PDS_VERSION_ID = PDS3 OBJECT = INSTRUMENT INSTRUMENT_HOST_ID = HST INSTRUMENT_ID = FOS OBJECT = INSTRUMENT_INFORMATION INSTRUMENT_NAME = "FAINT OBJECT SPECTROGRAPH" INSTRUMENT_TYPE = "SPECTROGRAPH" INSTRUMENT_DESC = " The information in this document is taken from [KINNEY1994]. Instrument Overview =================== The Faint Object Spectrograph has wavelength coverage on the blue side from 1150A to 5400A (FOS/BL), and on the red side from 1620A to 8500A (FOS/RD). There are both low spectral resolution (lambda/delta lambda ~ 250) and high resolution (lambda/delta lambda ~ 1300) modes, as discussed with in the FOS handbook. The brightest objects observable with FOS have magnitudes from V ~ 6 (for a G2V star) to V ~ 8 (for a B0V star or for an object with spectral shape of f nu proportional to nu^-1; see Table 1.3.1 for brightness limits of all gratings and spectral types. For magnitudes V ~ 20, the target counts are approximately the same as the detector dark counts (0.007 counts s^-1 diode^-1 on the blue side, and 0.01 counts s-1 diode^-1 on the red side) for a G2V star observed with the red side or for a B0V star observed with the blue side. These general traits of FOS blue side (FOS/BL) and red side (FOS/RD) are given in Table 1.0.1. The Faint Object Spectrograph has two Digicon detectors with independent optical paths. The Digicons operate by accelerating photoelectrons emitted by the transmissive photocathode onto a linear array of 512 diodes. The individual diodes are 0.31 seconds wide along the dispersion direction and 1.29 seconds tall perpendicular to the dispersion direction. The detectors are sensitive over the wavelength range from 1150A to 5400A (FOS/BL) and from 1620A to 8500A (FOS/RD). The quantum efficiencies of the two detectors are shown in Figure 1-1. The optical diagram for the FOS is given in Figure 1-2. The general characteristics of FOS/BL and FOS/RD are given in Table 1-2. Appendix H provides a comparison of the relative instrumental sensitivities of the FOS and GHRS Side 1. TIP: For the G130H spectral region the GHRS Side 1 G140L is normally more efficient than FOS/BL, especially when re-binned to FOS resolution. See Appendix H. Dispersers are available with both high spectral resolution (1 to 6A diode^-1, lambda/delta-lambda ~ 1300) and low spectral resolution (6 to 25A diode^-1, lambda/delta-lambda ~ 250). The actual spectral resolution depends on the point spread function of HST, the dispersion of the grating, the aperture used, and whether the target is physically extended. The brightest objects observable with FOS depend strongly upon the type of object and the combination of detector, spectral element, and aperture to be used; see Tables 1-10 and 1-11 for brightness limits for each detector and grating as a function of spectral type. Particle-induced FOS detector background is normally the dominant consideration in determining limits on faint sources that can be observed by FOS (see section 1.7). The mapping of the photocathode to the diode array is affected by the changing geomagnetic environment on orbit. An onboard real-time correction (the geomagnetic-image-motion, or GIM, correction) is applied routinely in all data-taking modes except ACQ/PEAK. The instrument has the ability to take spectra with high time resolution (>= 0.03 seconds, RAPID mode) and the ability to bin spectra in a periodic fashion (PERIOD mode). Although FOS originally had ultraviolet polarimetric capabilities with the FOS/BL G130H grating, the post- COSTAR environment allows polarimetry only for l ³ 1650A; that is only with gratings G190H, G270H, and G400H and both FOS/BL and FOS/RD. There is a large aperture for acquiring targets using on-board software (3.7² x 3.7², designation 4.3). Since the diode array extends only 1.3² in the Y-direction, this largest aperture has an effective collecting area of 3.7² x 1.3². Other apertures include several circular apertures with sizes 0.86² (1.0), 0.43² (0.5), and 0.26² (0.3); and paired square apertures with sizes 0.86² (1.0-PAIR), 0.43² (0.5-PAIR), 0.21² (0.25-PAIR), and 0.09² (0.1-PAIR), for isolating spatially resolved features and for measuring sky. In addition, a slit and several barred apertures are available (see Figure 1-3 and Table 1-4). FOS/BL sensitivity decreased by about 10% from launch until 1994.0, but has been stable since that time to the present. A dip in FOS/BL instrumental sensitivity to approximately 50% of pre-COSTAR levels centered at 2000A (which extends with an approximate Gaussian full width from 1600A to 2400A) appeared immediately post-servicing. The FOS/RD sensitivity is now generally stable to within 5%, but was observed to decrease rapidly in Cycles 1 and 2 in a highly wavelength dependent fashion between 1800A and 2100A, affecting gratings G190H, G160L, and to a lesser degree G270H. The flat fields for these 3 gratings changed little between early 1992 and mid- 1994. Flat fields have been obtained in either the large 3.7² x 1.3² aperture (4.3) or the 0.9² (1.0) aperture for the G190H, G160L, and the G270H gratings approximately quarterly beginning March, 1994 in order to monitor this effect. The sensitivity of both the blue and the red detectors is to be monitored approximately every 2 months in cycle 5. Please refer to Chapter 3 for a more complete characterization of the calibration status of the FOS. 3. INSTRUMENT PERFORMANCE AND CALIBRATIONS 3.1 Current Calibration Status After the deployment of COSTAR, an extensive calibration program was carried out during SMOV and Cycle 4, and the performance of the post-COSTAR FOS has been characterized in detail. In Cycle 5, we expect to maintain the routine calibration situation for the FOS. Our calibration program is divided into two parts: (1) a set of monitoring tests which aim to check the stability of the instrument performance, and (2) a set of specific tests designed to maximize the instrumental performance. As a general guide to the FOS calibration programs in Table 3-1 we provide a list of the calibration tests which are being performed in Cycle 5. For each calibration program we give the proposal ID, title, accuracy goal, number of orbits required, and comments which may include scheduling information. For detailed information regarding the Cycle 5 calibration program see Appendix I. The routine monitoring proposals are designed to monitor those aspects of the FOS performance that are known to show time variations. The focus test is conducted only once during the cycle because we know that the variations in the FOS focus are not large and do not affect the photometric accuracy of the data dramatically. The high voltage settings for the Digicons are also checked once a cycle. Since the FOS detectors are affected by external magnetic fields, the location of spectra on the photocathode and the FOS internal background observations will be conducted once every month. Similarly, the stability of the internal wavelength calibrations will be checked once every month. Some FOS detector/disperser combinations have shown temporal variations in their flat field structure during previous cycles. Several, though not all, flat fields will be monitored as frequently as every 2 months. The absolute photometric calibration of some spectral elements have shown modest temporal variations during Cycle 4; we will monitor these aspects of FOS sensitivity also once every 2 months. The special calibration proposals are designed to characterize those aspects of the FOS performance which have been specially requested by Cycle 5 GOs. The accuracies specified in Table 3-1 are the minimum goals of the Cycle 5 calibration program. The requirements for the success of the calibration program are slightly less stringent. We expect that the Cycle 6 calibration plan will be similar to that of Cycle 5. We expect to monitor instrument modes that have been previously characterized. We do not anticipate the expansion of the calibration beyond the areas described in the Cycle 5 plan. If the calibration accuracies specified here are insufficient to meet science program requirements, then proposers should be prepared to provide extra calibration time in their programs to meet their specific calibration needs. above. Table 3-1: Summary of FOS Cycle 5 Calibration Programs Proposal Title 6163 Focus, X-pitch, Y-pitch test 6165 Discriminator Test 6167 Dark Monitoring 6166 Y-base Monitoring and & 6236 Internal Wavelengths 6202 Spectral Flat Field Calibration 6203 Photometric Calibration 6206 Polarimetry Calibration 6204 Internal External Offset and Scattered Light Test 6205 Location of 1.0² aperture and the FOS PSF TOTAL TIME (including all executions) 3.2 Wavelength Calibration Unlike the IUE, all FOS wavelengths are vacuum wavelengths both below 2000A and Wavelength offsets between the internal calibration lamp and a known external point source are currently based on observations of the dwarf emission line star AU Mic that have been corrected for geomagnetically induced image drift (Kriss, Blair, and Davidsen 1992). On the red side, the mean offset between internal and external sources is +0.176 ± 0.105 diodes. On the blue side, the mean offset is -0.102 ± 0.100 diodes. These offsets are included in the pipeline reduction wavelength calibration. If the target is well-centered in the science aperture, velocity measurements based on single lines in FOS spectra have a limiting accuracy of roughly 20 km s-1 if wavelength calibrations are obtained at the same time as the science observations and with no filtergrating wheel motion between the science and the wavelength exposures. (See Appendix D for line lists and spectra of the comparison lamps for each detector/disperser combination.) If simultaneous wavelength calibrations are not obtained, the non-repeatability of order 0.35 diodes in the positioning of the filter-grating wheel will dominate the errors in the zero point of the wavelength scale (Hartig 1989). Visit 03 of the sample RPS2 FOS program in Appendix F presents an example of a science exposure and accompanying high-precision wavelength calibration. Wavelength Calibration Summary: - Filter-grating-wheel repeatability is of order 0.35 diodes. - FOS wavelength accuracy limit approximately 20 km/sec. - Precision wavelengths require centering £0.04² and science and wavecal exposures must be observed consecutively, i.e., with no move of filter-grating wheel in between. See example in Visit 03 of the sample RPS2 FOS program in Appendix F. 3.3 Absolute Photometry The post-COSTAR absolute photometric calibrations have been performed by observing the standard stars G191B2B (WD0501+527), BD+28D4211, BD+75D325, HZ-44, GD153, GD71, HZ-43, and BD+33D2642 in the 3.7² x 1.3² (4.3) and the 0.9² (1.0)apertures. Cycle 5 and 6 calibrations will be based primarily on BD+28D4211 and G191B2B. See Tables 3-2 and 3-3 for a summary of the detector/disperser/aperture combinations to be calibrated during Cycle 5. Observations of these spectrophotometric standard stars have been used to produce inverse sensitivity functions for all usable FOS detector/disperser combinations and all apertures except the 0.1-PAIR, 2.0-BAR, and 0.7X2.0-BAR. A highly precise target acquisition strategy (multi-stage ACQ/PEAK with pointing uncertainty of 0.04²) is used for these observations so that filter-grating wheel repeatability (0.10² = 0.35 diodes) is the dominant source of uncertainty in photocathode sampling. The standard star reference flux scale used for photometry is accurate to 1%. Limiting accuracies of FOS photometry are approximately 1.6% for FOS/BL and 2.0% for FOS/RD for wellcentered targets (accuracy £0.04²) in the 3.7² x 1.3² (4.3)and 0.9² (1.0) apertures. Modest temporal changes have been documented in the FOS/RD G190H and G270H sensitivity in the post-COSTAR period. The standard data reduction pipeline will include corrections for the influences of time dependence. As of May 1995 the sensitivities for all other detector/disperser combinations have been constant at the 2% level in the post-COSTAR epoch. The dominant error in FOS photometry is due to the location of the target in the aperture (see chapter 16 of the HST Data Handbook for the sources of photometric error). Centering accuracies of £0.08² are required for better than 5% photometric accuracies. Uncertainties in the instrumental magnetic deflection (the Y-base) required to direct photoelectrons from the photocathode to the diode array can affect the accuracy of photometry of extended sources observed through apertures comparable in size to the diodes (i.e., 0.9² (1.0) aperture and larger). Pre-COSTAR photometry was strongly dependent upon the combined effects of Optical Telescope Assembly (OTA) focus changes and time-dependent degradation of the FOS sensitivity. Nonetheless, limiting photometric accuracies similar to those quoted above are possible for properly re-calibrated pre-COSTAR data. Absolute Photometry Calibration Summary: - absolute reference system accuracy of 1%. - observational accuracies strongly dependent upon telescope jitter, target centering, aperture used, and stability of requisite instrumental magnetic deflection (Y-base). - limiting accuracies of approximately 1.6% (FOS/BL) and 2% (FOS/RD) require large apertures and precise centering (£0.04²). Observations of two hot spectrophotometric standard stars (G191B2B and BD+28D4211) are used to produce spectral flat fields for all usable FOS detector/disperser combinations. A highly precise target acquisition strategy (multi-stage ACQ/PEAK with pointing uncertainty of 0.04²) is used for these observations so that filter-grating wheel repeatability (0.10² = 0.35 diodes) is the dominant source of uncertainty in photocathode sampling. All post-COSTAR flat fields have been derived via the so-called superflat technique (described in Lindler et al CAL/FOS-088; flat field article by Keyes in HST Calibration Workshop, ed. Blades and Osmer (1994); and Part VI, Chapter 16 in the HST Data Handbook, ed. Baum (1994)). Appendix G provides figures showing flat field structure for most usable FOS detector/disperser combinations derived from Cycle 4 3.7² x 1.3² (4.3)aperture superflat observations. It must be emphasized that FOS flat field corrections are intended to remove photocathode granularity typically on the scale of 10 pixels or less. If high precision flat fields are required for scientific objectives, observers should attempt to attain the same pointing accuracy (described above) used for FOS flat field calibration observations so that the science target illuminates the same portion of the photocathode as was sampled by the calibration observations. During Cycle 4 several observations have been made to attempt to quantify the change in flat field granularity structure as a function of target mis-centering perpendicular to dispersion. Most pixel ranges in typical flat fields display deviations of 1-2% about the mean value of unity or about a local running mean. However, some substantial (5-50% deviations from local mean) features do occur. Initial evidence indicates that photocathode granularity in these strong features can change by 25% on the scale of a diode height (1.29²). Should such a feature occur in the vicinity of an important spectral line and target centering be less accurate than that of the flat field calibration observation, then the observed flux could be affected by a potentially unknown amount. We note that there is such a prominent (30%) feature in the 1500-1550 A range (affects C IV resonance doublet) of the FOS/BL G160L spectrum, which impacts all single apertures and all lower paired apertures, but is not present in any upper paired aperture G160L spectrum. Normally one epoch of flat field measurement is made per cycle with the 0.9² (1.0)aperture for each detector/disperser combination to be used. Additional individual detector/disperser/ aperture combinations are calibrated depending upon use or known flat field variability. Please refer to Tables 3-2 and 3-3 for a summary of planned Cycle 5 calibrations. 3.5 Sky Lines In the pre-COSTAR era substantial temporal variation in FOS/RD G190H was observed. Between launch in April, 1990 and November, 1991 degradation of the 1800-2100 A. region of the FOS/RD G190H and G160L flat fields proceeded at the approximate rate of 10% per year. The variation slowed in early 1992 and little change was noted in the pre-COSTAR era after November, 1992. Smaller changes also occurred on the same timescale at additional wavelengths for both gratings and in the FOS/RD G270H spectral region. Cycle 4 flat field observations have indicated the presence of some time-variability for these three spectral elements since March, 1994. These gratings will continue to be monitored at approximate two month intervals during Cycle 5. Since time dependence has been observed in the red side flat fields, red side data taken after January 1992 should be flat fielded with the data most appropriate to the observation. On-line reference guides accessed through the FOS WWW Homepage refer the user to the appropriate flat field. Red side data taken between October 1990 and January 1992 will be difficult to flat field because of the lack of time-dependent flat fields available in that interval. Flat Field Calibration Summary: - must use precise target centering (£0.04²) to achieve flat field accuracies appropriate for science goals of S/N³30. - deviations from STScI calibration flats as a function of distance perpendicular to dispersion not well calibrated at present. These deviations can be significant. The lines of geocoronal Lya l1216 and OI l1304 appear regularly in FOS spectra, with a width determined by the size of the aperture (see Table 1-4). Occasionally, when observing on the daylight side of the orbit, the additional sky lines of OI l1356 and of OII l2470 can also be seen. Second order Lya sometimes appears at l2432. 3.6 Polarimetry FOS/BL G130H polarimetry observation will not be feasible. Gratings G190H, G270H, and G400H will be usable with both FOS detectors. The 0.9² (1.0)aperture should be used to obtain the highest polarimetric precision. Only the FOS/RD 0.9² (1.0)aperture will be calibrated in Cycle 5. The instrumental polarization introduced by the COSTAR reflections is £2% and is wavelength dependent. Actual polarimetric accuracies depend on the measured signal, of course. For very bright sources, the Cycle 5 limiting linear polarization uncertainty goal is of the order of 0.2%. Our preliminary analysis of the Cycle 4 calibration indicates a linear polarization accuracy of approximately 0.7%, however. Naturally, for faint sources the accuracy is limited by the photon statistics obtained and by the intrinsic polarization of the object (see section 1.6). Please refer to the FOS WWW Homepage after 1 June 1995 for further updates on polarization accuracies and uncertainties. 3.7 Scattered Light Extensive modeling of scattered light which has its origin in the diffraction patterns of the gratings, the entrance apertures, and the micro-roughness of the gratings has been performed by M. Rosa (see Appendix C). Predictions from these models will be tested in Cycle 5 by comparison of FOS G130H and GHRS G140L observations of a single target. The Rosa modeling code is available at STScI for use as a post-observation parametric analysis tool. Please contact an Instrument Scientist for more information. A routine wavelength-independent correction for scattered light plus uncorrected background (dark) is performed in the post-observation data reduction pipeline for those gratings, such as G130H, which have regions of zero sensitivity to dispersed light. Some gratings have no regions of zero sensitivity and therefore no pipeline correction is made in these cases. 3.8 Detector Background (Dark) FOS background corrections applied in the standard pipeline reduction may be approximately 30% too small in some cases. Naturally, this affects only observations of faint sources. Approximate mean background count rates are 0.010 counts-sec-1-diode-1 for FOS/RD and 0.007 counts-sec-1-diode-1 for FOS/BL. The FOS team is presently analyzing all available background measures in order to derive corrected reference files and tables. New information will be posted in the ADVISORIES section of the FOS Homepage as appropriate. Background measures will continue to be made monthly throughout Cycles 5 and 6. 3.9 Dead Diodes Occasionally, one of the 512 diodes on the red or the blue side becomes very noisy or ceases to collect data. Since launch 3 diodes on the blue side and 2 diodes on the red side have stopped functioning. In addition, several diodes on each side have become noisy and have been disabled. With the inclusion of the diodes known to be dead or troublesome from pre-launch there are currently 26 FOS/BL diodes disabled and 15 disabled for FOS/RD. See Appendix B for the current disabled diodes list (several diodes are under scrutiny at present, but no new additions to the list have been made since December 1993). " END_OBJECT = INSTRUMENT_INFORMATION OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "BLADES&OSMER1994" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "BAUM1994" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "HARTIG1989" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "KRISSETAL1992" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "LINDLER&BOHLIN1986" END_OBJECT = INSTRUMENT_REFERENCE_INFO END_OBJECT = INSTRUMENT END