What's New with PUEO?

Jean-Luc Beuzit

Canada-France-Hawaii Telescope
Waimea, Hawaii
Electronic-mail: beuzit@cfht.hawaii.edu

Abstract:

The paper presents the recent developments of PUEO, the Adaptive Optics Bonnette (AOB) - currently one of the most pupular among CFHT instruments, with emphasis on the new infrared camera KIR and the point spread function reconstruction software package.

Introduction

 PUEO, the CFHT Adaptive Optics Bonnette (AOB) has been on the sky for two years, totaling over 150 nights of scheduled observing time. PUEO is routinely producing diffraction-limited images throughout near infrared to wavelengths as short as the J band (1.25 $\mu {\rm m}$) under reasonable seeing conditions. A significant number of papers have already been published reporting astrophysical results obtained with the system. New developments are still taking place around PUEO in order to increase the scientific return of adaptive optics (AO) observations. In particular, new software tools have been developed, and are now available to PUEO users, to estimate the quality of the correction achieved by the AOB and to retrieve the point spread function (PSF). Since December 1997, a new 1K $\times$ 1K high resolution near-infrared camera, KIR, is available to take full advantage of the diffraction-limited images produced by PUEO.

AOB Performance

 The results of the AOB commissioning have been extensively described in a recent paper by Rigaut et al. (1998). Please refer to this paper for more details on the AOB performance.

PSF Reconstruction

  One of the unique features of PUEO is that, for most observations, the PSF can be accurately retrieved from the wave-front sensor data and deformable mirror commands recorded during each data acquisition. Tools have been developed, and are now available to observers, to provide an accurate estimate of the AO-corrected, long-exposure PSF in the direction of the guide source when this guide source is of magnitude 13 or brighter (Véran et al. 1997; Thomas et al. 1998). The estimation/reconstruction process takes place in two steps, partly on-line, partly off-line.

At the telescope

For each exposure taken with PUEO, the following quantities are computed in real-time and displayed by the PSF monitoring tool, which is part of the data acquisition system (Figure 1):

• The uncorrected free atmospheric seeing.
• The AO-corrected, long-exposure, atmospheric PSF, i.e. the PSF that would have been obtained with a perfect telescope and optical train and a turbulence following the Kolmogorov model at high spatial frequencies. It therefore includes all the atmospheric residual aberrations, but not the non-common path aberrations, which mainly occur in the imagery channel, past the AO beam-splitter.

 

Figure 1:   PSF monitoring tool window

The PSF monitoring tool will also compute and save the structure function of the atmospheric corrected phase. For instance if the exposure is named 123456o.fits, the related structure function is saved as dph123456o.fits. These dph files must be taken by the observers together with their images because the actual PSF computation is based on their content.

Off-line reduction

The actual PSF differs from the atmospheric PSF due to the following effects:

• Uncorrected non-common path aberrations
• High spatial frequency aberrations not following the Kolmogorov model such as mirror seeing or polishing defects.

Usually, these effects can be considered as quasi-static and thus can be calibrated using the image of a point source, taken sometimes during the night, e.g. a photometric standard (Véran et al. 1997). CFHT provides a software called DPH2PSF that allows to compute the actual PSF from the atmospheric PSF. Because this software requires at least one reduced image of a point source, this computation is performed off-line by the observers. Of course, more reference images are useful, keeping in mind that for optimal PSF reconstruction the reference sources must be bright, to allow a good correction and to obtain high signal-to-noise ratios on the images. They should also be observed as close as possible in time to the object for which the PSF is required, to account for temperature variations or flexures. Of course, the dph files related to the acquisition of each reference source are also needed. The error on the estimated PSF is of the order of a few percents on the low and medium frequencies of the modulation transfer function (MTF) increasing up to ten percents at the highest spatial frequencies (Véran et al. 1997).

The DPH2PSF sotware is available either as an IDL procedure or as a stand-alone C package, along with more detailed information on the PSF reconstruction with PUEO, from the CFHT web page here

Observing in the Infrared: The KIR Camera

 KIR is a high resolution 1024 $\times$ 1024 near-infrared imaging camera dedicated to PUEO which has been developed as a collaborative effort between Université de Montréal, CFHT and Observatoire Midi-Pyrenées (Doyon et al. 1998).

Camera design

KIR is based on the 1K $\times$ 1K Rockwell Science Center HAWAII (HgCdTe Astronomical Wide Area Infrared Imaging) focal plane array, sensitive to radiation from 0.7 to 2.5 $\mu {\rm m}$. The camera consists in an LN2 cryostat which contains the detector, the fixed 0.67:1.0 transfer optics, a F/20 cold stop and a filter wheel (16 positions). The camera plate scale is 0.035''/pixel, yielding a total field-of-view of $36'' \times 36''$. The standard I, J, H, K and K' broad-band filters are provided, as well as several narrow-band filters (Table 3). A pre-amplifier and a shutter are mounted externally to the dewar. The system is driven by an SDSU/Leach CCD controller which is the system commonly used at CFHT for all visible and infrared detectors. Observers are provided with a simple and intuitive graphical user interface (GUI), called DetI, incorporated into the CFHT/Pegasus observing environment, through which they will configure the camera, control the data acquisition, monitor the data storage and do some pre-processing. Commands can be issued either from the GUI form or at the command line level. The latter feature is particularly useful as it allows relatively complex sequences of observations, such as dithering and mosaicing, to be programmed and prepared in advance within Cshell scripts. Different acquisition modes are available such as: one image per file, a cube of a given number of frames with the same integration time per file, or an image resulting from a given number of co-added exposures (power of 2 in number). The first light took place in September 1997 and the first astronomical observations were carried out on December 6, 1997. Figure 2 is a picture of KIR on PUEO during the last observing run and Figure 3 gives the design of the cryostat. The main characteristics of the camera are summarized in Table 1.

 

Figure 2:   KIR on PUEO

 

Figure 3:   Schematic view of the KIR camera.

 

$$\times$$ Detector format

Detector material HgCdTe

Readout mode Direct readout (4 channels)

$$\mu {\rm m}$$ Spectral range

$$\mu {\rm m}$$ Pixel pitch

Filling factor 100%

Operating temperature 77 K

$$\mu {\rm m}$$ > 0.9

$$\sim$$ Readout noise

Dark current 0.15 e-/s

Full well capacity 100000 e-

Mean quantum efficiency (detector only) 65%

Conversion factor 3.65 e-/ADU

Minimum integration time 0.1 s

Maximum integration time 1 hour

Readout time (full frame) 9 s

Plate scale 0.0352 arcsec/pixel

$$\times$$ Total field of view

Orientation on the sky North up, East left within 2 degrees

Filters

KIR is equipped with one filter wheel designed to hold 16 1/2 inch filters. Tables 2 and 3 list the broad-band and narrow-band filters that will be available for general use with KIR from July 1998 onwards. At the moment, Redeye filters are still used for KIR. In addition, visitor's filters can also be installed in KIR, provided that they are made available to CFHT staff well in advance of the corresponding observing runs.

 

$$\mu {\rm m} \mu {\rm m}$$ Filter T (%)

I 0.834 0.194 > 60

J 1.246 0.163 85

H 1.632 0.296 85

K 2.198 0.336 90

K' 2.115 0.350 92

 

$$\mu {\rm m} \mu {\rm m}$$ Filter T (%)

HeI_1 1.083 0.010 > 60

$$\gamma$$ 1.094 0.010 > 60

J continuum 1.207 0.015 68

[OII] 1.237 0.012 > 60

$$\beta$$ 1.282 0.012 > 60

H continuum 1.570 0.020 75

[FeII] 1.644 0.016 70

[FeII] continuum 1.690 0.018 70

H2 (1-0) 2.122 0.021 65

$$\gamma$$ 2.166 0.022 70

H2 (2-1) 2.248 0.022 65

K continuum 2.260 0.060 80

CO (2-0) 2.296 0.023 75

Performances

Image quality

Image quality of the camera was quantified using an array of 9.3 $\mu {\rm m}$ pin-holes located at the nominal f/20 input focus. Pin-hole images at all wavelengths show a well defined and symmetric Airy pattern, as expected for diffraction-limited optics. Since the predictions are that the Strehl ratio should be in excess of 90% at all wavelengths (> 90% in K), this requires Strehl ratio measurements with an accuracy of a few percent. At the time of writing, data analysis was still underway and the preliminary results obtained so far allow us to conclude that the Strehl ratio is well in excess of 90% at K. Optical distortion was tested using an array of 400 pinholes, arranged in a 20 $\times$ 20 grid, generated with the CFHT LAMA machine. The mask was located at the nominal f/20 input focus. Distortion is found to be lower than 0.3% over the whole field-of-view.

Flexures

We measured flexures of less than 0.06'' when going from zenith to 60 degrees in any direction, or less than 0.015'' per 1 hour exposure. This level of flexure is characteristic of what has been measured for the AOB (differential flexures between science path and WFS path), and means that KIR has a negligible contribution to the overall flexure.

Photometry and background emission

Table 4 gives the zero points, the throughput (from the top of the atmosphere to the detector) and the background equivalent magnitudes for the standard J, H and K broad-band filters, as obtained from the observation of photometric standards during one photometric night in March 98. These measurements have been obtained with the Redeye filters which are slightly different from the new KIR filters. New measurements with the KIR filters will be provided as soon as they become available.

 

Filter CWL FWHM Zero point Throughput Background

$$\mu {\rm m} \mu {\rm m}$$ (magnitude) (%) (mag/arcsec2)

J 1.25 0.29 23.6 20 16.3

H 1.65 0.31 23.1 26 14.6

K 2.20 0.40 22.6 27 11.6

 

Figure 4:   Limiting magnitude per pixel for broad-band filters.

Limiting Magnitudes

Figures 4 and 5 illustrate the limiting magnitudes per pixel reached with KIR, respectively for broad-band and narrow-band imaging, as a function of the total integration time, for S/N = 5. For each filter, the resulting image was obtained by combining several individual frames. The elementary integration time (per frame) depends on the filter and was thus selected in order to be limited by the background noise whenever possible. Typical elementary integration times are 3 minutes in J, 2 minutes in H, 1 minute in K and K' and 10 min for the narrow-band filters. Figure 6 gives the KIR detectivity in 1 hour for diffraction-limited point sources.

 

Figure 5:   Limiting magnitude per pixel for narrow-band filters.

 

Figure 6:   Limiting magnitude for a diffraction-limited source.

The read-out noise, dark current and frame rate

 The data acquisition electronics has a system gain of 3.65 e-/ADU. The read-out noise in the correlated double sampling (CDS) mode is currently $\sim$ 20 e- and is pick-up noise limited. A multiple sampling read-out (MSR) mode has recently been implemented and preliminary results are very encouraging. As expected, the read-out noise was reduced by a factor of 2 when averaging 4 read-outs at the beginning and at the end of the integration. The dark current is 0.15 e-/s. The whole detector can be read in 3.4 seconds which yields to a total time of 9 seconds to acquire and store a CDS image with a negligible integration time. The on-sky efficiency is therefore not very good as shown in Figure 7. We hope to be able to upgrade KIR soon with a new and faster array controller which will reduce the overheads by a significant amount.

 

Figure 7:   KIR on-sky efficiency.

Observing in the Visible

 

The FOCAM camera

The existing FOCAM camera can be used as a visible imager for the AOB, with two different configurations: the ``large'' field-of-view, coarse sampling, or, when mounted with its focal enlarger, the small field-of-view, fine sampling. Table 5 summarizes the plate scales and fields-of-view for two CFHT CCDs. It is important to remember that the actual field-of-view, for which images are substantially improved, will be truly determined by the isoplanatic patch size. The overall throughput of PUEO has been measured to be 70% over most of the visible range, excluding the beam-splitter. Three beam-splitters are available: 50%-50%, 10%-85% and 85%-10% (percentage of light sent respectively to the wavefront sensor and to the science detector). For objects brighter than R=10 and under good seeing conditions, images of FWHM down to 0.06''-0.10'' may be obtained.

 

CCD Plate scale Field-of-view

(arcsec/pixel) (arcsec)

STIS2 0.061 126

Loral3 0.044 90

The OASIS integral field spectrograph

 An integral field spectrograph has been designed and built at the Observatoire de Lyon for PUEO. The Optically Adaptive System for Imaging Spectroscopy, or OASIS, will offer various spectrographic and imaging modes. The OASIS available modes as for mid-98 are:

• Imaging (0.2''/pixel)
• TIGER spectroscopy (spatial sampling with lenslet array)

Integral Field Spectroscopy is envisioned as the appropriate spectroscopic tool for adaptive optics observations, the basic argument being that before correction, the observer does not know where to point. For instance long-slit spectroscopy, with a slit of 0.4'' width would require a knowledge of the position of features of interests (emission knots, galaxy nuclei, etc.) to an accuracy better than 0.1''. This is not possible before AO correction. Integral field spectroscopy does not impose this requirement. A two dimensional field-of-view allows a $\pm 1''$ uncertainty on positioning on the area of interest. See Bacon et al. in these proceedings for more details.

Fabry-Perot etalon

Figure 8 shows the scanning Fabry-Perot setup in the AOB focal enlarger. The enlarger optics doubles the image scale from f/20 to f/40. All the optics is located in a tube screwed to the top of the enlarger structure. The Fabry-Perot etalon is mounted on the bottom surface of the enlarger, preceding FOCAM in the optical path. The Fabry-Perot etalon is located in a converging f/40 beam, which is not an ideal configuration, the finesse being degraded by the de-collimation. The F-P configuration is currently undergoing tests.

 

Figure 8:   Fabry-Perot etalon set-up on AOB.

Future Plans and Ideas

 KIR shall soon be upgraded with a faster array controller (Section 4.3.5) to reduce overheads and increase efficiency on the sky. We are also studying the possibility of implementing a moving mechanism close to the input focal plane to bring in coronographic masks, double imaging beam splitter and slits, in order to do low-resolution spectroscopy with grisms mounted on the filter wheel. See also the proposal by Rouan et al. (these proceedings) of a high sensitivity integral field spectroscopic mode (R $\sim$1500) for IR, using a Fabry-Perot and a cold grism.

Many people were involved in the developments described in this paper. I would like to particularly thank G. Barrick, F. Beigbeder, S. Brau-Nogue, J.-C. Cuillandre, R. Doyon, S. Isani, B. Magrath, D. Nadeau, B. Starr, J. Thomas, P. Vallée, J.-P. Véran and F. Rigaut.