CCD Mosaic Imager Status

Since the last Newsletter, the CCD Mosaic Imager has been tested twice at the 0.9-m telescope. Because we are still developing the controllers and the associated software, only half of the full array of eight 2K x 4K CCDs was read out during these runs. Nevertheless, we were able to complete numerous on- telescope tests, and we summarize the principal accomplishments of the 0.9-m tests and subsequent work below.

1. The new 0.9-m corrector for the Mosaic is performing as expected over the field that we could image (half the detector), including the TV guide fields at the edge of the CCD Mosaic. We saw no evidence for a tilt across the large field.

2. The dual TV guide systems work extremely well, seeing down to about 17th magnitude in nearly full moon. Each TV views about 5.3' x 4.9' on the sky at the 0.9-m and the image quality is more than adequate to achieve excellent guiding.

3. The 14 position filter track worked flawlessly and repositioned as required to minimize "dust donut" residuals.

4. We have constructed a temporary R-band filter (5.8 inches square) in-house and we should have a commercially purchased set of BVR filters by the time you read this. We tested filter inserts that allow use of our 4 inch square filters, although these illuminate only 6K x 6K of the Mosaic at the 0.9-m (and probably 5K x 5K at the 4-m).

5. During the first run, the large focal plane shutter exhibited an offset of 0.07s from the commanded time. This was traced to a pinched air line (the shutter is pneumatically powered). The defective line has been replaced and modified to avoid another occurrence, but we have not had another on-telescope opportunity to verify that the problem was eliminated. Lab tests, though, show the timing errors to be negligible now.

6. Software control and interaction between the data acquisition system and the telescope has been verified. That is, filters, focus, and short telescope offsets can be commanded from the acquisition computer, and proper telescope information is being collected for use in the CCD headers.

7. Dewar hold time is acceptable at 17 hours.

In addition, work has progressed on the new 4-m corrector and its atmospheric dispersion compensator. The four large fused silica elements of the corrector have been fabricated in-house and multi-layer coated by Continental Optical. The coatings have been verified to perform very well across the entire optical band. From below 3500 Å to longward of 9500 Å , the combined losses on all four elements (eight surfaces) is never worse than 8%, and is typically about 6%. By comparison, a single-layer magnesium fluoride coating optimized for 5000 Å (a good compromise wavelength) would suffer losses of 15-20% over this range. Uncoated optics would approach 25% losses.

The 4-m Atmospheric Dispersion Compensator is expected to yield similar performance over its four surfaces (that is, an additional 3-4% loss). Fabrication of these elements and their coatings should be completed by the time you read this.

To assist prospective proposers in writing observing proposals, we present below the expected performance characteristics of the CCD Mosaic Imager. The Mosaic can be mounted at either the 4-m prime focus or the 0.9-m f/7.5 Cassegrain focus. Pixel scales and fields of view are compared with the current CCDs in use at the two foci:

                                   4-m PF       0.9-m f/7.5
				Mosaic  T2KB    Mosaic  T2KA
Pixel scale (arcsec/pixel)      0.26    0.47    0.43    0.68

Field of View (arcmin on edge)    36      16      59      23

The format of the Mosaic is 8192 x 8192 with 15 m pixels. It is constructed of eight separate CCDs, each 2048 x 4096. Gaps between CCDs are less than 1 mm (67 pixels). Multiple shifted exposures will be required to fill these in.

The filter track holds 14 filters, each 5.8 inches square. By next spring we expect to have standard B, V, R, and I broad-band filters and two narrow-band filters, probably Ha and redshifted H, each with a 70 Å bandpass. We have not yet finalized parameters for the two narrow-band filters. Filters will all be 12 mm thick to preserve the focus. TV guide fields are located north and south of the science field and are viewed through separate broad-band filters.

Each CCD is read out through a single amplifier. All the data streams are multiplexed back to the control computer where they are assembled into a single image. The entire mosaic can be read out in approximately 100 s. On-chip binning of pixels is supported; this will speed up readout accordingly.

Although the CCDs have not yet been fully characterized, we expect that the typical read noise will be about 8 . However, several of the CCDs show peculiar charge injection problems that raise the effective noise over parts of the chip to levels as high as 30 . In addition, small areas (< 15%) of two of the chips are compromised by significant traps. There are several bright or dark columns on each CCD; we expect that normal data processing will eliminate the effects of these. Because of a processing error in the CCD fabrication that results in poor charge transfer efficiency when the chips are very cold, they are operated at the unusually high temperature of -60C. This results in a dark signal of around 100-150 pixel hour.

The chips are thick and front-illuminated. They are not coated with any kind of fluorescent coating, so the quantum efficiency should be typical of front-illuminated devices, peaking at 50% at about 7000 Å and falling to less than 10% below 4500 Å and above 9000 Å. We have not yet measured system throughputs; users might expect the count rates at V, R, and I to be half of those measured with the thinned Tek chips typically used at these two telescopes, and significantly worse than that in the blue. Consequently, the current system is not recommended for use below 4500 . Although these CCDs will be in use during the spring 1997 semester, NOAO is working to replace these devices with thinned science-grade CCDs.

At both the 4-m and 0.9-m, new correctors have been designed and constructed to support the field of view and image quality requirements of the Mosaic imager, as mentioned above. The new 4-m corrector includes an atmospheric dispersion compensator, and it will be used with all CCD imaging at that focus. The 0.9-m corrector will be used with only the Mosaic, as it is incompatible with the gold guider. Image quality is expected to be excellent with both systems.

We will offer the Mosaic to users as soon as the system has been tested and characterized. At this writing, there are sufficient uncertainties in the Mosaic progress, that we cannot guarantee that we will offer the instrument during the spring 1997 semester. The primary subsystems undergoing major work are: the CCD controllers, CCD optimization, the 4-m corrector and its atmospheric dispersion corrector, and the user software. At this point, it seems unlikely that the Mosaic will be ready for users by February 1997, but use in May or June appears plausible. Check our Mosaic Web page at http://www.noao.edu/kpno/mosaic/mosaic.html for recent updates and revised status reports.

We have come up with the following plan for users interested in using the Mosaic during the spring semester. We divide potential Mosaic users into two classes: those whose science can be done with the current CCDs at the 4-m and 0.9-m, but who would achieve further benefit from use of the Mosaic; and those who would not be able to achieve their science goals without Mosaic. Proposers should consider both the areal and sampling advantages provided by Mosaic, in contrast to the lower quantum efficiency (particularly in the blue), poorer cosmetics, and higher readout noise and dark current of the Mosaic CCDs compared to the T2KA and T2KB CCDs. Users can refer to the expected Mosaic performance given above; any of the undersigned are available for phone or e-mail consultation. With the current Mosaic CCDs, it is likely that only projects requiring large areal coverage and not using blue filters will benefit from using Mosaic instead of T2KA/B.

For the first group of proposers, those able to do their science with T2KA at the 0.9-m or T2KB at the 4- m but desiring an "upgrade" to Mosaic, the standard proposal form should be filled out. For instrument, specify T2KA or T2KB followed by "(Mosaic if available)." It would be beneficial to include some discussion in the technical justification section of the proposal on how the Mosaic would help achieve the project's goals. We will block together the runs that would benefit from Mosaic. The schedule will show T2KA or T2KB for these runs. If it becomes evident before a scheduled block that Mosaic can be used, we will contact the users and convert the block to Mosaic on the schedule.

For the second group of proposers, those wanting time only if Mosaic is available, we will use a methodology similar to that being employed for the Phoenix IR spectrograph. To save the effort of writing a full proposal before it is known whether Mosaic will be available, we request simply an expression of interest by the proposal deadline of 30 September. This will be done by e-mail to kpno@noao.edu. The subject line should read " Mosaic expression of interest." The message should include the name(s) and e-mail address(es) of the investigators, a brief program title, the telescope (4-m and/or 0.9-m), the desired number of nights, and the desired lunar phase. Based on these expressions of interest, we will set aside a number of nights on the 4-m and 0.9-m late in the spring semester. Later, we will request more detailed proposals from those who have expressed interest. The timing of the proposal request will be determined by progress on Mosaic commissioning coupled with a desire to give proposers adequate notice. The proposals will then be reviewed and the nights assigned.

We trust that users will agree that the capabilities of Mosaic are sufficiently interesting to justify this extra effort in the proposal process, in order to make the instrument available for science observing as soon as possible.

Taft Armandroff, Todd Boroson, George Jacoby, Rich Reed (for the Mosaic Team)