The fundamental purpose of the scientific calibration of AXAF is to develop accurate predictions of the system response (in engineering units) to any scientifically interesting X-ray flux distribution. The basic strategy adopted to achieve this end is to develop physical performance models of each of the Observatory's scientific subsystems, and to constrain and verify these models by well-chosen ground measurements. The subsystem (telescope, grating and detector) models are then to be checked during calibration of the integrated system.
Figure 2: ACIS ground calibration flow. This paper focuses on the measurements listed in the shaded box.
This strategy is recapitulated at the ACIS instrument level: various instrument subsystem models are calibrated, and then checked during instrument-level performance tests. This structure is illustrated in Figure 2. The selection and calibration of individual ACIS CCD detectors is shown on the left half of Figure 2. Following fabrication at MIT Lincoln Laboratories, devices are subjected to a two-stage screening process, which is described in greater detail by Pivovaroff et al .
The remainder of the CCD subassembly calibration is divided into two phases. A subset of flight-quality devices has been characterized as transfer standards at the facilities of the Physikalisch-Technische Bundesanstalt (PTB) Laboratory at the Berliner Electronenspeichering-Gesellschaft für Synchrotronstrahlung (the BESSY synchrotron storage ring). The intensity of the undispersed synchrotron radiation at the PTB beamlines can be calculated to high accuracy (better than 1% in the spectral band 0.2 - 4 keV) . This radiation provides a primary standard for the absolute detection efficiency calibration of the transfer standard CCDs.
4 keV, we plan to use a Si(Li) detector as a transfer standard. This device and its use in ACIS CCD calibration are discussed by Manning et al.  We do not discuss it further here.
The calibration of flight detectors is performed at the MIT Center for Space Research. As is indicated in Figure 2, these measurements include determination of the energy-to-pulse-height relationship, the spectral resolution, and higher moments of the spectral response function. Dependence of the spectral response function with detector temperature, detector clock levels, and readout mode is also measured. Detection efficiency, especially as a function of position on the detector, is measured with reference to the calibrated transfer standard CCD detectors. Various other pertinent detector performance characteristics, as listed in Figure 2, are also measured.
Following subassembly calibration of individual CCDs, the detectors will be installed on the flight focal plane, which will then be integrated with the instrument electronics and other ACIS components. Key among these, from a calibration standpoint, are the ACIS optical blocking filters, which have been calibrated separately by our colleagues at the Pennsylvania State University. [9,10] The integrated ACIS instrument will be tested to verify selected aspects of the the subassembly calibration. Following the integrated instrument test, ACIS, along with other AXAF science instruments, will be calibrated with the AXAF High-Resolution Mirror Assembly at NASA's X-ray Calibration Facility (XRCF) at Marshall Space Flight Center.
The remainder of this paper describes the most important aspects of the CCD subassembly calibration listed in the shaded box in Figure 2: the spectral response characterization and the detection efficiency measurements.
Figure 3: Principal components of the CCD spectral response function. The (asymmetric) primary photopeaks, silicon K-escape and fluorescence features, and the extended low-energy tail are indicated. Note that the upturn in the tail at very low energies (E< 400 eV) is not due to electronic noise, but is a characteristic of the detector itself.