i=1 i=2 i=3 i=4 i=5 i=6

j=0 1.55e-03 -2.62e-02 3.02e-02 -1.22e-02 2.27e-03 -1.99e-04 6.59e-06
j=1 -6.42e-03 1.11e-01 -1.26e-01 4.90e-02 -8.74e-03 7.30e-04 -2.31e-05
j=2 3.08e-03 -4.07e-02 4.12e-02 -1.44e-02 2.60e-03 -2.33e-04 8.07e-06
j=3 -5.80e-04 5.89e-03 -5.11e-03 1.46e-03 -2.58e-04 2.58e-05 -1.00e-06
j=4 5.09e-05 -3.68e-04 2.21e-04 -1.98e-05 2.64e-06 -6.68e-07 4.17e-08
j=5 -2.14e-06 1.06e-05 -1.80e-06 -2.86e-06 6.01e-07 -3.18e-08 9.48e-11
j=6 3.50e-08 -1.28e-07 -4.37e-08 8.50e-08 -1.79e-08 1.26e-09 -2.69e-11
j=0 7.91e-04 -1.80e-02 2.35e-02 -1.05e-02 1.99e-03 -1.72e-04 5.54e-06
j=1 -3.97e-03 1.10e-01 -1.46e-01 6.41e-02 -1.17e-02 9.56e-04 -2.91e-05
j=2 1.75e-03 -4.86e-02 6.61e-02 -3.02e-02 5.89e-03 -5.20e-04 1.71e-05
j=3 -3.75e-04 1.12e-02 -1.54e-02 7.09e-03 -1.37e-03 1.19e-04 -3.87e-06
j=4 3.71e-05 -1.15e-03 1.59e-03 -7.29e-04 1.40e-04 -1.20e-05 3.86e-07
j=5 -1.68e-06 5.24e-05 -7.24e-05 3.32e-05 -6.35e-06 5.43e-07 -1.73e-08
j=6 2.83e-08 -8.81e-07 1.22e-06 -5.56e-07 1.06e-07 -9.08e-09 2.89e-10
j=0 3.60e-02 -6.98e-02 5.14e-02 -1.87e-02 3.40e-03 -2.94e-04 9.58e-06
j=1 -4.28e-02 2.21e-01 -2.56e-01 1.11e-01 -2.06e-02 1.70e-03 -5.10e-05
j=2 1.42e-02 -4.03e-02 4.13e-02 -2.04e-02 4.86e-03 -5.06e-04 1.87e-05
j=3 -2.16e-03 1.89e-03 -6.06e-04 8.42e-04 -4.11e-04 5.90e-05 -2.59e-06
j=4 1.67e-04 1.01e-04 -2.97e-04 8.07e-05 7.43e-06 -2.84e-06 1.58e-07
j=5 -6.35e-06 -1.12e-05 2.18e-05 -7.82e-06 5.87e-07 3.83e-08 -3.95e-09
j=6 9.45e-08 2.53e-07 -4.54e-07 1.77e-07 -2.05e-08 4.44e-10 2.69e-11
Table 5.8: k ji coefficients for the polynomial fit (Equation 5.24) for the angular response of NPD1 for
0.3 keV H.
Figure 5.10: The azimuthal responses of each sector of NPD1 measured for 0.3 keV H. The dashed
line shows the total angular coverage of the NPD1 sensor in the azimuthal plane. All responses are
normalized to have the peak maximum equal to 1.
ASPERA−3 NPD1
Dir 2 Dir 1 Dir 0
Figure 5.11: Angular response of the NPD1 sensor for 0.3 keV O given for the sectors Dir2, Dir1, Dir0
(from left to right). The format is the same as in Figure 5.9.
k ji i=0 i=1 i=2 i=3 i=4 i=5 i=6

j=0 2.32e-03 -3.85e-02 4.05e-02 -1.38e-02 2.11e-03 -1.51e-04 4.06e-06
j=1 -6.13e-03 6.23e-02 -5.17e-02 1.22e-02 -7.30e-04 -4.74e-05 4.41e-06
j=2 3.14e-03 -3.25e-02 2.87e-02 -8.23e-03 1.01e-03 -5.50e-05 1.10e-06
j=3 -6.27e-04 7.29e-03 -7.25e-03 2.54e-03 -4.19e-04 3.36e-05 -1.07e-06
j=4 5.83e-05 -7.38e-04 7.84e-04 -3.00e-04 5.48e-05 -4.81e-06 1.63e-07
j=5 -2.56e-06 3.56e-05 -4.00e-05 1.63e-05 -3.14e-06 2.86e-07 -9.87e-09
j=6 4.27e-08 -6.62e-07 7.84e-07 -3.36e-07 6.65e-08 -6.12e-09 2.12e-10
j=0 3.13e-04 1.75e-02 -2.84e-02 1.39e-02 -2.85e-03 2.58e-04 -8.49e-06
j=1 3.34e-04 -1.24e-01 1.85e-01 -8.65e-02 1.72e-02 -1.51e-03 4.87e-05
j=2 -6.05e-04 8.94e-02 -1.32e-01 6.10e-02 -1.19e-02 1.04e-03 -3.28e-05
j=3 6.65e-05 -1.67e-02 2.51e-02 -1.18e-02 2.39e-03 -2.13e-04 6.95e-06
j=4 9.86e-07 1.33e-03 -2.05e-03 9.92e-04 -2.06e-04 1.89e-05 -6.30e-07
j=5 -3.37e-07 -4.79e-05 7.65e-05 -3.80e-05 8.10e-06 -7.59e-07 2.58e-08
j=6 9.74e-09 6.38e-07 -1.07e-06 5.48e-07 -1.20e-07 1.14e-08 -3.93e-10
j=0 -1.14e-02 3.52e-02 -3.49e-02 1.41e-02 -2.76e-03 2.56e-04 -8.92e-06
j=1 5.55e-02 2.21e-01 -4.88e-01 2.79e-01 -6.05e-02 5.62e-03 -1.89e-04
j=2 -3.73e-02 -2.60e-02 1.35e-01 -9.23e-02 2.25e-02 -2.27e-03 8.19e-05
j=3 8.67e-03 -3.18e-03 -1.61e-02 1.35e-02 -3.61e-03 3.87e-04 -1.45e-05
j=4 -8.95e-04 6.67e-04 1.03e-03 -1.04e-03 2.94e-04 -3.26e-05 1.25e-06
j=5 4.20e-05 -3.70e-05 -3.54e-05 4.07e-05 -1.19e-05 1.33e-06 -5.15e-08
j=6 -7.31e-07 6.69e-07 5.16e-07 -6.33e-07 1.86e-07 -2.10e-08 8.16e-10
Table 5.9: k ji coefficients for the polynomial fit (Equation 5.24) for the angular response of NPD1 for
0.3 keV O.
Figure 5.12: The azimuthal responses of each sector of NPD1 measured for 0.3 keV O. The dashed
ASPERA−3 NPD2
Dir 2 Dir 1 Dir 0
α o , βo
α m , βm
Figure 5.13 shows the angular response of the NPD2 sensor for 1.3 keV H. The format
is the same as in Figure 5.9. The values of coefficients k ji for the fit in Figure 5.13 are in
Table 5.10.
Figure 5.14 shows the azimuthal profiles for 1.3 keV H of all three sectors of the NPD2
sensor plotted altogether. The dashed line shows the total angular coverage of the NPD2 sensor
in the azimuthal plane. All responses are normalized to have the peak maximum equal to 1.
k ji i=0 i=1 i=2 i=3 i=4 i=5 i=6

j=0 -3.06e-03 6.42e-02 -1.06e-02 -7.82e-03 2.70e-03 -2.98e-04 1.11e-05
j=1 3.28e-04 5.36e-02 -9.67e-02 4.60e-02 -9.01e-03 7.86e-04 -2.53e-05
j=2 3.56e-04 -2.15e-02 3.04e-02 -1.23e-02 2.24e-03 -1.93e-04 6.31e-06
j=3 -1.79e-04 4.70e-03 -5.00e-03 1.70e-03 -2.88e-04 2.45e-05 -8.25e-07
j=4 2.84e-05 -4.93e-04 4.33e-04 -1.27e-04 2.01e-05 -1.71e-06 5.99e-08
j=5 -1.77e-06 2.53e-05 -2.09e-05 5.95e-06 -9.09e-07 7.56e-08 -2.60e-09
j=6 3.80e-08 -5.03e-07 4.28e-07 -1.29e-07 2.00e-08 -1.60e-09 5.18e-11
j=0 6.58e-03 -3.82e-02 4.64e-02 -2.02e-02 3.90e-03 -3.44e-04 1.14e-05
j=1 -3.53e-02 8.96e-02 -9.31e-02 3.02e-02 -3.41e-03 1.04e-04 1.94e-06
j=2 2.16e-02 -1.63e-02 3.04e-03 8.13e-03 -3.27e-03 4.06e-04 -1.62e-05
j=3 -4.50e-03 3.00e-03 -7.68e-05 -2.03e-03 7.75e-04 -9.53e-05 3.80e-06
j=4 4.21e-04 -3.33e-04 8.40e-05 1.54e-04 -6.60e-05 8.40e-06 -3.40e-07
j=5 -1.83e-05 1.72e-05 -7.70e-06 -4.64e-06 2.43e-06 -3.24e-07 1.33e-08
j=6 3.02e-07 -3.30e-07 1.96e-07 3.99e-08 -3.17e-08 4.51e-09 -1.90e-10
j=0 3.04e-04 -3.35e-02 4.69e-02 -2.25e-02 4.48e-03 -3.94e-04 1.27e-05
j=1 -5.25e-03 2.51e-01 -3.45e-01 1.59e-01 -2.97e-02 2.44e-03 -7.40e-05
j=2 2.96e-03 -4.39e-02 5.07e-02 -2.24e-02 4.43e-03 -4.06e-04 1.40e-05
j=3 -4.03e-04 2.25e-03 -4.33e-04 -2.76e-05 -6.96e-05 1.74e-05 -9.95e-07
j=4 1.67e-05 7.88e-07 -2.65e-04 1.43e-04 -1.88e-05 3.30e-07 3.85e-08
j=5 1.05e-07 -2.36e-06 1.46e-05 -7.64e-06 1.06e-06 -3.19e-08 -1.24e-09
j=6 -1.31e-08 3.05e-08 -2.23e-07 1.18e-07 -1.58e-08 3.70e-10 2.57e-11
k
Table 5.10: ji coefficients for the polynomial fit (Equation 5.24) for the angular response of NPD2
for 1.3 keV H.
H O
Sensor Sector αo , deg βo , deg Ao , mm2 αo , deg βo , deg Ao , mm2
Dir0 25.0 1.9 11.80 27.0 0.9 11.66
NPD1 Dir1 0.0 2.0 13.03 −1.0 2.1 13.03
Dir2 −22.0 2.0 12.10 −29.0 1.3 11.46
Table 5.11: Angles αo , βo for different sectors, and corresponding effective aperture area for these
angles are given. αo , βo are calculated from the angular response for 0.3 keV H and 0.3 keV O of the
NPD1 sensor.
H
Sensor Sector αo , deg βo , deg Ao , mm2
Dir0 30.0 1.5 11.25
NPD2 Dir1 3.0 1.7 13.02
Dir2 −27.0 1.3 11.61
angles are given. αo , βo are calculated from the angular response for 1.3 keV H of the NPD2 sensor.
Table 5.11 shows the angles αo , βo which correspond to the maximum of the angular re-
sponse function of the NPD1 sensor for 0.3 keV H and 0.3 keV O. The effective aperture area
at these angles is also shown.
sponse function of the NPD2 sensor for 1.3 keV H. The effective aperture area at these angles
is also shown.
Tables 5.13 and 5.14 show the pure geometrical constant Go of the NPD1 and NPD2 sen-
sors, respectively, calculated according to Equation 5.3.
Go , ×10−3 , cm2 sr
H O
Sensor Dir0 Dir1 Dir2 Dir0 Dir1 Dir2
NPD1 8.76 11.45 9.34 8.94 10.45 7.13
Table 5.13: Pure geometrical factor Go computed from the angular response for 0.3 keV H and 0.3 keV
O of the NPD1 sensor. Efficiency ε is not included.
Go , ×10−3 , cm2 sr
H
Sensor Dir0 Dir1 Dir2
NPD2 8.98 12.77 9.06
Table 5.14: Pure geometrical factor Go computed from the angular response for 1.3 keV H of the NPD2
sensor. Efficiency ε is not included.
5.4.5 Geometrical factor

The total geometrical factors G = Go · ε(E) (in units of cm2 sr) of the NPD1 and NPD2 sensors
are shown in Tables 5.15 and 5.16, respectively.
NPD1 / ASPERA-3
Geometrical factor G, ×10−4
E, keV 5.0 3.0 1.3 0.7 0.5 0.3
Dir0 − 9.05 2.58 0.56 − 0.02
O Dir1 − 8.29 2.51 0.62 − 0.04
Dir2 − 7.35 2.25 0.53 − 0.03
Dir0 – 10.43 5.34 2.33 – 0.42
H Dir1 – 6.76 3.83 1.93 − 0.48
Dir2 – 13.03 6.87 3.11 − 0.61
Table 5.15: The absolute geometrical factor of the ASPERA-3 / NPD1 sensor, computed from the fit
function of the efficiency of NPD1 and pure geometrical factors (Table 5.13).
NPD2 / ASPERA-3
Geometrical factor fit G, ×10−4
E, keV 5.0 3.0 1.3 0.7 0.5 0.3
Dir0 11.46 6.82 1.85 0.45 0.18 0.03
O Dir1 12.64 9.01 3.00 0.79 0.31 0.05
Dir2 15.11 9.64 2.55 0.54 0.19 0.03
Dir0 7.65 6.46 3.61 1.75 1.05 0.39
H Dir1 7.54 6.21 3.55 1.86 1.19 0.52
Dir2 10.36 8.77 4.96 2.45 1.48 0.57
function of the efficiency of NPD2 and pure geometrical factors (Table 5.14).

Figures 5.15, 5.16 show TOF spectra obtained by the NPD1 sensor for H and O ion beams,
respectively, for different energies. Figures 5.17, 5.18 show TOF spectra obtained by the NPD2
sensor for H and O ion beams, respectively, for different energies.
Figure 5.15: The NPD1 TOF spectra for H beams of different energies.
Figure 5.16: The NPD1 TOF spectra for O beams of different energies.
Figure 5.17: The NPD2 TOF spectra for H beams of different energies. The reason for the structure in
the 0.3 keV TOF spectrum (at ∼300 ns) is as yet unclear. It can perhaps be due to electronic disturbances.
Figures 5.19 and 5.20 show the TOF spectrum maximum as a function of the incident ion
beam energy for the NPD1 and NPD2 sensors, respectively. The error bars show FWHM of the
TOF spectra. The dashed lines show ideal theoretical dependencies T M = f (E) (Equation 5.15)
with no energy loss, i.e., ∆Eloss =0. The solid lines show the best fits assuming the TOF distance
L equal to 80 mm. These curves can be expressed by Equation 5.15.
Figure 5.19: The NPD1 TOF spectrum maximum as a function of the initial energy Eo . Open squares
and solid circles show experimental data for H and O beams, respectively. The dashed lines depict the
dependences with no energy loss, ∆Eloss = 0. The solid lines show the best fit.
Table 5.17 shows ∆Eloss as a function of the initial particle energy Eo for both NPD sensors.
The incident ion energy loss during interaction with the Start surface is clearly energy depen-
dent.
Figures 5.21 and 5.22 show the energy loss ∆Eloss (panels a) and the relative energy loss
K = ∆Eloss /E (panels b) of the NPD1 respectively NPD2 sensor graphically.
Figure 5.23 shows the estimation of the energy resolution of the NPD1 and NPD2 sensors.
The energy resolution is defined as double FWHM of the TOF spectra, normalized to the TOF
peaks.
m ∆Eloss m ∆Eloss
O 1.766Eo0.777 O 1.365Eo0.820
NPD1 H 0.132Eo1.134 NPD2 H 0.112Eo1.154
Table 5.17: Energy loss ∆Eloss as a function of the initial energy Eo and ion species.
Figure 5.20: The NPD2 TOF spectrum maximum as a function of the initial energy Eo . Open squares
and solid circles show experimental data for H and O beam, respectively. The dashed lines depict the
dependences with no energy loss, ∆Eloss = 0. The solid lines show the best fit.
a) b)
Figure 5.21: (a) Energy loss ∆Eloss and (b) relative energy loss K = ∆Eloss /E are plotted as a function
of the initial energy Eo for NPD1.
a) b)
a) b)
Figure 5.23: Energy resolution ∆E/E as a function of the initial energy Eo is plotted for different ion
species; for a) the NPD1 and b) the NPD2 sensors. The dashed and dotted lines are to guide the eye.

As seen in Figures 5.19 and 5.20, TOF for H and O overlap in the range of ∼260 - 650 ns for
the energy range of 100 eV - 10 keV. Particles with TOF in the range of >650 ns are assumed to
be O. Particles with TOF less than 260 ns are assumed to be H. Table 5.18 shows TOF ranges
in which O and H species can be identified directly from TOF measurements. PH analysis is
required to distinguish between the species inside this TOF range.
ASPERA-3
NPD1 NPD2
T M (RAW) T M (BIN) T M (RAW) T M (BIN)
H < 261 ns < 7 ch < 268 ns < 7 ch
O > 663 ns > 10 ch > 665 ns > 10 ch
Table 5.18: Identification of O and H species from TOF measurements. TOF values are shown for two
modes, RAW and BIN. T M shows the TOF in ns in the RAW mode and the channel number in the BIN
mode.
Table 5.19 shows the values of N indicating the channel number above which mostly O
atoms give a considerable tail in the pulse height distribution, and PO , PH indicating probability
for O and H atoms respectively to produce stop signals of a magnitude higher than level N.
Channel N is defined in such a way as to minimize the value of PPHO .
ASPERA-3
NPD1 NPD2
Dir0 Dir1 Dir2 Dir0 Dir1 Dir2
N 9 11 10 12 9 11
PH , ×10−2 0.03 0.00 0.38 0.93 0.12 0.33
PO , ×10−2 0.63 0.08 5.42 6.11 1.25 7.62
Table 5.19: Value N indicates a number of a channel above which mostly O atoms give a considerable
tail in pulse height distribution. PH and PO are the probabilities for H and O, respectively, to produce
stop signals of a magnitude higher than level N.
Figures 5.24 and 5.25 show pulse height distributions from 0.3 keV H (the red curve) and
3.0 keV O (the black curve) for NPD1 and that from 0.3 keV H (the red curve) and 5.0 keV O
(the black curve) for the NPD2 sensor, respectively, in the logarithmic scale.
5.4.8 Heater and temperature sensor characterization

The heater is for the Start surface conditioning (Section 4.2.2). The heater increases tempe-
rature of the Start surface by 50◦ over the ambient temperature. It takes ∼2 hours to reach
the equilibrium maximum temperature of the Start surface. Temperature sensors were verified
after the NPD integration with the ASPERA-3 package during TVAC test (Section 4.6).
5.4.9 Dark noise

Electromagnetic disturbances during communication of the DPU with the spacecraft were mea-
sured after the NPD integration with the ASPERA-3 package. Electromagnetically induced
noise at biased MCP detectors is 620 c/s for non-correlated counts. TOF coincidence totally
removes this noise.
ASPERA−3 NPD1
H Dir0
O
N
Dir1
Dir2
Figure 5.24: Pulse height distributions from 0.3 keV H (the red curve) and 3.0 keV O (the black curve)
are shown for sectors Dir0, Dir1, Dir2 (from top to bottom) of the ASPERA-3 / NPD1. Plots are shown
in the logarithmic scale. Both distributions are normalized to have the same area equal to 1. The dashed
lines show channel N values. The threshold for the NPD1 Dir1 was erroneously set to channel 11.
ASPERA−3 NPD2
H Dir0
O
N
Dir1
Dir2
Figure 5.25: Pulse height distributions from H and O of about the same velocity are shown for all
sectors of the ASPERA-3 / NPD2. The format is the same as in Figure 5.24.
5.5 ASPERA-4 / NPD calibration results

5.5.1 Calibration objectives
The ASPERA-4 / NPD calibration objectives are given in Table 5.20.
N Calibrated Ion beam Comment

parameter/ function Energy, keV Species
1 MCP characterization 5.0 O Determine nominal bias of the Start /
Stop MCP detectors.
2 Efficiency measurements (10.0), 5.0, 3.0, H, O Energy / mass dependent efficiency ε
1.3, 0.7, 0.5, for sectors Dir0/Dir1/Dir2
0.3, ( 0.1)
3 Angular response (10.0), 5.0, 3.0, H, O Azimuthal and elevation resolution
1.3, 0.7, 0.5, The geometrical constant Go estima-
0.3, ( 0.1) tion
4 Geometrical factor G (10.0), 5.0, 3.0, H, O Derived from:
1.3, 0.7, 0.5, 1. the angular response
0.3, ( 0.1) 2. the efficiency measurements
5 Energy and mass resolu- 5.0, 3.0, 1.3, H, O TOF and PH distribution functions
tion 0.7, 0.5, 0.3
6 Heater and temperature Start surface heating
sensor characterization
7 EMC related disturbances Functional test on the ASPERA-4 level
Table 5.20: The ASPERA-4 / NPD calibration objectives.

The MCP assemblies were tested up to 3100 V bias. The working point, i.e., nominal MCP
bias, was defined. Figure 5.26 shows the Start, Stop0, Stop1, Stop2 MCP detectors count versus
MCP bias for the NPD1 (left panel) and NPD2 (right panel) sensors.
Figure 5.26: The Start, Stop0, Stop1, Stop2 MCP detector count versus MCP bias for the NPD1 (left
panel) and NPD2 (right panel) sensors. The dashed lines depict the working points.

Though the saturation was not reached, the nominal bias was chosen at the point where the
curve derivative decreased. The lower bias was chosen to reduce a risk of discharge.
5.5.3 Efficiency measurements

Accumulation time t was set to 1.0 s. Efficiency measurements are performed for azimuthal
αm , and elevation βm angles. Table 5.21 shows the angles αm and βm , the pin-hole effective
area, and the values of the angular response function at these angles for both the NPD1 and
NPD2 sensors.
Dir0 40.0 0.0 10.0 0.64
NPD1 Dir1 0.0 0.0 13.0 0.88
Dir2 −40.0 0.0 10.0 0.58
Dir0 35.0 0.0 10.7 0.86
NPD2 Dir1 0.0 0.0 13.0 0.93
Dir2 −35.0 0.0 10.7 0.87
Table 5.21: Azimuth and elevation angles αm , βm at which the efficiency was measured. A(αm , βm ) the
pin-hole effective area at these angles. f (αm , βm ) is a value of the function defining the shape of the
angular response, calculated from the angular response measurements for 1.3 keV H for both the NPD1
and NPD2 sensors (Section 5.5.4).
Efficiency εm was calculated according to Equations 5.11 and 5.13:

CT otal (αm , βm )
εm = q ·
I
q XX
= ·γ· Ci j (αm , βm )
I i j
where i, j - the indexes over coordinates in the plane perpendicular to the beam, γ is defined
by Equation 5.12, Ageom = A(αm , βm ).
Then the efficiency was recomputed for angles αo , βo by Equation 5.14. Tables 5.22 and
5.23 show the efficiency ε (αo , βo ) of NPD1 and NPD2, respectively.
ASPERA-4 / NPD1
E, keV 10.0 5.0 3.0 1.3 0.7 0.5 0.3 0.1
Dir0 18.39 10.49 13.94 4.60 0.57 0.33 − −
O Dir1 12.29 10.04 5.77 1.82 0.50 0.18 0.04 −
Dir2 13.88 7.56 5.77 1.54 0.26 0.13 − −
Dir0 − 7.54 16.27 5.80 3.58 2.27 0.48 −
H Dir1 − 4.45 6.61 3.31 1.69 1.19 0.44 0.04
Dir2 − 5.24 6.83 2.93 1.43 1.02 0.39 −
ASPERA-4 / NPD2
E, keV 5.0 3.0 1.3 0.7 0.5 0.3 0.1
Dir0 9.41 6.44 1.19 0.21 0.03 0.01 −
O Dir1 6.45 4.83 1.24 0.26 0.04 0.03 −
Dir2 15.02 11.89 4.72 0.69 0.08 0.03 −
Dir0 5.27 4.89 2.38 2.37 0.47 0.18 −
H Dir1 3.89 3.99 2.31 1.53 0.54 0.17 0.01
Dir2 9.17 8.62 4.95 1.38 1.20 0.47 −
Figures 5.27 and 5.28 show the NPD sensors’ efficiency profiles together with the best
fittings by the function
ln (y) = a0 + a1 · x + a2 · ln (x) (5.25)
where x = ln (E), y = 104 · ε, and E an incident ion energy in units of eV.

The plots presented are valid for the energy range of 0.3 - 5.0 keV for H, and the energy
range of 0.3 - 10.0 keV for O. Tables 5.24 and 5.25 show the coefficients for the efficiency
fit-function for NPD1 and NPD2, respectively.
NPD1 / ASPERA-4
H O
a0 a1 a2 a0 a1 a2
Dir0 -35.27 -4.02 35.81 -68.98 -7.01 63.49
Dir1 -29.37 -3.42 30.29 -57.65 -5.36 51.43
Dir2 -27.05 -2.81 26.87 -63.18 -5.68 55.23
NPD2 / ASPERA-4
H O
a0 a1 a2 a0 a1 a2
Dir0 -36.60 -3.83 35.33 -95.06 -9.18 84.09
Dir1 -27.67 -2.58 26.15 -61.80 -5.55 53.98
Dir2 -32.29 -3.44 31.96 -87.42 -8.83 79.46
5.5.4 Angular response

Examples of the angular responses for 1.3 keV H and 1.3 keV O are shown only. The an-
gular responses have been also measured for several other energies (see Table 5.20) for both
NPD sensors. The values of coefficients k ji for the respective fits are collected in the file
vex_npd1_npd2_angular response.dat and available on the ASPERA-3 / ASPERA-4 server.
ASPERA−4 NPD1
Figure 5.27: Absolute efficiency ε(αo , βo ) of NPD1 is plotted in dependency on the ion mass and energy
for different sectors. Least-square fitting functions are shown by lines.
Figures 5.29 and 5.31 show the angular responses of the NPD1 sensor for 1.3 keV H and
1.3 keV O, respectively. Each figure shows the angular response of the sectors Dir2, Dir1, Dir0
(from left to right). The upper row is the measured azimuth-elevation response. The second
row from the top shows the fitting of the measured and smoothed data, normalized to have the
peak magnitude equal to 1. The contours show the polynomial fits based on the function
X
f (x, y) = k ji · xi · y j (5.26)
where x and y are azimuth and elevation angles scaled as x= α/5+10, y = β + 6. k ji are
coefficients.
Functions are valid in the range of α [-45, 45], and β [-4, 6]. The third row shows the azimuthal
response at β=βo and the bottom row shows the elevation response at α=αo . Diamonds show
the measured and smoothed data. The fittings are shown by thick lines. FWHM is given for
each profile. The values of k ji for the fit in Figure 5.29 are in Table 5.26. The values of k ji for
the fit in Figure 5.31 are in Table 5.27.
Figures 5.30 and 5.32 show the azimuthal profiles for respectively 1.3 keV H and 1.3 keV
O of all three sectors of the NPD1 sensor plotted altogether. The dashed lines show the total
ASPERA−4 NPD2
Figure 5.28: Absolute efficiency ε(αo , βo ) of NPD2 is plotted in dependency on the ion mass and energy
for different sectors. Least-square fitting functions are shown by lines.
angular coverage of the NPD1 sensor in the azimuthal plane. All responses are normalized to
have the peak maximum equal to 1.
ASPERA−4 NPD1
Dir 2 Dir 1 Dir 0
α o , βo
α m , βm
Figure 5.29: Angular response of the NPD1 sensor for 1.3 keV H given for the sectors Dir2, Dir1,
Dir0 (from left to right). The upper row shows the measured azimuth-elevation response. The second
row from the top shows the fitting of the measured and smoothed data. The third and fourth rows
show the cross-sections of the angular response function at αo and βo , respectively. FWHM is given for
each profile. Two red dots depict the angular positions corresponding to the position αm , βm at which
the efficiency measurements have been performed, and to the position αo , βo of the maximum of the
azimuth-elevation response function.
Dir j i=0 i=1 i=2 i=3 i=4 i=5 i=6

0 -4.52e-04 6.90e-03 -7.35e-03 2.82e-03 -4.35e-04 2.89e-05 -6.92e-07
1 1.60e-03 -2.63e-02 2.94e-02 -1.24e-02 2.12e-03 -1.57e-04 4.22e-06
2 -1.19e-03 1.92e-02 -2.15e-02 8.96e-03 -1.44e-03 9.81e-05 -2.39e-06
0 3 2.66e-04 -3.78e-03 4.05e-03 -1.68e-03 2.59e-04 -1.63e-05 3.57e-07
4 -2.47e-05 2.91e-04 -2.92e-04 1.24e-04 -1.79e-05 9.93e-07 -1.67e-08
5 9.72e-07 -8.12e-06 6.76e-06 -3.07e-06 3.82e-07 -1.19e-08 -1.82e-10
6 -1.31e-08 3.65e-08 1.24e-08 2.04e-09 1.84e-09 -4.14e-10 1.98e-11
0 -1.09e-04 3.53e-04 1.03e-05 -4.93e-04 1.19e-04 -9.61e-06 2.58e-07
1 -5.38e-03 1.02e-01 -1.11e-01 4.07e-02 -6.02e-03 3.84e-04 -8.84e-06
2 2.38e-03 -3.98e-02 4.00e-02 -1.45e-02 2.15e-03 -1.39e-04 3.27e-06
1 3 -4.47e-04 6.88e-03 -6.65e-03 2.42e-03 -3.53e-04 2.19e-05 -4.87e-07
4 3.91e-05 -5.63e-04 5.25e-04 -1.90e-04 2.66e-05 -1.55e-06 3.11e-08
5 -1.58e-06 2.17e-05 -1.94e-05 6.92e-06 -9.21e-07 4.89e-08 -8.32e-10
6 2.42e-08 -3.18e-07 2.74e-07 -9.55e-08 1.20e-08 -5.65e-10 7.09e-12
0 1.18e-03 -1.91e-02 1.91e-02 -6.10e-03 7.41e-04 -3.51e-05 4.64e-07
1 -1.38e-02 2.30e-01 -2.37e-01 7.65e-02 -9.40e-03 4.51e-04 -6.12e-06
2 4.67e-03 -7.04e-02 6.80e-02 -2.08e-02 2.58e-03 -1.33e-04 2.22e-06
2 3 -6.44e-04 8.72e-03 -7.87e-03 2.30e-03 -2.98e-04 1.74e-05 -3.71e-07
4 4.41e-05 -5.36e-04 4.48e-04 -1.24e-04 1.73e-05 -1.18e-06 3.07e-08
5 -1.50e-06 1.64e-05 -1.26e-05 3.30e-06 -5.07e-07 4.08e-08 -1.24e-09
6 2.01e-08 -2.03e-07 1.43e-07 -3.52e-08 6.22e-09 -5.86e-10 1.99e-11
k
for 1.3 keV H.
ASPERA−4 NPD1
Dir 2 Dir 1 Dir 0
Dir j i=0 i=1 i=2 i=3 i=4 i=5 i=6

0 -1.47e-03 2.34e-02 -2.54e-02 9.26e-03 -1.40e-03 9.36e-05 -2.29e-06
1 6.26e-03 -9.75e-02 1.03e-01 -3.57e-02 5.08e-03 -3.14e-04 6.96e-06
2 -4.02e-03 6.75e-02 -7.36e-02 2.63e-02 -3.81e-03 2.39e-04 -5.38e-06
0 3 8.70e-04 -1.51e-02 1.67e-02 -6.03e-03 8.79e-04 -5.53e-05 1.25e-06
4 -8.25e-05 1.45e-03 -1.63e-03 6.02e-04 -8.86e-05 5.60e-06 -1.26e-07
5 3.47e-06 -6.22e-05 7.08e-05 -2.66e-05 3.96e-06 -2.51e-07 5.68e-09
6 -5.32e-08 9.64e-07 -1.11e-06 4.28e-07 -6.42e-08 4.08e-09 -9.21e-11
0 -9.63e-05 1.58e-03 -1.99e-03 8.26e-04 -1.82e-04 1.78e-05 -6.10e-07
1 8.90e-04 3.46e-03 -1.25e-02 7.76e-03 -1.23e-03 6.90e-05 -1.12e-06
2 -1.34e-03 2.30e-02 -2.44e-02 7.96e-03 -1.23e-03 9.11e-05 -2.57e-06
1 3 3.19e-04 -6.60e-03 7.37e-03 -2.44e-03 3.77e-04 -2.76e-05 7.71e-07
4 -3.20e-05 7.11e-04 -8.10e-04 2.71e-04 -4.21e-05 3.10e-06 -8.64e-08
5 1.45e-06 -3.33e-05 3.84e-05 -1.30e-05 2.04e-06 -1.50e-07 4.19e-09
6 -2.42e-08 5.67e-07 -6.60e-07 2.28e-07 -3.59e-08 2.64e-09 -7.35e-11
0 1.06e-03 -3.14e-02 3.98e-02 -1.62e-02 2.71e-03 -2.00e-04 5.44e-06
1 -1.55e-02 3.32e-01 -3.73e-01 1.36e-01 -2.02e-02 1.31e-03 -3.08e-05
2 4.94e-03 -8.51e-02 8.42e-02 -2.64e-02 3.45e-03 -1.96e-04 3.97e-06
2 3 -6.60e-04 7.89e-03 -5.66e-03 9.33e-04 -1.15e-05 -6.77e-06 3.46e-07
4 4.48e-05 -2.85e-04 -1.22e-05 1.19e-04 -3.07e-05 2.78e-06 -8.55e-08
5 -1.53e-06 1.50e-06 1.34e-05 -1.00e-05 2.08e-06 -1.72e-07 4.98e-09
6 2.07e-08 8.16e-08 -3.43e-07 2.10e-07 -4.13e-08 3.30e-09 -9.37e-11
k
for 1.3 keV O.
ASPERA−4 NPD2
Dir 2 Dir 1 Dir 0
Figures 5.33 and 5.35 show the angular responses of the NPD2 sensor for 1.3 keV H and
1.3 keV O, respectively. The format is the same as in Figure 5.29. The values of coefficients k ji
for the fit in Figure 5.33 are in Table 5.28. The values of coefficients k ji for the fit in Figure 5.35
are in Table 5.29.
Figures 5.34 and 5.36 show the normalized azimuthal profiles for respectively 1.3 keV H
and 1.3 keV O of all three sectors of the NPD2 sensor plotted altogether. The dashed lines
show the total angular coverage of the NPD2 sensor in the azimuthal plane.
Dir j i=0 i=1 i=2 i=3 i=4 i=5 i=6

0 -4.66e-03 2.52e-02 -2.62e-02 1.00e-02 -1.65e-03 1.22e-04 -3.32e-06
1 2.11e-02 -3.65e-02 1.35e-02 -1.49e-03 -1.82e-04 3.84e-05 -1.59e-06
2 -1.39e-02 2.10e-02 -4.68e-03 -2.39e-04 3.30e-04 -4.48e-05 1.73e-06
0 3 3.24e-03 -3.84e-03 -1.74e-04 5.81e-04 -1.69e-04 1.76e-05 -6.06e-07
4 -3.37e-04 3.08e-04 1.13e-04 -9.79e-05 2.37e-05 -2.27e-06 7.44e-08
5 1.60e-05 -1.03e-05 -9.79e-06 6.29e-06 -1.37e-06 1.23e-07 -3.88e-09
6 -2.81e-07 1.03e-07 2.53e-07 -1.40e-07 2.84e-08 -2.42e-09 7.36e-11
0 -4.74e-03 -1.19e-02 1.34e-02 -3.53e-03 3.72e-04 -1.43e-05 5.98e-08
1 1.70e-04 3.69e-02 -4.81e-02 2.51e-02 -4.80e-03 3.82e-04 -1.08e-05
2 9.07e-03 -1.09e-02 6.16e-03 -2.75e-03 5.25e-04 -4.23e-05 1.22e-06
1 3 -2.53e-03 3.25e-03 -8.32e-04 5.46e-05 1.57e-05 -2.73e-06 1.14e-07
4 2.60e-04 -3.97e-04 6.19e-05 2.57e-05 -9.21e-06 9.77e-07 -3.38e-08
5 -1.17e-05 2.02e-05 -1.77e-06 -2.39e-06 7.01e-07 -6.94e-08 2.30e-09
6 1.93e-07 -3.68e-07 1.39e-08 5.80e-08 -1.58e-08 1.51e-09 -4.93e-11
0 5.97e-04 -8.31e-03 1.12e-02 -4.47e-03 6.53e-04 -3.90e-05 8.01e-07
1 -4.61e-03 1.57e-01 -2.26e-01 9.99e-02 -1.68e-02 1.21e-03 -3.15e-05
2 9.79e-09 -2.02e-02 4.47e-02 -2.19e-02 3.98e-03 -3.09e-04 8.62e-06
2 3 1.90e-04 -6.09e-04 -3.44e-03 2.18e-03 -4.59e-04 3.98e-05 -1.22e-06
4 -2.40e-05 2.05e-04 1.37e-04 -1.32e-04 3.24e-05 -3.07e-06 9.96e-08
5 1.12e-06 -1.06e-05 -3.52e-06 4.84e-06 -1.29e-06 1.27e-07 -4.24e-09
6 -1.83e-08 1.72e-07 5.33e-08 -7.82e-08 2.12e-08 -2.11e-09 7.08e-11
k
for 1.3 keV H.
ASPERA−4 NPD2
Dir 2 Dir 1 Dir 0
Dir j i=0 i=1 i=2 i=3 i=4 i=5 i=6

0 -5.76e-03 9.96e-02 -1.03e-01 3.83e-02 -6.36e-03 4.87e-04 -1.39e-05
1 5.66e-03 -9.95e-02 7.12e-02 -2.30e-02 4.03e-03 -3.45e-04 1.10e-05
2 -1.81e-03 3.03e-02 -7.54e-03 3.67e-04 -1.22e-04 2.65e-05 -1.34e-06
0 3 2.21e-04 -2.80e-03 -2.96e-03 1.74e-03 -2.82e-04 1.78e-05 -3.72e-07
4 -8.94e-06 -2.22e-05 6.55e-04 -3.14e-04 5.13e-05 -3.48e-06 8.40e-08
5 -1.01e-07 1.31e-05 -4.44e-05 2.00e-05 -3.25e-06 2.25e-07 -5.59e-09
6 9.40e-09 -4.07e-07 9.80e-07 -4.27e-07 6.94e-08 -4.82e-09 1.21e-10
0 1.10e-02 4.33e-01 -3.64e-01 1.10e-01 -1.56e-02 1.04e-03 -2.67e-05
1 2.01e-02 -4.19e-01 3.26e-01 -8.89e-02 1.12e-02 -6.68e-04 1.54e-05
2 -5.95e-03 1.15e-01 -7.92e-02 1.78e-02 -1.61e-03 5.23e-05 -1.34e-07
1 3 7.19e-04 -1.20e-02 5.16e-03 2.01e-04 -2.62e-04 3.07e-05 -1.06e-06
4 -4.08e-05 5.41e-04 1.00e-04 -2.29e-04 5.72e-05 -5.18e-06 1.59e-07
5 1.02e-06 -9.09e-06 -1.96e-05 1.55e-05 -3.36e-06 2.86e-07 -8.52e-09
6 -7.96e-09 -6.74e-10 4.84e-07 -3.15e-07 6.45e-08 -5.35e-09 1.57e-10
0 3.77e-02 -1.00e-01 9.23e-02 -3.25e-02 5.14e-03 -3.73e-04 1.01e-05
1 -5.19e-02 3.85e-01 -4.18e-01 1.53e-01 -2.43e-02 1.74e-03 -4.65e-05
2 1.91e-02 -1.21e-01 1.19e-01 -3.67e-02 5.28e-03 -3.63e-04 9.62e-06
2 3 -3.25e-03 1.74e-02 -1.53e-02 3.79e-03 -4.54e-04 2.83e-05 -7.29e-07
4 2.80e-04 -1.34e-03 1.08e-03 -2.19e-04 2.02e-05 -1.00e-06 2.35e-08
5 -1.17e-05 5.27e-05 -4.08e-05 7.29e-06 -5.18e-07 1.70e-08 -2.81e-10
6 1.87e-07 -8.23e-07 6.29e-07 -1.08e-07 6.74e-09 -1.49e-10 7.94e-13
k
for 1.3 keV O.
H O
Sensor Sector αo , deg βo , deg mm2
Ao , αo , deg βo , deg Ao , mm2
Dir0 27.0 0.9 11.62 23.0 0.6 12.00
NPD1 Dir1 0.0 1.1 13.04 0.0 0.7 13.04
Dir2 −29.0 1.5 11.40 −30.0 0.8 11.29
Dir0 32.0 1.1 11.06 36.0 0.7 10.55
NPD2 Dir1 5.0 0.9 12.99 −1.0 0.4 13.04
Dir2 −27.0 0.6 11.62 −25.0 0.2 11.82
angles are given. αo , βo are calculated from the angular response for 1.3 keV H and 1.3 keV O for both
NPD1 and NPD2 sensors.
sponse functions of both NPD sensors for 1.3 keV H and 1.3 keV O. The effective aperture
area at these angles is also shown.
Table 5.31 shows the pure geometrical constant Go of both the NPD1 and NPD2 sensors,
calculated according to Equation 5.3.
Go , ×10−3 , cm2 sr
H O
Sensor Dir0 Dir1 Dir2 Dir0 Dir1 Dir2
NPD1 6.94 10.33 6.71 6.98 10.10 6.01
NPD2 6.79 12.10 7.56 6.43 11.36 6.25
Table 5.31: Pure geometrical factor Go computed from the angular response for 1.3 keV H for both
NPD1 and NPD2 sensors. Efficiency ε is not included.
5.5.5 Geometrical factor

The total geometrical factors G = Go · ε(E) (in units of cm2 sr) of the NPD1 and NPD2 sensors
are shown in Tables 5.32 and 5.33, respectively.
NPD1 / ASPERA-4
E, keV 10.0 5.0 3.0 1.3 0.7 0.5 0.3 0.1
Dir0 11.82 10.20 7.42 2.30 0.49 0.10 − −
O Dir1 13.79 9.80 6.65 1.85 0.51 0.20 0.03 −
Dir2 8.01 5.87 3.83 0.83 0.20 0.07 − −
Dir0 − 9.18 7.79 4.47 2.13 1.27 0.49 −
H Dir1 − 6.47 5.33 3.34 1.90 1.13 0.50 0.03
Dir2 − 4.40 3.80 2.01 1.08 0.69 0.31 −
Table 5.32: The absolute geometrical factors of the ASPERA-4 / NPD1 sensor, computed from the fit
functions of the efficiency of NPD1 and pure geometrical factors of NPD1 derived from the angular
responses for respective energies.
NPD2 / ASPERA-4
E, keV 5.0 3.0 1.3 0.7 0.5 0.3 0.1
Dir0 7.99 4.37 0.84 0.14 0.05 0.003 −
O Dir1 9.17 5.40 1.28 0.32 0.13 0.02 −
Dir2 14.19 8.73 2.06 0.35 0.11 0.01 −
Dir0 4.67 3.79 1.84 0.90 0.56 0.21 −
H Dir1 7.42 5.64 2.57 1.26 0.70 0.28 −
Dir2 7.29 6.29 3.14 1.40 0.79 0.30 −
function of the efficiency of NPD2 and pure geometrical factors of NPD2 derived from the angular
responses for respective energies.

Figures 5.37, 5.38 show the TOF spectra obtained by the NPD1 sensor for H and O ion beams,
respectively, for different energies. Figures 5.39, 5.40 show the TOF spectra obtained by the
NPD2 sensor for H and O ion beams, respectively, for different energies.
Figure 5.37: The NPD1 TOF spectra for H beams of different energies. The reason for the structures in
the 0.3 keV TOF spectrum (at ∼250 ns) and in the 0.5 keV TOF spectrum (at ∼200 ns) is as yet unclear.
It can perhaps be due to electronic disturbances.
Figure 5.39: The NPD2 TOF spectra for H beams of different energies.
Figures 5.41 and 5.42 show the TOF spectrum maximum as a function of the incident ion
beam energy for the NPD1 and NPD2 sensors, respectively. The error bars show FWHM of the
TOF spectra. The dashed lines show ideal theoretical dependencies T M = f (E) (Equation 5.15)
with no energy loss, i.e., ∆Eloss =0. The solid lines show the best fits, assuming the TOF
distance L equal to 80 mm. These curves can be expressed by Equation 5.15.
Figure 5.41: NPD1 TOF spectrum maximum as a function of the initial energy Eo . Open squares
and solid circles show experimental data for H and O beam, respectively. The dashed lines depict the
dependence with no energy loss, ∆Eloss = 0. The solid lines show the best fit.
Table 5.34 shows ∆Eloss as a function of the initial particle energy Eo for both NPD sensors.
The incident ion energy loss during interaction with the Start surface is clearly energy depen-
dent.
Figures 5.43 and 5.44 show the energy loss ∆Eloss (panels a) and the relative energy loss
K = ∆Eloss /E (panels b) of the NPD1 respectively NPD2 sensor graphically.
Figure 5.45 shows the estimation of the energy resolution of the NPD1 and NPD2 sensors.
The energy resolution is defined as double FWHM of the TOF spectra, normalized to the TOF
peaks.
m ∆Eloss m ∆Eloss
O 1.781Eo0.763 O 0.843Eo0.873
NPD1 H 0.082Eo1.146 NPD2 H 0.124Eo1.128
Table 5.34: The energy loss ∆Eloss as a function of the initial energy Eo and ion species.
Figure 5.42: NPD2 TOF spectrum maximum as a function of the initial energy Eo . Open squares
and solid circles show experimental data for H and O beams, respectively. The dashed lines depict the
dependence with no energy loss, ∆Eloss = 0. The solid lines show the best fit.
a) b)
a) b)
a) b)
Figure 5.45: Energy resolution ∆E/E as a function of the initial energy Eo is plotted for different ion
species; for a) the NPD1 and b) the NPD2 sensors. The dashed and dotted lines are to guide the eye.

As seen in Figures 5.41 and 5.42, TOF for H and O overlaps in the range of ∼260 - 650 ns for
the energy range of 100 eV - 10 keV. Particles with TOF in the range of >650 ns are assumed to
be O. Particles with TOF less than 260 ns are assumed to be H. Table 5.35 shows TOF ranges
in which O and H species can be identified directly from TOF measurements. PH analysis is
required to distinguish between the species inside this TOF range.
ASPERA-4
NPD1 NPD2
T M (RAW) T M (BIN) T M (RAW) T M (BIN)
H < 258 ns < 7 ch < 268 ns < 7 ch
O > 629 ns > 10 ch > 654 ns > 10 ch
Table 5.35: Identification of O and H species from TOF measurements. TOF values are shown for two
modes, RAW and BIN. T M shows the TOF in ns in the RAW mode and the channel number in the BIN
mode.
Table 5.36 shows the values of N indicating the channel number above which mostly O
atoms give a considerable tail in the pulse height distribution, and PH , PO indicating probability
for H and O atoms respectively to produce stop signals of a magnitude higher than level N.
Channel N is defined in such a way as to minimize the value of PPHO .
ASPERA-4
NPD1 NPD2
Dir0 Dir1 Dir2 Dir0 Dir1 Dir2
N 7 11 10 11 10 5
PH , ×10−2 1.16 8.34 1.27 4.00 0.26 0.06
PO , ×10−2 3.18 10.70 5.99 6.00 4.45 2.10
Table 5.36: Value N indicates a number of a channel above which mostly O atoms give a considerable
tail in pulse height distribution. PH and PO are the probabilities for H and O, respectively, to produce
stop signals of a magnitude higher than level N.
Figures 5.46 and 5.47 show pulse height distributions from 0.3 keV H (the red curve) and
5.0 keV O (the black curve) for the NPD1 and NPD2 sensors, respectively, in the logarithmic
scale.
5.5.8 Heater and temperature sensor characterization

The heater is for the Start surface conditioning (Section 4.2.2). The heater increases tempe-
rature of the Start surface by 50◦ over the ambient temperature. It takes ∼2 hours to reach
the equilibrium maximum temperature of the Start surface. Temperature sensors were verified
after the NPD integration with the ASPERA-4 package during TVAC test (Section 4.6).
5.5.9 Dark noise

Electromagnetic disturbances during communication of the DPU with the spacecraft were mea-
sured after the NPD integration with the ASPERA-4 package. Electromagnetically induced
noise at biased MCP detectors is 620 c/s for non-correlated counts. TOF coincidence totally
removes this noise.
ASPERA−4 NPD1
H Dir0
O
N
Dir1
Dir2
Figure 5.46: Pulse height distributions from 0.3 keV H (the red curve) and 5.0 keV O (the black curve)
are shown for sectors Dir0, Dir1, Dir2 (from top to bottom) of the ASPERA-4 / NPD1. Plots are shown
in the logarithmic scale. Both distributions are normalized to have the same area equal to 1. The dashed
lines show channel N values.
ASPERA−4 NPD2
H Dir0
O
N
Dir1
Dir2
Figure 5.47: Pulse height distributions from H and O of about the same velocity are shown for the
ASPERA-4 / NPD2 sensor. The format is the same as in Figure 5.46.
Chapter 6
Scientific results. The NPD

measurements at Mars.
The ASPERA-3 / NPD is the first ENA sensor flown to Mars. In the meantime, the ENA
environment of Mars was described using numerical simulations and models, based on Mars –
solar wind interaction (see Section 1.3). In this chapter the observations and scientific findings
by the ASPERA-3 / NPD are reviewed. One of the most important results from the NPD
measurements is an observation of the ENA jet (cone) and its dynamics (Futaana et al., 2006a;
Grigoriev et al., 2006), discussed in Sections 6.1, 6.2. In Section 6.3 other scientific results
from the ASPERA-3 / NPD are briefly overviewed.
6.1 Subsolar ENA jet
6.1.1 Introduction
Futaana, Barabash, Grigoriev et al. (2006a) reported on the intensive ENA emission from the
subsolar region of Mars, so-called ENA jet/cone, detected by the NPD instrument.
This study uses data recorded between May 24 and July 1, 2004, when the orbit and the
attitude of Mars Express were optimal for investigating ENA emissions from the subsolar re-
gion. The NPD operations lasted ∼30 minutes, starting ∼20 minutes after the periapsis. During
these time intervals, the Mars Express was nadir pointing and the ASPERA-3 scanner rotation
axis pointed towards the center of Mars. NPD was running in the BIN mode. The TOF spectra
observed by the NPD1 are used in this study since the subsolar region was out of the NPD2
field-of-view.
The NPD sensor is an open system, and thus UV photons entering the instrument result
in non-correlated counts on the Start and Stop units. The non-correlated count rate on Start is
∼10 kHz and on each Stop ∼300 Hz. These non-correlated signals result in random correlated
TOF signals as a background level. Since the TOF distribution of this signal is basically con-
stant over the entire TOF window of NPD, we can estimate the background level from the TOF
spectrum obtained. To recover absolute flux, this random correlated signal must be subtracted
from the recorded signal.
Figure 6.1 shows a schematic view of the NPD configuration and viewing directions relative
to the spacecraft body.
119
120 CHAPTER 6. SCIENTIFIC RESULTS. THE NPD MEASUREMENTS AT MARS.
Mars (nadir pointing) Mars (nadir pointing)
Satellite Velocity Satellite Velocity
NPD-1 NPD-2 NPD-2 NPD-1

Dir-0 Dir-0
Dir-1 Dir-1
Dir-2 Dir-2
Zb
Zb
Xb
Yb
Figure 6.1: The configuration of the NPD instruments on board Mars Express seen from (left) the −Yb
direction and (right) the −Xb direction. The satellite coordinate system is used. Two NPD instruments
are mounted on the Zb plane and each has a 9◦ × 90◦ FOV. The FOV inclines ±15◦ in the Yb direction for
the NPD1 and NPD2 respectively. The attitude mode of the satellite during the observations discussed
in this paper was nadir pointing, i.e., Zb pointed toward the center of Mars. From Futaana et al. (2006a).
6.1.2 Observations
Figure 6.2 shows the Mars Express trajectory on June 2, 2004 (Orbit 466) in the cylindrical
coordinate system, based on the Mars-Sun orbit (MSO) coordinate system, in which the x-axis
is aligned from Mars to the Sun, the z-axis is perpendicular to the Martian orbital plane toward
ecliptic north and the y-axis completes the right-handed system. In the cylindrical coordinate
system, the horizontal axis corresponds to the x-directionp of the MSO coordinates, and the
vertical axis is the distance from the Mars-Sun line (r = y2 + z2 ).
2004/06/02
10000
r=sqrt(y^2+z^2) [km]
8000
6000
06:30
4000 06:00 06:20
06:10
2000
0
-10000 -5000 0 5000 10000
x [km]
Figure 6.2: The Mars Express trajectory on June 2, 2004, (Orbit-466) in the cylindrical coordinate
system: the horizontal axis corresponds to the x-direction of the MSO coordinates and the vertical axis
is the distance from the Mars-Sun line. The dashed lines are the modeled BS and MPB (Vignes et al.,
2000). From Futaana et al. (2006a).
The pericenter height is 265 km (05:40 UT). NPD operated between 06:03 and 06:31 UT.
Mars Express was located in the magnetosheath region at the start of the operations and moved
to the solar wind region after a BS crossing at 06:05 UT as deduced from data recorded by the
6.1. SUBSOLAR ENA JET 121
ASPERA-3 IMA and ELS sensors. The BS was located closer to the planet than the modeled
BS (shown in Figure 6.2).
06:05
06:10
06:15
06:20
06:25
06:30
Figure 6.3: The viewing geometry of the NPD1 on June 2, 2004. The first and second columns are 3-D
representations of the NPD1 FOV at different times seen from different view points. The green and blue
fans are the Dir1 and Dir2 FOVs, respectively. The third column is a fish-eye projection of the same
geometry (see text). The red sphere represents Mars and the green and blue rectangles show the Dir1
and Dir2 FOVs, respectively. The yellow dots are the directions toward points of 200 km, 500 km, 1000
km and 3397 km (=1R M ) above the subsolar point. From Futaana et al. (2006a).
The viewing geometry of the NPD1 is shown in Figure 6.3. The first and second columns
are 3-D representations of the NPD1 FOV at different times along the Mars Express orbit (white
ellipse) seen from different view points. The green and blue fans are the Dir1 and Dir2 FOVs
and the yellow dots correspond to the subsolar point. The third column is a fish-eye projection
of the same geometry, i.e. an image as it would be seen by an observer located on the spacecraft
with the boresight direction aligned toward the nadir. The red sphere represents Mars and the
green and blue rectangles show the Dir1 and Dir2 FOVs, respectively. The yellow dots indicate
the directions toward points of 200 km, 500 km, 1000 km and 3397 km (=1R M ) above the
subsolar point.
Figure 6.4 shows measurements from the NPD1 sensor. The first and second panels from
top are TOF spectrograms of Dir1 and Dir2, respectively. The right axis gives the correspond-
ing energy per atomic mass 1 amu. Intense ENA signals with TOF less than 400 ns, correspond-
ing to >200 eV, are seen. The third panel shows the count rate of the Dir1 and Dir2 integrated
over the TOF from 124.5 to 387 ns, which corresponds to the energy range 0.34-3.2 keV/amu.
0.0 0.2 0.4 0.6 0.8 1.0
Count rate [Counts/sec/nsec] 2004/06/02

NPD-1 DIR-1
1000
Energy [eV/amu]
100
TOF [ns]
1000
100
06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00 06:30:00 06:35:00
NPD-1 DIR-2
1000
Energy [eV/amu]
100
TOF [ns]
1000
100
06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00 06:30:00 06:35:00
250
Time of Flight [124.5 - 387 ns]
200
Count rate
[#/sec]
150 DIR 1
100
50 DIR 2
0
06:00:00 06:05:00 06:10:00 06:15:00 06:20:00 06:25:00 06:30:00 06:35:00
Figure 6.4: The NPD measurements between 06:03 and 06:31 UT on June 2, 2004. The first and second
panels are TOF spectrograms of the Dir1 and Dir2, respectively. The right axis gives the corresponding
energy per mass. The third panel shows the count rates of the Dir1 and Dir2 integrated over the TOF
range of 124.5-387 ns (0.34-3.2 keV/amu). From Futaana et al. (2006a).
Figure 6.5 shows the TOF spectra of the NPD1/Dir1 averaged over the time periods 06:10-
06:15 and 06:25-06:30 UT on June 2, 2004. The peak appears at an energy of ∼1.5 keV/amu.
The total ENA count rate decreased as the spacecraft moved away from Mars (Figure 6.4).
As seen in Figure 6.3, the FOVs of the NPD1/Dir1 and Dir2 covered the subsolar region during
the observations.
The typical integrated count rate was ∼200 counts/s (Figure 6.4). The random correlated
count (a background level), is estimated to be ∼ 0.48 counts/s per 1-ns bin from the spectrum
(dashed line in Figure 6.5), which corresponds to ∼130 counts/s in TOF range 124.5-387 ns.
The effective ENA signal is then ∼70 counts/s. Using the geometrical factor of the instrument
(G0 · ǫ ∼ (9.78 − 17.1)×10−5 cm2 sr for 0.7-1.3 keV hydrogen atoms, Chapter 5), the ENA dif-
ferential flux, J, is calculated as J =(4-7)×105 cm−2 sr−1 s−1 , assuming the observed ENAs are
Energy [eV/amu]
1000 100
1.0
0.8
Count rate [#/sec/nsec]

0.6
2004/06/02 06:10-06:15
0.4
2004/06/02 06:25-06:30
0.2
0.0
100 1000
Time of flight [ns]
Figure 6.5: TOF spectra of the NPD1/Dir1 obtained during the time period 06:10-06:15 and 06:25-
06:30 on June 2, 2004. The dashed line, 0.48 counts/s per 1-ns bin, depicts the background level of the
period 06:10-06:15. From Futaana et al. (2006a).
2004/05/30
10000
r=sqrt(y^2+z^2) [km]
8000
6000
11:20
4000 10:50 11:10

11:00
2000
0
-10000 -5000 0 5000 10000
x [km]
Figure 6.6: The trajectory of Mars Express (Orbit 456) in the same format as in Figure 6.2. From Fu-
taana et al. (2006a).
hydrogen.
Another example of a similar NPD observation is discussed below. The observation was
conducted on May 30, 2004. The orbit and the FOVs geometry are shown in Figures 6.6 and
6.7, respectively. They are almost the same as in the previous example from June 2. Figure
6.8 shows the NPD1 measurement. Instead of a gradual decrease of the count rate as in the
previous example, NPD observed a sharp decrease in the count rate at ∼11:03 UT within a
time scale of the order of 1 minute. Figure 6.9 shows the TOF spectra at 10:56-11:00 (before
the decrease) and at 11:12-11:16 (after the decrease). Their shapes are very similar to those
obtained on June 2 (Figure 6.5).
Another characteristic signature is a short enhancement in the ENA count rate, observed
at around 11:10 UT. Since the FOV pointed at the subsolar region and the TOF spectrum of
these enhanced ENAs was almost the same as that observed before the sharp decrease, they
probably originated from the same source and reflect a reconstruction of the interaction region
(see Section 6.1.3).
10:55
11:00
11:05
11:10
11:15
Figure 6.7: The viewing geometry of the NPD1 on May 30, 2004. The format is the same as in Figure
6.3. From Futaana et al. (2006a).
We searched through all of the TOF spectrograms obtained by the NPD1 between May 24,
2004, and July 1, 2004, when the trajectories and the attitude of the Mars Express were nearly
the same, which means almost the same vantage points and observation directions. A total of 38
orbits were available, and the NPD observed an intense ENA flux in 36 cases. We used all the
TOF spectrograms such as Figures 6.4 and 6.8 to divide the observations into two categories.
The sharp changes in an ENA flux were observed in 23 orbits (64%), and the gradual decreases
(i.e., no sudden change of an ENA flux) were seen in 13 orbits (36%).
6.1.3 Discussion
The ENA flux depends strongly on the position of the satellite, even though the FOVs of the
NPD were always viewing toward the subsolar region. This observational fact can be explained
in terms of a highly directional ENA emission around the subsolar region, i.e. a subsolar ENA
jet (or cone). Figure 6.10 is a schematic representation of a concept of the subsolar ENA jet.
Such an ENA flux can be detected when the sensor is within this ENA jet (case (a) in the
0.0 0.2 0.4 0.6 0.8 1.0
Count rate [Counts/sec/nsec] 2004/05/30

NPD-1 DIR-1
1000
Energy [eV/amu]
100
TOF [ns]
1000
100
10:50:00 10:55:00 11:00:00 11:05:00 11:10:00 11:15:00 11:20:00
NPD-1 DIR-2
1000
Energy [eV/amu]
100
TOF [ns]
1000
100
10:50:00 10:55:00 11:00:00 11:05:00 11:10:00 11:15:00 11:20:00
250
Time of Flight [124.5 - 387 ns]
200
Count rate
[#/sec]
150 DIR 1
100
50 DIR 2
0
10:50:00 10:55:00 11:00:00 11:05:00 11:10:00 11:15:00 11:20:00
Figure 6.8: The NPD measurements recorded between 10:51 and 11:18 UT on May 30, 2004. The
format is the same as in Figure 6.4. From Futaana et al. (2006a).
figure). As soon as the spacecraft leaves the jet (case (b) in the figure), the ENAs cannot be
detected even though the instrument FOV covers the source region.
The region of the ENA generation and direction of ENA movement point to three possi-
ble origins, namely solar wind protons and ENAs back-scattered from the Martian exosphere
(Kallio and Barabash, 2001), shocked solar wind protons charge-exchanged on the Martian
hydrogen corona (Holmström et al., 2002), and planetary hydrogen photo-ionized, accelerated
and charge-exchanged in the exosphere (Lichtenegger et al., 2002). These processes are all
related to interactions between the solar wind and the Martian upper atmosphere, and the ENA
generation is expected to be large in the subsolar region because the solar wind can penetrate
to much lower altitudes, reaching altitudes with high neutral gas density.
Since back-scattered ENAs are expected to be emitted isotropically from the Martian upper
atmosphere (Kallio and Barabash, 2001; Futaana et al., 2006b), they are unlikely to result in
the ENA jet generation.
We consider now shocked solar wind just above the IMB, which is defined as an envelope
of the void of the solar wind (Lundin et al., 2004). There the stream lines of the shocked solar
wind are highly deflected, and an ENA jet can be formed by charge exchange with the dense
atmosphere. This scenario is possible because the solar wind can actually reach altitudes as low
as 300 km in the subsolar region, as evidenced by the recent results of the ASPERA-3 ion mass
spectrometer (Lundin et al., 2004). The size of the source, i.e., a region of the most intense
ENA production by shocked solar wind charge exchanged with the exosphere, must be at least
on the order of the scale height of the hydrogen, i.e., several hundred km. ENA production
from the shocked solar wind is described in Kallio et al. (1997) and Holmström et al. (2002) in
more detail. However the proton flux model has a singularity at the subsolar point, and cannot
Energy [eV/amu]
1000 100
1.0
Count rate [#/sec/nsec] 0.8
0.6 2004/05/30 10:56-11:00
0.4
2004/05/30 11:12-11:16
0.2
0.0
100 1000
Time of flight [ns]
Figure 6.9: TOF spectra of the NPD1/Dir1 averaged over the time interval 10:56-11:00 and 11:12-11:16
on May 30, 2004. From Futaana et al. (2006a).
be used to simulate ENA productions there, while the ENA generation must be highest in the
subsolar region.
The accelerated planetary protons, originated from exospheric hydrogen photo-ionization
and then converted to ENAs through the charge exchange process, have been considered by
Lichtenegger et al. (2002). Planetary protons transported to the magnetosheath are accelerated
by the convection electric field. Such accelerated protons could form an ENA jet after they
experience charge exchanges in the subsolar region. Somewhat similar jets of planetary oxygen
ENAs were predicted by Barabash et al. (2002). Lichtenegger et al. (2002) concluded that the
ENA flux produced by this mechanism would be as high as the shocked solar wind ENAs.
These authors also compared the energy spectra of these two mechanisms, and concluded that
careful analysis of the spectra can distinguish the accelerated planetary ENAs from the shocked
solar wind ENAs. In the present instance it is difficult to discuss the spectra in detail because
they are too coarse due to the observation mode employed.
The observation geometry over all 38 events was rather similar in the MSO coordinate sys-
tem. During the observations, the NPD FOV plane was close to the ecliptic plane. Therefore,
only a limited range of angles (< 30◦ ) relative to the ecliptic plane was sampled, and we cannot
thus conclude whether we are dealing with a jet or a cone. Existing MHD models of the solar
wind – Mars interaction (e.g. Tanaka and Murawski, 1997) provide a rather cylindrically sym-
metric pattern of the solar wind proton flow around Mars, and thus one would be in favor of
the conical ENA jet. On the other hand hybrid models (e.g., Brecht, 1997b) show asymmetric
features of the proton flow in association with the interplanetary convection electric field direc-
tion (E = −VSW × BIMF, where VSW and BIMF are the solar wind velocity and the interplanetary
magnetic field, respectively). The analysis has not taken the electromagnetic direction into
account because of the absence of a magnetometer aboard Mars Express. However, since the
observations cover a long period (∼40 days), the distribution of the IMF direction can be con-
sidered to be random. Therefore, a conical ENA jet geometry is rather plausible, and Figure
6.10 is considered as a cross-section of the axi-symmetric geometry of the conical ENA jet.
The ENA count rate was found to decrease very rapidly (∼several tens of seconds) in two
thirds of the orbits, as shown in Figure 6.8. This implies that the source of the ENA jet/cone
Bow shock
IMB
Solar wind
stream line
Source of
ENA jet
ENA jet
(b) ENAs are not

detectable even
though the FOV
covers the source
(a)
region
ENAs are detectable
Figure 6.10: Geometry of the subsolar ENA jet. The subsolar jet can be detected when the sensor is
within it (case a). As soon as the spacecraft leaves the jet (case b) the ENAs cannot be detected even
though the instrument FOV covers the source region. From Futaana et al. (2006a).
should be confined in space. We do not have a clear explanation as to what controls its size
and directionality. Existence of such a compact and directional ENA jet also raises a question
with regard to the validity of current MHD models. In all MHD models, the subsolar point is
a stagnation point with almost zero bulk velocity. The associated ENA emissions should be
rather isotropic due to the high temperature of shocked solar wind. However the observations
indicate highly anisotropic emission with solar wind energy. MHD models may not be valid
in describing the Martian subsolar region, because the system size becomes comparable to the
proton gyro-radius there. Therefore, one needs more detailed 3-D hybrid models dedicated
particularly to this domain in order to investigate the physics of the sharp decrease of the ENA
signal.
What causes the short bursty enhancements in the ENA signal such as were observed at
11:10 UT, May 30 (Figure 6.8)? These ENA bursts cannot be explained by changes in the
observational geometry, which remained the same during the observation. The jet dynamics
are discussed in Section 6.2.
6.1.4 Summary
The NPD, an ENA sensor of the ASPERA-3 experiment on board Mars Express, detected in-
tense fluxes of ENAs emitted from the subsolar exosphere. Typical ENA fluxes were
(4-7) ×105 cm−2 sr−1 s−1 in the energy range of 0.3-3 keV/amu. These ENAs are likely to be
generated through charge exchange between the shocked solar wind protons and the Martian
exosphere in the subsolar region, where the solar wind plasma penetrates to its lowest altitude
and where the neutral gas density is high.

As the satellite moved away from Mars, the observed ENA signal decreased even though
the subsolar region remained within the NPD FOV. These decreases are gradual in one third
of the orbits, while the decreases are very sharp (on a scale of a few tens of seconds) in two
thirds of the orbits in question. This behavior can be explained if the spacecraft crossed a
spatially constrained ENA jet/cone, as shown in Figure 6.10. Such anisotropic ENAs indicate
that the solar wind flow around the subsolar region is also highly anisotropic, and this cannot
be explained in the frame of the existing MHD models.
6.2 Observations of the Martian subsolar ENA jet oscillations

6.2.1 Introduction
The initial ASPERA-3 / NPD observations revealed that an ENA flux is emitted anisotropically
from the subsolar region of Mars (Futaana et al. (2006a); Section 6.1). This section reports on
observed oscillations of the ENA jet (Grigoriev et al., 2006).
The data set, used in this analysis, consists of 38 orbits recorded by the NPD1 detector
during the period May 24 - July 1, 2004 when the orbit and the attitude of Mars Express
were optimal to investigate ENA emissions from the subsolar region of Mars. During this
observation period the NPD sensor was switched into BIN mode. Full spectrum accumulation
time was 1 s.
The basic characteristics of the ENA jet/cone established earlier (see Section 6.1) are as
follows: averaged energy ∼1.5 keV/amu, energy range is from 0.34 to 3.0 keV/amu. ENA
differential flux is estimated to be j∼(4-7)×105 cm−2 sr−1 s−1 .
6.2.2 Observation geometry

A typical Mars Express spacecraft orbital position (orbit 499, on June 11, 2004) during the
measurement cycle is shown in Figure 6.11. The orbital period during this time was about
6.5 hr. The NPD operation started ∼20 minutes after the periapsis and lasted for ∼30 minutes.
The sensors were operational while Mars Express was moving on the dayside outbound through
the magnetosheath region through the bow shock and the region of undisturbed solar wind.
Figure 6.11(a) is in the cylindrical coordinate system, described in Section 6.1.2, scaled in
Martian radii (R M ). Dashed curved lines show the MPB and the bow shock locations modeled
by Vignes et al. (2000). A dotted line shows the Mars Express trajectory during the observation
period. The orbital part where the NPD sensor was switched on (11:50 – 12:17 UT) is depicted
as a thick line; diamonds along this line represent time intervals of 10 min. Figure 6.11(b)
shows the spacecraft location projected on a YZ MS O plane. Mars Express crosses the ecliptic
plane, shown by a dot-dashed line, at the beginning of every measurement session.
Mars Express was nadir-pointing during the observations near the perigee. The plane where
ASPERA-3 was mounted was perpendicular to the nadir, facing the planet (Figure 6.1). In this
configuration, the symmetry axis between the channels Dir0 of the NPD1 and NPD2 detectors
was pointing towards the center of Mars.
The NPD1 FOV was covering the subsolar region of Mars. Figure 6.12(a) shows the sensor
observation geometry. In this plot, the spacecraft-centered ecliptical coordinate system of the
epoch 2000 is employed. The northern hemisphere (positive ecliptic latitudes) is shown on
the left side of the figure and the southern hemisphere (negative ecliptic latitudes) is shown on
the right side. The vantage point is the origin of the coordinate system. The Martian limb is
6.2. OBSERVATIONS OF THE MARTIAN SUBSOLAR ENA JET OSCILLATIONS 129
a) b)
ecliptic
plane
Figure 6.11: a) Mars Express trajectory on June 11, 2004 (orbit 499). The cylindrical coordinate system
based on the MSO reference frame is used. The thick line depicts the measurement interval. Dashed
lines are the modeled bow shock (BS) and magnetic pileup boundary (MPB, Vignes et al., 2000). b) Mars
Express trajectory, projected in YZ MS O plane (seen from the Sun direction). Adapted from Grigoriev
et al. (2006).
MEX North NPD FOV MEX South NPD FOV DOY 163
a) 2004−06−11 11:50:00 2004−06−11 11:50:00
[Scanner at 90 deg] [Scanner at 90 deg]
0 Sun 0 b)
Mars
330 30 330 30
30 MARS 30 MARS NPD1

300 60 300 60
0
60 60 NPD1 2 1
270 90 270 90
NPD2
240 120 240 120
210 150 210 150

180 180
Figure 6.12: a) NPD sensor observation geometry in the spacecraft-centered ecliptical coordinate sys-
tem taken for the epoch 2000. The Martian limb is shown by an ellipse. The elongated fans correspond
to both NPD FOVs. The black and dark grey polygons depict viewing directions Dir1 and Dir2, respec-
tively, of the NPD1 sensor. The filled dots show the location of the centers of the Sun (black) and Mars
(light grey). b) A sketch (not in scale) emphasizes Mars Express location and the viewing direction of
the NPD1 during the measurements. The black and dark grey numbered sectors correspond to the NPD1
channels Dir1 and Dir2 FOVs. A sun symbol shows the Sun direction. From Grigoriev et al. (2006).
shown by an ellipse. It is partially visible on both hemispheres. The elongated fans show the
NPD1 and NPD2 FOV. The black and dark grey polygons depict viewing directions Dir1 and
Dir2, respectively, of the NPD1 sensor. The filled dots show the location of the centers of the
Sun (black) and Mars (light grey). Figure 6.12(b) represents a sketch that clarifies the NPD1
observation geometry, emphasizing the sensor’s location and attitude during the measurement.
The black and grey sectors correspond to the detector channels Dir1 and Dir2 FOV. These
channels were observing the Martian subsolar region and therefore suited for investigation of
subsolar ENA jet emission.
6.2.3 ENA jet fluctuation observation

The energy-time spectrograms obtained by the sensor channels Dir1 and Dir2 are shown in
Figure 6.13(a, b). The data was recorded during 11:50-12:17 UT on June 11, 2004 (orbit 499).
The count rate is color coded. The color bar, on the right side, is scaled in counts/read-out/ns
or counts/s/ns, assuming 1 ns TOF resolution of the sensor’s electronics. The left vertical axis
gives the TOF of particles in ns. The right vertical axis is the energy scale, assuming particles
as hydrogen. The jet peak energy is ∼2.0 keV/amu in this case. One clearly sees oscillations
of the ENA flux.
a) cnt/s/ns
1.00
1000
Energy [eV / amu]

100
TOF, ns
Dir - 1
0.10
1000
100
10000 0.01
11:52 11:54 11:56 11:58 12:00 12:02 12:04 12:06 12:08 12:10 12:12 12:14 12:16
b) cnt/s/ns
1.00
1000
Energy [eV / amu]

100
TOF, ns
Dir - 2
0.10
1000
100
10000 0.01
11:52 11:54 11:56 11:58 12:00 12:02 12:04 12:06 12:08 12:10 12:12 12:14 12:16
c) 140
120
100
[count / s]
Dir - 1
80
60
40
20
0 11:52
11:52 11:54
11:54 11:56
11:56 11:58
11:58 12:00
12:00 12:02
12:02 12:04
12:04 12:06
12:06 12:08
12:08 12:10
12:10 12:12
12:12 12:14
12:14 12:16
12:16
140
d)
120
100
[count / s]
Dir - 1
80
60
40
20 ∆Τ
0 11:52 11:54 11:56 11:58 12:00 12:02 12:04 12:06 12:08 12:10 12:12 12:14 12:16
11:52 11:54 11:56 11:58 12:00 12:02 12:04 12:06 12:08 12:10 12:12 12:14 12:16
4/6/11 Time, UT
Figure 6.13: Panels a,b: TOF spectra of the NPD1 Dir1 and Dir2 channels. The corresponding energy
is shown in the right axis assuming the mass 1 amu. Horizontal blue lines specify the TOF window ξ
= 99 – 374 ns, which covers the ENA jet energy range. Panels c,d: integrated over the TOF window ξ
correlated count rate. Fast variations are identified by red arrows. Slow variations are depicted by thin
vertical black lines. Different low pass filter window is used: 40 s for panel c and 90 s for panel d. The
peak-to-peak period is marked as ∆T. From Grigoriev et al. (2006).
In these panels, UV photon induced background noise has been subtracted from the spec-
tra. As with any ENA instrument, NPD possesses a residual sensitivity to UV flux. This causes
a background in the correlated count rate, which does not depend on the TOF. Therefore, to
obtain the signal associated with ENAs only, we subtracted the UV-induced background, as-
sumed to be equal to the count rate at the last TOF bin, where no ENAs are expected due to
low sensitivity of the NPD sensor to low-energy ENA.
In order to investigate signal variations, we defined a TOF window ξ as follows: the ξ
consists of 6 TOF bins (99 – 374 ns), entirely covering the ENA jet energy range (0.34 - 5.1
keV/amu). The ξ window is marked by a pair of blue horizontal lines in panels (a) and (b).
Two panels, Figure 6.13(c, d), show the correlated count of the channel Dir1, integrated
over ξ (grey line). Thick blue lines show the data, smoothed using the low-pass filter with 40 s
(panel (c)) and 90 s (panel (d)) running averages.
As a general behaviour, the ENA jet intensity increases while the spacecraft enters the
ENA jet region and decays as the spacecraft leaves the ENA jet region. At the same time, some
periodic fluctuations of the signal appear during the first 10 - 12 minutes after the measurements
start. The zoomed-in view of these signal variations is shown in Figure 6.14. Arrows emphasize
the peaks. The peak-to-peak period is 150 - 180 s with the FWHM ∼50 s. The FWHM of a peak
is calculated between the peak maximum and a background reference level, which is estimated
separately for each peak. The ENA jet long-term intensity variation is thus taken into account.
The variations of the signal reach 30% of the maximum value. These are referred to as fast
oscillations.
As the spacecraft moves further, another type of signal variation occurs. Two high-intensity
peaks appear at approximately 12:04 and 12:08, as seen in Figure 6.13. These variations are
different from the fast oscillations described earlier. The peak-to-peak period, ∆T, is ∼270 s,
and the FWHM is about 110 - 140 s. The variation depth is about 50%. This type of long-term
signal variation is referred to as a slow variation.
90
80
70
[count / s]
Dir - 1
60
50
40
30
20 11:52 11:54 11:56 11:58 12:00 12:02
11:52 11:54 11:56 11:58 12:00 12:02

4/6/11 Time, UT
Figure 6.14: Typical flux variations due to fast oscillations. The picture is in the same format as
Figure 6.13(c). The black arrows show the peaks of the fast oscillations. From Grigoriev et al. (2006).
The peak-to-peak times of the slow variation events do not depend on the window size of
running average taken in a range of 20 to 120 s. That is reasonable, since a shorter running
window retains statistical fluctuations while a longer than 120 s window smoothes out the
oscillations. In case of the fast variations, changes in the steps of the running average length
within a range of 20 to 50 s do not affect the periodicity of the variations.
6.2.4 Statistics on the intensity variations

Criteria for selecting fast and slow variation events are as follows: In case of fast oscillations
signal depth amplitude variation must reach at least 20 - 30%. The number of peaks observed
during an observation period is at least 3, with ∆T unchanged. Peak-to-peak period ∆T varies
in the range of 40 - 180 s. In the case of slow variations, signal depth amplitude change is
∼50-80%. The number of peaks detected during an observation period is 2 or more, and ∆T is
∼300 s ±15%.
Figure 6.15 shows the statistical distributions of the peak-to-peak times for two different
groups of ENA jet flux intensity variations. The histogram bin steps are equal to 20 sec and
60 sec for fast and slow oscillations, respectively. Each distribution is normalized to the maxi-
mum of the occurrence.
fast slow
∆Τ
Figure 6.15: The peak-to-peak ∆T period distributions of both fast (dark-grey colored) and slow (light-
grey colored) groups are shown. From Grigoriev et al. (2006).
The fast and slow oscillation distributions are shadowed by dark grey and light grey, res-
pectively. The peak of the fast oscillation distribution is at ∼50 sec and FWHM is ∼30 s. These
variations are periodic and are detected in every orbit. The signal peak-to-valley ratio can
reach 20 - 30% of the maximum flux intensity. The peak of the slow oscillation distribution is
at ∼300 s and FWHM is ∼30-80 s. The variations of the signal reach 60 - 80% of the maximum
value. The slow variations are found in 15 of 38 orbits.
The distribution of FWHM of the signal peaks in the case of slow ENA jet variations is
plotted in Figure 6.16. The histogram bin step equals to 40 s. Evidently, the majority of the
detected peaks has FWHM of ∼160 s. The shape and location of the peaks are highly variable
from orbit to orbit. The number of peaks detected in one orbit is generally two – they occur in
30% of all orbits. Multiple peaks (three or more) appear only in 10% of the orbits.
Figure 6.17 shows the peak occurrence dependence on the angle between the −Y-axis and
the position of the Mars Express spacecraft, defined as ϕ (positive if counting clockwise).
There is no preferable direction of the intensified ENA flux. The peaks are detected with equal
probability from different vantage points within the NPD observation region.
6.2.5 Discussion
NPD detected intense ENA fluxes with energies in the range 0.3-3.2 keV/amu when the FOV
of NPD was pointed toward the subsolar region (Futaana et al., 2006a). In all cases, the ENA
flux intensity variations were detected. These variations can be categorized into two modes,
namely fast and slow variations. The fast variations occur with a typical period of 28.5-200 sec
and are detected on each orbit. The slow variations have a period of 143-500 sec and occur in
40% of the cases. We now discuss the possible generation mechanisms of the ENA jet intensity
variations.
The MGS magnetometer team (Espley et al., 2004) reported low-frequency (period ∼25 s)
magnetic field oscillations in the different plasma regions of Mars. One possible interpretation
of the source of oscillations on the dayside of the magnetosheath is mirror mode instabilities.
The important characteristic of the mirror modes is an anticorrelation between the magnetic
field strength and the plasma density. Therefore, these kinds of magnetic field oscillations can
represent one of the mechanisms of ENA jet fast oscillation generation. This is because the
Figure 6.16: The distribution of FWHM of peaks Figure 6.17: Location of the spacecraft during
in case of slow ENA jet variations. FWHM of ENA peak detection and its occurrence is shown.
a peak is expressed in seconds (X-axis). Y-axis Angular distance between the −Y-axis and the
marks the relative occurrence of cases. The his- Mars Express spacecraft position is shown by ϕ
togram binning size is 40 s. From Grigoriev et al. (see text for details). ϕ is in degrees (X-axis).
(2006). Y-axis marks the peak occurrence. From Grig-
oriev et al. (2006).
change of plasma density in the magnetosheath is reflected by the change in ENA production-
rate through the change in the occurrence of the charge exchange process. However the pre-
dominant period of the main fluctuation of the magnetic field seems to be shorter than the
period of the ENA fast oscillations (28.5-200 sec), derived from the NPD observations. The
mirror-mode instability is thus a less plausible mechanism to explain the fast oscillations.
Another possible mechanism is the global oscillations of the IMB. Futaana et al. (2006c)
discussed an example of a global response of the Martian plasma environment to an interplane-
tary shock. They used an ENA jet observation that shows an extremely intensive flux followed
by a sudden decrease of the flux intensity. Using simultaneous plasma data, they concluded
that the decrease is caused by the fast IMB displacement due to an interplanetary shock. They
also reported oscillations of the ENA jet flux (period ∼1 min) just after the abrupt decrease,
regarded in this study as fast oscillations. However, to the best of our knowledge, there have
been neither theoretical nor experimental reports on global oscillations of the Martian plasma
obstacle.
Two peaks in the ENA jet fast oscillation distribution are apparent in our data set: ∼50 and
∼80 sec (Figure 6.15). Remarkably, these periods coincide with the typical time periods of the
electron flux intensity oscillations observed by the ELS / ASPERA-3 instrument (between 50
and 100 s, Winningham et al., 2006). Such oscillations of electron fluxes have been observed in
different regions of Mars, but the dayside magnetosheath is one of the most active regions. The
frequency range encompasses the typical O+ ion gyro-frequency in the magnetosheath (Espley
et al., 2004). Cyclotron instability of the newly created planetary ion beams generates Alfvén
waves with a frequency close to the local gyro-frequency (in the spacecraft frame of reference).
The associated plasma density variations result in ENA flux oscillation. While it seems that the
observed fast ENA oscillations are connected to the detected electron flux oscillations (plasma
is quasi-neutral), it is not known whether the wave energy is sufficient to cause 20-30% plasma
density variations.
The observed slow variations have a characteristic peak-to-peak time ∼300 s. As shown
in Section 6.2.4, the number of peaks in the case of slow ENA flux variations detected in one
orbit is typically 2. It is rare for more peaks to occurs. Such differences between the ENA jet
flux fluctuation signatures imply that these two types of fluctuations have different production
mechanisms.
It is important to emphasize that the observed variations can reflect either temporal vari-
ations of the generation region or spatial structure of the jet itself. Because of the spacecraft
motion we cannot resolve this ambiguity. In the case of the temporal jet variations, such be-
havior can be explained by bursty intensification of the ENA source as a result of the temporal
changes in the upstream conditions, mostly the solar wind dynamic pressure. Holmström et al.
(2002) investigated the deflected solar wind ENA flux dependence on the solar wind and exo-
sphere parameters. They concluded that the exospheric density is the most important parameter
controlling the ENA flux. The IMB position is dependent on the exospheric density. Hence,
if the ENA generation region moves closer to the planet, then the solar wind ions can reach
denser part of the Martian exosphere, creating higher ENA flux.
The waves on the IMB resulting from Kelvin-Helmholtz instability could, in principle,
cause the observed oscillations. However Penz et al. (2004) considered the development of
the Kelvin-Helmholtz instability on the Martian boundary in one-fluid approximation and con-
cluded that the subsolar ionopause of Mars is stable with respect to the Kelvin-Helmholtz
instability.
The vantage points at the moments when the NPD1 sensor detected peaks of slow ENA
jet intensity variations are distributed uniformly over the whole observation region, as shown
in Figure 6.17. Meanwhile, the observation period is long (∼40 days) and therefore the distri-
bution of IMF direction and magnitude, as well as solar wind parameters, can be considered
random. It is very likely that the ENA generation region is very sensitive to the exterior and
is changeable under unsteady solar wind dynamic pressure and IMF direction. Unfortunately,
lack of undisturbed solar wind measurements during the NPD observations (single point mea-
surements) as well as in situ IMF measurements prevent us from making such a conclusion.
There is also a certain possibility that the slow variations of the ENA jet represent spatial
structures of the jet itself. The majority of the detected peaks have the FWHM about 160 s,
corresponding to 4-5◦ of an angular width of the ENA stream, taking into account the spacecraft
velocity ∼3 km/s and the distance to the generation region 6 2R M . If we assume existence of
ENA "sub-jets" with an angular width of 4-5◦ , the observed slow variations can possibly be
explained. This concept is shown in Figure 6.18. Such ENA sub-jets are likely to appear
if the ENA generation source under certain conditions becomes patchy. What can cause the
ENA generation source patchiness? Existence of small-scale magnetic field structures called
magnetic flux ropes (e.g., Russell and Elphic, 1979; Elphic and Russell, 1983) is one possibility
for such effects. The existence of these flux ropes at the Martian ionopause has been reported
by Vignes et al. (2004). The magnetic character of the flux ropes at Mars appears chaotic.
At the same time, these magnetic structures are stationary in time relative to the speed of the
spacecraft. The width of the flux ropes is of the order of a few tens of kilometers. Most of the
flux ropes have been detected at high solar zenith angles. A smaller number of flux ropes has
been identified for solar zenith angles lower than 20◦ (Vignes et al., 2004). Meanwhile, size of
the region of the most intensive ENA generation was estimated to be ∼6400 km (Futaana et al.,
2006c), i.e., within solar zenith angles lower than 40◦ . So, the existence of magnetic flux ropes
is likely to affect the transport of planetary protons to the magnetosheath and consequently,
the ENA production. Again, the mechanism of ENA formation through charge exchange of
accelerated protons in the magnetosheath (Lichtenegger et al., 2002) is one of the most probable
ENA jet generation mechanisms (Futaana et al., 2006a).
More detailed study of the ENA jet behavior should be the subject of future investigations.
The next step would be to study the morphology of the ENA jet as a function of the upstream
MPB BS
solar wind
stream line
ENA jet
00
11
11
00
00
11
00
11
00
11
Source of
1010
00
11
00
11
00
11
11
00
ENA jet(cone) 00
11
00
11
00
11
00
11 1010
00
11 101011111
00
11 00000
00
11
000000
111111 0
1 00000
11111Detected ENA
00
11
000000
111111 0
1 00000
11111 oscillations(subjets)
00
11
000000
111111 101011111
00000
00
11
000000
111111 00000
11111
00
11
000000
111111 101011111
00000
00
11
00
11 10 ∆Τ
00
11
0010
11
ENA jet
Figure 6.18: The schematic illustration of spatially resolved ENA sub-jets. From Grigoriev et al.
(2006).
conditions such as the solar wind dynamic pressure and IMF direction. But to resolve the
nature of the ENA jet oscillations and related dynamics of the Martian induced magnetosphere
the key investigation would be multipoint measurements at Mars. These would allow us to
make correct correlations of in situ measurements of the local plasma parameters and remote
observations of the ENA flux emerging from the interaction region.
6.2.6 Summary
We presented observations of the Martian subsolar ENA jet intensity oscillations, made by
NPD. We categorized the flux intensity variations into two groups. First group is fast oscilla-
tions with a characteristic peak-to-peak time ∆T ∼50 s. These flux variations possess a periodic
structure and are observed in all orbits. The second group is slow variations, whose time scale
is of the order of ∼300 s. These variations appear in 40% of all ENA jet observations.
The fast oscillations are possibly formed by low frequency ion waves (period ∼50-100 sec)
at the subsolar region. The observed period correlates with electron oscillation period detected
by ELS which, in turn, is close to the oxygen gyro-period. However the energetics of the
process is not clear.
The origin of the slow variations is a mystery. They might be related to temporal variations
of the solar wind dynamic pressure. On the other hand, the slow oscillations can also be
explained by a sub-jet structure as shown in Figure 6.18. Such structures could possibly appear
due to the existence of magnetic flux ropes in the Martian ionosphere.
6.3 Other results by the ASPERA-3 / NPD
6.3.1 Global response of Martian plasma environment to an interplanetary struc-

ture: From ENA and plasma observations at Mars
The paper by Futaana, Barabash, Grigoriev et al. (2006c) presents a case study of the subsolar
ENA jet (Futaana et al., 2006a). The detailed investigations of the subsolar ENA jet would
advance the understanding of the dynamics involved in the interaction between the solar wind
and Mars. The event considered was interpreted as a global response of the Martian plasma
environment (in a time scale of ∼10 s) to a change in the solar wind conditions.
The measurements were conducted by the NPD sensor on June 7, 2004. The NPD data is
treated together with in situ data of the ELS and IMA spectrometers. Figure 6.19 shows the
NPD TOF spectrum (upper panel) and the ELS electron energy spectrum and fluxes (middle
and lower panels), respectively. Three notable signatures of this unique event (at 13:58 UT)
are discussed. The ENA flux is ∼ 5 times higher than the nominal subsolar ENA jet flux of
4-7 ×105 cm−2 sr−1 s−1 (Futaana et al., 2006a). The jet outer boundary is very abrupt. It took
∼10 s for Mars Express to cross this boundary, while the typical time scale is about 1 min. Ad-
ditionally, there are three quasi-perioduc flux enhancements observed right after the spacecraft
left the jet region.
Figure 6.19: The observations from the NPD and ELS. Panel (a) is the count rate of the subsolar ENA
flux integrated over the TOF window (50-1900 ns) for the directions Dir1 (green) and Dir2 (blue line) of
the NPD1. Panel (b) shows the energy-time spectrogram observed by the ELS. Panel (c) shows the time
series of counts within the energy range 10-50, 50-130, 150-500, 500-2000 and 2000-15000 eV from
the ELS. The ENA flux is several times higher than the nominal subsolar ENA jet flux. Abrupt depletion
of ENA fluxes occurred at 13:58:40 UT. Following are quasi-perioduc variations in flux observed right
after the spacecraft left the jet region. Adapted from Futaana et al. (2006c).
6.3. OTHER RESULTS BY THE ASPERA-3 / NPD 137
This unusually high ENA flux can be interpreted as a result of a compressed Martian plasma
obstacle. During this time interval, the Mars Global Surveyor (MGS) magnetometer recorded
a strong increase of magnetic pressure inside the IMB. This means that the solar wind dy-
namic pressure outside the IMB increased by several times during this event, assuming that the
magnetic field pressure in the induced magnetosphere region balances the incident solar wind
dynamic pressure (Crider et al., 2003). The BS and IMB locations, derived from the plasma
spectrometers data, were much closer to the planet than the average locations (Vignes et al.,
2000).
The abrupt flux change is most likely to be caused by a quasi-perpendicular interplanetary
shock, which led to a reconfiguration of the plasma obstacle shape and location, as illustrated
in Figure 6.20. Consequently, the ENA jet generation region was relocated: it moved closer to
the planet.
It has been suggested that the quasi-periodic ENA jet flux enhancements are due to a global
vibration of the Martian plasma obstacle, caused by the interplanetary shock hit. Such vibra-
tions can result in ENA flux variations as the ENA jet generation region displaces back and
forth.
From the time difference between the changes of the Martian plasma obstacle and the ENA
jet signal, the Martian subsolar ENA jet generation region size is estimated to be as large
as ∼2R M .
Figure 6.20: Illustration of an interpretation of the interplanetary shock interaction with the Martian
plasma environment: (a) before and (b) after the shock crossing of the planetary ENA jet source. The
source region of the subsolar ENA jet, which is very close to the IMB, was pushed toward the planet
due to the high dynamic pressure of the interplanetary shock. The spacecraft went out of the subsolar
ENA jet very quickly due to this reconfiguration of the Martian plasma obstacle. From Futaana et al.
(2006c).
6.3.2 The Hydrogen Exospheric Density Profile Measured with

ASPERA-3 / NPD
Galli et al. (2006c) discussed the exobase temperature and density of Mars during a period of
low solar activity, derived from the NPD measurements of Lyman-α limb emission from the
exospheric hydrogen of Mars. These parameters are crucial for the models of atmospheric loss
and ENA production (e.g., Lichtenegger et al., 2002; Barabash et al., 2002). The hydrogen
exobase density and temperature were estimated to be nH = 1010 m−3 , and T > 600 K, respec-
tively.
The analysis was based on the Lyman-α limb emission measurements at high altitudes.
NPD was switched into the TOF mode during the measurements. The data obtained during
the observation period (12:00 – 13:10 UT, on April 25, 2004) were integrated over 5 minutes
intervals and a height of the UV-induced background in the TOF spectra for each measurement
was retrieved. The NPD count rates, caused by UV photons, were converted into fluxes of
UV emission of the neutral hydrogen, integrated along the line-of-sight through the Martian
exosphere. The NPD sensor sensitivity to Lyman-α was calibrated using the all-sky UV map,
based on the UV measurements by the Solar Wind Anisotropies (SWAN) experiment on the So-
lar and Heliospheric Observatory (SOHO) (NASA, 2006, http://sohowww.nascom.nasa.gov).
Consequently, the non-planetary UV background was subtracted from the TOF spectra.
The measurements were interpreted by means of a numerical exosphere model, which
yields a hydrogen density profile nH (r) for a given exobase density and temperature. It is a
one-dimensional Monte-Carlo model (Wurz and Lammer, 2003), for which the exobase den-
sity and temperature are the two free parameters. The model assumed an exobase location at
220 km (Lichtenegger et al., 2006) above the surface, a fully thermalized hydrogen atom pop-
ulation at the exobase and a spherically symmetric distribution of neutral hydrogen with one
single constant temperature. The modeled density profile was related to UV emission by means
of a radiation transport equation. The UV emission was assumed to be entirely due to resonant
scattering on hydrogen atoms.
The model was verified by evaluating the measurements by the Mariner 6 and 7 space-
craft (Anderson and Hord, 1971). The values of exospheric density and temperature derived
with the Monte-Carlo model (Wurz and Lammer, 2003) were compared with results obtained
by Anderson and Hord (1971). A good agreement within a factor of 2 was shown.
Afterwards, the model was used to evaluate the data obtained with the NPD sensor. A single
hydrogen component exobase was assumed. The resulting exobase density of nH = 6 × 109 m−3
was found to be 5 times lower and the temperature of T < 1000 K to be 2 times higher if com-
pared with values derived by Anderson and Hord (1971). Figure 6.21 shows the measurements
from both the NPD (from 2004) and Mariner 6 and Mariner 7 (from 1969) together with the
modeled profiles. The difference can be explained either by different observation geometry or
by asymmetry in the exobase due to high spatial variability (Holmström, 2006) or just by a
generally thinner exobase.
So high exospheric temperature derived from the NPD measurements appears to be in-
consistent with exospheric temperatures (< 250 K) inferred from aero-braking maneuvers of
recent spacecraft like Mars Odyssey (Lichtenegger et al., 2006). This difference can be ex-
plained by the presence of a cold bulk component and a tenuous extended hot population of
atoms that have been heated up in photo-dissociation and dissociative recombination processes
(Lichtenegger et al., 2006). If the two-component exobase approach is valid, then it is possible
to assume that the Lyman-α airglow measurements are mostly affected by the hot component,
while the measurements performed during aero-braking are affected by the hydrogen cold com-
ponent of the exobase.
A two-component exobase approach was employed for evaluation of the NPD data; i.e.,
an exobase with two hydrogen components of distinct density and temperature was assumed.
Figure 6.21: Measured column densities of neutral hydrogen (symbols) plotted against tangential height
hi . The thick lines represent modeled temperature and the exobase density. The most notable difference
between the NPD measurements from 2004 (triangles) and the Mariner 6 data (rectangles) and Mariner
7 data (diamonds) is the diminished exosphere density. From Galli et al. (2006c).
A linear combination of two one-component models with different parameters was used. Re-
sulted hot component remained the same as in the single component case, while a cold com-
ponent was calculated to possess the exobase density nH,cold = 1.2 × 1010 m−3 and temperature
Tcold = 180 K.
A theoretical model (Krasnopolsky, 2002) of the Martian exosphere yields exobase densi-
ties of nH = 2 × 1011 m−3 for a low solar activity, which is more than one order of magnitude
away from the NPD observation. It is concluded that the limb emission measurement is do-
minated by a hydrogen component that is considerably hotter than the bulk temperature at the
exobase, and the Martian hydrogen exosphere is considerably thinner than previously modeled
for low solar activity (e.g., Krasnopolsky, 2002).
6.3.3 Energetic Hydrogen and Oxygen Atoms Observed on

the Nightside of Mars
The paper by Galli et al. (2006a) presents measurements of hydrogen and oxygen ENAs on
the nightside of Mars, performed by the NPD sensor. A global picture of the Martian ENA
environment can improve our knowledge of interaction processes between the solar wind and
the Martian atmosphere, as well as allowing us to infer the physical properties of the exosphere
and to quantify atmospheric loss. Energy spectra of different H-ENA populations and the maps
of H-ENA outflow are presented. The differential flux of H-ENAs can reach 105 cm−2 sr−1 s−1 .
The global atmospheric loss rate of hydrogen and oxygen due to production of ENAs is esti-
mated to be < 1 g/s.
In this study only H and O species are considered due to their abundance and because
the sensor favors detection of them. The study also attempts to distinguish between planetary
and solar wind hydrogen ENAs by comparing the ENA differential intensity energy spectra
Figure 6.22: The NPD2 measurements (channel Dir2) on the nightside of Mars at the IMB, with FOV
directed close to the tangential direction to the IMB surface. The measured integral intensity reaches
several 105 cm−2 sr−1 s−1 . The measured (thin line) and the reconstructed (bold line) TOF spectrum is
shown in upper panel. The lower panel shows the differential intensity energy spectrum (solid line) that
corresponds to the reconstructed TOF signal, and a two-component power law distribution fit (dashed
line). From Galli et al. (2006a).
measured in the region inside the IMB, at the IMB, in the subsolar region of Mars and in the
magnetosheath region (X = -2.5 R M ).
Figure 6.22 shows the TOF and energy spectra obtained by the NPD2 sensor when the
spacecraft was crossing the IMB. The position of the IMB refers to the plasma model of Kallio
et al. (1997). Spectra of the tailward ENAs, obtained inside the IMB and at the IMB, can be
approximated by two-compoinent power law with a roll-over at 1.1 keV (dashed line). The
spectrum of the dayside solar wind can be parametrized by the Maxwell-Boltzmann distribu-
tion. Unfortunately, even though the energy spectrum of the nightside ENAs is much broader
than that of the solar wind ENAs, it was found to be impossible to distinguish between the
contributions of the solar wind and planetary protons to the typical ENA spectrum in general.
The reason is that the ENA energy spectrum is a convolution of charge exchange processes
along the line-of-sight intersecting several regions of varying temperature, density and flow di-
rections. In the case of measurements in the magnetosheath (Figure 6.23), the energy spectrum
has a bi-modal structure with a peak at 0.35 keV reproducible as slowed down and therma-
lized, neutralized solar wind with a thermal spread of 20 - 200 eV and bulk velocity < 300
km/s (dotted line in the energy spectrum). The other part is likely to be approximated by a
two-component power law fit with a roll-over at 2.8 keV (dashed line). That part of the energy
spectrum is concluded to be due to the planetary pick-up protons.
Figure 6.23: The NPD1 measurements (channel Dir2) in the magnetosheath. The measured (thin line)
and the reconstructed (bold line) TOF spectrum is shown in upper panel. The lower panel shows the
reconstructed differential intensity energy spectrum that corresponds to the reconstructed TOF signal.
The dashed and dotted lines are fit curves for a maxwellian and a two-component power law distribution.
From Galli et al. (2006a).
The ENA measurements were used to derive global ENA maps to be compared with theo-
retical models of solar wind ENAs (Gunell et al., 2006) and of planetary ENAs (Lichtenegger
et al., 2002). The measured fluxes were found to be one or two orders of magnitude lower than
the predicted ones. Such disagreement can result from the difference between the present exo-
sphere properties and those assumed for the hydrogen exosphere models. As input parameters,
the models used the hydrogen exobase density and temperature, predicted by Krasnopolsky
and Gladstone (1996) for solar minimum, namely nH = 1012 m−3 and T = 200 K. The recent
UV Lyman-α limb emission measurements by ASPERA-3 / NPD indicated much thinner and
hotter hydrogen exosphere (Galli et al., 2006c), i.e., nH = 1010 m−3 and T > 600 K. The spatial
distribution and the integral intensity of the solar wind model from Gunell et al. (2006), being
recalculated with hotter and less denser exobase, match the observations. The ENA flux inten-
sity of a few 105 cm−2 sr−1 s−1 is consistent with solar wind ENAs. Therefore, it was concluded
that the majority of the ENAs detected inside the IMB are due to the solar wind protons; the
derived ENA maps give a representative image of the ENA flow at the Martian nightside for
solar minimum conditions. Model calculations (Lichtenegger et al., 2002) also indicate that
the ratio of planetary-to-solar wind ENA fluxes should not exceed 20% at any time during the
solar cycle.
Simulations predict significant oxygen ENA flow in the Martian tail (e.g., Barabash et al.,
2002; Gunell et al., 2006). In contrast to theoretical predictions, no intense oxygen ENA signals
over the integral intensity of 2 × 104 cm−2 sr−1 s−1 (detection limit) have been detected. This is
one-two orders of magnitude lower that the simulations.
Finally, the total ENA production rates for both hydrogen and oxygen at Mars are computed
to be 2-3 ×1024 s−1 and <1022 s−1 , respectively. Consequently, the corresponding atmospheric
loss rate is estimated to be less than 1 g/s. It is comparable with an ion escape rate of a few g/s,
as reported recently by Barabash et al. (2007a).
6.3.4 First ENA observations at Mars: ENA emissions from the Martian upper
atmosphere
The paper by Futaana et al. (2006b) reports on the observations of backscattered H-ENAs (so-
called ENA albedo), detected for the first time near the subsolar point of Mars. Knowledge
about the particle flux and the energy flux of the backscattered ENAs provides information
concerning energy and momentum transport from the solar wind into the Martian upper atmo-
sphere. Basic characteristics of the observed ENA albedo are: H-ENAs are emitted from the
broad region of Martian upper atmosphere, within solar zenith angles of 30◦ in the energy range
of 0.2 - 2.3 keV with a peak energy of ∼1.1 keV, differential flux is ∼1.5 - 2.0 ×106 cm−2 sr−1 s−1 ,
energy flux is ∼ 9.5 ×109 eV cm−2 s−1 .
The precipitating particle flux generating this ENA albedo was estimated to be about 0.94 −
1.26 × 107 cm−2 s−1 . The distribution and the energy of the ENA albedo was found to agree well
with the results of a Monte-Carlo simulation by Kallio and Barabash (2001). The simulated
flux is, however, lower by a factor of 3-4 than the detected one, likely due to different solar
wind conditions, e.g., higher solar wind flux. Solar wind conditions were estimated from the
IMA / ASPERA-3 in situ measurements in the shocked solar wind as 1.3 - 1.8 times higher
than those used by Kallio and Barabash (2001) in simulations. At the same time the derived
value could be overestimated, if the MHD shock theories are considered. The lower altitude of
the IMB at Mars may also result in higher ENA albedo flux. Finally, part of the shocked solar
wind protons can penetrate below the IMB due to finite gyro-radius, interact with the dense
part of the exosphere and be scattered back as neutrals, contributing to the ENA albedo. It was
shown that evaluation of the effects of this mechanism (also known as proton–ENA albedo)
requires more precise models of the interaction between the solar wind and Mars.
The ENA albedo maps were suggested as a tool to investigate the distribution of the precip-
itating flux. Such maps can provide information concerning energy and momentum transport
from the solar wind into the Martian upper atmosphere.
6.3.5 Direct Measurements of Energetic Neutral Hydrogen in the Interplanetary

Medium
Galli et al. (2006b) reports on the NPD sensor observations of interplanetary ENAs within the
energy range 200 eV to 10 keV over time range of 10 months, during the Mars Express cruise
phase and also in orbit around Mars. Imaging these ENAs and deriving their energy spectra
would reduce our uncertainties about the interaction of the heliosphere with the local interstellar
medium. The paper also presents and discusses the energy spectra of the non-planetary ENA
signal as well as the ENA flux spatial distribution.
In order to select only measurements of ENAs that have no Martian origin, observations
with the NPD FOV as far as 45◦ away from the Martian limb are considered. Several tens of
energy spectra in the data set were identified to attribute to energetic hydrogen atoms originat-
ing in the inner heliosheath. The spatial flux distribution in general supports this conclusion.
The following analysis showed that these ENAs are observed for all ecliptic longitudes with
a typical flow direction between 90◦ and 180◦ away from the direction of the interstellar gas
flow. The integrated ENA fluxes are found to be enhanced by a factor of two in the anti-apex
direction.
The energy spectra of the detected ENAs are all very similar and are best described by a
two-component power law. Figure 6.24 shows a typical energy spectrum of ENA flux of non-
planetary origin. On average the energy spectra have a roll-over at E = 770 eV, a first and a
second slope can be parameterized as f (E) ∼ E −1.61 and f (E) ∼ E −3.31 , respectively. The ENA
flux integrated over an energy range of 0.2 to 10 keV varies on a time scale of months in a
range of 5 ×103 to 1.5 ×105 cm−2 sr−1 s−1 . The lower level of the flux intensity is defined by the
detection threshold.
Interference
Figure 6.24: Typical example of a non-planetary ENA signal, measured during cruise phase. The
top panel shows the original TOF spectrum (averaged over 86 minutes of data) and the reconstructed
TOF signal after background removal and noise filtering. Narrow lines at TOF bin 112 and 229 are
spikes introduced by the sensor electronics. They were removed from the TOF spectrum before further
analysis. The bottom panel shows the energy spectrum calculated from the reconstructed TOF signal
above, and a fit to this spectrum (dotted line). From Galli et al. (2006b).
Several alternative sources of the observed signal were considered. Solar wind-induced sig-
nal is ruled out, as the spectra are obtained when the NPD FOV was away from the Sun. Also,
similar spectra were observed in the Martian eclipse. Planetary origin ENAs have a roll-over of
the spectra above 1 keV and fluxes, reaching a few by 106 cm−2 sr−1 s−1 (see Section 6.3.3) and
are, therefore, disregarded. Neutralized ions of co-rotating interaction regions (CIR) cannot
be taken into account either, as the ion fluxes needed in the CIRs to account for the observed
ENAs fluxes are three orders of magnitude higher than actual CIR fluxes. Signals related to UV
bright stars cannot be related with these interplanetary ENAs, as the background spectra caused
by UV photons are very different from the particle signal spectra, and we found no correlation
of the ENA signal with nearby UV bright stars coming into the FOV of the NPD sensor.
The similarity of the energy spectra can be due to smearing of the temporal variations in
the ENA flux at the location of origin, i.e., the termination shock, by the time they arrive at
Mars. The traveling time over such distances can be expected to vary from tenths of a year to
several years.
6.3.6 Energetic Neutral Atoms from the Heliosheath

Wurz et al. (2006) continues analysis of the interplanetary ENA observations, reported by Galli
et al. (2006b), following the idea of heliospheric origin of the ENA signal. Figure 6.25 shows
energy spectra obtained by the NPD together with the data sets from the HENA instrument
on IMAGE (Kallenbach et al., 2005) and the High-Energy Suprathermal Time-of-Flight sen-
sor (HSTOF) of the Charge Element and Isotope Analysis System (CELIAS) on the SOHO
(Hilchenbach et al., 1998). All three data sets showed good agreement in a comparison of
energy spectra.
The ENA measurements were also compared with the theoretical heliospheric models (Mc-
Comas et al., 2004; Fahr and Scherer, 2004; Chalov et al., 2003), which predict energy spectra
of ENAs produced in the heliospheric interface region and travelling inward to Earth orbit.
Interestingly, all these models predict ENA fluxes which are one – two orders of magnitude
lower than the measured ENA fluxes. So far, none of the considered models can be chosen to
describe the observed data.
Figure 6.25: Modeled H-ENA energy spectra together with the measurements. The red dashed line
and the dashed-dotted black line are from McComas et al. (2004) for a strong and a weak termination
shock (TS), respectively; the dotted blue line is data from a two-shock model (Zank, private comm.);
the dash-dot-dot-dot green line shows data from Fahr and Scherer (2004); and the long-dashed orange
curve shows data from Chalov et al. (2003). Data points are from the NPD sensor of ASPERA-3 on
Mars Express (green triangles), from the HENA instrument on IMAGE (red dots), and from the HSTOF
sensor of the CELIAS instrument on SOHO (open squares). From Wurz et al. (2006).
Chapter 7
Summary and future prospects
The thesis is focused mainly on the development of NPD instrumentation. In chapter 1 it re-
views the ENA basics including ENA production mechanisms and classification, and the ENA
environments of Mars and Venus. It describes different ENA sources at Mars, namely neut-
ral solar wind, shocked solar wind, accelerated planetary protons, back-scattered ENAs (ENA
albedo), and sputtered atmospheric neutrals. The Venus ENA environment resembles that of
Mars, though the different masses of the planets result in different scale heights. Consequently,
the exospheric density profiles of the planets are different, making ENA fluxes at Venus gene-
rally lower than at Mars. NPD and NPI are the first instruments to conduct investigations of
the ENA environment at Mars and Venus.
Chapter 2 describes the principles of ENA detection along with ENA detecting instrumen-
tation. It presents different concepts to measure ENAs in different energy ranges. The new
concept developed for NPD utilizes the surface interaction technique, making it possible to
partially cover MENA and LENA energy ranges. The low energy detection limit of the sensor
is defined as >100 eV because of the interaction of an incoming ENA with two surfaces.
Chapter 3 describes the ASPERA-3 and ASPERA-4 plasma and neutral particle packages,
comprising the NPD sensors. Each package is a compact, highly integrated instrument inclu-
ding a scanner that allows sufficient coverage even on a 3-axis stabilized spacecraft platform.
An internal digital processing unit provides data acquisition, preprocessing, compression, and
packaging of telemetry packets.
Chapter 4 presents the NPD development process. Sensor integration, verification and tes-
ting were performed at IRF, Kiruna, Sweden. The chapter also describes the NPD mechanical
design. It is important to emphasize that the sensor is of low density design in comparison
with other sensors with similar characteristics, due to empty space in the field-free zone and
especially in the electronics compartment. The latter can be used to accommodate a high
voltage power supply internally, so the sensor can be used as a stand-alone unit for future
missions. The NPD data is of a new type, requiring development of dedicated software. A quick
look data display ’DVP’ with a graphical user interface based on Interactive Data Language
(IDL) was created. It is described in Appendix A.3.
Chapter 5 gives the calibration results for the NPD sensors. The sensors have a compara-
tively large geometrical factor with efficiency reaching 16% for ENAs of an energy of 6 keV.
NPD is capable of distinguishing between low energy oxygen and high energy hydrogen atoms
by TOF analysis only. The absence of high fluxes of high energy (> 6 keV) O-ENAs at Mars
makes this feature very important. The Start surface conditioning was shown to be not so
critical.
145
146 CHAPTER 7. SUMMARY AND FUTURE PROSPECTS
The sensor possesses an open architecture, hence UV photon suppression is based on the
Start surface coating optimization along with the TOF coincidence scheme. Eventually, the
background in TOF spectra induced by UV photons leads to a decrease in the signal-to-noise ra-
tio. However this feature can be used, and NPD can operate as a UV detector, if it is absolutely
calibrated with an UV source, most promising against hydrogen Lyman-α. NPD was in-flight
UV calibrated against interstellar sources and well-known Lyman-α fluxes scattered on inter-
planetary hydrogen. The UV photon rejection factor of the sensor was estimated to be ∼10−8 .
Chapter 6 shows the first scientific results at Mars. For the first time several ENA sources
were diagnosed, i.e., H-ENA jet/cone, back-scattered ENA (H-ENA albedo), interplanetary H-
ENAs. The Martian ENA jet/cone dynamics were also investigated and jet oscillations with
two different frequencies were identified. Again, due to single point measurements and the
spacecraft movement, it is not possible to resolve an ambiguity whether the flux has temporal
or spacial variations. To counter this, remote vantage points not affected by the spacecraft
motion would be the ideal place for planetary ENA imaging on future missions. Spectra of the
different ENA populations were obtained in various regions of Mars, including the subsolar
region, the magnetic pile-up boundary, the magnetosheath, and the eclipse region. Martian
exobase profiling was also performed using NPD as a UV detector. Correlation of the NPD
measurements with ELS and IMA spectrometers data allowed us to investigate the dynamics
of the entire Martian plasma environment. We confirmed that the latter is sensitive to the
upstream solar wind conditions.
Finally, we summarize the main lessons learnt following the NPD design, development,
operations, and data analysis. The ENA detection concept based on the surface interaction
processes is well suited for neutral particle detection in an energy range partly covering the
LENA and MENA regions, i.e., 100 eV – few keV. There is a very promising possibility of
distinguishing between hydrogen and oxygen ENAs by means of TOF analysis only (in a cer-
tain energy range), in particular keeping in mind that no other techniques (e.g., pulse height
analysis) are required. The UV photon-related count rate results in a rise of a background level
in TOF spectra. Direct exposure to the Sun and Venus dayside limb saturates the Start and Stop
counters. Therefore, the Sun and Venus surface avoidance techniques must be implemented to
avoid a premature aging of MCPs. The scanner was not operated during the first several years
of the Mars Express mission due to an unexpectedly cold environment. The drawback of this
is that NPD – NPI data inter-comparison is somewhat difficult, as the NPD field-of-view plane
is deflected 15◦ off the NPI field-of-view plane. In such a configuration both ENA sensors only
pointed in the same direction when the spacecraft was changing attitude. Because of different
instrument accommodation on Venus Express the ASPERA-4 scanner had to operate from the
very beginning to avoid exposure to the Sun.
If making priorities for future development of an NPD-based instrument, we would empha-
size the angular resolution. It can be improved down to 10◦ by mapping secondary electrons
yielded from the Start surface to a position-sensitive detector.
Appendix A
NPD data processing
A.1 In addition to the NPD operation modes

The RAW mode. Table A.1 shows a bit allocation in a 32-bit event, generated by the DigTOF
electronics in the RAW mode.
bit position number of bits parameter

31 : 25 7 spare
24 : 22 3 coincidence flag (COIN)
21 : 20 2 direction (DIR)
19 : 12 8 Pulse Height (PH)
11 : 0 12 TOF (0.5 ns resolution)
Table A.1: The RAW mode event bit layout.
Both the DigTOF electronics and the DPU have a 16-bit wide bus. Each 32-bit event
generated in DigTOF is stored as two 16-bit values, which are composed back to 32-bit value
in the DPU after data are transfered from DigTOF. A disadvantage of having a data size twice
the data bus wide is possible data packet corruption during transfer from DigTOF to DPU. Each
data entry is stored in a data packet sequentially, upper 16-bit, lower 16-bit word. Readout of
the RAW data buffer of 512 values is carried out by the DPU through 2 × 512 read accesses to
the respective 16-bit register of the DigTOF electronics. In total 1024 readouts are performed
to obtain a data packet. There is a small probability for a read signal to be generated due to
disturbances induced in the data bus, e.g., on a read-line. An additional fake read access results
in a shift in the sequence of 16-bit words obtained by the DPU and consequently in the wrong
data order, thus making the data packet unusable. DigTOF generates a certain hexadecimal
value of ’0x5555’ for every read access exceeding 1024. Thus, special attention was paid to
keeping the signal lines between the DPU and DigTOF clean from electromagnetic noise.
No data reduction is performed in the RAW mode because this mode was intended to be
used for calibration purposes. Telemetry packets can be compressed by a lossy-less RICE
algorithm in order to decrease the bit-rate. Optimization of RICE conversion during coding
and decoding of telemetry packets is done in the DPU by splitting each 32-bit event into two
16-bit words, extracting certain information from the 32-bit event and forming two streams of
data, according to Table A.2.
147
148 APPENDIX A. NPD DATA PROCESSING
stream 1 stream 2
bit position number of bits parameter bit position number of bits parameter
15 : 14 2 DIR 7:0 8 PH
13 : 11 3 COIN — — —
10 : 0 11 TOF (1 ns) — — —
Table A.2: Structure of data streams in a telemetry packet containing RAW data is shown. Data streams
are formed in order to optimize RICE (de-)compression.
The BIN mode. The TOF and pulse height of valid events are stored in compressed form
in the BIN mode. Time-of-flight values are stored in logarithmically divided 16 TOF bins.
Figure A.1 shows the size of the TOF bins and their location in the TOF space.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 TOF step
12 16 20 25 31 40 50 63 78 99 124 156 195 245 307 386 TOF window size [ns]
50 62 78 99 124 156 196 246 309 387 488 610 767 962 1207 1515 1900 TOF [ns]
Figure A.1: TOF bins definition in the BIN mode
Pulse heights are stored in 16 linearly-divided PH bins, converted from 8-bit values to 4-bit
by a 4 bit shift, i.e., scaled in 16 steps.
The data array consists of 16× 16× 3 = 768 16-bit counters addressed by TOF, pulse height,
and direction values, respectively. Each valid event binned according to its TOF, pulse height
and sector number increases by one a corresponding counter in the data array. The 16-bit
counter values can be log-compressed (see Section A.2) to 8-bit values and RICE compressed
by the DPU.
In order to decrease the bit-rate further, the 16T OF × 16PH matrices can be compressed to
XT OF × YPH independently for each direction according to Table A.3.
Matrix No. X Y interpretation

0 16 16 no compression of pulse height data
1 16 2 on board H and O discrimination
2 16 1 pulse height information is not used
Table A.3: Reduction matrices in BIN mode
The default matrix is No.=1, corresponding X=16, Y=2. In that case discrimination between
H and O ENAs is done on board by the DPU. The reduction of 16 mass steps Mi to 2 is done
according to the next algorithm.




 0 if T i < 7

1 if T i > 10



Mi = 




 0 if 7 6 T i 6 10, M 6 P

1 if 7 6 T i 6 10, M > P

where T i is a TOF step number, N the threshold, shows the PH channel number defined in
Section 5.2.6. The principle of mass identification is discussed in Section 5.2.6.
Figure A.2 shows the algorithm graphically. The plot shows a TOF × PH matrix obtained
in the BIN mode with 16 TOF steps as columns and 16 PH steps as rows.
A.2. LOG-COMPRESSION ALGORITHM 149
All events outside the TOF interval 7 – 10 are treated according to their TOF values only.
Events with TOF values less than 7 (more energetic) are treated as H, events with TOF va-
lues higher than 10 are treated as O. To distinguish between species within the TOF window
7 – 10 where both H and O events may occur, the PH analysis is used (see Section 5.2.6). The
threshold N, separating areas where mostly H or O events occur, is commandable and is set
separately for each direction. N is defined in on-ground calibrations. The calibrated thresholds
are shown in Tables 5.19 and 5.36.
TOF
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0
1
2
3
4
H H+O O
5
6
7
PH
8
9 N
10
11
12
H+O
13
14
15
Figure A.2: PH reduction. TOF × PH matrix is shown. The black thick line separates two regions of
the matrix. Events with TOF values less than 7 are treated as H, events with TOF values higher than 10
are treated as O. To distinguish between species within the TOF window 7 – 10 steps the PH analysis is
used.
A.2 Log-compression algorithm

A 16-bit value can be log-compressed to 8 bits. The compressed value comprises a 4-bit expo-
nent (e) and a 4-bit mantissa (m), and can be computed from e and m according to the following
rules: 
 m,

 if e < 1 (no compression for count 632),
count = 
 (m + 16) · 2e−1 , if e > 1.

A.3 NPD data display

I wrote a quick look data display DV P was written by myself in IDL to display the NPD data.
It is object-oriented and has a graphical user interface. Several examples of the data display
window showing the NPD / ASPERA-4 data obtained in different modes are presented below.
Initially, DV P was used to monitor the NPD status during TVAC tests. Then the program was
further developed to display the NPD data obtained during flight. DVP can be used to plot not
only housekeeping information (e.g., monitors of the temperature sensors, low voltages, bias
HV, and a number of the NPD status registers) but even science data. Figure A.3 shows the
TOF spectra and distributions, and linear counter data obtained by NPD1 in the TOF mode.
Controlling buttons allow manipulation of the data files.
Figure A.4 shows the TOF and PH spectra as well as non-validated event counts (left panel)
obtained by NPD1 in the RAW mode. The right panel shows the corresponding TOF and PH
distributions.
Figure A.5 depicts house-keeping data obtained by NPD2 during one of the measurement
sessions. The house-keeping data displayed includes several bias HV monitors, the respective
HV reference settings, and the temperature profile.
Figure A.3: DVP display. The TOF spectra and distributions obtained by the three sectors of the NPD1
(from top to bottom) in the TOF mode. The linear counter data is shown in the lower panel.
A.3. NPD DATA DISPLAY 151
Figure A.4: DVP display. The NPD1 TOF and pulse height spectra for all three sectors (left panel
from top to bottom) obtained in the RAW mode. Right panel shows the corresponding TOF and PH
distributions. Lower left panel displays linear counters.
Figure A.5: DVP display. The house-keeping data collected by NPD2 during one of the measurement
sessions with bias HV ramping up and down is displayed. The bias HV monitors of the deflector, Start,
and Stop units are shown in the upper panels (from top to bottom). The corresponding HV references
and the temperature profile are shown next.
This page is left in-

tentionally blank
Bibliography
Bibliography
Acuña, M. H., J. E. P. Connerney, P. Wasilewski, R. P. Lin, K. A. Anderson, C. W. Carlson,
J. McFadden, D. W. Curtis, D. Mitchell, H. Reme, C. Mazelle, J. A. Sauvaud, C. d’Uston,
A. Cros, J. L. Medale, S. J. Bauer, P. Cloutier, M. Mayhew, D. Winterhalter, and N. F. Ness,
1998. Magnetic field and plasma observations at Mars: Initial results of the Mars Global
Surveyor mission, Science, 279, 1676–1680.
Amsif, A., J. Dandouras, and E. C. Roelof, 1997. Modeling the production and the imaging of
energetic neutral atoms from Titan’s exosphere, J. Geophys. Res., 102(A10), 22,169–22,182,
doi:10.1029/97JA01597.
Anderson, D. E., and C. W. Hord, 1971. Mariner 6 and 7 ultraviolet spectrometer experiment:
Analysis of hydrogen lyman-alpha data, J. Geophys. Res., 76, 6666–6673.
Barabash, S., 1995. Satellite observations of the plasma-neutral coupling near Mars and the
Earth, Ph.D. thesis, Swedish Institute of Space Physics, Kiruna, Sweden.
Barabash, S., O. Norberg, R. Lundin, S. Olsen, K. Lundin, P. C. Brandt, E. C. Roelof, C. J.

Chase, B. H. Mauk, H. Koskinen, and J. Rynö, 1998. Energetic neutral atom imager on
the Swedish microsatellite Astrid, in Measurement Techniques in Space Plasmas: Fields,
Geophysical Monograph, vol. 103, edited by R. F. Pfaff, J. E. Borovsky, and D. T. Young,
pp. 257–262, AGU, Washington, DC.
Barabash, S., M. Holmström, A. Lukyanov, and E. Kallio, 2002. Energetic neutral

atoms at Mars. 4. Imaging of planetary oxygen, J. Geophys. Res., 107(A10), 1280, doi:
10.1029/2001JA000326.
Barabash, S., R. Lundin, H. Andersson, J. Gimholt, M. Holmström, O. Norberg, M. Ya-

mauchi, K. Asamura, A. J. Coates, D. R. Linder, D. O. Kataria, C. C. Curtis, K. C. Hsieh,
B. R. Sandel, A. Fedorov, A. Grigoriev, E. Budnik, M. Grande, M. Carter, D. H. Reading,
H. Koskinen, E. Kallio, P. Riihela, T. Säles, J. Kozyra, N. Krupp, S. Livi, J. Woch, J. Luh-
mann, S. McKenna-Lawlor, S. Orsini, R. Cerulli-Irelli, M. Maggi, A. Morbidini, A. Mura,
A. Milillo, E. Roelof, D. Williams, J.-A. Sauvaud, J.-J. Thocaven, T. Moreau, D. Winning-
ham, R. Frahm, J. Scherrer, J. Sharber, P. Wurz, and P. Bochsler, 2004. ASPERA-3: Anal-
yser of space plasmas and energetic ions for Mars Express, in Mars Express: The scientific
payload, vol. SP-1240, edited by A. Wilson, pp. 121–139, ESA Special Publication.
Barabash, S., R. Lundin, H. Andersson, K. Brinkfeldt, A. Grigoriev, H. Gunell, M. Holmström,

M. Yamauchi, K. Asamura, P. Bochsler, P. Wurz, R. Cerulli-Irelli, A. Mura, A. Milillo,
153
154 BIBLIOGRAPHY
M. Maggi, S. Orsini, A. J. Coates, D. R. Linder, D. O. Kataria, C. C. Curtis, K. C. Hsieh,

B. R. Sandel, R. A. Frahm, J. R. Sharber, J. D. Winningham, M. Grande, E. Kallio, H. Kosk-
inen, P. Riihelä, W. Schmidt, T. Säles, J. U. Kozyra, N. Krupp, J. Woch, S. Livi, J. G. Luh-
mann, S. Mckenna-Lawlor, E. C. Roelof, D. J. Williams, J.-A. Sauvaud, A. Fedorov, and J.-J.
Thocaven, 2006. The Analyser of Space Plasmas and Energetic Atoms (ASPERA-3) for the
Mars Express Mission, Space Sci. Rev., 126, 113–164, doi:10.1007/s11214-006-9124-8.
Barabash, S., A. Fedorov, R. Lundin, and J.-A. Sauvaud, 2007a. Martian Atmospheric Erosion
Rates, Science, 315(5811), 501–503, doi:10.1126/science.1134358.
Barabash, S., J.-A. Sauvaud, H. Gunell, H. Andersson, A. Grigoriev, K. Brinkfeldt, M. Holm-

ström, R. Lundin, M. Yamauchi, K. Asamura, W. Baumjohann, T. Zhang, A. J. Coates, D. R.
Linder, D. O. Kataria, C. C. Curtis, K. C. Hsieh, B. R. Sandel, A. Fedorov, C. Mazelle, J.-J.
Thocaven, M. Grande, H. E. J. Koskinen, E. Kallio, T. Säles, P. Riihela, J. Kozyra, N. Krupp,
J. Woch, J. Luhmann, S. McKenna-Lawlor, S. Orsini, R. Cerulli-Irelli, A. Mura, A. Milillo,
M. Maggi, E. Roelof, P. Brandt, C. T. Russell, K. Szego, D. Winningham, R. A. Frahm,
J. Scherrer, J. R. Sharber, P. Wurz, and P. Bochsler, 2007b. The Analyser of Space Plas-
mas and Energetic Atoms (ASPERA-4) for the Venus Express Mission, Planet. Space Sci.,
doi:10.1016/j.pss.2007.01.014, in press.
Biernat, H. K., N. V. Erkaev, and C. J. Farrugia, 1999. Aspects of MHD flow about Venus, J.
Geophys. Res., 104, 12,617–12,626, doi:10.1029/1999JA900032.
Biernat, H. K., N. V. Erkaev, and C. J. Farrugia, 2001. MHD effects in the Venus magnetosheath
including mass loading, Adv. Space Res., 28, 833–839.
Bransden, B. H., and M. R. C. McDowell, Charge Exchange and the Theory of Ion-Atom
Collisions, Oxford University Press, New York, 1992.
Brecht, S. H., 1997a. Solar wind proton deposition into the Martian atmosphere, J. Geophys.
Res., 102(A6), 11,287–11,294.
Brecht, S. H., 1997b. Hybrid simulations of the magnetic topology of Mars, J. Geophys. Res.,
102(A3), 4743–4750.
Brinkfeldt, K., P. Enoksson, M. Wieser, and S. Barabash, 2006a. Modelling of Microshutters

for Space Physics Applications, Eurosensors XX, Göteborg, Sweden.
Brinkfeldt, K., H. Gunell, P. C. Brandt, S. Barabash, R. A. Frahm, J. D. Winningham, E. Kallio,

M. Holmström, Y. Futaana, A. Ekenbäck, R. Lundin, H. Andersson, M. Yamauchi, A. Grig-
oriev, J. R. Sharber, J. Scherrer, A. J. Coates, D. R. Linder, D. O. Kataria, H. Koskinen,
T. Säles, P. Riihela, W. Schmidt, J. Kozyra, J. Luhmann, E. Roelof, D. Williams, S. Livi,
C. C. Curtis, K. C. Hsieh, B. R. Sandel, M. Grande, M. Carter, J.-A. Sauvaud, A. Fe-
dorov, J.-J. Thocaven, S. McKenna-Lawler, S. Orsini, R. Cerulli-Irelli, M. Maggi, P. Wurz,
P. Bochsler, N. Krupp, J. Woch, M. Fraenz, K. Asamura, and C. Dierker, 2006b. First ENA
observations at Mars: Solar wind ENAs on the nightside, Icarus, 182, 439–447.
Chalov, S. V., H. J. Fahr, and V. V. Izmodenov, 2003. Evolution of pickup proton spectra in the
inner heliosheath and their diagnostics by energetic neutral atom fluxes, J. Geophys. Res.,
108(A6), 1266, doi:10.1029/2002JA009492.
BIBLIOGRAPHY 155
Chase, C. J., and E. C. Roelof, 1995. Extracting evolving structures from global magneto-
spheric images via model fitting and video visualization, Johns Hopkins APL Technical Di-
gest., 16(2), 111–122.
Collier, M. R., T. E. Moore, K. W. Ogilvie, D. Chornay, J. Keller, S. Boardsen, J. Burch, B. E.

Marji, M.-C. Fok, S. Fuselier, A. Ghielmetti, B. Giles, D. Hamilton, B. Peko, J. Quinn,
E. Roelof, T. Stephen, G. Wilson, and P. Wurz, 2001. Observations of neutral atoms from
the solar wind, J. Geophys. Res., 106, 24,893–24,906.
Crider, D. H., D. Vignes, A. M. Krymskii, T. K. Breus, N. F. Ness, D. L. Mitchell, J. A. Slavin,

and M. H. Acuña, 2003. A proxy for determining solar wind dynamic pressure at Mars using
Mars Global Surveyor data, J. Geophys. Res., 108(A12), 1461, doi:10.1029/2003JA009875.
Curtis, C. C., and K. C. Hsieh, 1989. Remote Sensing of Planetary Magnetospheres: Imaging
via Energetic Neutral Atoms, in Solar System Plasma Physics, Geophysical Monograph,
vol. 54, edited by J. H. Waite, Jr., J. L. Burch, and R. L. Moore, pp. 247–251, American
Geophysical Union, Washington, DC.
Dandouras, J., and A. Amsif, 1999. Production and imaging of energetic neutral atoms from
Titan’s exosphere: a 3-D model, Planet. Space Sci., 47, 1355–1369.
Donahue, T. M., and R. E. Hartle, 1992. Solar cycle variations in H(+) and D(+) densities in
the Venus ionosphere - Implications for escape, Geophys. Res. Lett., 19, 2449–2452.
Elphic, R., and C. Russell, 1983. Global characteristics of magnetic flux ropes in the Venus
ionosphere, J. Geophys. Res., 88, 2993.
Espley, J. R., P. A. Cloutier, D. A. Brain, D. H. Crider, and M. H. Acuña, 2004. Observations

of low-frequency magnetic oscillations in the Martian magnetosheath and magnetic pileup
region and and tail, J. Geophys. Res., 109(A0), 7213, doi:10.1029/2003JA010193.
Fahr, H.-J., and K. Scherer, 2004. Energetic neutral atom fluxes from the heliosheath varying
with the activity phase of the solar cycle, Astrophys. Space Sci. Trans., 1, 3–15.
Fedorov, A., E. Budnik, J.-A. Sauvaud, C. Mazelle, S. Barabash, R. Lundin, M. Acuña,

M. Holmström, A. Grigoriev, M. Yamauchi, H. Andersson, J.-J. Thocaven, D. Winning-
ham, R. Frahm, J. R. Sharber, J. Scherrer, A. J. Coates, D. R. Linder, D. O. Kataria,
E. Kallio, H. Koskinen, T. Säles, P. Riihelä, W. Schmidt, J. Kozyra, J. Luhmann, E. Roelof,
D. Williams, S. Livi, C. C. Curtis, K. C. Hsieh, B. R. Sandel, M. Grande, M. Carter,
S. McKenna-Lawler, S. Orsini, R. Cerulli-Irelli, M. Maggi, P. Wurz, P. Bochsler, N. Krupp,
J. Woch, M. Fränz, K. Asamura, and C. Dierker, 2006. Structure of the Martian wake,
Icarus, 182, 329–336, doi:10.1016/j.icarus.2005.09.021.
Fok, M.-C., T. E. Moore, M. R. Collier, and T. Tanaka, 2004. Neutral atom imaging of
solar wind interaction with the Earth and Venus, J. Geophys. Res., 109(A18), 1206, doi:
10.1029/2003JA010094.
Funsten, H., D. J. McComas, K. R. Moore, and E. E. Scime, 1995. Low Energy Neutral
Atom Imaging Techniques for Remote Observations of the Magnetosphere, J. Spacecraft
and Rockets, 32, 899.
156 BIBLIOGRAPHY
Funsten, H. O., D. J. McComas, and M. A. Gruntman, 1998. Neutral Atom Imaging: UV

Rejection Techniques, in Measurement Techniques in Space Plasmas: Fields, Geophysical
Monograph, vol. 103, edited by R. F. Pfaff, J. E. Borovsky, and D. T. Young, pp. 251–256,
AGU, Washington, DC.
Futaana, Y., S. Barabash, A. Grigoriev, M. Holmström, E. Kallio, P. C:son Brandt, H. Gunell,

K. Brinkfeld, R. Lundin, H. Andersson, M. Yamauchi, S. McKenna-Lawlor, J. D. Win-
ningham, R. A. Frahm, J. R. Sharber, J. Scherrer, A. J. Coates, D. R. Linder, D. O.
Kataria, T. Säles, P. Riihela, W. Schmidt, H. Koskinen, J. Kozyra, J. Luhmann, E. Roelof,
D. Williams, S. Livi, C. C. Curtis, K. C. Hsieh, B. R. Sandel, M. Grande, M. Carter,
J.-A. Sauvaud, A. Fedorov, J.-J. Thocaven, S. Orsini, R. Cerulli-Irelli, M. Maggi,
P. Wurz, P. Bochsler, N. Krupp, J. Woch, M. Fraenz, K. Asamura, and C. Dierker,
2006a. First ENA observations at Mars: Subsolar ENA jet, Icarus, 182, 413–423, doi:
10.1016/j.icarus.2005.08.024.
Futaana, Y., S. Barabash, A. Grigoriev, M. Holmström, E. Kallio, P. C:son Brandt, H. Gunell,

K. Brinkfeld, R. Lundin, H. Andersson, M. Yamauchi, S. McKenna-Lawlor, J. D. Win-
ningham, R. A. Frahm, J. R. Sharber, J. Scherrer, A. J. Coates, D. R. Linder, D. O.
Kataria, T. Säles, P. Riihela, W. Schmidt, H. Koskinen, J. Kozyra, J. Luhmann, E. Roelof,
D. Williams, S. Livi, C. C. Curtis, K. C. Hsieh, B. R. Sandel, M. Grande, M. Carter, J.-
A. Sauvaud, A. Fedorov, J.-J. Thocaven, S. Orsini, R. Cerulli-Irelli, M. Maggi, P. Wurz,
P. Bochsler, A. Galli, N. Krupp, J. Woch, M. Fraenz, K. Asamura, and C. Dierker, 2006b.
First ENA observations at Mars: ENA emissions from the Martian upper atmosphere, Icarus,
182, 424–430, doi:10.1016/j.icarus.2005.09.019.
Futaana, Y., S. Barabash, A. Grigoriev, D. Winningham, R. Frahm, and R. Lundin, 2006c.

Global Response of Martian Plasma Environment to an Interplanetary Structure: From ENA
and Plasma Observations at Mars, Space Sci. Rev., 126, 315–332, doi:10.1007/s11214-006-
9026-9.
Galli, A., P. Wurz, S. Barabash, A. Grigoriev, H. Gunell, R. Lundin, M. Holmström, and

A. Fedorov, 2006a. Energetic Hydrogen and Oxygen Atoms Observed on the Nightside
of Mars, Space Sci. Rev., 126, 267–297, doi:10.1007/s11214-006-9088-8.
Galli, A., P. Wurz, S. Barabash, A. Grigoriev, R. Lundin, Y. Futaana, H. Gunell, M. Holm-

ström, E. C. Roelof, C. C. Curtis, K. C. Hsieh, A. Fedorov, D. Winningham, R. A. Frahm,
R. Cerulli-Irelli, P. Bochsler, N. Krupp, J. Woch, and M. Fraenz, 2006b. Direct Measure-
ments of Energetic Neutral Hydrogen in the Interplanetary Medium, Astrophys. J., 644,
1317–1325, doi:10.1086/503765.
Galli, A., P. Wurz, H. Lammer, H. I. M. Lichtenegger, R. Lundin, S. Barabash, A. Grigoriev,

M. Holmström, and H. Gunell, 2006c. The Hydrogen Exospheric Density Profile Measured
with ASPERA-3/NPD, Space Sci. Rev., 126, 447–467, doi:10.1007/s11214-006-9089-7.
Grande, M., 1997. Investigation of magnetospheric interactions with the Hermean surface,
Adv. Space Res., 19(10), 1609–1614, doi:10.1016/S0273-1177(97)00374-8.
Grigoriev, A., Y. Futaana, S. Barabash, and A. Fedorov, 2006. Observations of the Martian
Subsolar ENA Jet Oscillations, Space Sci. Rev., 126, 299–313, doi:10.1007/s11214-006-
9121-y.
BIBLIOGRAPHY 157
Gruntman, M., 1993. A new technique for in situ measurement of the composition of neutral
gas in interplanetary space, Planet. Space Sci., 41(4), 307–319.
Gruntman, M., 1997. Energetic neutral atom imaging of space plasmas, Rev. Sci. Instrum.,
68(10), 3617–3656.
Gunell, H., M. Holmström, H. K. Biernat, and N. V. Erkaev, 2005. Planetary ENA

Imaging: Venus and a comparison with Mars, Planet. Space Sci., 53, 433–441, doi:
10.1016/j.pss.2004.07.021.
Gunell, H., M. Holmström, S. Barabash, E. Kallio, P. Janhunen, A. F. Nagy, and Y. Ma, 2006.
Planetary ENA imaging: Effects of different interaction models for Mars, Planet. Space Sci.,
54, 117–131, doi:10.1016/j.pss.2005.04.002.
Hilchenbach, M., K. C. Hsieh, D. Hovestadt, B. Klecker, H. Gruenwaldt, P. Bochsler, F. M.

Ipavich, A. Buergi, E. Moebius, F. Gliem, W. I. Axford, H. Balsiger, W. Bornemann, M. A.
Coplan, A. B. Galvin, J. Geiss, G. Gloeckler, S. Hefti, D. L. Judge, R. Kallenbach, P. Laev-
erenz, M. A. Lee, S. Livi, G. G. Managadze, E. Marsch, M. Neugebauer, H. S. Ogawa, K.-U.
Reiche, M. Scholer, M. I. Verigin, B. Wilken, and P. Wurz, 1998. Detection of 55-80 keV
Hydrogen Atoms of Heliospheric Origin by CELIAS/HSTOF on SOHO, Astrophys. J., 503,
916–921, doi:10.1086/306022.
Holmström, M., 2006. Asymmetries in Mars’ Exosphere: Implications for X-ray and ENA
Imaging, Space Sci. Rev., 126, 435–445, doi:10.1007/s11214-006-9036-7.
Holmström, M., S. Barabash, and E. Kallio, 2002. Energetic neutral atoms at Mars. 1. Imaging
of solar wind protons, J. Geophys. Res., 107(A10), 1277, doi:10.1029/2001JA000325.
Hsieh, K. C., B. R. Sandel, V. A. Drake, and R. S. King, 1991. H Lyman-α transmittance of

thin C and Si/C foils for keV particle detectors, Nuclear Instruments and Methods in Physics
Research B, 61, 187–193, doi:10.1016/0168-583X(91)95460-U.
Jans, S., Ionization of energetic neutral atoms for application in space instrumentation, Diplo-
marbeit, Universität Bern, 2000.
Johnstone, A. D., Alsop, C., Burge, S., Carter, P. J., Coates, A. J., Coker, A. J., Fazakerley,
A. N., Grande, M., Gowen, R. A., Gurgiolo, C., Hancock, B. K., Narheim, B., Preece, A.,
Sheather, P. H., Winningham, J. D., Woodliffe, and R. D., 1997. PEACE: a Plasma Electron
and Current Experiment, Space Sci. Rev., 79, 351–398.
Kallenbach, R., M. Hilchenbach, S. V. Chalov, J. A. Le Roux, and K. Bamert, 2005. On the

”injection problem” at the solar wind termination shock, Astron. & Astrophys., 439, 1–22,
doi:10.1051/0004-6361:20052874.
Kallio, E., 1996. An empirical model of the solar wind flow around Mars, J. Geophys. Res.,
101(A5), 11,133–11,147.
Kallio, E., and S. Barabash, 2000. On the elastic and inelastic collisions between precipitating
energetic hydrogen atoms and Martian atmospheric neutrals, J. Geophys. Res., 105(A11),
24,973–24,996, doi:10.1029/2000JA900077.
158 BIBLIOGRAPHY
Kallio, E., and S. Barabash, 2001. Atmospheric effects of precipitating energetic hy-
drogen atoms on the Martian atmosphere, J. Geophys. Res., 106(A1), 165–177, doi:
10.1029/2000JA002003.
Kallio, E., and P. Janhunen, 2001. Atmospheric effects of proton precipitation in the Martian at-
mosphere and its connection to the Mars–solar wind interaction, J. Geophys. Res., 106(A4),
5617–5634, doi:10.1029/2000JA000239.
Kallio, E., and P. Janhunen, 2002. Ion escape from Mars in a quasi-neutral hybrid model, J.
Geophys. Res., 107(A3), 1035, doi:10.1029/2001JA000090.
Kallio, E., J. G. Luhmann, and S. Barabash, 1997. Charge exchange near Mars: The solar
wind absorption and energetic neutral atom production, J. Geophys. Res., 102(A10), 22,183–
22,197, doi:10.1029/97JA01662.
Kallio, E., S. Barabash, K. Brinkfeldt, H. Gunell, M. Holmström, Y. Futaana, W. Schmidt,

T. Säles, H, Koskinen, P. Riihelä, R. Lundin, H. Andersson, M. Yamauchi, A. Grigoriev,
J. D. Winningham, R. A. Frahm, J. R. Sharber, J. Scherrer, A. J. Coates, D. R. Linder, D. O.
Kataria, J. Kozyra, J. G. Luhmann, E. Roelof, D. Williams, S. Livi, P. C. Brandt, C. C. Curtis,
K. Hsieh, B. R. Sandel, M. Grande, M. Carter, J.-A. Sauvaud, A. Fedorov, J.-J. Thocaven,
S. McKenna-Lawler, S. Orsini, R. Cerulli-Irelli, M. Maggi, P. Wurz, P. Bochsler, N. Krupp,
J. Woch, M. Fränz, K. Asamura, and C. Dierker, 2006. Energetic neutral atoms (ENA)
at Mars: Properties of the hydrogen atoms produced upstream of the Martian bow shock
and implications for ENA sounding technique around non-magnetized planets, Icarus, 182,
448–463.
Kass, D. M., 1999. Change in the Martian Atmosphere, Ph.D. thesis, California Institute of
Technology, California, United States.
Krasnopolsky, V. A., 2002. Mars’ upper atmosphere and ionosphere at low, medium, and
high solar activities: Implications for evolution of water, J. Geophys. Res., 107(E12), 5128,
doi:10.1029/2001JE001809.
Krasnopolsky, V. A., and G. R. Gladstone, 1996. Helium on Mars : EUVE and PHOBOS data
and implications for Mars’ evolution, J. Geophys. Res., 101(A7), 15,765–15,772.
Lichtenegger, H., H. Lammer, and W. Stumptner, 2002. Energetic neutral atoms at Mars. 3.
Flux and energy distributions of planetary energetic H atoms, J. Geophys. Res., 107(A10),
1279, doi:10.1029/2001JA000322.
Lichtenegger, H. I. M., H. Lammer, Y. N. Kulikov, S. Kazeminejad, G. H. Molina-Cuberos,

R. Rodrigo, B. Kazeminejad, and G. Kirchengast, 2006. Effects of Low and High Ener-
getic Neutral Atoms on Martian and Venusian Dayside Exospheric Temperature Estimations,
Space Sci. Rev., 126, 469–501, doi:10.1007/s11214-006-9082-1.
Luhmann, J. G., 1992. Comparative studies of the solar wind interaction with weakly magne-
tized planets, Advances in Space Research, 12, 191–203, doi:10.1016/0273-1177(92)90331-
Q.
Luhmann, J. G., and J. U. Kozyra, 1991. Dayside pickup oxygen ion precipitation at Venus and
Mars: Spatial distributions, energy deposition and consequences, J. Geophys. Res., 96(A4),
5457–5467.
BIBLIOGRAPHY 159
Lukyanov, A., S. Barabash, and M. Holmström, 2004. Energetic neutral atom imaging at
Mercury, Adv. Space Res., 33, 1890–1898, doi:10.1016/j.asr.2003.05.035.
Lundin, R., A. Zakharov, R. Pellinen, H. Borg, B. Hultqvist, N. Pissarenko, E. M. Dubinin,

S. W. Barabash, I. Liede, and H. Koskinen, 1989. First measurements of the ionospheric
plasma escape from Mars, Nature, 341, 609–612.
Lundin, R., S. Barabash, H. Andersson, M. Holmström, A. Grigoriev, M. Yamauchi, J.-A.

Sauvaud, A. Fedorov, E. Budnik, J.-J. Thocaven, D. Winningham, R. Frahm, J. Scherrer,
J. Sharber, K. Asamura, H. Hayakawa, A. Coates, D. R. Linder, C. Curtis, K. C. Hsieh,
B. R. Sandel, M. Grande, M. Carter, D. H. Reading, H. Koskinen, E. Kallio, P. Riihela,
W. Schmidt, T. Säles, J. Kozyra, N. Krupp, J. Woch, J. Luhmann, S. McKenna-Lawler,
R. Cerulli-Irelli, S. Orsini, M. Maggi, A. Mura, A. Milillo, E. Roelof, D. Williams, S. Livi,
P. Brandt, P. Wurz, and P. Bochsler, 2004. Solar wind-induced atmospheric erosion at
Mars: First results from ASPERA-3 on Mars Express, Science, 305, 1933–1936, doi:DOI:
10.1126/science.1101860.
Ma, Y., A. F. Nagy, K. C. Hansen, D. L. DeZeeuw, T. I. Gombosi, and K. G. Powell, 2002.

Three-dimensional multispecies MHD studies of the solar wind interaction with Mars in the
presence of crustal fields, J. Geophys. Res., 107(A10), 1282, doi:10.1029/2002JA009293.
Massetti, S., S. Orsini, A. Milillo, A. Mura, E. de Angelis, H. Lammer, and P. Wurz, 2003.
Mapping of the cusp plasma precipitation on the surface of Mercury, Icarus, 166, 229–237,
doi:10.1016/j.icarus.2003.08.005.
McComas, D., F. Allegrini, P. Bochsler, M. Bzowski, M. Collier, H. Fahr, H. Fichtner, P. Frisch,

H. Funsten, S. Fuselier, G. Gloeckler, M. Gruntman, V. Izmodenov, P. Knappenberger,
M. Lee, S. Livi, D. Mitchell, E. Möbius, T. Moore, D. Reisenfeld, E. Roelof, N. Schwadron,
M. Wieser, M. Witte, P. Wurz, and G. Zank, 2004. The Interstellar Boundary Explorer
(IBEX), in Physics of the Outer Heliosphere, American Institute of Physics Conference Se-
ries, vol. 719, edited by V. Florinski, N. V. Pogorelov, and G. P. Zank, pp. 162–181, AIP
Conference Proceedings, doi:10.1063/1.1809514.
McComas, D. J., B. L. Barraclough, R. C. Elphic, H. O. Funsten III, and M. F. Thomsen, 1991.

Magnetospheric imaging with low-energy neutral atoms, in Proceedings of the National
Academy of Sciences, vol. 88, pp. 9598–9602.
McComas, D. J., H. O. Funsten, and E. E. Scime, 1998. Advances in Low Energy Neutral
Atom Imaging, in Measurement Techniques in Space Plasmas: Fields, Geophysical Mono-
graph, vol. 103, edited by R. F. Pfaff, J. E. Borovsky, and D. T. Young, pp. 275–280, AGU,
Washington, DC.
McEntire, R. W., and D. G. Mitchell, 1989. Instrumentation for Global Magnetospheric Imag-
ing via Energetic Neutral Atoms, in Solar System Plasma Physics, edited by J. H. Waite, Jr.,
J. L. Burch, and R. L. Moore, 54, pp. 69–80.
Mitchell, D. G., S. M. Krimigis, A. F. Cheng, S. E. Jaskulek, E. P. Keath, B. H. Mauk, R. W.

McEntire, E. C. Roelof, C. E. Schlemm, B. E. Tossman, and D. J. Williams, 1998. The
Imaging Neutral Camera for the Cassini Mission to Saturn and Titan, in Measurement Tech-
niques in Space Plasmas: Fields, Geophysical Monograph, vol. 103, edited by R. F. Pfaff,
J. E. Borovsky, and D. T. Young, pp. 281–287, AGU, Washington, DC.
160 BIBLIOGRAPHY
Mitchell, D. G., S. E. Jaskulek, C. E. Schlemm, E. P. Keath, R. E. Thompson, B. E. Tossman,

J. D. Boldt, J. R. Hayes, G. B. Andrews, N. Paschalidis, D. C. Hamilton, R. A. Lundgren,
E. O. Tums, P. Wilson IV, H. D. Voss, D. Prentice, K. C. Hsieh, C. C. Curtis, and F. R.
Powell, 2000. High Energy Neutral Atom (HENA) Imager for the IMAGE Mission, Space
Sci. Rev., 91(1-2), 67–112.
Moore, T. E., D. J. Chornay, M. R. Collier, F. A. Herrero, J. Johnson, M. A. Johnson, J. W.

Keller, J. F. Laudadio, J. F. Lobell, K. W. Ogilvie, P. Rozmarynowski, S. A. Fuselier,
A. G. Ghielmetti, E. Hertzberg, D. C. Hamilton, R. Lundgren, P. Wilson, P. Walpole, T. M.
Stephen, B. L. Peko, B. Van Zyl, P. Wurz, J. M. Quinn, and G. R. Wilson, 2000. The Low
Energy Neutral Atom Imager for IMAGE, Space Sci. Rev., 91(1-2), 155–195.
Norberg, O., M. Yamauchi, R. Lundin, S. Olsen, H. Borg, S. Barabash, M. Hirahara, T. Mukai,

and H. Hayakawa, 1998. The Ion Mass Imager on the Planet-B spacecraft, Earth, Planets,
and Space, 50, 199–205.
Penz, T., N. Erkaev, H. Biernat, H. Lammer, U. Amerstorfer, H. Gunell, E. Kallio, S. Barabash,

S. Orsini, A. Milillo, and W. Baumjohann, 2004. Ion loss on Mars caused by the Kelvin-
Helmholtz instability, Planet. Space Sci., 52, 1157–1167, doi:10.1016/j.pss.2004.06.001.
Perez, J. D., M.-C. Fok, and T. E. Moore, 2000. Deconvolution of Energetic Neutral Atom
Images of the Earth’s Magnetosphere, Space Sci. Rev., 91, 421–436.
Pollock, C. J., K. Asamura, J. Baldonado, M. M. Balkey, P. Barker, J. L. Burch, E. J. Korpela,

J. Cravens, G. Dirks, M.-C. Fok, H. O. Funsten, M. Grande, M. Gruntman, J. Hanley, J.-M.
Jahn, M. Jenkins, M. Lampton, M. Marckwordt, D. J. McComas, T. Mukai, G. Penegor,
S. Pope, S. Ritzau, M. L. Schattenburg, E. Scime, R. Skoug, W. Spurgeon, T. Stecklein,
S. Storms, C. Urdiales, P. Valek, J. T. M. van Beek, S. E. Weidner, M. Wüest, M. K. Young,
and C. Zinsmeyer, 2000. Medium Energy Neutral Atom (MENA) Imager for the IMAGE
Mission, Space Sci. Rev., 91(1-2), 113–154.
Roelof, E. C., and A. J. Skinner, 2000. Extraction of ion distributions from magnetospheric
ENA and EUV images, Space Sci. Rev., 91, 437–459.
Rosenbauer, H., N. Shutte, A. Galeev, K. Gringauz, and I. Apathy, 1989. Ions of Martian origin
and plasma sheet in the Martian magnetosphere - Initial results of the TAUS experiment,
Nature, 341, 612–614, doi:10.1038/341612a0.
Russell, C., and R. Elphic, 1979. Observation of magnetic flux ropes in the Venus ionosphere,
Nature, 279, 618–620.
Tanaka, T., and K. Murawski, 1997. Three-dimensional MHD simulation of the solar wind
interaction with the ionosphere of Venus: Results of two-component reacting plasma simu-
lation, J. Geophys. Res., 102(A9), 19,805–19,821.
Vignes, D., C. Mazelle, H. Rème, M. H. Acuña, J. E. P. Connerney, R. P. Lin, D. L. Mitchell,

P. Cloutier, D. H. Crider, and N. F. Ness, 2000. The solar wind interaction with Mars:
locations and shapes of the Bow Shock and the Magnetic Pile-up Boundary from the obser-
vations of the MAG/ER experiment on board Mars Global Surveyor, Geophys. Res. Lett.,
27(1), 49–52, doi:10.1029/1999GL010703.
BIBLIOGRAPHY 161
Vignes, D., M. Acuña, J. Connerney, D. Crider, H. Rème, and C. Mazelle, 2004. Magnetic Flux
Ropes in the Martian Atmosphere: Global Characteristics, Space Sci. Rev., 111, 223–231,
doi:10.1023/B:SPAC.0000032716.21619.f2.
Wieser, M., S. Barabash, M. Emanuelsson, K. Brinkfeldt, and P. Enoksson, 2006. Microme-

chanical shutter based mass spectrometers for charged and neutral particles, in Proceedings
Book of the International Conference on Future Perspectives of Space Plasma and Particle
Instrumentation and International Collaborations, Rikkyo, Japan.
Williams, D. J., E. C. Roelof, and D. G. Mitchell, 1992. Global magnetospheric imaging,

Reviews of Geophysics, 30(3), 183–208.
Winningham, J. D., R. A. Frahm, J. R. Sharber, A. J. Coates, D. R. Linder, Y. Soobiah,

E. Kallio, J. R. Espley, R. Lundin, S. Barabash, M. Holmström, H. Andersson, M. Ya-
mauchi, A. Grigoriev, J. R. Scherrer, S. J. Jeffers, D. O. Kataria, J. Kozyra, J. Luhmann,
E. Roelof, D. Williams, S. Livi, C. C. Curtis, K. C. Hsieh, B. R. Sandel, H. Koskinen,
T. Säles, P. Riihela, W. Schmidt, M. Grande, M. Carter, J.-A. Sauvaud, A. Fedorov, J.-J.
Thocaven, S. McKenna-Lawler, S. Orsini, R. Cerulli-Irelli, M. Maggi, P. Wurz, P. Bochsler,
N. Krupp, J. Woch, M. Fraenz, K. Asamura, and C. Dierker, 2006. Electron oscillations in the
induced Martian magnetosphere, Icarus, 182, 360–370, doi:10.1016/j.icarus.2005.10.033.
Witte, M., H. Rosenbauer, E. Keppler, H. Fahr, P. Hemmerich, H. Lauche, A. Loidl, and

R. Zwick, 1992. The interstellar neutral-gas experiment on ULYSSES, Astron. Astrophys.
Suppl. Ser., 92, 333–348.
Wiza, J. L., 1979. Microchannel plate detectors, Nuclear Instruments & Methods, 162, 587–
601.
Wurz, P., 2000. Detection of energetic neutral atoms, in The Outer Heliosphere: Beyond
the Planets, edited by K. Scherer, H. Fichtner, and E. Marsch, pp. 251–288, Copernicus-
Gesellschaft, Katlenburg-Lindau, Germany.
Wurz, P., and H. Lammer, 2003. Monte-Carlo simulation of Mercury’s exosphere, Icarus, 164
(1), 1–13.
Wurz, P., A. Galli, S. Barabash, and A. Grigoriev, 2006. Energetic Neutral Atoms from the
Heliosheath, in Physics of the Inner Heliosheath, American Institute of Physics Conference
Series, vol. 858, edited by J. Heerikhuisen, V. Florinski, G. P. Zank, and N. V. Pogorelov,
pp. 269–275, AIP Conference Proceedings, doi:10.1063/1.2359338.
Ziegler, J. F., J. P. Biersack, and U. Littmark, The Stopping and Range of Ions in Solids, Perg-
amon Press, New York, 1985.
162 BIBLIOGRAPHY
Glossary of Acronyms
1D one-dimensional MOS FET Metal-Oxide-Semiconductor
2D two-dimensional Field-Effect Transistor
AIV assembly, integration, and MPB Magnetic Pile-up Boundary
verification MSO Mars-Sun orbit
ADC Analog-to-Digit Converter NPI Neutral Particle Imager
ASPERA Analyzer of Space Plasma NPD Neutral Particle Detector
and EneRgetic Atoms NSW Neutral Solar Wind
AU Astronomical Unit OSR Optical secondary reflection
BS Bow Shock PIPPI Prelude In Planetary Particle Imaging
CELIAS Charge, Element, and Isotope PH Pulse Height
Analysis System SEY Secondary Electron Yield
CEM Channel Electron Multiplier SOHO Solar and Heliospheric Observatory
CIR Co-rotating Interaction Region SRAM Static Random Access memory
DAC Digit-to-Analog Converter SRIM Stopping and Range of Ions in Matter
DC Direct current SWAN Solar Wind Anisotropies
DigTOF Digital Time-of-Flight electronics SZA Solar Zenith Angle
DPU Digital Processing Unit TC Telecommand
EEPROM Electrically Erasable Program- TDC Time-to-Digit Converter
mable Read Only Memory TOF Time-of-Flight
ELS ELectron Spectrometer TS Termination Shock
EMC Electromagnetic compatibility TVAC Termo-vacuum
ENA Energetic neutral atom UV Ultra Violet
ESA European Space Agency VLENA Very low-energy neutral atoms
EUV Extreme Ultra Violet
FEE Front-end Electronics
FOV Field-of-View
FPGA Field Programmable Gate Array
FWHM Full Width at Half Maximum
HENA High-energy neutral atoms
HK Housekeeping
HSTOF High-Energy Suprathermal
Time-of-Flight
HV High Voltage
HV PS High Voltage Power Supply
IDL Interactive Data Language
IMA Ion Mass Analyzer
IMAGE Imager for Magnetopause-to-
Aurora Global Exploration
IMB Induced magnetosphere boundary
IMF Interplanetary magnetic field
IMI Ion Mass Imager
IP Ionopause
LENA Low-energy neutral atoms
LOS Line-of-Sight
MCP MicroChannel Plate
MEMS Micro-Electro-Mechanical System
MENA Medium-energy neutral atoms
MGS Mars Global Surveyor
MHD Magnetohydrodynamics
MLI Multi-layer insulation
Acknowledgments
First of all I express my sincere gratitude to my supervisor Professor Stas Barabash, for his sup-
port, continuous guidance, and criticism during the ASPERA-3 and ASPERA-4 projects, and
for his inexhaustible patience during the writing phase of this dissertation. I appreciate highly
his vast knowledge, expertise, and understanding which added considerably to my graduate
experience.
I am deeply indebted to Andrei Fedorov who introduced me to and made me interested in
the fields of space plasma physics and experimental science.
I thank Mats Holmström, Tima Sergienko, Martin Wieser, and Masatoshi Yamauchi for
scientific support which helped enrich my experience. Special thanks go to Futaana Yoshifumi,
who supported me with advanced knowledge on how to write and revise a scientific paper.
I am very grateful to Rick McGregor for the uphill work of proof-reading this thesis.
I would also like to thank the Swedish National Graduate School of Space Technology
which provided financial support.
I thank the electronics lab staff involved in development, building and verification of the
NPD sensors. Thanks also to the IRF workshop which manufactured different parts of the NPD
sensor, and especially to Tero Saarijärvi for performing the tough job of making tens of holes
with a thread for M1.2 screws in several NPD chassis.
I acknowledge the ASPERA team for making it possible to build the NPD instruments.
I am grateful to Leif Kalla for advice on software development, and to Grigory Nikulin for
fruitful discussions about IDL tricks.
I am very thankful to my wife Daria whose love, patience, moral support, and editing
assistance have sustained me, particularly in those many days in which I spent more time with
my computer than with my family. I am in debt to my parents for their understanding, unending
patience and encouragement when it was most required.
Finally, I say thank you to all colleagues and friends of mine at IRF in Kiruna for keeping
a friendly work environment here.
163

Full Text 01

Uploaded by

Copyright:

Available Formats

Full Text 01

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Full Text 01

Uploaded by

Copyright:

Available Formats

The Neutral Particle Detector

on the Mars and Venus Express missions

Swedish Institute of Space Physics

IRF Scientific Report 290

Nyckelord: ENA-avbildning, exosfär, magnetosfär, Mars, Venus, solvindens växelverkan

1 Energetic neutral atoms in space 3

2 Energetic neutral atoms detection 15

3 The ASPERA-3 and ASPERA-4 experiments 31

4 The Neutral Particle Detector (NPD) 41

5 The NPD calibrations 63

6 Scientific results. The NPD measurements at Mars. 119

7 Summary and future prospects 145

A NPD data processing 147

Glossary of Acronyms 161

Energetic neutral atoms in space

1.1 Production mechanisms

1.1.1 Charge exchange

A+energetic + Mcold → Aenergetic + Mcold

Atmospheric sputtering. The atmospheric sputtering mechanism with reference to Mars is

1.3 ENAs at non-magnetized planets

ENA source Energy, Flux, cm−2 sr−1 s−1 keV−1 Species

Induced magnetosphere boundary

• Accelerated planetary ions

• Back-scattered hydrogen (ENA albedo)

1.3.2 ENA environment of Venus

Energetic neutral atoms detection

2.1 ENA imaging

Image plane Image plane

Object plane Object plane

2.2 Principle functions of ENA instruments

2.3 Deflection systems

2.5 ENA detection and analysis: Instrument examples

2.5.2 Surface interaction

Figure 2.8: Schematic view of a neutral-to-negative ion conversion process.

Figure 2.10: ENA deflection on a reflection surface.

SSD sensors Deflectors / collimators

SSD sensor card

Retarding grid Imaging MCP

Secondary ion emission

Secondary electron emission

2.5.3 High frequency shutters

The ASPERA-3 and ASPERA-4

3.1 Scientific objectives

Scientific objectives Associated measure- Measurement requirements

Table 3.1: The ASPERA-3 scientific objectives

Table 3.1, together with the corresponding instrument requirements.

- What is the mass composition of the escaping plasma?

3.2 Instrument overview

tion of ultra-light, low-power electron sensors. It is a standard top-hat electrostatic analyzer in

Parameter NPI NPD ELS IMA

3.2.1 The Ion Mass Analyzer (IMA)

3.2.2 The Electron Spectrometer (ELS)

3.2.3 The Neutral Particle Imager (NPI)

3.2.4 The Digital Processing Unit (DPU) and the scanner

The Neutral Particle Detector (NPD)

4.1 The measurement technique

Start Surface Stop Surface

b) Stop MCP secondary

4.2 NPD mechanical design

Figure 4.3: The 3D model of the NPD sensor.