The Advanced X-ray Astrophysics Facility (AXAF) is an orbiting x-ray telescope being developed as part of NASA's Great Observatories program for launch in 1998.[1] AXAF will provide high-angular-resolution imaging spectroscopy and transmission grating spectroscopy in the 0.1-10 keV band. The MIT Center for Space Research, together with the MIT Lincoln Laboratory and Pennsylvania State University, are developing the AXAF CCD Imaging Spectrometer (ACIS), one of AXAF'S two scientific focal plane instruments. The detectors are described by Burke, et al.[2] Although on-board x-ray sources are provided, extremely accurate calibration of the CCDs before launch in a laboratory is essential to meet the stringent demands of the AXAF scientific objectives; for example, the goal for detection efficiency knowledge accuracy is of order 1%. This paper describes the laboratory calibration x-ray sources and environment used for this purpose, which were designed, built, and operated by CCD Laboratory at MIT's Center for Space Research.
The ACIS instrument calibration program, summarized in another contribution to these proceedings (Bautz et al; 2808-16), includes measurements of the CCD energy scale and spectral response function, as well as absolute measurements of X-ray CCD detection efficiency as a function of X-ray energy and detector position. In this paper, we focus on the X-ray sources developed for the absolute quantum efficiency measurements. Our strategy for making these measurements is to calibrate reference transfer standards at a synchrotron light source whose spectral radiant intensity is known absolutely. We have used the facilities of the PTB Laboratory at BESSY[4] for this purpose. We then measure in our laboratory the detection efficiency of the flight detectors, relative to that of the absolutely calibrated transfer standards, using the X-ray sources described here.
The principal objective is to provide a stable x-ray source of known
energy and flux onto an ACIS CCD. There are major three requirements:
(1) To avoid pileup, the detected event flux should not exceed 6400
counts during an exposure (typically either 3.1 or 7.1 sec). (2) For
resolution measurements, the selection of energies should include
about 10 known lines ranging from about 200 eV to 10 keV. (3) For
quantum efficiency measurements, the flux should be stable to within 1
percent with a smooth and repeatable uniformity. There are three
major constraints: (1) radiation damage protection requires a detector
temperature of -120 degrees C. (2) Low temperature conditions require
a vacuum of around 1
to avoid potentially damaging
buildup of ice. (3) AXAF contamination control requirements imply
cleanliness standards.
In general, a good x-ray source of known energies are characteristic
x-rays resulting from deexcitation of atoms after an inner shell
ionization. Target atoms with atomic number Z between 6 and 32 can
yield k alpha x-rays between 277 eV and 9875 eV, respectively. Ions,
electrons, and photons can induce inner shell ionizations with
sufficient energy, although the efficiencies can widely vary. The
limitation for efficient high energy x-ray production (at high Z) is
to provide an ionizing radiation whose energy is higher than the
desired ionization energy (typically around 10 keV). The limitation
for efficient low energy production (at low Z) is to induce
ionizations near the material surface so that emerging characteristic
x-rays are not greatly self-absorbed. For example, the self
absorption length for emerging aluminum k alpha x-rays is 10 microns,
and any ionizing radiation penetrating much further becomes redundant.
The low energy efficiency is further lowered by a fluorescent yield
scaling as Z
.
There are three candidates for the ionizing radiation: (1) Ions have
short penetration lengths but are undesirable since MeV energies are
required for a reasonable x-ray yield. Such sources are typically
unstable nuclei with large Z (e.g. alpha particles from 
Cm)
having many other decay products which contaminate the source spectrum
with out-of-band radiation. (2) Photons have good x-ray yields when
the photon energy is not exceeding larger than the ionization energy.
However, stable low energy x-ray sources are difficult to find which
do not incorporate an optical blocking filter, making ionization of
low Z targets difficult. (3) Electrons have smaller ionization
cross-sections, however they can be copiously produced. They are most
efficient when the electron energy is within a factor of two of the
ionization energy.
Our source selection combines two methods permitting inner shell ionizations in the range from Z=6 (277 eV) to Z=32 (9875 eV). (1) A commercial x-ray tube using a 15 kV electron beam on a molybdenum target produces a continuum spectrum which efficiently ionizes targets from Z=13 to Z=32. (2) A newly developed tritium excitation source produces a continuum of electrons from 0 to 18 keV which efficiently ionizes targets from Z=6 to Z=13. The advantage of the high energy x-ray source (HEXS) is ease of use and construction; the disadvantage is beam current instability and inefficiency at low energies. The advantage of the tritium source is low energy capability and stability; the disadvantage is tritium outgassing and contamination. Quantum efficiency (QE) is measured by by comparing the detected x-ray flux between a ``flight'' CCD with unknown QE versus the detected x-ray flux of a ``reference'' CCD with known QE. Both CCDs are mounted together on a translation stage which can toggle each underneath an x-ray source. The absolute calibration of the reference CCD is discussed in another paper (2808-16). The remainder of this paper discusses in further detail the high energy x-ray source (section 2), the tritium source (section 3), and the entire calibration assembly (section 4).