The ACIS reference CCDs were calibrated against the SSD at five primary
energies, corresponding to the K
lines of Si (1.74 keV), P
(2.01), Ti (4.51), Mn (5.89), and Cu (8.03). The Mn lines are
generated by radioactive decay of a 10 
Ci 
Fe source with an
Al collimator followed by a 2 mm Al aperture. When used with the
calibration aperture, this produces a circular spot of irradiation on
the CCD approximately 320 pixels (7.7 mm) in diameter.
The other lines are produced by irradiating pure samples of the target
element with continuum X-ray emission from a commercial X-ray tube
with a molybdenum anode, operated at 15 kV. This radiation
efficiently ionizes the target material's inner shell, yielding
fluorescent X-ray emission at the characteristic K lines of each
material used. This source, referred to as the High Energy X-ray
Source (HEXS) is described elsewhere in these
proceedings[4]. Depending on the angle between the Mo anode
and the surface normal of the target element, the radiation pattern
reaching the CCD has a width varying from approximately 13 to 17 mm.
The current was adjusted on the commercial X-ray tube to optimize the
flux rate for efficient counting with losses due to pulse pile-up limited
to 
or less. Pile-up corrections have been measured as a function
of count rate, assuming the X-ray flux is linear with tube current.
For each reference CCD, the calibrated points above 4 keV will be used
to model the depletion depth, which is the key parameter governing
detection efficiency for the higher energy X-rays. The depletion
depth has been estimated to be 47 
m for one reference CCD (device
w34c3). The flight chips will be operated with higher gate voltages
and are estimated to have a depletion depth of order 70 
m,
increasing the detection efficiency for the higher energies. Other
detection parameters such as the CCD oxide layer, which are more
important for low energy X-rays (
4 keV), are determined from the
PTB/BESSY calibrations.