WFPC2 Cycle 7 Closure Report
S. Baggett, J. Biretta, S. Casertano, S. Gonzaga, I. Heyer, M. McMaster, C. O'Dea, A. Schultz, B. Whitmore, M. S. Wiggs
Abstract
This report describes in detail the WFPC2 observations used to maintain and improve the quality of WFPC2 calibrations during Cycle 7 and their status as of November 1999. Also included are the WFPC2 programs executed during the NIC3 Campaign in 1998 and summaries of the Cycle 6 proposal analyses completed since the writing of the Cycle 6 Closure report.
1. Overview
As discussed in the Cycle 7 Calibration Plan for WFPC2 (ISR 97-06, Casertano et al.), the major goals of the calibration program during Cycle 7 were to verify the stability of the instrument and to continue pursuing the remaining issues limiting the photometric accuracy. Cycle 7, at about 24 months long (June 1997 to June 1999), was longer than previous cycles. The original plan called for a norminal 1 year Cycle, however, this was later extended because 1) the Servicing Mission verifications in early 1997 required more orbits than originally estimated and 2), the NICMOS observing program was accelerated in order to complete as many observations as possible before the cryogen ran out (this included nearly 1000 extra orbits accepted during a special call for NICMOS proposals in 1997). For WFPC2, the increased Cycle length translated to an increase in emphasis on the monitoring proposals, which were of course extended as necessary to provide coverage throughout the entire term of Cycle 7 and a smooth transition into the Cycle 8.
Monitoring Programs
As in previous Cycles, the monitoring program revolved around the monthly decontamination (decon) procedures used to remove contaminants from the CCD windows and to anneal hot pixels (proposal 7619 ). Each decon was flanked by the Photometric Monitoring program (7618 ), consisting of observations of the standard white dwarf GRW+70d5824 which provided regular measurements of the photometric throughput, the contamination state of the CCD windows, the PSF properties at a variety of wavelengths, and the OTA focus. Regular internal observations were also used for monitoring: weekly darks (7619, 7620, 7621, 7712, 7713), to produce dark reference files and hot pixel lists; weekly biases, INTFLATs, and K-spots (7619 , 7622 , 7623 ) as a check of the camera's optics, electronics chain, and the pixel-to-pixel response in the visible and to produce bias reference files; Earth flats (7625.pro ), to map any changes in the flatfields; and UV flats (7624 ) to monitor the pixel-to-pixel response in the UV.
Most of the internal programs were continuations from previous Cycles; however, in response to behaviour noted in the instrument during previous cycles, two new monitors were implemented in Cycle 7. The first was the set of Supplemental Darks (7621, 7712, 7713 ; multiple proposals were required for ease in implementation and scheduling). The supplemental darks are short, 1000 sec darks taken daily on a low-priority, non-interference basis, up to 3 darks per day, intended as an aid in identifying hot pixels. The other new monitoring program was the Astrometric Monitor program (7627 ), to allow tracking of the WFPC2 chip positions in the focal plane: previous measurements revealed that a gradual shift of ~1 pixel total has occurred since early 1994.
Special Programs
Three special calibration programs were planned for Cycle 7: two to help improve the photometric accuracy of WFPC2 and one to help expand the PSF library holdings. The Photometric Characterization (7628 ) repeated some of the previous cycle's NGC 2100 imaging, to check for temporal changes in the zeropoints, and included new observations of NGC2419 to expand the coverage to bright red giants. In addition, the usual filter sweep was performed using the primary white dwarf standard GRW+70d5824 and the field standards in ω Cen. The second special calibration program was the CTE Characterization (7630 ); these data were taken to provide a complete set of observations with which to explore the parameter space of the `long vs short' anomaly (stars appear fainter in shorter exposures) and refine the flux and background-level dependent aperture corrections. Finally, the PSF Characterization program (7629 ) obtained observations of ω Cen in the "wide UBVR" filters (F300W, F450W, F606W, F702W) and I filter (F814W); these filters are frequently used for their high throughput but are not as well characterized as the standard UBVR set (F336W, F439W, F555W, F675W).
Cycle 6 Carry-Overs and NIC3-Campaign WFPC2 Programs
We also take this opportunity to close out the remaining Cycle 6 programs that had not been completed at the time of the Cycle 6 Closure Report (ISR 98-01, Baggett et al.) and to include the WFPC2 calibration programs implemented as a result of the NIC3 observing campaign in 1998; the programs are summarized in Table 2. Results from those proposals are presented here, if possible, under their Cycle 7 counterpart proposals (e.g., Earth Flat results from Cycle 6 are included in the Earth Flat results from Cycle 7). The Cycle 6 carry-overs include the Visflat Monitor (6906 , results are included with 7623 Internal Flats), the Earth Flats (7909 , results included with 7625 Cycle 7 Earth Flats) program, the Photometric Transformation (6935 , results included with 7628 Photometric Characterization), the UV Throughput (6936), the PSF Characterization (6938 , results are included with 7629 Cycle 7 PSF Characterization, the Post-NIC3 Campaign Focus Check (7925 , results included with 7618 Photometric Monitoring), and the Long Decons (8049 , results included with 7619 Decontaminations).
2. Format of this document
Table 1 presents a summary of the Cycle 7 calibration proposals; Table 2 presents a summary of the Cycle 6 carry-overs and the WFPC2 programs run during the NIC3 campaign in 1998. The tables include proposal titles and numbers, frequency with which the program was executed, estimated resources (actual external orbits used included in brackets), any products of the analysis, accuracy of the results, and general notes. The remainder of this document consists of detailed descriptions of each calibration program in a standard format, designed for ease in viewing electronically as well as on paper. For each program, the left side of the page ("Plan") contains the original description of the planned observations, their purpose and expected results; the right side of the page ("Results") describes any modifications to the Plan, details on the execution, actual resources used (as obtained from PRESTO statistics), results achieved, a timeline of activity, and plans for any future continuation of the program. If the program is not complete, an estimate is given of the resources that will be necessary for completion. Finally, a detailed bibliography is provided, listing any new documents since the last closure report as well as pointers to items of general interest.
Table 1: Summary of Cycle 7 Calibration Programs.
|
7621 + |
|||||||
|
Placeholder; later withdrawn in favor of Cycle 8 program 8053. |
|||||||
|
In pogress. Flats for end of cycle; results included as part of 7625. |
|||||||
Orbits listed in Estimated Time column are orbits requested and used; [] marks actual number used when different from the initial estimate.
Table 2: Summary of Carry-Over Cycle 6 and WFPC2 NIC3 Campaign Programs.
|
Done. Included in results for 7623 (Internal Flats) |
|||||||
|
Done. Included in results for 7625 (Cycle 7 Earth Flats). |
|||||||
|
In progress. Included in results for 7628 (Photometric Characterization). |
|||||||
|
Done. Included in results for 7629 (Cycle 7 PSF Characterization). |
|||||||
|
Withdrawn in favor of 8054 (Cycle 7 LRF Calibration). |
|||||||
|
Program deferred to Cycle 8 ( 8053 ). |
|||||||
|
In progress. Included in results for 7627 (Cycle 7 Astrometric Monitor). |
|||||||
|
Done. Included in results for 7618 (Photometric Monitor). |
|||||||
|
Done. Included in results for 7619 (Decontamination proposal). |
|||||||
Proposal ID 7618: WFPC2 Cycle 7: Photometric Monitor
|
Regular external check of instrumental stability. Based on Cycle 6 program 6902. |
Three extra orbits were added to allow for special monitoring done during the NIC3 campaign in June 1998. |
||
|
Standard star GRW+70d5824 is observed before and after a decontamination (decon) using three different strategies: (1) F170W in all four chips to monitor contamination in the far UV; (2) F439W, F555W, F814W on the PC to monitor focus; (3) F160BW, F218W, F255W, F300W, F336W, F439W, F555W, F675W, F814W in a different chip each month. Observations are taken after each decon and before every other decon, resulting in 36 orbits for 24 decon cycles. |
No problems. Due to changing target visibility windows during the year, occasional exposures had to be trimmed to maintain single-orbit visits. |
||
|
Instrument Handbook, reports at monthly TIPS meetings, WWW (sensitivity trends); updates in UV sensitivity variation used in SYNPHOT. |
Photometric monitoring results (see Figure 1) were presented in the WFPC2 Handbook , at TIPS meetings, and on WWW . The data were also used in Longterm Photometric Stability study ISR 98-03 (Baggett & Gonzaga) and in focus monitoring (Lallo). |
||
|
Overall discrepancies between the results of this test need to be measured to better than 2% and are expected to be less than 1% rms. This has been the case in Cycles 4 through 6. The point of the test is to measure this variation. Focus measurements have an expected accuracy of 1.5 micron, and a goal of 1 micron; the uncertainty in the focus determination is dominated by external factors, such as OTA breathing. |
The standard star countrates have remained stable to about 1%. Longterm photometric monitoring (ISR 98-03, Baggett & Gonzaga) found that typical fluctuations are ~2% or less peak to peak over 4 years in filters longwards of and including F336W. At the same time, the UV throughput has gradually evolved, with post-decon countrates increasing in some filters (e.g., 12% in F160BW+PC and ~9% in F170W+PC), while decreasing in other filters (e.g., ~3% in F255W+PC; see Figure 2). In addition, contaminant growth rates have slowed slightly for some UV filters (e.g., ~1%/day to 0.5%/day in F160BW on PC). In general, data taken with filters redward of F336W do not require a contamination correction based on day since decon however, they may require a small zeropoint correction. The Observatory WWW pages present the focus monitoring results (see also Figure 4, below); ~1 micron RMS of secondary mirror motion is typically achieved (Lallo). |
||
|
Needs to be scheduled shortly before and after decontaminations (up to 5 days). |
|||
Figure 1. Photometric monitoring results for PC1 and WF3, from Feb. 1994 through Sep. 99, taken from WWW photometric monitoring memo (Gonzaga et al.,). Note the restoration to "normal" throughput in the UV after each decon as well as the gradual longterm trends.
Figure 2. Contamination-corrected normalized countrates for PC and WF2 as a function of Modified Julian Date (1994-1998), illustrating the longterm photometric changes in WFPC2 (figure from Longterm Photometric Stability study ISR 98-03 (Baggett & Gonzaga)). The cause of the discontinuity (near Feb 1995, MJD 49750) remains elusive.
Figure 3. History of the PC focus since Jan 1994. The focus values have been corrected for breathing (single-orbit timescale) effects using the Hershey model; superimposed on the data are the secondary mirror moves and the best fit exponential (figure taken from Observatory WWW pages (Lallo)).
Proposal ID 7619: WFPC2 Cycle 7 Decontamination (includes results from 8049 Long Decons during NIC3 Campaign )
|
UV blocking contaminants are removed, and hot pixels cured, by warming the CCDs to +20C for six hours. |
|||
|
The decontamination (decon) itself is implemented via the DECON mode, in which the TECs are turned off and the CCD and heatpipe heaters are turned on to warm the detectors and window surfaces. Keeping WFPC2 warm for ~6 hours has been shown in previous Cycles to be sufficient to remove the contaminants and anneal many hot pixels.; continuation of 6-hour decons is anticipated for Cycle 7. The observations taken before and after each decon consist of: 4 biases (2 at each gain), 4 INTFLATs (2 at each gain), 2 K-spots (both gain 15, one short and one long, optimized for PC and WF), and 5 darks (gain 7, clocks off). To minimize time-dependent effects, each set of internals will be grouped within 2 days and performed no more than 1 day before the decon and no later that 12 hours after the decon. To prevent irretrievable loss of the critical pre-decon hot pixel status information, the darks will be executed as a non-interruptible sequence at least 30 minutes after any other WFPC2 activity. |
Routine decons were successfully performed once every ~28 days. Due to SMOV in 1997 and the NIC3 campaign in 1998, not all requested internal orbits were needed. In June 1998, the secondary mirror was moved to bring NIC3 into focus, a NIC3 Observing "Campaign". The mirror move of course resulted in moving WFPC2 out of focus, so no WFPC2 science programs were scheduled during the campaign. This opportunity was used to test whether very long decons (four decons, each 24 hours at +22C, proposal 8049) would reduce the CTE effects - it did not (Casertano, priv.comm.; also see results from CTE monitor 7929 , which was run a few weeks after the NIC3 campaign's end). |
||
|
Those obtained from use of darks, biases and other intenals (see Proposals 7620 and 7622). |
Dark reference files for pipeline (Wiggs et al.); hotpixel lists for WWW (Wiggs & Casertano); updates to WFPC2 History File on WWW (Baggett & Wiggs). |
||
|
Objective is to insure that the contaminants are periodically removed from the CCD window. Biases, darks and other internals taken with this proposal are used in generating reference files (see Proposals 7620 and 7622). |
N/A. Periodic removal of contaminants and annealing of hot pixels was successful. Table 3 below lists all decon procedures done to date. Figure 1 in Photometric Monitoring proposal (7618) section illustrates the restoration to nominal throughput after each decon. |
||
|
Requires scheduling at 4 week intervals. Dark frames taken before decons must be protected from residual images. |
|||
Table 3. WFPC2 Decontamination Dates and Parameters, taken from the WFPC2 History Memo on the WWW (maintained by Baggett and Wiggs). Column labeled `t' is length of time chips are kept warm; time given reflects the cooldown start time, cameras are ready for science ~3.5 hours later.
Proposal ID 7620: WFPC2 Cycle 7 Standard Darks
|
Measure dark current on individual pixels and identify hot pixels at frequent intervals. |
More internal orbits were required than initially requested, due to the extension of Cycle 7. |
||
|
Every week, five 1800s exposures are taken with the shutter closed. The length of the exposures is chosen to fit within an occultation period. The weekly frequency is required because of the high formation rate of new hot pixels (several tens per CCD per day). Five darks a week are required for cosmic ray rejection, to counterbalance losses due to residual images, and to improve the noise of individual measurements. Even with these measures, some weeks no usable darks will be available because of residual images. Normally this results only in a longer-than-usual gap in the hot pixel lists, but in a decontamination week, information on pixels that became hot and then annealed would be lost irretrievably. For this reason, pre-decon darks are to be executed NON-INT and at least 30 minutes after any WFPC2 activity (see Proposal 7619). Normal darks do not need to be protected in this fashion. |
No problems, proposal executed as planned. Note that additional information on hot pixels can be obtained from the dark frames taken via the S upplemental Darks program (7621, 7712, 7713) . The supplemental darks are archived only and not used to generate dark reference files. |
||
|
Weekly dark frames delivered to CDBS and monthly tables of hot pixels on the Web. |
Reference files delivered to CDBS roughly every week (Wiggs et al.), accessible via Starview or WWW Reference File listing . WWW hotpixel lists are maintained as well. (Wiggs et al.). |
||
|
The required accuracy for darks is about 1 e-/hour (single-pixel rms) for the vast majority of science applications. The expected accuracy in a typical superdark is 0.7 e-/hour for normal pixels. The need for regular dark frames is driven by systematic effects, such as dark glow (a spatially and temporally variable component of dark signal) and hot pixels, which cause errors that may exceed these limits significantly. |
The typical superdark accuracy is ~0.7 e-/hour. The low-level dark current has been steadily increasing: over ~5 years, the dark current in the chip centers has increased by a factor of ~2.2 in the WF chips and ~1.3 in PC (TIR 98-03; Baggett et al.). The increase is included in the reference files; the dark current remains a small portion of the total image noise. An additional change noted is an increase in the number of permanent hot pixels (Casertano, priv.comm.). Over the two years since the previous super-dark, the number of permanent hotpixels has increased by a factor ~2.5 at all intensity levels - though these permanently hot pixels still represent only a very small fraction (~0.2%) of the total pixels. These permanent hot pixels are likely caused by radiation damage to the detector pixels. |
||
Proposal ID 7621 , 7712 , 7713: WFPC2 Cycle 7 Supplemental Darks
|
This program is designed to provide up to three short (1000s) darks per day, to be used primarily for the identification of hot pixels. Shorter darks are used so that observations can fit into almost any occultation period, making automatic scheduling feasible. Supplemental darks will be taken at low priority, and only when there is no other requirement for that specific occultation period. This program is complementary with 7620, Standard Darks, whose longer individual observations are better suited to produce high-quality pipeline darks and superdarks, and are also carried out at higher priority. Note that hot pixels are often a cause of concern for relatively short science programs, since they can mimic or mask key features of the observations, and about 400 new hot pixels per CCD are formed between executions of the Standard Darks program (7620). These observations will be made available as a service to the GO community, and there is no plan to use them in our standard analysis and products. This program is only feasible starting in Cycle 7, thanks to the Solid State Recorder. |
No problems. These darks are rated as low priority, to be taken on a non-interference basis as time allows. However, generally the schedule allowed for all 3 supplemental darks to be taken each day. |
||
|
2016 internal orbits (occultation periods) maximum, depending on availability of suitable opportunities; program runs in no-interference, low-priority mode. The proposal will be designed to obtain up to 3 darks per day over Cycle 7, provided there are no other scheduling conflicts. |
|||
|
There is no plan to use these darks in the standard WFPC2 analysis tasks or products. Data are taken and archived, available to the user community via Starview . |
|||
|
Must be designed to allow automatic scheduling at low priority. Three program IDs required due to large number of visits (one dark per visit). |
|||
Proposal ID 7622: WFPC2 Cycle 7 Internal Monitor
|
Verification of short-term instrument stability for both gain settings. |
|||
|
The internal observations will consist of 8 biases (4 at each gain) and 4 INTFLATs (2 at each gain). The entire set should be run once per week, except for decon weeks, on a non-interference basis. This proposal is similar to the Cycle 6 Internal Monitor (6905). |
Due to SMOV in 1997 and the NIC3 campaign in 1998, not all requested internal orbits were needed. |
||
|
Superbiases delivered yearly to CDBS; TIPS reports on possible buildup of contaminants on the CCD windows (worms) as well as gain ratio stability, based on INTFLATs. A Technical Instrument Report will be issued if significant changes occur. |
Results were presented at TIPS meetings. Updated bias files have been installed in CDBS (Gonzaga et al.), accessible via Starview or WWW Reference File listing . No significant differences were found between the old and new biases. |
||
|
Approximately 120 bias frames will be used for each superbias pipeline reference file, generated once per year; accuracy is required to be better than 1.5 e-/pix, expected to be 0.8 e-/pix. |
Sets of 120 bias frames are used to generate reference files; the resulting accuracy is ~0.5 e-/pixel ( Properties of WFPC2 Biases, (O'Dea et al., ISR 97-04). Statistics of recent biases are given in the table below. |
||
Table 4. Statistics (in DN) of recent superbias files; pedigree column lists epoch of bias frames used in the superbias.
Proposal ID 7623: WFPC2 Cycle 7 Internal Flats
|
Monitor the pixel to pixel flatfield response and the VISFLAT lamp degradation as well as detect any possible changes due to contamination. This program is a combination and continuation of the Cycle 6 VISFLAT and INTFLAT Monitor proposals (6906, 6907, respectively). |
|||
|
This proposal contains an INTFLAT filter sweep, a VISFLAT mini-sweep, linearity tests, and monitoring images. Monitoring is carried out by taking INTFLATs with the photometric filter set after each decon. The VISFLAT mini-sweeps (before and after decon, twice during the cycle) will include the photometric filter set at gain 7, plus the linear ramp filter FR533N at both gains to test the camera linearity. The INTFLAT sweep, taken within a two-week period, includes almost all filters, some with both blades and gains. The linearity test will be done at both gains and blades using F555W, and an additional set with one blade and gain with clocks on. |
Due to SMOV in 1997 and the NIC3 campaign in 1998, not all internal orbits were needed. |
||
|
TIPS report, Technical Instrument Report if any significant variations are observed. |
Results have been presented at TIPS meetings as well as in ISR 99-01 Internal Flatfield Monitoring (O'Dea, Mutchler, & Wiggs) and in TIR 98-02 Analysis of Excess Charge in WFPC2 Overscans (Mutchler, O'Dea, & Biretta). |
||
|
Assuming Cycle 7 results will be similar to those from previous cycles, the VISFLATs should be stable to better than 1%, both in overall level and spatial variations (after correcting for lamp degradation), and contamination effects should be < 1%. For the INTFLATs, the signal-to-noise ratio per pixel is estimated to be similar to the VISFLATs, but the spatial variations in the illumination pattern are much larger. However, the INTFLATs will provide a baseline comparison of INTFLAT vs VISFLAT, in the event of a complete failure of the CAL channel system. Temporal variations in the flatfields can be monitored at the 1% level. Gain ratios should be stable to better than 0.1%. |
A detailed comparison of VISFLATs and INTFLATs taken from 1994 to 1998 has been presented in ISR 99-01 Internal Flatfield Monitoring II. Stability of the Lamps, Flats and Gains (O'Dea, Mutchler, & Wiggs). The analysis showed that 1) the VISFLAT lamp continues to degrade (see Figure 4 below), 2) the INTFLAT lamps are quite stable, having increased in output by only 1.3% in 4 years, 3) the relative gain ratios are stable in WF2 and WF4, although there is a steady 0.3 % drift in the WF3 gain ratio, and a jump in the PC ratio of 0.7% during the servicing mission, 4) the pixel-to-pixel response of the chips is very stable, 5) there are a few small variations in the VISFLAT fields (donuts and blemishes) due to dust on optical surfaces either moving or appearing. The effects of contamination (location of "worms") are stable with time. The VISFLAT data were also used in an effort to further understand charge trapping effects in the WFPC2 CCDs (TIR 98-02, Mutchler, O'Dea, & Biretta). A significant signal above the nominal bias level was found in the VISFLAT overscan columns, a signal which decays to zero over several pixels. The amount of overscan charge appears independent of image counts (at least at high count levels), which is consistent with it being due to CTE traps in the serial register though other causes can not be ruled out. The amount of excess charge has also been increasing slowly with time (doubling since installation of WFPC2, see Figure 5), possibly due to long-term radiation damage in the CCD . |
||
|
The observations, especially VISFLATs, should be scheduled into a minimum of lamp cycles in order to help preserve the lamp's lifetime. To minimize filter wheel usage during the INTFLAT portion of the proposal, each filter will be cycled through the two shutters and two gains. Each visit will contain just a few filters, in order to allow flexibility in scheduling; however, the INTFLAT sweep should be completed over a time span of no more than two weeks. |
Overall, the results from the VISFLATs and INTFLATs are in good agreement. If the VISFLAT lamps become unusable, the INTFLAT lamps would provide useful diagnostics of all properties except possibly for the stability of the flat fields on large scales which is affected by the variations in the Carley bulbs. |
||
Figure 4. VISFLAT lamp intensity as a function of cumulative lamp cycles (from ISR 99-01 (O'Dea et al.) )
Figure 5. Excess overscan flux as a percentage of the incident flux for all four WFPC2 CCDs. Data from both 1994 (solid line) and 1998 (dashed line) are plotted to illustrate that the amount of deferred charge found in the WFPC2 overscans is increasing over time (from TIR 98-02, Mutchler, O'Dea, & Biretta; a paper copy of the report is available upon request to help@stsci.edu).
Proposal ID 7624: WFPC2 Cycle 7 UV Flat Field Monitor
Figure 6. UV flat field statistics from April 1994 to June 1999, in four filters for the four cameras (Baggett). Countrates are averages of the central 300x300 pixels, normalized to the October 1994 set of UV flats (F185W images are normalized to Aug 1995 data due to lack of F185W data in Oct 1994). UV flats taken more than 10 days after a decon procedure are not included.
