The ultimate angular resolution of any telescope is given by the diffraction limit, θ_d = λ/D, where λ is the wavelength and D is the telescope aperture. For Chandra’s 1.2 m aperture at 5 keV (λ = 0.25 nm), θ_d turns out to be 40 micro-arcsec, some 12,000 times smaller than Chandra’s actual (and still unsurpassed in the x-ray regime) angular point-spread function size of 0.5 arcsec. This would be a remarkable angular resolution for such a compact optical system, even exceeding the performance of the Event Horizon Telescope with an aperture that is essentially the size of the Earth. Why isn’t Chandra’s resolution better?
1. Chandra’s mirror shells are not quite diffraction limited, by they are not too far off for soft x rays (E ~ 1 keV). Making them perfect would only gain an improvement of a factor of a few at best.
2. Making the resolution better requires smaller detector pixels. Chandra’s CCD pixels are 24 micron in size. Even if pixels could be shrunk to 1 micron, this would still require increasing the focal length from 10 m to 5 km.
3. Most importantly: By Fermat’s theorem, achieving diffraction-limited performance requires all optical paths from source to image planes be the same length to within a small fraction of the wavelength. However, Chandra’s Wolter Type 1 mirror prescription decidedly does not follow Fermat’s requirement for multiple shells, so that even if a 5000 m focal length was implemented, only a 100× resolution improvement would result. Basically, each shell is its own telescope, with an aperture given by its projected annulus.
A new Wolter telescope design. We recently published a novel X-ray telescope design based on the well-understood Wolter Type 2 prescription [1]. This design preserves the advantages of the traditional Type 1 prescription—including a wide field of view and energy band, compact size and large effective area due to high packing density—while matching the path lengths of all shells and correcting for the chromatic aberration inherent in grazing-incidence telescopes operating below the critical angle. Such a telescope would have some 10^3 to 10^6 times better resolving power than Chandra, allowing direct X-ray imaging of some of the most highly energetic, compact and violent regions of the Universe. Just a couple of revolutionary science examples would be resolving the fine structure of quasar jets down to the event horizon and testing dark matter clumping via speckles produced by gravitational lensing.
Needed: Diffraction-limited mirror segments (almost there with single crystal silicon/Zhang group at GSFC), improved metrology for near-conical mirrors to improve the polish/metrology cycle, x-ray metrology, stable and sensitive alignment and mounting technologies.