12.4 Observing Too-Bright Objects with STIS

As described in Section 7.7, the STIS MAMA detectors are subject to damage at high local and global count rates. The MAMA detectors also suffer uncorrectable non-linearity at similar count rates (see Section 7.5.4). There are therefore configuration-specific count rate limits for all observations that use the MAMA detectors; sources brighter than allowed by the limits cannot be observed in that configuration.

The STIS CCDs are not subject to the same bright object constraints, as the CCD cannot be damaged by observations of bright sources. At high accumulated count/pix levels, however, the CCD saturates and charge bleeds along the columns. When CCDGAIN=4, the saturated counts can be recovered by summing over the pixels bled into, and this spatially integrated count rate remains linear with exposure level (see STIS ISR 1999-05). This is not true for CCDGAIN=1. The charge bleeding resulting from saturation does somewhat complicate flat fielding and removal of cosmic rays, however, as it changes the locations of some of the observed counts.  As described previously (see Section 7.3.2), CCD saturation can be avoided by keeping exposure times short when observing bright targets. Spatial scanning with the CCD (Section 12.12) may also be employed.  The minimum exposure time for CCD observations (0.1 second) dictates the maximum source brightness which can be observed without saturating.

The only way to use STIS to observe a source that is too bright is to use a configuration or observing mode which reduces the flux from the target (globally and/or locally), bringing it into the observable regime. The options available to achieve this reduction are:

• Use a smaller slit to reduce the transmitted light for spectroscopic observations (see Section 13.4—you will find there the percent flux transmitted through each slit as a function of wavelength).
• Select a more appropriate grating or filter configuration. The solution may be a configuration with higher resolving power if it is the local limit which is being violated, or a configuration that covers a smaller spectral range if the global limit is being violated. In more extreme cases, you may be forced to choose a grating (filter) that covers an entirely different region of the spectrum. Note that if you are observing in first order in the NUV, you can consider using the CCD NUV first-order spectroscopic modes G230LB and G230MB (see Section 4.1.6).
• Use a neutral-density-filtered full aperture. The neutral-density filters are described in Section 5.4; they produce attenuations ranging by factors from 10^–1 to 10^–6. Note, however, that the ND filters are located in the slit wheel. Thus, all supported ND full-filtered exposures will be slitless; i.e., you cannot use a slit and an ND full filter together. Similarly, you cannot use a ND full filter and another filter in imaging mode. Also note that the NDQ1, NDQ2, NDQ3, and NDQ4 filters are four distinct quadrants of a single filter, all of which are simultaneously imaged. Note that NDQ4 is of little use, as any target that requires this filter is too close to the NDQ1 quadrant to pass bright-object screening.
• Use one of the echelle or long calibration slits which contain neutral-density filters. Supported neutral-density slits for the echelles are 0.2X0.05ND (with ND=2.0) and 0.3X0.05ND (with ND=3.0), where if ND=x, the flux is attenuated by approximately 10^–x. Supported neutral-density long-slits that can be used for first order or echelle observations are 31X0.05NDA (with ND=0.4), 31X0.05NDB (with ND=0.8), and 31X0.05NDC (with ND=1.2). Use of these long slits with an echelle grating will cause order overlap problems (see Section 12.2), but for point sources the order separation may be adequate for many science programs. Use of the neutral-density long-slits with the PRISM remains "available-but-unsupported" at this time.
• Use spatial scanning (CCD only) to spread the exposure over a larger spatial extent (see Section 12.12).