Energetic Neutral Atoms (ENA) in Space. Mike Gruntman, Energetic neutral atom imaging of space plasmas, Review of Scientific Instruments, 1997.

This is a review paper published in the leading archival journal on physics instrumentation Review of Scientific Instruments (M. Gruntman, Energetic neutral atom imaging of space plasmas, Vol. 68, No. 10, pp. 3617-3656) in 1997. Download the full text of the article in pdf format. Several references to Mike Gruntman's publications are also linked directly to downloadable pdf files.

Experimental techniques and instrumentation for space plasma imaging in fluxes of energetic neutral atoms (ENAs) are reviewed. ENAs are born in charge exchange collisions between space plasma energetic ions and background neutral gas. ENAs are ubiquitous in the space environment and their energies are in the range from a few eV up to >100 keV. Contrary to charged particles, ENAs can travel large distances through space with minimal disturbance, and by recording ENA fluxes as a function of observational direction, one can reconstruct a global image of a planetary magnetosphere or the heliosphere. Plasma ion energy distribution and ion composition can be remotely established by measuring ENA energies and masses. ENA imaging opens a new window on various phenomena in space plasmas with a promise to qualitatively improve our understanding of global magnetospheric and heliospheric processes. At first we review ENA fluxes in space and their properties, and present a brief history of ENA experimental studies and the evolution of experimental approaches. The concepts of ENA imaging and particle identification are considered and followed by comparison with corpuscular diagnostics of fusion plasmas. Basic ENA techniques and instrument components are then described in detail and critically evaluated; performance characteristics, limitations, and requirements to key instrumental elements are discussed. And finally, representative ENA instruments are shown, and promising instrumental approaches are identified.

In the beginning of the space age, in the 1950s and early 1960s, many space experiments were simply an extension of the measurements performed in the physics laboratory, although under very unusual conditions of spacecraft. Severe limitations of available power, requirements of small mass and size, remote control and data acquisition through limited telemetry called for new approaches to designing and building space instruments. An emphasis on reliability—one cannot go and replace a fuse or correct axis alignment on a spacecraft in orbit—was another important new requirement. Space experiments were initially considered by many in the physics community as an unconventional application of simplified laboratory techniques. (As Samuel Johnson put it, ‘‘A horse that can count to ten is a wonderful horse, not a wonderful mathematician.’’) However with time, space instruments evolved into a highly specialized area of scientific instrumentation pioneering new measurement techniques and leading instrument development in such areas as plasma analyzers, particle- and photon-counting position-sensitive detectors, and many others. In this article we review a new emerging field of space experiments and instrumentation:
imaging of space plasmas in fluxes of energetic neutral atoms.

The interaction between charged and neutral particles is a common phenomenon in space plasmas. Whenever an energetic ion undergoes a charge exchange process in a collision with a neutral background atom, an energetic neutral atom (ENA) is born. Ion-electron recombination and neutral atom acceleration by the solar gravitation may also contribute to an ENA population under certain conditions. ENAs are ubiquitous in space environment and their study opens a new window on various phenomena in space plasmas with a promise to qualitatively improve our understanding of global magnetospheric and heliospheric processes. However ENAs have remained poorly explored due to enormous experimental difficulties.

ENAs, contrary to charged particles, can travel large distances through space with minimal changes without undergoing further interaction with plasma. ENA measurements are recognized as a powerful tool to remotely study various global plasma objects in space [1–8,9,10,11,12]. By recording ENA fluxes as a function of observational direction, one can reconstruct a global image of the object of interest, thus the term ‘‘ENA imaging,’’ first introduced in 1984 for imaging from outside [13] and from inside [14] of the magnetosphere. Plasma ion energy distribution and ion composition can be remotely established by measuring ENA energies and masses. ENA imaging usually means not only determining ENA flux angular distribution but also ENA energies and masses. An ENA imaging experiment would ideally produce a set of images of a plasma object in ENAs of different masses and in different energy ranges.

Protons are the most abundant component of space plasma ions. Unlike other space plasma ions (e.g., He⁺ and O⁺), protons cannot be imaged optically, which makes ENAs in many cases the only tool to study processes of interest remotely. The definition of an ‘‘energetic’’ particle was originally limited by a minimum energy of several keV, but new experimental techniques have significantly lowered the energy threshold. For the purpose of this article, the ENAs occupy the energy range from few eV up to several hundred keV.

Any object that contains energetic ions and background neutral gas can be imaged in ENA fluxes. Important examples are planetary magnetospheres and the heliosphere. (The heliosphere, the region containing expanding solar wind plasma, is about 200 AU in size; 1 AU = 1 astronomical unit = 1.5´10¹³ cm is the distance between the earth and the sun.) Planetary magnetospheres are filled with plasma and the solar wind plasma fills interplanetary space in the heliosphere. Being far from thermodynamic equilibrium, space plasmas are characterized by wildly varying populations of energetic ions. Neutral atom background around the Earth is provided by the terrestrial exosphere that contains escaping hydrogen atoms [15–17]; the extended hydrogen geocorona was observed and measured many times [18,19].

The presence of a global population of neutral atoms in interplanetary space is less known. The interstellar gas from the local interstellar medium (LISM) permeates the heliosphere. If one excludes the sun, planets, and other celestial bodies, then 98%–99% of the mass of matter filling the heliosphere is represented by neutral atoms with only the remaining 1%–2% of matter being plasma [10]. Outside the heliosphere, interstellar space is filled by dilute interstellar gas with varying degrees of ionization. Hence ENAs are born in the planetary environment and in interplanetary and interstellar space as well.

Almost 3 decades of instrument development made ENA imaging of space plasmas possible: the phase of practical implementation has been finally achieved. Simple ENA instruments were recently flown on GEOTAIL [20], CRRES [21], and ASTRID [22,23] missions around the Earth. A sophisticated first large size ENA camera will perform imaging of the Saturn’s magnetosphere on the Cassini mission to be launched in October 1997 [24]. A dedicated space mission to globally image the terrestrial magnetosphere was studied for several years [25–27]. A medium-class explorer (MIDEX) mission IMAGE was recently selected by NASA to perform such imaging; it is presently under preparation for launch in January, 2000. Another ENA experiment ISENA [28] was unfortunately lost with the SAC-B spacecraft during launch in October 1996. A number of other experiments to image planetary magnetospheres and the heliosphere have been proposed and are currently at different stages of development.

ENA images and their evolution in time promise a breakthrough in the understanding of fundamental global processes in space. Conventional in situ measurements of local plasma parameters are inherently limited in their capabilities. Some plasma regions of interest are too far away to be conveniently visited by spacecraft. For example, only remote observations are capable of providing continuous monitoring of the time-varying size and shape of the global heliosphere [6–9,29].

Experimental studies of planetary magnetospheres face difficulties of another kind. The measurements of magnetospheric plasma are performed from fast moving (2–9 km/s) spacecraft, and they cannot unambiguously distinguish between temporal and spatial variations of plasma parameters [1,5]. Consequently even the simultaneous measurements from several spacecraft are inherently insufficient for reconstruction of complex global magnetospheric processes and require heavy reliance on often simplified and incomplete models.

Accurate understanding of the global magnetospheric processes has become especially important with the growing realization of possible adverse effects of space environment on many technological systems, both on the ground and in space. Communications, TV broadcasting, world-wide navigation, and national security applications that include advance warning, reconnaissance, and nonproliferation compliance monitoring, are increasingly dependent on space deployed technical assets. Our ability of predicting the magnetospheric conditions, especially during geomagnetic storms, remains disappointing [30]. NASA, the Air Force, and National Oceanic and Atmospheric Administration (NOAA) are currently working to establish a national space weather service [31,32].

Although first dedicated direct ENA measurements were attempted in the late 1960s, enormous experimental difficulties prevented detailed study of ENAs. ENA fluxes are very weak, sometimes <1 cm^–2 s^–1, and the realistic approach to their direct detection is based on particle interaction with solid surfaces, e.g., electron emission. Ultraviolet (UV) and extreme ultraviolet (EUV) photons interact with surfaces often in a similar way, and the background EUV/UV photon fluxes are 3–7 orders of magnitude higher than those of ENAs. Therefore background photon-induced count rate of a conventional secondary electron multiplier would be 2–6 orders of magnitude higher than the ENA count rate. Such inhospitable conditions make ENA measurement an exceptionally challenging task.

There are similarities and there are essential differences between corpuscular diagnostics of hot plasmas in the laboratory and ENA diagnostics of space plasmas. Neutral atom emissions from magnetically confined fusion plasmas were used efficiently to determine plasma ion temperature [33–36]. The photon-to-ENA ratios in space are not unlike those from fusion plasmas. An important difference between conditions in space and those in the laboratory is that fusion plasma processes are of relatively short duration, <1 – 10 s, with high ENA fluxes, while ENA fluxes are very low in space but it is possible to accumulate the signal much longer. For example, the desired temporal resolution for study of important magnetospheric ring current is in the 5–15 min range. The fusion plasmas have become ‘‘ENA-thick’’ with the increasing density, which limits passive corpuscular diagnostics to the study of the plasma edges. In space, the objects are usually ‘‘ENA-thin,’’ and ENAs can travel large distances without much disturbance.

The goal of this article is to review and critically evaluate experimental techniques and instrumentation for space plasma imaging in ENA fluxes, covering the energy range from a few eV up to >100 keV. ENA instrumentation was traditionally divided into two groups corresponding to ‘‘high’’ and ‘‘low’’ energies with a new ‘‘ultralow’’ energy group emerging. (Due to the lack of a better phrase we will be using hereafter such awkward terms as ‘‘low-energy ENA’’ and ‘‘high-energy ENA.’’) The division between high- and low-energy ENAs, although never clearly defined, results from use of different approaches and instrument components for suppression of EUV/UV radiation. For example, solid state detectors and thin-film filters are used in high-energy ENA instruments, while many other components, such as ultrathin foils and microchannel plate detectors, are common for both groups. ENA instruments employ many approaches and techniques widely used in space ion analyzers which is an exceptionally developed and advanced area of space instrumentation [37–49]; the ion analyzers are beyond the scope of this article.

At first we review ENA fluxes in space and their properties, and present a brief history of ENA experimental study and the evolution of experimental approaches. The concepts of ENA imaging and particle identification are considered and followed by comparison with corpuscular diagnostics of fusion plasmas. Basic ENA techniques and instrument components are then described in detail and critically evaluated; performance characteristics, limitations, and requirements to key instrumental elements are discussed. And finally, representative ENA instruments are shown and promising new approaches and developments are identified.

ENA fluxes come from different ion populations with different compositions, flux levels, and energy, spatial, and temporal dependencies. ENAs are formed in charge exchange collisions (Fig. 1) between energetic plasma ions and neutral gas atoms

where I₁⁺ is an ion of species ‘‘1’’ and A₂ is an atom of species ‘‘2.’’ Species ‘‘1’’ and ‘‘2’’ may be identical (e.g., H⁺ + H → H + H⁺), and a simultaneous exchange of two electrons is possible (He⁺⁺ + He → He + He⁺⁺). The initial velocity of an energetic particle is only slightly changed in a charge exchange collision [50]. Proton–hydrogen charge-exchange collisions are often the most important process in space plasma: they occur at large impact parameters with only a small momentum exchange between collision partners. For many practically important applications ENAs can be assumed to be born exactly with the ion momentum.

FIG. 1. Charge exchange collision between an energetic plasma ion and a neutral gas atom; I₁⁺ is an ion of species ‘‘1’’ and A₂ is an atom of species ‘‘2.’’

After the birth of an ENA, its trajectory is defined by the initial velocity and gravitational forces only. With a few exceptions, gravitation can be disregarded, and one can assume that the ENA preserves both the direction and magnitude of the energetic ion velocity before the charge-exchange collision. As ENAs travel in space, they may be lost in charge exchange, electron collisions, and photoionization.

Only few species are important for ENA formation. Neutral gas in the heliosphere consists of hydrogen (~90%) and helium (~10%) atoms. Atomic hydrogen dominates the neutral particle environment around the earth from an altitude of 600 km and a few thousands km during periods of minimum and maximum solar activity, respectively. Other important neutral species around the earth are helium and oxygen atoms. Magnetospheric plasma consists mostly of protons with some helium, oxygen, and sulfur (found at Jupiter) ions. The protons are the major component of the interstellar and solar wind plasmas; the latter contains also ~5% of double-charged helium ions (alpha particles).

Charge exchange cross sections important for ENA production are readily available [51]. A cross section energy dependence for proton charge exchange on hydrogen and helium atoms is shown in Fig. 2. The difference in cross sections reflects the fact that charge exchange is of a resonance type for proton–hydrogen collisions and requires overcoming an energy threshold in proton–helium collisions. Since the background neutral hydrogen is usually much more abundant than helium, the charge exchange on hydrogen atoms would dominate hydrogen ENA production for energies <10– 100 keV. The charge exchange on helium may become however important for energies >100 keV.

FIG. 2. Cross section energy dependence for proton charge exchange on hydrogen and helium atoms.

The ENA measurements would allow one to study the ion population if neutral gas parameters are known, and conversely the neutral atom population characteristics could be obtained if the ion parameters are known. Some information is usually available on both the ions and neutral gas, but often we have a relatively good knowledge of only one component, for example neutral particle environment around the earth. Let us consider now the sources of major magnetospheric and heliospheric ENA fluxes.

Magnetospheres are objects formed by the solar wind plasma flow around planets with intrinsic magnetic field (Mercury, Earth, Jupiter, Saturn, Uranus, and Neptune). Magnetic field presents an obstacle for the highly supersonic solar wind plasma flow, and a bow shock is formed in front of the planet. The size and shape of the magnetospheres are determined by the strength and orientation of the magnetic field, which is usually compressed at the sunward side and significantly stretched at the nightside (magnetospheric tail). Planets without intrinsic magnetic field (Venus, Mars, and Moon) may sometime form magnetosphere-like plasma structures around the upper atmospheres that are ionized by solar EUV and x-ray radiation. The earth’s magnetic field dominates the terrestrial magnetosphere and efficiently traps charged particles (Fig. 3), which may be accelerated up to very high energies [52]. The ionospheric and solar wind plasmas find a way to leak and fill the magnetosphere with ions and electrons.

FIG. 3. Planet’s magnetic field dominates the magnetosphere and traps charged particles, which may be accelerated up to very high energies. Trapped charged particles gyrate about magnetic field lines, participate in drift motion, and form radiation belts. (After Ref. 52.)

Various aspects of magnetospheric physics are discussed in a number of publications [53–68]. The earth’s magnetosphere is shown schematically in Fig. 4. The bow shock is located approximately at the distance of 12–18 R_E (R_E » 6370 km is the earth’s radius). A long magnetospheric tail may stretch far beyond the Moon’s orbit (~380 000 km) in the antisolar direction. Processes in the magnetospheric tail play an important role in transport of energized plasma toward the earth during magnetospheric disturbances. The orbits of practically all earth-orbiting satellites are inside the magnetosphere. Low-earth orbits are those with altitudes below a few thousand km. A large number of communications and direct broadcasting satellites are at geosynchronous (geostationary) orbit with a one-day period. This orbit is at approximately a 36 000 km altitude which is well within the magnetosphere.

FIG. 4. Three-dimensional cutaway view of the terrestrial magnetosphere showing various currents, fields, and plasma regions. (After Ref. 197.)

A magnetosphere is a very complicated object with plasma parameters varying wildly from one region to another. Solar wind carries frozen-in interplanetary magnetic field, about ~5 nT near the earth, which changes its direction occasionally. Interplanetary magnetic field interaction with the geomagnetic field is important for energy transfer to the magnetosphere. Many processes involving magnetic field occur in the boundary region between the plasma of the geospheric origin and the solar wind plasma, the geopause [67]. The geopause is at a geocentric distance of ~10 R_E on the upstream side. Variations of the solar wind pressure and magnetic field result in change of the magnetosphere’s size, shape, electric current patterns, and ion flows and lead to energization of electrons and ions and development of various instabilities [53,58,69].

The solar-terrestrial link through the interaction of the solar wind with the magnetosphere makes magnetospheric conditions strongly dependent on the solar activity. The most prominent manifestation of this link is large nonrecurrent geomagnetic storms, which are believed to be triggered by relatively dense clouds of plasma ejected from the solar surface, the so called coronal mass ejections, impinging on the magnetosphere [70–76]. A magnetospheric storm is accompanied by an increase of a ring current [5,69,77,78], which produces perturbations up to 1% of the magnetic field at the earth’s surface. The ring current is an electric current flowing around the earth due to the presence of energetic ions in the magnetosphere; its decay occurs largely through charge exchange on background neutral atoms resulting in production of ENA fluxes [69,77,79].

Magnetospheres of other planets have been studied by flyby spacecraft (Mercury, Jupiter, Saturn, and Uranus), in some cases (Venus and Mars) by orbiters and landers, and by astronomical observations and radio emission detections from earth. There are detailed descriptions of the magnetospheres of Mercury [80–82], Venus [81,83], Mars [81,84], Jupiter [85,86], Saturn [87] and Uranus [88]. The magnetosphere of Jupiter is currently being explored by the Galileo orbiter, and the magnetosphere of Saturn will be studied by the forthcoming Cassini mission.

The emphasis of the study of the terrestrial magnetosphere has shifted from discovering the new magnetospheric features to attempts to explain how the magnetosphere works. Such shift is driven not only by the maturing of the field but by the practical requirements as well. An important goal is to construct a magnetosphere model that will provide capabilities of predicting magnetosphere’s behavior in response to solar disturbances, in particular characteristics of geomagnetic storms and substorms. Although a clear understanding of the relation between the disturbed space weather conditions and adverse effects on various technological systems has yet to be achieved [71,89,90], significantly improved knowledge of geomagnetic storm processes is needed to better understand and to reduce storm-related damage.

On the ground, major magnetospheric disturbances cause changes in the geomagnetic field, which in turn lead to the induction currents in long conductors [91]. Effects of geomagnetic disturbances include disruptions of cable communications, which were observed in telegraph lines since the middle of the 19th century [91,92]; interference with navigational systems such as LORAN and OMEGA and high-frequency (HF) and ultrahigh frequency (UHF) communications; various effects (leading to blackouts of large areas) on power distribution systems [93]; corrosion of pipelines [94,95]; and interference in high-resolution global positioning system (GPS) technology. Magnetospheric storms significantly increase precipitation of energetic particles, which poses health hazards to airline crews and passengers at high altitudes on polar routes; the crews of high-altitude reconnaissance planes may also be affected.

In space, magnetospheric disturbances damage and reduce lifetime of satellites [96,97], both at low-earth and geosynchronous orbits. In particular, excessive charging of spacecraft surfaces may cause irreparable damage to space systems [98,99]. Major geomagnetic storms lead to heating and expansion of the upper atmosphere. The atmospheric expansion may lead to a significant increase of atmospheric drag on low-altitude satellites and cause their premature reentry. Establishment of a national space weather service requires significantly improved understanding of adverse effects of the disturbed space weather conditions on technological systems [89,90,97] as well as capabilities of advance warning and storm prediction [30–32].

Plasmas in different regions of the terrestrial magnetosphere, radiation belts, cusps, plasmasphere, magnetosheath, plasma mantle, plasma sheet, and polar wind, are characterized by widely varying parameters, and these regions actively interact with each other. For example, magnetosheath plasma has temperatures of 0.1–1 keV, proton energies in the magnetospheric tail proton streams are 10–100 eV, while ions in the ring current have energies between few hundreds eV up to hundreds of keV. The macroscale characteristics of the magnetospheric regions and their boundaries and how they interact to define global characteristics of the magnetosphere are not well known [5,100] resulting in a long list of unanswered specific questions [101].

The uppermost part of a planetary atmosphere, the exosphere [15–17], provides neutral collision partners for ENA production. The exospheric hydrogen atom population was extensively studied around earth by the Dynamic Explorer (DE-1) satellite [102,103]. The number density distribution, which can be assumed to be spherically symmetric close to the earth, is shown in Fig. 5 [102]. At larger geocentric distances, solar radiation pressure (in resonance hydrogen H I Ly-α line, 1216 Å) would produce an asymmetry of hydrogen spatial distribution [104]. Abundant background neutral gas can be found in the magnetospheres of other planets as well [105].

FIG. 5. Radial profile of the spherically symmetric geocoronal hydrogen number density. (After Ref. 102.)

The composition of magnetospheric ENAs is largely determined by the composition of energetic ions. Hydrogen (H), helium (He), and oxygen (O) ENAs have been identified by the first experimental ENA composition measurement in the terrestrial magnetosphere [106]. One expects to find also sulfur (S) ENAs in Jupiter’s magnetosphere (sulfur is abundant in the plasma torus as a result of volcanic activity on Jupiter’s moon Io).

Extensive computer modeling [1,2,5,79,107–118] predicts magnetospheric ENA fluxes escaping the earth’s magnetosphere outward in the range 0.01–10 cm^-2 s^-1 sr^-1 keV^-1. For example, the ENA flux at 40 keV energy is ~10² ´(d/R_E)² cm^-2 s^-1 keV^-1 at large distance, d, from the earth [5]. A typical angular resolution in ENA imaging instruments (an angular size of imaging pixels) is 5°´5°»8´10^-3 sr, or smaller, and the expected magnetospheric fluxes are in the range 10^-4–10^-1 cm^-2 s^-1 keV^-1 per pixel.

An important type of ENAs can precipitate toward the earth surface [118–123], for example ENAs born in charge exchange of ring current ions (Fig. 6). The ENAs may reach low altitudes where they are re-ionized by charge exchange and newly born energetic ions are trapped by the magnetic field [124]. The precipitating ENAs can be studied from low-earth orbit spacecraft and were recently measured by the CRRES [21,125] and ASTRID [22,23,118] satellites. The characteristics of precipitating ENAs are important for verification and testing of our theories of nocturnal thermospheric heating, low latitude aurorae, formation of a low-altitude ion belt at low latitudes, particle precipitation, and escaping neutrals. The precipitating ENA fluxes may be as high as 10⁴ cm^-2 s^-1 sr^-1 keV^-1 [28,111,121].

FIG. 6. ENAs produced in charge exchange of ring current ions. The neutrals travel on straight line trajectories, mostly outwards, but a fraction impinges on the upper atmosphere, and depending on species and energy, some will then re-ionize, and near the equator become temporarily trapped. (After Ref. 120.)

The measurements of precipitating ENA fluxes are possible only at altitudes higher than a certain level where the effect of collisions with ionospheric and atmospheric ions and neutrals is minimal. The collisions with atmospheric neutrals are the most important limitation [123], and the computer simulations [123] showed that ENA measurements would be possible during solar maximum at altitudes higher than 600–700 km with less than 10% of ENAs affected. Actually the disturbance of ENA fluxes would be smaller since the effect of the collisions in the computer simulation [123] was overstated: the scattering angles of most of the ENAs that experienced an elastic collision would be smaller than the angular resolution of an ENA instrument. In addition, some elastic collision cross sections [126] were calculated in the angular range where the elastic scattering model assumption was not valid. One can expect that the measurements of precipitating ENAs would be possible at altitudes as low as 500–600 km.

The interaction between the sun and LISM is manifested by the buildup of a heliosphere [127–131]. The sun is a source of the highly supersonic flow of plasma called the solar wind [132–135]. Solar wind expands into the LISM which is filled with partially ionized interstellar gas, interstellar magnetic field, and cosmic rays. The LISM is a medium with a small but finite pressure. The dynamic pressure of the expanding solar wind flow decreases with distance from the sun, and at a certain distance the solar wind expansion must be stopped. The cavity containing the solar wind is called the heliosphere. The dynamic pressure of the solar wind varies by a factor of 2 during the 11-year solar cycle [136], thus resulting in variations of the size and shape (heliosphere ‘‘breathing’’) of the heliosphere [136,137].

The heliosphere is a complicated phenomenon where solar wind and interstellar plasmas, interstellar gas, magnetic field, and cosmic rays play prominent roles. The structure of the heliosphere and its boundary, as well as properties of the LISM, are of fundamental interest and importance and the available experimental data are scarce and indirect. The heliosphere provides a unique opportunity to study in detail the only accessible example of a commonplace but fundamental astrophysical phenomenon—the formation of an astrosphere. A self-consistent model of the stationary heliosphere has yet to be built and some aspects of the interaction, for example temporal variations and instabilities, are not satisfactorily understood even on the qualitative level. The physics of the LISM is also poorly understood [138–140].

FIG. 7. Two-shock model of the solar wind interaction with the local interstellar medium (LISM). (TS) termination shock; (HP) heliopause; (BS) bow shock; (CR) cosmic rays; [ISG(P)] interstellar gas (plasma); (B) magnetic field.

Various heliospheric models were proposed for different solar wind plasma and LISM parameters [128–130,141–147]. A possible heliospheric structure is shown in Fig. 7 for the most advanced and quantitatively developed two-shock model [146,148–150]. A supersonic flow of the solar wind plasma terminates at a solar wind shock front [termination shock (TS)] beyond which its kinetic energy is largely converted into thermal energy of the subsonic plasma. A supersonic flow of the interstellar plasma (‘‘interstellar wind’’) is stopped at the bow shock (BS).

The interstellar wind blows from a direction with ecliptic longitude 252° and latitude 17° and with a velocity 26 km/s = 5.5 AU/year [151–154], and the heliosphere is likely to be elongated along this velocity vector (Fig. 7). Solar wind anisotropy [135] as well as the presence of the (currently unknown) interstellar magnetic field (typical value 0.1–1 nT) would result in deviations from cylindrical symmetry. The earth is positioned in the upwind direction in the beginning of June each year. Neutral interstellar gas consists mostly (80%–90%) of hydrogen atoms with a number density 0.05– 0.2 cm^-3; the remaining atoms are helium with the addition of traces of heavier elements (O, Ne, Ar, etc.) [138–140,155]. Interstellar gas is partially ionized [138,140,156], and the interstellar plasma is believed to have a number density 0.02–0.1 cm^-3. The degree of interstellar gas ionization would affect the morphology of the heliospheric interface region [150].

The estimates of the size of the heliosphere vary between 70 and 120 AU. The closest possible position of the termination shock is ‘‘pushed’’ steadily away from the sun by the Voyager 1 spacecraft, which was at 65.2 AU on January 1, 1997, and which continues to move in approximately the upwind direction with the speed of 3.5 AU/year. Voyager 1 did not cross the termination shock yet. Another distant spacecraft, Pioneer 10, was at 66.6 AU downwind from the sun on January 1, 1997, and it moves with the speed of 2.7 AU/year. The Pioneer 10 scientific mission was terminated several months ago due to decreasing capabilities of its power system.

Production of heliospheric ENAs requires background neutral gas. Two major sources of background neutral particles in the heliosphere are provided by the interstellar gas penetrating the solar system and, in the sun’s vicinity (<0.5 AU), solar wind plasma neutralization on interplanetary dust. Planets provide localized sources of thermally escaping neutral atoms. The lifetime of a hydrogen atom with respect to ionization is about 20 days at 1 AU. Hydrogen would form a cloud with a radius of 0.01 AU around the earth. Thus the highly localized planetary neutrals can be disregarded, when global populations of heliospheric ENAs are considered. However ENAs emitted by giant planets (Jupiter and Saturn) may substantially contribute to a global population of heliospheric neutral minor constituents, viz., atomic oxygen [157].

Number density of interstellar gas is so small (~0.1 cm^-3) that atom mean free path with respect a collision is larger than the expected size of the heliosphere. Therefore interstellar neutral atoms can be treated as individual particles moving under the forces of the solar gravitational attraction and solar radiation pressure [158]. The radiation pressure approximately counterbalances solar gravitation for hydrogen atoms but is unimportant for helium. Atom loss occurs due to ionization by charge exchange with the solar wind ions, solar EUV photoionization, and collisions with solar wind electrons, the latter process being important only close to the sun (<1 AU).

A concept of interstellar hydrogen and helium atom penetration of the heliosphere is supported by extensive experimental observations. The techniques to study interstellar gas include observation of interplanetary glow (resonant scattering of the solar radiation in H I 1216 Å and He I 584 Å lines) [141–144,158–162], detection of pickup ions in the solar wind [163–166], direct detection of interstellar helium flux [153,154], and astronomical observations of the nearby interstellar medium [151,152,167,168]. An important minor heliospheric constituent, oxygen, can be from interstellar [169,170] and magnetospheric [157] sources. The number density of interstellar neutrals filling the heliosphere is in the 0.01–0.1 cm^-3 range.

Another source provides neutral gas in the sun’s immediate vicinity (<0.5 AU). Interplanetary space is filled with a population of interplanetary, or zodiacal dust that tends to congregate toward the sun due to the Poynting–Robertson effect. The surface layer (<500 Å) of the dust grains is quickly saturated by the bombarding solar wind ions, which leads to desorption of neutral atoms and molecules from the surface to maintain equilibrium [12,142,171,172,173]. The estimates of this neutral particle source suffer from a large uncertainty in the dust population and details of the outgassing process, but the neutral particle number density is unlikely to exceed 0.01 cm^-3.

Neutral interstellar gas (ISG), which serves as a background gas for ENA-producing charge-exchange collisions in the heliosphere, can be directly detected by ENA instruments. Fluxes of interstellar atoms can thus be called ISG ENA fluxes. Direct detection of ENA fluxes of interstellar helium atoms accelerated by the solar gravitation was recently demonstrated for the first time by the GAS experiment on the Ulysses spacecraft [153,154,174–176]. The helium flux is 10⁴–10⁵ cm^-2 s^-1 with atom energies in the 30–100 eV range.

Different radiation pressures and ionization rates lead to differences in properties of interstellar atoms in the heliosphere. In particular, different interstellar species would have different fluxes and velocity distribution functions at the same observation point [10]. Expected fluxes of interstellar neutrals at 1 AU vary from 10⁴ cm^-2 s^-1 down to 10^-1 cm^-2 s^-1 and their velocities vary from 10 to 70 km/s for an observer moving with the earth along its orbit around the sun (orbital velocity of the earth is ~29.8 km/s). The Ulysses instrument GAS is not capable of mass identification, but an alternative detection technology based on surface conversion to negative ions would be capable of ISG ENA mass analysis [10,177]. Thus direct in situ measurement of interstellar hydrogen, deuterium, and oxygen atoms is possible at 1 AU but has not been realized yet. An accurate direct measurement of the interstellar deuterium-to-hydrogen ratio may potentially provide important constraints on Big Bang cosmology, and oxygen is important to the theory of stellar formation and evolution [10].

The understanding of the heliosphere penetration by interstellar neutrals [158] led to the development of a concept of a permanently existing neutral component in the solar wind. This neutral solar wind (NSW) is believed to be born in charge exchange between the solar wind ions and interplanetary neutrals [178]. Solar wind plasma recombination contributes only a small fraction to the neutral component [12,142,179]. The NSW atoms move in the antisunward direction with approximately a solar wind velocity (300–800 km/s). At 1 AU, NSW consists of neutral hydrogen and helium atoms with an estimated flux of 10³–10⁴ cm^-2 s^-1 depending on the observer position at the earth’s orbit [12]. The NSW constitutes a 10^-5–10^-4 fraction of the solar wind at 1 AU. As the solar wind expands toward the boundaries of the heliosphere, the NSW fraction would increase to 10%–20% and play an important role in the shaping the global heliosphere [149,180]. The NSW flux may be significantly larger when relatively cold solar material is occasionally ejected from the sun in the coronal mass ejection events.

Due to the earth’s orbital motion, the NSW flux would be seen as coming from the direction several degrees off the sun for an observer moving with the earth [181,182]. Detailed computer simulations show that the NSW flux is confined within a few degree filed-of-view (FOV) [183]. An experiment to measure NSW was prepared in the early 1980s, but not flown yet [182].

If the supersonic flow of the solar wind plasma terminates at a solar wind shock front (TS), the plasma flow kinetic energy is largely converted into thermal energy of the subsonic plasma (Fig. 7). There is a certain probability for hot (T>100 eV) protons of the postshock solar wind plasma to charge exchange on background ISG between the termination shock and the heliopause and give rise to creation of fast hydrogen atoms. These atoms, called low-energy heliospheric ENAs, were predicted [184,185] in 1963; their characteristics were studied theoretically [9,129,142,186–188] but never explored experimentally.

Low-energy heliospheric ENAs are probably the only messengers born beyond the solar wind termination shock capable of reaching the inner solar system with minimal changes. The expected ENA flux is highly anisotropic: the flux increases with the decrease of distance from the shock to the sun, and the intensity and energy distribution of ENAs are very sensitive to the details of the interaction of the solar wind with the LISM, the parameters of the LISM, and the characteristics of the distant solar wind [9]. For a termination shock at 80 AU from the sun, the total expected ENA flux is 200 cm^-2 s^-1 sr^-1 from the upwind direction, and atoms are in the energy range 100–800 eV. The ENA flux smoothly changes with the angle of observation, decreasing by a factor of 2 at 40° from the upwind direction [9].

The strong dependence of ENA characteristics on the heliospheric properties makes ENA measurement an ideal direct method to remotely study the distant boundaries of the heliosphere [9]. Only a remote technique can provide a global view of the structure and dynamics of the heliosphere. The heliosphere ENA imaging will become especially important when a Voyager 1 spacecraft one day crosses the termination shock. Voyager in situ measurement will allow ‘‘calibration’’ of the remote observations in one point-direction: the measurements of low-energy heliospheric ENAs would reliably establish the shape of the heliosphere on the basis of the distance to the termination shock to be determined by Voyager 1. The next opportunity to obtain data from the heliospheric interface by in situ measurements from another spacecraft (planned Interstellar Probe) may not come earlier than year 2020.

Space plasma in the heliosphere and at its boundary is not in equilibrium and different processes result in distinctive populations of highly energetic ions, which, after charge exchange, would produce high-energy heliospheric ENAs (from 10 keV up to >1 MeV).

Heliospheric neutrals are ionized and picked up by the solar wind flow. After reaching the termination shock these pickup ions are believed to be accelerated to high energies and can reenter the heliosphere as cosmic rays, conventionally called anomalous cosmic rays (ACRs) [189]. Charge exchange of ACR ions produces ENAs whose detection would reveal details of ACR production and acceleration at the heliospheric boundary.

Shocks in plasma efficiently accelerate ions to high energies [190,191]. Various shocks travel through the heliospheric plasma and serve a source of energetic ions and correspondingly high-energy ENAs. Complex shock structures in the solar wind include merged interaction regions (MIRs) and corotating interaction regions (CIRs). High-energy heliospheric ENAs are generated by charge exchange of ACR ions, MIR, and CIR shock-accelerated ions, quite-time interplanetary ions (QTIP), and energetic solar particles (ESPs) [6,8,11]. Computer simulations predict complicated dependence of ENA characteristics on the direction of observation [6,8]; upper limits of ENA fluxes do not exceed 10^-2, – 10^-3 cm^-2 s^-1 sr^-1 keV^-1 and decrease with increasing atom energy. Theoretical models predict the maximum of the ACR-produced high-energy ENA fluxes from the heliospheric tail (downwind) region [192], which is exactly opposite to the direction where the maximum low-energy ENA flux is expected. The analysis of the experimental data from the CELIAS instrument [49] on the SOHO spacecraft may produce first experimental data on high-energy heliospheric ENAs in the near future [193].

While there are a number of publications on different aspects of the history of solar-terrestrial and magnetospheric physics[65,66,194–198], the story of ENA experimental study has never been told in detail. The presence of ENAs in the terrestrial environment was reliably established for the first time in 1950 by optical recording of Doppler-shifted hydrogen Balmer Ha emission (6563 A) in an aurora [199,200]. The precipitating hydrogen ENAs are born in the charge exchange between energetic protons and neutrals of the upper atmosphere and exosphere. Balmer Ha emission is in the visible wavelength range. It can be optically detected from the ground, and auroral emissions were extensively used to study the characteristics of energetic particles [194,201]. While hydrogen Balmer lines were observed in the auroral regions since the late 1930s [202], it was not until 1950 that a Doppler-shifted Ha line was unambiguously explained with the presence of hydrogen ENAs [199].

The importance of ENA production processes for the magnetosphere was understood [203] by noting that the proton–hydrogen atom charge exchange cross section was rather high for collision velocities less than the electron velocity in a Bohr orbit, i.e., for protons with energies <25 keV (Fig. 2). Charge exchange determines many important properties of geomagnetic storms. The "main phase" of a geomagnetic storm may last from 12 to 24 h, and it is characterized by a weakening of the geomagnetic field. The main phase is usually followed by a "recovery phase," when a gradual field recovery toward the initial undisturbed value of the geomagnetic field is observed. The recovery time constant may be 1 day or sometimes longer.

It was suggested for the first time in 1959 that charge exchange between the magnetic storm protons and neutral atmospheric hydrogen atoms provided the mechanism for the recovery phase [77]. The charge exchange process leads to the production of fast hydrogen atoms and observation of such atoms was first proposed in 1961 as a tool to study the proton ring current present during a magnetic storm [78]. A source of ENAs beyond the magnetospheric boundary, viz., charge exchange between the solar wind and the escaping hydrogen geocorona, was also identified for the first time [78]. The concept of imaging the magnetospheric ring current in ENA fluxes from outside [13] and, "in a limited fashion," from inside [14] was introduced much later in 1984.

The presence of atomic hydrogen in interplanetary space was first derived [185] in 1963 from sounding rocket measurements [204] of Doppler-broadened hydrogen Ly-a (1216 A) radiation (see also review of the early study of extraterrestrial Ly-a radiation [205]). It was recognized since the late 1950s that Doppler shift measurements could distinguish between the telluric (geocorona) and interplanetary hydrogen [206]. The emerging concept of the heliosphere [184] was extended in 1963 by the suggestion that about half of the solar wind protons would reenter the solar cavity in the form of hydrogen ENAs (with 3/4 of the initial solar wind velocity) as a result of processes at and beyond the solar wind termination region [185]. It was established later that an interplanetary glow in the hydrogen and helium resonance lines was produced by resonant scattering of the solar radiation by interstellar gas directly entering the solar system [141–144,158–161]. The "returning" neutral solar wind flux [185] is believed to be significantly smaller and highly anisotropic [9].

It was also suggested in the early 1960s that a large number of neutral atoms could be present in the solar wind as transients due to ejection of solar matter in violent events [207,208]. Neutral hydrogen atoms in solar prominences are observed optically, and it was argued that they may reach 1 AU. The follow-on calculations [209,210] showed that most of the neutral atoms would not survive travel to 1 AU because of ionization by solar EUV radiation and electron collisions. The neutral component of the solar wind is born mostly in charge exchange between the solar wind ions and interstellar gas filling the heliosphere [12,142,178].

Direct ENA measurements promised exceptional scientific return, but the necessary instrument development was only started in the late 1960s by Bernstein and co-workers [119,211–213] at TRW Systems Inc., Redondo Beach, California. Direct, in situ measurement of ENA fluxes in space was first attempted on April 25, 1968 in a pioneer rocket experiment [119]. The first dedicated ENA instrument was launched on a Nike–Tomahawk sounding rocket from Fort Churchill, Manitoba, Canada. This experiment was followed by the launch of a similar instrument on a Javelin sounding rocket on March 7, 1970 to an altitude of 840 km at Wallops Island off the coast of Virginia [214]. The experiments detected hydrogen ENA fluxes in the range 10⁵–10⁹ cm^-2 s^-1 sr^-1 with energies between 1 and 12 keV. The reported ENA fluxes were considered by many as excessively high, however these results have never been directly challenged in the literature.

The first ENA instrument [212] is shown in Fig. 8. It had many features and introduced many components and techniques that would later on become widely used by various modern ENA instruments. The instrument was based on foil-stripping of ENAs and subsequent analysis of the resulting positive ions. Electrostatic deflection plates were used to remove incident protons and electrons with energy <25 keV at the entrance (the instrument also had an additional 100 G magnetic field to remove electrons with energies >50 keV). The deflection plates served to define the solid angle of the instrument. ENAs were stripped passing an ultrathin (2 mg/cm²) carbon foil mounted on an 80% transparent grid. A hemispherical energy analyzer, which focused the ions in one dimension, was used for energy analysis. The stripped ENAs, protons, in a selected energy range, passed through the analyzer and were counted by a channel electron multiplier (CEM). Use of two additional identical instrument sections without deflecting voltage and without an ultrathin foil allowed the simultaneous measurement of proton fluxes and the monitoring of the background count rate during the experiment, respectively.

FIG. 8. Schematic representation of the first ENA instrument flown on a sounding rocket in 1968. The instrument consists of charged particle deflector, ultrathin carbon foil, electrostatic analyzer, and ion detector. (After Ref. 212).

An attempt to measure ENAs was made in the RIEP experiment (a Russian acronym for "registerer of intensity of electrons and protons") on the Soviet Mars-3 interplanetary mission (launched May 28, 1971; entered low Mars orbit on December 2, 1971). The experiment was designed to measure the energy distribution of plasma ions and electrons in the Mars’ environment as well as in the solar wind during interplanetary coast. The RIEP instrument consisted of eight separate cylindrical electrostatic energy-per-charge analyzers, each followed by a CEM to count particles [215]. Each analyzer unit was designed to measure charged particles of a selected energy-per-charge ratio. Two ultrathin carbon foils (150 A or ~3.5 mg/cm²) were installed in front of two of the eight analyzers. A comparison of count rates from analyzers with and without foil while measuring particles with the same energy-per-charge ratio was expected to provide information on high-intensity neutral atom fluxes. No charged particle deflectors were used in front of the ultrathin foils, and the experiment failed to establish ENA fluxes, which are usually relatively weak.

Another ENA instrument, a slotted-disk velocity selector, was successfully built and mechanically and electrically tested in a rocket flight in 1975 [181]. This narrow (FOV) instrument, which demonstrated efficient rejection of charged particles and photons, was especially suitable for measurement of ENAs with velocities <500 km/s (E<1.3 keV/nucleon). It is interesting (see below) that measurements of the interstellar helium flux and of the neutral component in the solar wind were considered as possible applications [181]. Apparently due to large size, mass, and power consumption as well as the torque exerted on a spacecraft (the instrument included at least two disks 16.24 cm in diameter on a shaft 74 cm long spinning at 4.5´10⁴ rpm) this instrument was never used for ENA measurements, and the technique development was discontinued.

The interest to the concept of a mechanically moving ENA velocity selector was recently revived by suggestion to use unconventional high-frequency mechanical shutters mounted on ceramic piezoelectric crystals [216]. Such an approach seems to be especially promising for the study of low-energy ENAs, but further development is needed to demonstrate its feasibility. The initial ENA measurements [119,214] were not repeated and/or independently verified, and they were largely ignored by the space community. Exceptionally strong EUV/UV radiation background was identified as a major obstacle for reliable ENA measurements in space, and experimental difficulties were perceived as insurmountable by many at that time.

Three groups accepted the experimental challenge and continued independent development of dedicated ENA instrumentation in the late 1970s. A group at the Max-Planck-Institute for Aeronomy (MPAe), Lindau, Germany targeted direct in situ detection of interstellar helium flowing into the solar system [174,175]. (The experiment was suggested for the first time in 1972 in a proposal by H. Rosenbauer, H. Fahr, and W. Feldman.) Another group at the Space Research Institute (IKI), Moscow, USSR, planned to measure the neutral component in the solar wind and heliospheric ENAs [177,182,217] It is interesting that the Moscow group initially considered also direct, in situ detection of the interstellar helium flux [218] but ultimately decided to concentrate on the neutral solar wind (NSW) and heliospheric ENAs. The NSW experiment was actively supported by a group at Space Research Center in Warsaw, Poland, which also theoretically studied heliospheric ENAs and ENAs produced in the giant-planet magnetospheres [219–221]. A third group at the University of Arizona focused on the possibilities of measuring ENAs in geospace [222].

The GAS instrument [175] to directly detect fluxes of interstellar helium (E = 30–120 eV) had a dramatic history of being completely redesigned and built within a record 3 month period in an "almost super-human effort," [176] which had become necessary to realize the original experimental concept on the European-built Ulysses probe after cancellation of the U.S. spacecraft. The GAS was then successfully flown [153,154,176] on Ulysses which was launched in 1990 after many delays [223], including the one caused by the Space Shuttle Challenger explosion. The neutral helium instrument is based on secondary ion emission from a specially prepared surface [175,176]. The experiment has produced unique data on interstellar helium characteristics and ENAs emitted from Jupiter’s Io torus and continues operating successfully [154].

A new ultrathin-foil based "direct-exposure" technique, which does not require stripping of incoming ENAs, was developed by the Moscow group to detect the neutral solar wind [177,182,217]. NSW measurements can be performed from an interplanetary or high-apogee earth-orbiting spacecraft by pointing the instrument several degrees off the sun. The possibility of taking advantage of the aberration caused by the earth’s motion around the sun was first suggested in 1975 [181] and was independently "rediscovered" later [177,182]. The NSW instrument was built for the Soviet Relikt-2 mission which was originally planned to be launched in 1987 [182]. The initial experiment also included measurements of ENAs from the heliospheric interface and ENAs emitted from the terrestrial magnetosphere. Detection of ENAs escaping Jupiter and Saturn was also expected [182,220,224]. The Relikt-2 mission was postponed many times, and is still awaiting launch (now scheduled for 1999).

The NSW instrument included a diffraction filter [182,225] to significantly increase the ENA-to-EUV/UV photon ratio in the sensor. The filter was optional for the planned measurements [182], and it was not fully developed, when the instrument was built. A new diffraction filter technology is emerging [226,227], and a new generation of diffraction filters for ENA instruments is currently being developed and evaluated (see Sec. VI F) [228,229,230,231]. An introduction of the diffraction filters opens the way to take full advantage of a highly efficient direct-exposure technique.

ENA instrument development in the late 1970s and early 1980s did not attract much attention in the space community, and did not enjoy enthusiastic support of the funding agencies. However, the importance of ENAs for mass, energy, and momentum transport in space was established and new opportunities offered by remote ENA imaging in separating spatial variations from temporal ones in space plasmas were gradually recognized.

Several measurements of large fluxes of ions (E<10 keV) near the equator at altitudes below 600 km had been reported since the early 1960s [232]. Fluxes of high-energy ions (0.25<E<1.5 MeV) were measured later at low altitudes by the German AZUR spacecraft [124,233], and the ions with energies down to 10 keV were detected there in 1973 [234]. Theoretical calculations predicted a short lifetime for such low altitude protons near the equator due to collisions with atmospheric particles. The proton loss thus required an injection of protons in a limited region below 600 km. The required proton source to compensate proton loss due to interaction with atmosphere must also be atmosphere dependent. The explanation, found in 1972 [124], suggested that trapped energetic ions in the ring current at much higher altitudes produce ENAs in charge exchange, and a fraction of these ENAs reaches low altitudes where they are re-ionized by charge exchange and are consequently trapped by magnetic field (Fig. 6).

In the heliosphere, the neutral solar wind is believed to provide a transport mechanism similar to that of magnetospheric ENAs [180]. As the solar wind expands toward the boundaries of the heliosphere, the neutral fraction in the solar wind gradually increases to 10%–20% at the termination shock (Fig. 7). While the solar wind plasma flow is terminated by the shock, the solar wind ENAs easily penetrate the region of the heliospheric interface and enter the LISM, sometimes called "very local" interstellar medium (VLISM). The solar wind ENAs thus interact with the approaching local interstellar plasma via charge exchange with plasma protons. Hence, the boundary of the region of the sun’s influence, the solar system "frontier," extends further into "pristine" interstellar medium. The significance of this neutral solar wind effect [180] on LISM was recently confirmed by detailed computer simulations of the heliospheric interface [149,150].

The interest in ENA characteristics and instrument development began to grow after the serendipitous discovery of ENAs made by energetic particle instruments on several spacecraft. Energetic particle instruments are usually based on solid state detectors and often capable of discriminating against electrons. However, such instruments generally cannot distinguish between a charged particle and an ENA. Therefore an energetic particle detector would efficiently serve as an ENA detector only in the absence of the normally abundant ions. Only high-energy ENAs (E>10–20 keV/nucleon) can be detected by such instruments.

Analysis [235] of inconsistencies in interpretation of energetic particle measurements by the IMP 7 and 8 satellites (at ~ 30–35 R_E from the Earth; 1 R_E = 6378 km is the Earth radius) led to the conclusion that a certain fraction of counts during periods of very low fluxes [236] was caused by radiation belt-produced ENAs with energies 0.3–0.5 MeV.

Both energetic ions and neutrals were detected during the magnetic storm in 1982 by the SEEP instrument on the S81-1 spacecraft at low altitude. A double charge-exchange mechanism (Fig. 6) was invoked to explain the observations, and it was also suggested that measurements of energetic neutral atoms at low altitude "might be able to image, in a limited fashion, the integral ion intensities of the ring current as a function of latitude and longitude" [14].

ENA fluxes of nonterrestrial origin were detected on a Voyager 1 spacecraft during flybys of Jupiter [237] and Saturn [238] n March 1979 and November 1980, respectively. The low energy charged particle (LECP) instrument on Voyager-1 included a silicon detector to accumulate counts from eight separate directions in the ecliptic plane [239]. The detector was designed to measure ions (electrons were swept away by magnetic field) with energies >40 keV.

During the approach to Jupiter, when Voyager 1 was still outside the gigantic Jovian magnetosphere, an excess count rate was measured in the sector containing the planet in its FOV. The detector characteristics limit the possible sources of the excess counts to energetic ions, x rays, and ENAs. No energetic ion fluxes with required intensity and energy were expected at the location of measurements since there were no magnetic field lines connecting to Jupiter [237]. It is known that x rays can be generated in planetary magnetospheres by precipitating energetic electrons, usually in polar regions. Consideration of x-ray generation showed that the required fluxes of electrons were several orders of magnitude higher than those found in the Jovian magnetosphere. The conclusion of data analysis was that "the only remaining possibility of explaining the excess counts ... is energetic neutral atoms" [237]. The energy dependence of the observed ENA spectra was similar to the one established for energetic ions in the magnetosphere during the close flyby.

A similar excess count rate was measured one and a half years later during the flyby of Saturn [238]. Here again the observed count rate, if interpreted in terms of x rays, cannot be reasonably related to precipitating magnetospheric electrons. It was concluded that "charge exchange of energetic ions with satellite tori is an important loss mechanism at Saturn as well as at Jupiter" [238].

Possible ENA signatures in the experimental data obtained by energetic particle instruments [240] on IMP 7,8 and ISEE 1 were analyzed in 1982 [241,242]. Detection of ENAs (E~50 keV) was unambiguously established, and their source was identified as the ring current in the terrestrial magnetosphere [107]. Coarse spatial information on ENA-producing regions was derived and ENA energy distribution and mass composition were determined. The ENA measurements were made from positions where magnetic field lines allowed only negligible fluxes of energetic ions, so that the detectors counted only ENAs. The analysis [107] of these measurements was a major milestone in validating the idea that the global magnetospheric processes can be efficiently studied remotely by measuring ENAs. It was also suggested that ENA imaging could be used to study the magnetospheres of Jupiter and Saturn [107].

The follow-on analysis of the ISEE 1 data demonstrated the powerful potential of the ENA detection as an imaging technique by reconstructing the first ENA global image of the storm-time ring current (at E~50 keV) [79]. The instrument [240] on ISEE 1 was capable of measuring the incoming ENA flux, and imaging was performed by a combination of spacecraft spin and instrument axis scanning by a moving motorized platform. A procedure for computer simulation of all-sky ENA images was established and the theoretically predicted images were compared with the ring current image obtained by ISEE 1 at a radial distance of 2.6 RE during a 5 min observation.

The analysis of excessive count rates [237,238] detected by Voyager 1 during Jupiter and Saturn flybys allowed one to determine some ENA characteristics in the vicinity of the giant planets [108]. Further computer simulations of expected ENA emissions demonstrated the efficiency of ENA imaging as a tool to study the magnetospheres of Jupiter and Saturn from flyby spacecraft and orbiters [109,110,224,243].

Use of energetic particle instruments for detection of high-energy ENA fluxes in the absence of ions became an established experimental technique. The storm-time ENA images of the polar cap [244] were recently obtained by the CEPPAD experiment [245] on the POLAR spacecraft. The instrument cannot distinguish between ions and neutrals, and the ENA images were recorded during the portions of the spacecraft orbit where the fluxes of the charged particles were very low.

An instrument with a dedicated high-energy ENA (20<E<1500 keV) detection channel was flown on the CRRES satellite in 1991 [21,125]. The neutral channel that consisted of a magnetic deflector followed by a solid-state detector measured ENA fluxes precipitating to low altitudes at the equatorial regions. A small geometrical factor put limit to the imaging capabilities of the instrument. A conceptually similar ENA instrument was recently flown on ASTRID [22,23]. A more sophisticated high energy particle (HEP) Instrument (10–100 keV ENAs) was launched on the GEOTAIL spacecraft in 1992 [20].

The current phase in the study of ENAs in space plasmas is characterized by extensive computer simulations of ENA images, novel instrument development, and the preparation and planning of a number of dedicated ENA space experiments. (MG: remember that this was written and published in 1997, before IMAGE, Cassini, TWINS, and IBEX missions.) The concept of global ENA imaging of the magnetosphere is firmly established [1–5,100,111,112]. Computer simulations of low-energy [113–115] and high-energy [111,116,118,246,247] ENA imaging emphasized the importance of ENA measurements throughout a wide energy range from few eV up to several hundred keV. Various aspects of heliospheric ENA global imaging were theoretically studied for various species and energy ranges [6–8,9,10,11,12,157,192,248]. The first ENA detections obtained from planetary magnetospheres in the early 1980s triggered new interest in developing dedicated ENA instrumentation: various techniques for detection and imaging of low- and high-energy ENAs were proposed and their laboratory evaluation was begun [243,249,250–252,253,254]. Several new research groups have entered the ENA-imaging field since then and significant improvements of the known experimental techniques as well as a number of innovative approaches were proposed, especially for low-energy ENAs [4,10,125,216,225,228,229,230,254–258].

NASA recently selected an IMAGE mission to perform comprehensive imaging of the terrestrial magnetosphere in EUV, FUV, and ENA fluxes. A sophisticated first large size ENA camera INCA [24] on Cassini will soon perform ENA imaging of the Saturnian magnetosphere and the exosphere of the Saturnian moon Titan [259].

An energetic ion gyrates about a magnetic field line in space (Fig. 3). When charge exchange occurs, the resulting ENA is liberated from the magnetic field and, as a stone from a slingshot, it moves straight away from the point of its birth. Reconstruction of a global ENA image requires measuring ENA flux dependence on the direction of observation: the instrument has ideally to determine the trajectory (flight direction) of each individual ENA, identify its mass and energy, and accumulate the image in the memory. A sequence of ENA images (for different masses and energies) would allow direct observation of global plasma dynamics, e.g., development and decay of a ring current during magnetic storms and variation of the heliosphere’s size and shape during the solar cycle.

Some plasma ions, for example He⁺ and O⁺, can be imaged optically by registering resonantly scattered solar photons at λ = 304 and 834 Å, respectively [246,247,260–263]. Unlike He⁺ and O⁺, protons, the most abundant component of space plasmas cannot be imaged optically, which makes ENAs in many cases the only tool to study processes of interest remotely. ENA imaging, complemented when possible by imaging in the EUV [5,247,264] and FUV [264,265] spectral ranges as well as x-ray imaging [266,267] and radio sounding [268], promises a breakthrough in our understanding of plasma processes in and dynamics of the magnetosphere and heliosphere. (Feasibility of locating and monitoring the position of the earth’s magnetopause and plasmapause by a radio wave sounder from a high-altitude satellite is presently being intensely debated [268–271].)

An instrument deflector-collimator would define the FOV and prevent ions and electrons from entering the instrument. The required mass resolution is usually modest: to distinguish among hydrogen (1 amu), helium (4), and oxygen (16) atoms. One can also expect sulfur (32) ENAs in the Jupiter’s magnetosphere.

ENA images can be obtained either from a three-axis stabilized spacecraft or from a spinning spacecraft. The requirements of ambient ion deflection favor a slitlike instrument aperture for imaging in one dimension only (Sec. VI B). The imaging in the second dimension can be obtained by using spacecraft spinning. In that case one records one-dimensional images for consecutive orientations of the spacecraft, which may be only a few degrees apart, as the spacecraft spins about its axis (instantaneous orientation, or attitude, of a spacecraft is usually known with high precision). A two-dimensional composite image is then reconstructed from a set of one-dimensional images. A one-dimensional imager on a spinning platform has become a favorite configuration for ENA experiments.

Background EUV/UV photons may trigger MCP detectors in ENA instruments either directly or via photoelectron emission from foils and other exposed surfaces. The spectral range of concern is usually limited to wavelengths λ < 1400 Å. An exposure to direct solar light will ‘‘blind’’ and may permanently damage an unprotected instrument. A special photometric unit, 1 R = 1 Rayleigh = 10⁶/4p cm^-2 s^-1 sr^-1, is used to describe diffuse photon fluxes. If one applies the same unit to neutral particle fluxes, then an ENA flux of 1 cm^-2 s^-1 sr^-1 would correspond to ~10^-5 R. Thus the expected magnetospheric and heliospheric ENA fluxes are in the 10^-8–1 R range.

A strong background EUV/UV radiation makes space an exceptionally inhospitable place for ENA measurements. The major EUV/UV source in the magnetosphere is the dayglow at the sunlit side and the nightglow at the night side. The glow arises from scattering of sunlight (including multiple scattering of photons in sometimes optically thick environment) and emission associated with various collisional processes in the upper atmosphere, exosphere, and ionosphere [17,19]. Typical dayglow and nightglow spectra contain a number of EUV and UV spectral lines [272,273]. At 600 km altitude, the total dayglow intensities are 54 kR and 25 kR for observations down and up, respectively [272]. The most prominent lines are hydrogen H I 1216 Å and oxygen O I 1304 Å with significant contribution of other helium, oxygen, and nitrogen spectral lines. The most important spectral lines of nightglow are hydrogen H I 1216 Å and helium He I 584 Å [273]. The nightglow intensity is 3600 R, which is 4–11 orders of magnitude higher than expected ENA fluxes.

Interplanetary EUV/UV glow is produced by resonant scattering of solar radiation by heliospheric hydrogen and helium atoms. The glow brightness depends on the direction of observation and varies between 500 and 1000 R in H I 1216 Å and between 1 and 10 R in He I 584 Å at 1 AU.

An ENA image of a plasma object ideally consists of a set of images obtained for different particle masses (e.g., H, He, and O) in different energy ranges (e.g., 1–5, 5–10, and 10–20 keV, ...). ENA imaging thus includes two interrelated tasks, viz., obtaining the object’s image and ENA identification (mass, energy, and velocity). We will consider these two tasks separately, and then demonstrate (Sec. VII) how they are combined in the instruments. The images of plasma objects can be obtained in two ways: observing an object from a remote vantage point outside (external imaging) and observing from within a plasma object, the inside-out internal imaging (Fig. 9). The ENA emitting plasma can be assumed to be ‘‘ENA-thin’’ with the exception of the measurements at low-earth orbit, where multiple charge exchange may become important.

FIG. 9. Two imaging geometries: imaging from inside-out and imaging from outside.

Imaging from outside can be performed from a spacecraft flying by a planetary magnetosphere, or from a spacecraft in a high-apogee or high-altitude orbit. The high-apogee configuration is illustrated in Fig. 10. The most advanced ENA imaging missions, such as Cassini at Saturn, utilize external outside viewing geometry. An example of the inside-out internal imaging is observation of the heliosphere (Fig. 7) from the earth’s orbit and study of precipitating magnetospheric ENA fluxes (Fig. 6) at low-earth orbit. The measurements of the latter type were performed on CRRES [21] and ASTRID [22,23] missions.

FIG. 10. imaging of the terrestrial magnetosphere from a hypothetical high-apogee polar orbit. Magnetic field lines crossing the magnetic equator at three and five earth radii (L=3,5) are shown for reference. (After Ref. 5.)

The ENA flux j_ENA,i(cm^-2 s^-1 sr^-1 keV^-1) of a given species i from a given direction s (Fig. 11) is

where j_i(s,E) is the directional differential flux of parent ions, n_k(s) is the number density of neutral species k of the background gas, s_ik(E) is the energy-dependent charge exchange cross section between ions of species i and neutrals of species k, the factor

allows for extinction of ENAs on their way from a point of birth to an observation point O, and b is the ENA loss rate (charge exchange, electron impacts, and photoionization). The ENA flux reaching the observer from a given direction is thus determined by an integral along the line-of-sight s (Fig. 11), and the flux contains information on the velocity distribution function of ions along the line-of-sight.

FIG. 11. Geometry of remote sensing of an ion population by measuring ENA fluxes from an observation point O.

An ENA image is reconstructed by measuring the ENA flux dependence on the direction of observation. The recording of ENA images is complicated by fast motion of the spacecraft and by differences in ENA velocities. The spacecraft motion puts the limit on possible image accumulation time, which may be in conflict with the desired statistical accuracy. The observer motion may also become advantageous if it allows derivation of the ENA velocities due to the aberration effect. For large distances between of the spacecraft and plasma object, high-velocity, and slower ENAs simultaneously emitted by the same plasma region would be detected at different times. This time difference varies from 1 to 15 min for earth magnetosphere imaging from a high-altitude spacecraft to more than a year for the imaging of the heliospheric boundary from 1 AU.

An ENA image is a projection of a three-dimensional ENA-emission plasma object on a two-dimensional image plane. The interpretation of images thus becomes model dependent. Image inversion is a specialized area [1,2,5] that is beyond the scope of this article. We only note that forward modeling is often used as a method of choice of treating ENA images, that is varying the free parameters to achieve the best fit of the model predictions to experimentally obtained images [1,2,5,79,116]. It is important that ion distribution functions may not be entirely independent in different magnetospheric regions. Energetic ion motion in the magnetosphere is constrained by magnetic field geometry, which allows one to relate ion characteristics in spatially separated areas. Significant progress was achieved in applying discrete inverse theory to optical geophysical images [264], which may also be used for some aspects of ENA imaging. Inversion of ENA images requires accurate knowledge of the instrument characteristics such as the FOV and detection efficiencies. Future ENA experiments will also perform simultaneous observations from multiple spatially separated spacecraft to achieve tomographic imaging of planetary magnetospheres.

The ENA fluxes are usually weak and the observation time is limited. Consequently a number of counts in an image pixel is mostly small and, as a result, significant statistical noise due to the random nature of the particle flux is often experienced. This noise is present even in the absence of the detector intrinsic noise. The trade off between the observation time and the desired image angular resolution and ‘‘photometric’’ accuracy is one of the most important goals of the modeling of an ENA imaging experiment.

A weak ENA flux cannot be collected and concentrated by diffracting and/or refracting elements as it is done in optics. In this respect, ENA imaging is similar to hard x-ray imaging [274,275]. Reconstruction of the trajectories of individual ENAs can be achieved by a two-point trajectory extrapolation, when the location of particle entry into the instrument is connected by a straight line with the point of ENA impact on the instrument ‘‘focal plane,’’ i.e., the plane where the ENA image is formed.

The point of ENA impact can be determined by a position-sensitive detector. The simplest way to determine the entry point is to restrict the entry area by a mechanical aperture: this is a pinhole camera [Fig. 12(A)]. In a one-dimensional imaging system a slit would play a role of a pinhole. Another approach is based on placing an ultrathin foil at the instrument entrance (see Sec. VI A) and determining the position of ENA entry by measuring electron emission from the foil caused by the passing ENA. Since an ENA flight direction may be changed in the foil passage (scattering), the latter configuration allows reconstruction of the ENA trajectory inside the instrument after the entrance foil.

FIG. 12. Pinhole (A) and coded-aperture (B) imaging cameras.

While a simple pinhole camera can be used for ENA imaging in any energy range, thin-foil cameras are limited to high-energy ENAs. The direction of high-energy ENA arrival practically coincides with the particle trajectory inside the instrument, and the latter trajectory can be used for ENA image reconstruction. A low-energy ENA would significantly scatter in the foil. Consequently, the imaging in low-energy ENA fluxes requires a pinhole camera with a disadvantage of low instrument throughput. In the context of imaging, the division of particles in high-energy and low-energy ENAs depends on the foil thickness and the required angular resolution.

A pinhole camera provides excellent imaging but requires relatively high ENA fluxes. ENA fluxes in space are weak, and an increase of the instrument’s geometrical throughput, or geometrical factor, is of paramount importance. The greater the desired angular resolution, the smaller the pinhole should be for the same area of the image in the focal plane. The speed of the image forming system may not suffice to produce an image of the desired photometric quality during a given exposure time. The situation may be also aggravated by the detector noise independent of the object image. (This noise should be distinguished from the noise due to random nature of the incoming particle flux, which is object-dependent).

The coded-aperture technique was proposed simultaneously and independently in the mid-1980s by two groups [249,276] to enhance performance characteristics of ENA instrumentation. One group suggested installing coded-aperture masks at the instrument entrance to increase geometrical throughput [112,276–278]. Another group argued that although a coded aperture was not superior to a pinhole in terms of the signal-to-noise ratio for magnetosphere imaging, the coded-aperture camera would allow efficient use of the same sensor for imaging simultaneously in both low- and high-energy ENA fluxes [249].

The coded-aperture technique is based on simultaneous use of a number of pinholes. The images formed by the pinholes mix (multiplex), overlap and superimpose (not matching each other) and produce picture at the instrument’s image plane non recognizable at first glance [Fig. 12(B)]. If a pinhole pattern is selected in a proper way, then it is possible to unscramble the resulting image by postprocessing, while minimizing errors and artifacts, and produce the image of the object.

Image multiplexing was pioneered in 1968 [279,280] and used many times in space and laboratory experiments [274,275,281–288]. Uniformly redundant arrays (URAs) [283,289] of pinholes were found to provide the most efficient aperture coding with minimal imaging artifacts. URAs are based on pseudorandom sequences [290,291] widely used in the communications technology. A conceptually similar technique with multiplexing in time instead of space is successfully used in neutron scattering [292] and molecular beam [293,294] experiments.

A detailed comparison [295] between pinhole and coded-aperture cameras is summarized in Table I. Three factors are most important in the selection of the ENA imaging system. First, the magnetospheric and heliospheric images are expected to be widely distributed structures with slowly varying brightness rather than starlike objects. Second, the detector noise (with the exception of the statistical noise due to random nature of particle fluxes) is efficiently suppressed by coincidence requirements employed by ENA identification techniques. Third, EUV/UV background radiation may overload MCP detectors in the sensor.

The coded-aperture technique is advantageous when used for the detection of starlike objects and in the presence of the object-independent detector noise. Pinhole cameras seem to be superior for ENA imaging since they are simpler, more tolerant to background EUV/UV, and free of the problems with object partial coding, while the coded-aperture technique is prone to artifacts [295]. Additionally, even small manufacturing imperfections of coded-aperture masks may result in very complex imaging artifacts [296,297]. A pinhole-type camera is thus an imaging configuration of choice, at least for the experiments in the near future.

ENA diagnostics is a well-established approach to study fusion plasma characteristics [33–36,298–300]. Such corpuscular diagnostics have been successfully used in various forms since the early 1960s [33,301] at many magnetically confined plasma machines, e.g., T-3,4 [301], DOUBLET III [302], PLT [34,303], TFTR [304], 2XIIB [300], JET [305,306], ASDEX [307], JT-60 [308,309], RFX [310], MST [311], and TORTUR [312,313]. The energy range of the measured neutrals extends from a few hundred eV [303,314] up to 0.1–1.0 MeV [304,306,309].

Most of the fusion plasma ENA analyzers in the energy range 0.5–100 keV are based on ionizing the neutrals in special stripping cells and subsequent analysis of positive ions. A stripping cell is usually filled with gas [301,302,308,311,314] or plasma [315]. Pulsed gas and pulsed plasma targets have been also used to increase stripping efficiency while minimizing the load on vacuum pumps. One of the major advantages of gas targets is minimal disturbance of the neutral atom energy distribution (compare with energy losses and scattering in ultrathin foils: see Sec. VI D 2). Gas targets can also be used in space instruments that do not require continuous longterm operation, such as the solar EUV optics-free spectrometers [316–318]. For ENA imaging in space, the use of gas targets is unlikely.

The instruments for detection of fusion plasma-emitted ENAs with E<500 eV are usually based on mechanical choppers in various configurations [303,307,310] although electron impact ionization has also been used [319]. The chopper approach is conceptually not unlike the slotted-disk velocity selector developed for space neutrals [181]. Conventional mechanical choppers are unacceptably large and heavy for space instruments, however recent suggestion [216] to mechanically modulate the incoming ENA flux by miniature transmission gratings may open a way for this technique in space instruments.

Neutral atom surface conversion to negative ions was recently implemented for fusion plasma diagnostics [312,313]. A similar approach for detection of ultralow-energy ENAs in space [10,177] has tremendous promise when fully developed and will be discussed in some detail below (Section VI G 2). Similarities between ENA imaging of space and fusion plasmas led to the proposals of application of recent technological advances in space instruments to fusion plasma diagnostics [320].

Presently all ENA instruments can be roughly divided into three partially overlapping major groups (Fig. 13): instruments for high-energy ENAs (E>10 keV/nucleon), for low-energy ENAs (100 eV<E<50 keV), and for ultralow energy ENAs (E<500 eV). Various aspects of ENA instrumentation were reviewed in the past [4,10,111,112,216,248,250,256,295,321]. In this section we will start with a brief description of common approaches in ENA instruments, and then consider in some detail major instrument components and techniques. Then in Sec. VII we will consider representative instrument designs illustrating the state of the art and new directions in ENA instrumentation.

FIG. 13. Energy ranges of three major ENA instrument groups.

An ENA instrument consists of a deflector-collimator followed by a sensor performing detection and identification of incoming neutral particles. The deflector prevents ambient charged particles from entering the sensor, but leaves the incoming EUV/UV background radiation unaffected. The upper energy limit of the instruments is determined by the deflector ability to separate ENAs from energetic ions by preventing the latter from entering the sensor (usually 100– 300 keV/e). Cosmic ray particles (penetrating particles) are rejected by using sensor capabilities of measuring ENA energy and/or velocity.

Let us first demonstrate some concepts of ENA detection and identification on a simplified generic sensor found in various forms in many ENA instruments. This sensor consists of an ultrathin (20–200 Å) foil (UTF) at the entrance, electrostatic mirror (EM), and detectors D₁ and D₂ (Fig. 14). An ENA penetrates the foil and hits the detector D₂. The electrons emitted from the foil by ENA passage are accelerated and transported to the detector D₁. By measuring coordinates of the electron impinging on D₁ (X₁,Y₁) and particle on D₂ (X₂,Y₂), one can determine the particle’s trajectory between the foil and the detector D₂.

FIG. 14. Schematic of a generic sensor illustrating techniques used in many ENA instruments. (UTF) ultrathin foil; (EM) electrostatic mirror, (D₁) and (D₂) detectors capable of determining coordinates (X,Y) of particle impinging on the sensitive surface and accurately fixing the moment of particle registration.

The particle velocity is determined by measuring a time interval (time of flight) between electron detection by D₁ and particle detection by D₂. If D₂ is a solid-state detector, then the particle total energy can also be established. Simultaneous measurement of particle velocity and energy allows unambiguous mass identification. The ability of thin-foil time-of-flight (TOF) spectrometers to distinguish among incoming monoenergetic particles with different masses improves with the increasing particle energy. For example masses 1, 2, and 4 can be separated at such low energies as 3 keV [322,323], while separation of masses 12 and 16 would require energy ~25 keV [37].

TOF spectrometers are exceptionally efficient in rejecting detector noise counts. A valid ENA event (detection) requires signals from several detectors within well-defined narrow (100–200 ns) time gates (coincidence requirement). Background photons and stray particles would produce detector noise count rates proportional to the instrument geometrical factor and correspondingly proportional to the incoming ENA fluxes. The noise counts are random, and there is a certain probability that they produce a false ‘‘ENA event’’ by triggering the detectors within the coincidence time gates. The rate of double and triple random coincidences is proportional to the square and cube of the incoming flux, respectively. The ENA signal is directly proportional to the incoming ENA flux. This square or cube proportionality of the random coincidence rates allows extraction of a weak ENA signal from the superior detector noise count rate [217]. For example, a triple coincidence random noise due to background EUV/UV radiation is expected to be about seven events per year in the ENA instrument on Cassini [24].

Particles with very high energies (>1 MeV), typical for nuclear physics experiments, are detected with almost 100% efficiency by TOF spectrometers. The energies of ENAs and electrons emitted from thin foils are much smaller, and the detection efficiencies can be anywhere between a fraction of a percent and 100%. Electron yields from foils may also vary from a hundredth to a few electrons per incident ENA. A coincidence technique offers a unique opportunity to determine, in the absence of noise, instrument absolute detection efficiency without independent knowledge of the incoming calibration flux [217]. This feature of the coincidence technique was used for the first time in the 1920s by Geiger and Werner [324] who determined the efficiency (<100%) of human observers in counting scintillations.

Schematics of high-energy, low-energy, and ultralow-energy ENA sensors are shown in Figs. 15, 16, and 17, respectively.

A high-energy ENA instrument can be protected against background radiation by a thin (~0.1 mm) film (TF) filter, which significantly reduces the EUV/UV photon flux into the instrument (Fig. 15). A protective thin-film filter at the instrument entrance would modify energy and angular distributions of incoming ENAs. Thin-film filters and solid-state detectors determine an effective energy threshold (>10 keV/nucleon) for such instruments [250]. Electron emission from the film, or from an additional ultrathin foil, can be used to provide a START signal for a TOF analyzer and to determine the position (X₁,Y₁) of the point of ENA entry. The detector D provides a STOP signal and particle position (X₂,Y₂). A solid-state detector D would also measure the total energy of the particle. Thus one can determine the particle’s trajectory, velocity, and energy (and mass).

FIG. 15. Schematic of high-energy ENA detection.

Conceptually similar TOF spectrometers are often used for particle identification in nuclear physics [325]. ENA energies are insufficient for thin solid-state DE detectors, as used in particle DE-E telescopes [38,40], therefore ultrathin foils are needed for START signal generation. The ultrathin foils were introduced in TOF spectrometers initially for study of particles with E>100 keV [326–328]; this energy limit was lowered later down to ~10– 20 keV [37] and eventually down to 600 eV [217].

Particle absorption, straggling, and scattering prohibit use of protective thin-film filters for detection of ENAs with E<10 keV. Two ways to register low-energy ENAs by filterless, open-type instruments have been developed (Fig. 16): (A) to convert ENAs to ions by stripping in an ultrathin foil with subsequent deflection (and separation from the background photons), analysis and detection of ions [4,212,321]; and (B) to expose the instrument directly to both the EUV/UV radiation and ENA flux and to detect ENAs by the TOF technique [182,217].

FIG. 16. Schematic of low-energy ENA detection by conversion of ENAs to ions (A) and by a direct-exposure time-of-flight technique (B).