QUANTUM CALORIMETRY - X-ray astronomy

X-rays reveal the hot universe

Since its beginnings in the 1960s, x-ray astronomy has played a large part in revealing the universe's unexpectedly hot and energetic side. (See David Helfand's article, "X-Rays from the Rest of the Universe," PHYSICS TODAY, November 1995, page 58.) Today, we know that most of the universe's nonexotic matter (at least in the forms we can detect) exists as gas at temperatures of tens of millions of degrees. Clusters of galaxies --- the largest gravitationally bound structures in the universe --- hold huge quantities of this gas, whose combined mass in a typical cluster outweighs that of the constituent galaxies by up to a factor of five. Essentially invisible except at x-ray wavelengths, the hot intracluster medium traces a cluster's gravitational potential well, whose main source of mass is neither hot gas nor galaxies, but the still-mysterious dark matter.

X-rays have also provided a way to study physics in the immediate vicinity of black holes (see Roger Blandford and Neil Gehrels's article, "Revisiting the Black Hole," PHYSICS TODAY, June 1999, page 40) and neutron stars (see Lars Bildsten and Tod Strohmayer's article, "New Views of Neutron Stars," PHYSICS TODAY, February 1999, page 40).

To see why x-rays reveal the extreme environments around these two kinds of compact star, consider the Stefan Boltzmann equation,

P = A σ T⁴

which gives the radiative power P of a black body in terms of its radiating area A, its temperature T, and the Stefan Boltzmann constant σ. This equation, combined with the inverse-square law, tells us that small, faraway things will be too faint to see unless they are very hot. For example, a celestial object --- a neutron star, active galactic nucleus, or supernova remnant, say --- can be 10¹² times smaller in area or 10⁶ times farther away and still be as visible, provided it's also 10³ times hotter and that we're looking in the right waveband! Fortunately, the gravitational potential near neutron stars and black holes provides more than enough energy to raise the temperature of infalling matter to x-ray emitting levels.

By measuring this energetic radiation, we can study regions of extreme temperature and gravitational and magnetic fields that cannot be reproduced on Earth. Spectroscopy can be an effective tool in this endeavor, especially for the 0.1--10 keV x-ray band, corresponding to temperatures from 10⁶ to 10⁸K. At these temperatures, hydrogen and helium are fully ionized, but bremsstrahlung is still inefficient, and the dominant radiation consists of collisionally excited characteristic lines of partially ionized heavy elements. These lines provide a wealth of diagnostics on the elemental abundances and physical conditions in the gas, and measurements of Doppler shifts and linewidths provide invaluable information about the gas's motion.

However, the detailed physical information that the spectra contain can only be extracted by a high-resolution spectrometer like a calorimeter. In the precalorimeter era, the choice between wavelength dispersive devices (diffraction gratings and Bragg crystal spectrometers) and nondispersive spectrometers (proportional counters and solid-state detectors) presented a dilemma. Dispersive spectrometers offer very good energy resolution, but at low throughput. Nondispersive spectrometers, on the other hand, have very high efficiency, but relatively poor resolution.

Gratings have further disadvantages. Their good energy resolution requires optics with very high angular resolution, but, to date, the design of x-ray telescopes has involved a trade-off between angular resolution and collecting area, putting gratings at a sensitivity disadvantage. Furthermore, gratings disperse the spectrum --- into several orders --- across a position-sensitive detector. But when the x-ray emitting object is itself extended, untangling the dispersed spectra is a daunting prospect.

As a result of these considerations, most astronomical x-ray spectroscopy so far has been carried out with nondispersive detectors that directly determine the photon energy by measuring the amount of ionization produced by each event. Unlike gratings, the energy resolution of these devices is not affected by the acceptance angle, so they can have very high throughput. However, statistical fluctuations in the fraction of a photon's energy that goes into ionization from one event to the next fundamentally limit the resolution.

For silicon, the theoretical limit is about 100 eV at 6 keV, a value that is closely approached by modern x-ray charge-coupled devices. An examination of figure 1, which shows modeled x-ray emission around the iron K lines from a 7x10⁷K plasma as seen with 1, 10, and 100 eV resolution, makes it clear that much higher resolution is needed to take full advantage of the information available. Indeed, the need for higher resolution is what inspired our quest for a better detector.