The High Energy Focusing Telescope (HEFT)

HEFT Science Objectives

The hard X-ray band at 10–100 keV is ideal for the study of non-thermal processes, as the sky is not dominated by thermal emission from stars and diffuse astrophysical plasma, unlike lower-energy bands. Examples of non-thermal phenomena detectable in hard X-ray include the so-called diffuse X-ray background, the up-scattering of lower-energy photons by relativistic electrons in various energetic environments, and various nuclear and relativistic processes observable in stellar remnants, including young supernova remnants, and pulsar and neutron star environments. Many of these phenomena are of high astrophysical significance. Yet, current telescope technologies in hard X-ray, with their limited sensitivity and angular and spectral resolution, limit our ability to study these phenomena in fine detail. It is the goal to further study these non-thermal processes that motivates us to develop the High Energy Focusing Telescope.

The following passages are excerpts from graduate student Hubert Chen's thesis on HEFT.

The X-ray background

On 1962- 6-19, the first X-ray astronomy experiment onboard an Aerobee rocket discovered that the X-ray sky is filled with an ambient diffuse emission that is spatially fairly homogeneous, at 1.7 photons/s/cm²/sr within 1.5–6.2 keV (2–8 Å) (Giacconi et al., 1962). This diffuse emission is known as the X-ray background (XRB). Surveys in recent years in soft X-rays (0.2–10 keV), with much improved angular resolution enabled by the focusing telescopes Chandra and XMM-Newton, have revealed that most of the X-ray background at those energies are actually previously unresolved point sources, from about 90% of the XRB at 0.5–2.0 keV, to 80–90% at 2–6 keV and 50–70% at 6–10 keV (Brandt and Hasinger, 2005). Over 70% of these point sources are found to be active galactic nuclei (AGNs)—the nuclei of ‘active galaxies’, which emit a large amount of radiation across a broad range of frequencies, in comparison to normal galaxies, such as our own Milky Way. However, at higher energies, the resolved fraction of the X-ray background remains low: observations with IBIS on INTEGRAL, the newest hard X-ray astronomical instrument currently in orbit, resolve only 3% of the XRB into point sources at 20–50 keV (Krivonos et al., 2005). It is believed that highly obscured types of AGNs, e.g., Type-II Seyfert galaxies, contribute to the XRB at high energies. Yet, due to the lack of instrument in the hard X-ray band with high angular resolution, observational evidence remains elusive. This is significant, because the spectral density of energy flux (i.e., flux per unit frequency, νFν) for the XRB is concentrated roughly in the range 20–40 keV (Fabian and Barcons, 1992), and at these energies, the XRB is as yet only 3% resolved. In other words, we have yet to account for the complete origin of the XRB at the frequencies where its effect is the most prominent. If the conjecture that highly obscured AGNs make up the majority of the XRB at these energies, then by resolving this population of hard X-ray sources, we would advance our understanding of these interesting sources tremendously. Of course, such a feat requires that hard X-ray telescopes with high resolving power be available. Unfortunately, the current generation of astronomical instruments in hard X-ray does not yet allow us to achieve such a feat.

Core-collapse supernova remnants

Supernovae can be classified into two main categories—Type Ia supernovae are those formed from the collapse of carbon-nitrogen-oxygen white dwarfs, when they accrete matter beyond the Chandrasekhar mass limit of 1.4 M; core-collapse supernovae are those formed from stellar explosions, when stars use up their fuel for fusion and reach the end of their lives. In the scenario of a core-collapse supernova, the stellar explosion ejects the outer layers of the star into the surrounding medium, forming an expanding shell of ejecta. In contrast, the inner layers shrink inwards due to gravity and form a compact object—either a neutron star or a black hole. Whether a neutron star or a black hole is formed depends on the amount of matter that falls inwards to form the compact object; this in turn depends on the radius at which the infalling matter and ejecting matter separate. We call this critical radius the ‘mass cut’.

As much as astronomers have studied core-collapse supernovae and their remnants (SNRs), we still do not know much about the mass cut, and consequently, the initial mass of neutron stars and black holes. According to present theories of core-collapse supernovae, there is an important diagnostic tool that can tell us the location of this mass cut: at the mass cut, an ample supply of alpha particles at extremely high temperatures (above 3.5 × 109 K) fuses into carbon-12 and heavier nuclei to form a sequence of isotopes (up to nickel-56) with equal and even numbers of protons and neutrons, called the alpha nuclei. A prominent member of the alpha nuclei is the radioactive isotope titanium-44, with a half-life of t½ = (58.9 ± 0.3) years (Ahmad et al., 2006). It decays, by way of scandium-44 (t½ = 3.97 hours), to the stable isotope calcium-44. Because of the relatively long lifetime of Ti-44, we should be able to detect the radioactive emission of Ti-44 in young core-collapse supernova remnants, such as the three-century old Cassiopeia A (Cas A) and the twenty-year old SNR 1987A. In fact, previous observations of Cas A by Comptel onboard the Compton Gamma-Ray Observatory (CGRO) has detected a 1.156 MeV nuclear spectral line, produced when newly created Ca-44 de-excites to its nuclear ground state. Further observations of Cas A by the Phoswich Detection System (PDS) onboard BeppoSAX also detected excess line emission above a continuum in the hard X-ray band, corresponding to two spectral lines of equal strength at 67.9 and 78.4 keV of Sc-44, as it de-excites to its ground state (Vink et al., 2001). Subsequent observations with IBIS/ISGRI onboard INTEGRAL confirmed the BeppoSAX detection, and also spectrally resolved the two lines (Renaud et al., 2006). If we are able to spatially resolve these nuclear line emissions associated with the decay of Ti-44 in nearby, young, core-collapse SNRs, we would be able to tell the location of the mass cut, and thus, shed light on the initial mass of neutron stars and black holes.

Unfortunately, none of the gamma-ray and hard X-ray instruments in operation today has sufficient angular resolution to spatially resolve Cas A and other nearby young supernova remnants: Cas A and the Crab Nebula are each about 6′ across, which is much smaller than the angular resolution of Comptel on CGRO (1.7°–4.4° FWHM), PDS on BeppoSAX (1.3° FWHM), and IBIS/ISGRI on INTEGRAL (12′). While the angular resolution of MeV-range astronomical instruments are still nowhere near the size of nearby young SNRs, the hard X-ray band shows better promise, as shown by the improvement from PDS on BeppoSAX (in operation from 1996 to 2002) to IBIS/ISGRI on INTEGRAL (in orbit since 2002). Thus, the goal of spatially resolving Ti-44 in nearby SNRs by detecting either or both of the emission lines at 67.9 and 78.4 keV still awaits a new hard X-ray telescope with arcminute or better angular resolution.

Inverse Compton scattering of cosmic background photons

Many large celestial systems are populated with relativistic particles (mostly electrons, positrons and protons). These relativistic particles interact with photons from the cosmic microwave background (CMB) via inverse Compton scattering. As a result, the relativistic particles transfer energy to the CMB photons, upscattering the latter from the microwave band to the X-ray band (IC/CMB). Some of these large celestial systems, such as many clusters of galaxies and jets of active galaxies, are also filled with a net magnetic field. In these cases, the relativistic particles also release energy via interaction with the magnetic field and the emission of synchrotron radiation at radio frequencies. Oftentimes, however, critical numbers such as the number density of relativistic particles, ne, or the strength of the magnetic field, B, are unknown. Yet, these values are important quantities that tell us the energy budget of the host system. Because these systems are amongst the largest-scale systems in the universe, the understanding of these systems are important in the study of large-scale structures.

Through radio observations, one can measure the strength of synchrotron emission from these systems. However, because synchrotron emission depends on both ne and B, one cannot untangle these quantities without making some conventional assumption, such as minimum energy or the equipartition of energy density between particles and the magnetic field. In contrast, IC/CMB emission depends only on ne. Thus, we can deduce ne via observations of IC/CMB in the X-ray band, without making the previously mentioned assumptions. By combining IC/CMB observations with radio observations, one can also deduce B unambiguously, as well as quantities such as the energy density of the host system (Harris and Grindlay, 1979).

Recent observations of active galaxies, relic radio lobes and clusters of galaxies have shown that minimum energy and equipartition are not always valid assumptions about these celestial systems. This shows the importance of X-ray observations of IC/CMB emission. Unfortunately, current observations of galaxy clusters and AGNs in the soft X-ray band are often dominated by their much stronger thermal emission, making the measurement of IC/CMB emission either impossible, or at best imprecise. Also, with the typically µG-level field strength in these environments, relativistic particles that emit IC/CMB radiation in soft X-ray emit synchrotron radiation at radio frequencies below about 10 MHz—a band inaccessible from Earth due to reflection by the ionosphere. To mitigate the problem of contamination by thermal emission in the soft X-ray band, and to measure the IC/CMB emission of the same particles whose synchrotron emission we observe at hundreds of MHz, one must study these systems in the hard X-ray band. Such effort is again limited by current technologies in hard X-ray astronomy. With the sizes of many clusters of galaxies on the order of a few arcminutes (comparable to the size of a young SNR mentioned above), the current generation of hard X-ray instruments are unable to spatially resolve these systems, making IC/CMB observations as yet impossible.

Cyclotron Resonance Scattering Features

In an X-ray binary star system, a compact object and a normal star companion are gravitationally bound to each other. Particles escape from the companion star either in a stellar wind or through Roche lobe overflow. These particles accrete onto the compact object; in the process, the charged particles accelerate and emit X-rays. In X-ray binary systems where the compact object is a highly magnetized neutron star, the strong magnetic field around the neutron star confines the charged particles to travel in helical trajectories along magnetic field lines, due to the Lorentz force. When one considers the motion of a charged particle in the direction perpendicular to the field lines, one finds that its energy is quantized in Landau levels. The energies En of these levels are uniformly spaced, as in a quantum harmonic oscillator:

En = (n + ½) ℏ ω / (1 + z),
where ℏ is the reduced Planck constant, ω the (angular) cyclotron frequency, and z the gravitational redshift for the neutron star. The magnetic field strength B and the charge-to-mass ratio q / m of the particle in motion determine the cyclotron frequency, ω = B q / m (in SI units), and thus the separation between successive Landau levels. Much like bound electrons within an atom, charged particles in a strong magnetic field transition between the Landau levels via photon-induced excitations and de-excitations; in the process, they preferrentially absorb and scatter photons of energies equal to the energy differences between Landau levels. Because Landau levels are uniformly spaced, the energy differences between levels are always multiples of ℏ B q / m / (1 + z). For electrons (q = e, m = me) travelling in the magnetosphere of a highly magnetized neutron star with a typical field strength of 10¹² gauss, the energy differences are multiples of ℏ B q / m / (1 + z) = ℏ B e / me / (1 + z) = 12 keV / (1 + z), well within the hard X-ray band. The absorption and scattering of photons at these discrete energies form observable absorption lines in the hard X-ray spectrum of certain X-ray binaries. We call these lines cyclotron resonance scattering features (CRSFs).

Because the energies of CRSFs are a pure function of the magnetic field strength (for electrons), CRSFs provide the only known direct diagnostic of the magnetic field at the surface of neutron stars and pulsars, where the field is strongest, due to the convergence of field lines at the magnetic poles. As the field strength decreases with distance from the neutron star surface, so does the energies of absorption and scattering of photons by electrons. Thus, while the centroid energy of a CRSF tells us the magnetic field strength at the surface of the compact object, the energy profile of the absorption line provides further information on the structure of the magnetic field. To date, CRSFs have been discovered in 13 accreting X-ray binary systems (see Coburn et al., 2002, for a representative sample). In addition, a putative 73 keV emission line may exist in the spectrum of the Crab pulsar (even though it is not in a binary system), possibly also due to cyclotron emission (Ling et al., 1979, amongst others). With the current generation of X-ray astrophysical instruments, the spectral resolution is only sufficient in detecting a deficiency of counts in a spectrum of an X-ray binary, and thus the precence of a CRSF. One obtains the centroid frequency of the absorption line by fitting the line to a theoretical model line profile, with various assumptions on the structure of the field. For further progress in our understanding of the strong magnetic fields around neutron stars and pulsars, there is a need for a new hard X-ray instrument with much higher spectral resolution than the present ones. With increased spectral resolution, we would be able to resolve the line profiles of CRSFs and study the structure of these magnetic fields.

Non-thermal emission in the Crab Nebula

The Crab Nebula, with a strong X-ray pulsar within a young supernova remnant, is a valuable natural laboratory for the study of pulsars and their non-thermal emission. According to magnetohydrodynamic (MHD) models of the Crab Nebula (Pelling et al., 1987), the pulsar dissipates the energy of its rotation via the acceleration and ejection of particles. These accelerated particles become relativistic; they travel away from the pulsar in bulk motion and form a stellar (or pulsar) wind. Being a supersonic flow of plasma, the pulsar wind produces an MHD stand-off shock as it comes into contact with the ambient medium of supernova ejecta. The supersonic flow decelerates through the shock, transferring its energy to particles in the ambient medium. As a result, the postshock medium is compressed, increasing the densities of both charged particles and magnetic field lines in the postshock plasma. With the strengthened magnetic field, the high-density and high-energy charged particles emit synchrotron radiation, which we observe in the Crab Nebula from the optical band through the X-ray band.

In the Crab Nebula, synchrotron emission shows structures dependent on both the position and energy of emission. The pulsar wind is not isotropic; rather, it flows preferrentially in the equatorial plane, perpendicular to the spin axis of the pulsar. As a result, synchrotron emission from the postshock medium is also the most prominent in the equatorial plane; it appears as a torus around the central pulsar in X-ray observations. The inner perimeter of this torus marks the position of the stand-off shock, about 10″ across in soft X-ray observations. In contrast, the outer extent of the toroidal emission varies with the band of observation—as the postshock distance increases, emission at high energies decreases faster than at low energies, due to energy loss by the emitting particles through synchrotron radiation. Thus, information on the extent of this emission as a function of energy helps us to characterize the pulsar wind, and to constrain current MHD models of the Crab pulsar.

In soft X-rays, the Chandra X-ray Observatory has performed direct imaging of the Crab Nebula with subarcsecond resolution, showing structures of the toroidal emission zone in spectacular detail. In contrast, this central part of the Crab Nebula has only been imaged indirectly in the hard X-ray band, up to 64 keV and with an angular resolution of 15″ (Pelling et al., 1987). These hard X-ray images are reconstructions by deconvolving one-dimensional scans of the Crab Nebula with a non-imaging instrument. They show the toroidal emission to span about 1.5′ in half-power diameter at hard X-ray energies (22–64 keV). For continuing progress in this work, first done in the 1980s, both direct imaging in the hard X-ray band and imaging at energies beyond 64 keV are much desirable.

Focusing opens a new window in hard X-ray astronomy

As evident from the various examples detailed above, many non-thermal astrophysical processes create phenomena in the hard X-ray band with interesting spatial features at the arcminute level, and sometimes with discrete spectral features as well. The set of resolvable targets in hard X-rays is restricted by the limiting angular and spectral resolutions of the current generation of non-focusing hard X-ray astrophysical instruments. Progress in this field relies on improvements in instrument technologies on several fronts—higher angular resolution, higher spectral resolution, as well as higher sensitivity (that is, higher signal-to-noise ratio and lower minimum detectable flux). Our development of HEFT is a first step towards these goals.

To illustrate the improvement enabled by focusing telescope technologies in hard X-ray, shown here are images of the Crab Nebula in various energy bands, as seen with the most sensitive telescopes in operation today. Each image is 6′ wide. Note that the angular resolution of IBIS on INTEGRAL is 12′, twice the width of these images. Here is the Crab Nebula in various energy bands again, including a hard X-ray image from the preliminary analysis of HEFT data taken during its 2005 observation run. The angular resolution of HEFT is about 1.5′.