LIRC II User's Guide: oberving techniques

Lick Infrared Camera User's Guide

Observing Technqiues

Basic techniques for making calibration frames (darks, flat-fields, standard stars, and sky frames) and observing sources (source frames, nodding, dithering, and mosaicing, and guided vs. unguided exposures) are outlined below. The following notes are generally useful for any observing program.

Note that, due to controller electronics, actual exposures times are slightly longer than indicated by the data-taking system. Consult the users' logbook, the Lirctop log, or a support scientist, if you are not sure of the length of the delay.

A program's particular requirements will dictate the specifics of the observing strategy, but an observer will typically emerge from his or her observing run with some combination of source frames, sky frames, standard stars, darks, flats, and bias frames. See Reduction Techniques for suggestions on reducing LIRC-II data.

Near-IR data can be gathered using methods similar to those used with CCDs, as long as the background is sky-dominated rather than thermally-dominated. However, IR techniques differ from the visual in some important respects:

Exposures are generally sky background limited and relatively short. Observations usually consist of multiple exposures of the same target, which are later combined.
Because of the strong sky contribution, sky subtraction is of fundamental importance. Sky frames are generally produced at the same rate as target frames, and near them in space and time.
Near-infrared sky brightness can vary significantly on short time scales, requiring frequent sky frames.
Dithering for sky subtraction and to remove bad pixels is a common practice.
Observations are typically made without a shutter (LIRC-II has none), and sometimes without guiding.
Twilight and/or night sky exposures are customarily used for flat-fielding, but must be coadded and appropriately filtered to create useful flat-field frames.
Dark current and readnoise are significantly higher than with CCDs-- approximately 2e^- pixel^-1 second^-1 and 50e^- respectively, for LIRC-II.
Note that a minimum signal of about 1000 DN is required to ensure that the detector is not read-noise limited (ie. photon shot noise will be (1000 DN S 12e^-/DN)^1/2 or about 110e^- rms of shotnoise in the signal vs. approximately 50e^- readnoise).

Darks

The dark current in the NICMOS-3 array is small (about 2 e^- pixel^-1 sec^-1), but significant for some applications, and should be measured on a nightly basis so that it may be subtracted from the source and sky measurements. To reduce noise, a relatively large number (e.g. ten or more) of dark frames should be averaged together.

Darks are made by setting the filter and lens wheels to their dark positions, so that the array sees only the cold metal of the wheels. All three wheels can be set by selecting the `dark' command in the motor control program.

Ideally, each dark frame should be of the same duration as the source and sky measurements to which it will be applied, but may be scaled with caution-- remembering not to scale the bias. We suggest taking at least ten darks for each exposure time, and combining them with a median algorithm.

Darks can be taken in twilight or during the day, as long as the dome is closed and lights turned out. However, the dark current may change very slightly (<1%) during the course of a night. One can reduce this source of error by making two sets of dark frames, evening and morning, and applying them to the first and second halves of the night, respectively.

Lirctop includes a procedure, under the `calibrations' option, for automatically taking multiple dark exposures. The procedure takes up to ten sets of dark frames without intervention. Each set can have a different integration time and a different number of exposures.

`Zero-second' darks, or bias frames, may be used as a separate bias measurement for baseline subtraction.

Flat-Fields

Flat-field calibration frames are necessary to remove pixel-to-pixel variations across the LIRC-II array. The ideal flat field is one which is illuminated in the same way as, and close in color to, the image to which it will be applied.

Using the evening and morning twilight sky to take flat-field frames makes efficient use of the available observing time, and generally provides the best flatness for data reduction. The telescope is opened to the twilight sky and a series of exposures made through each filter that will be used for observing. To permit the removal of stars from flat-field frames, move the telescope ten or more arcseconds between exposures. Lirctop includes a procedure, under its `calibrations' option, which partly automates the taking of twilight flats.

The appropriate position of the TUB diagonal mirror while taking flats will depend on which TV camera you have chosen. If you anticipate using both TV cameras, complete sets of flats should be made at both mirror positions.

At K-band, the best flat field frame will be the difference of two twilight frames taken with the same integration time. Since the twighlight sky is changing, this removes thermal emission coming from the telescope, leaving the flat-field frame with only the far-field illumination.

However, it is not always possible to obtain twilight flats, or at least an adequate number of them. Twilight progresses rapidly, making them difficult to complete, especially if observing through several bandpasses or using both TV cameras. A series of relatively blank night-sky fields, slightly offset from one another, and later combined through a median filter or clipping algorithm to remove stars, provides a very good flat field.

If the source fields are relatively sparse, an average value of the sky, taken from several source frames, offset from one another, may themselves be used to create flat-field and sky-subtraction frames (Cowie, Gardner, Lilly, and McLean, 1990).

Standard Stars

Tables of photometric standard stars with their near IR magnitudes can be found in IR Standard Stars. Sources for these standards are the Kitt Peak National Observatory (KPNO) Infrared Observing Handbook (1987); Elias et al. (1982); and Zuckerman and Becklin (1987).

Both the KPNO and Elias standards are bright enough to saturate the detector relatively quickly unless they are trailed on a short exposure. For rough estimates of count rates, see Instrument Characteristics. Do not defocus stars to prevent satuaration; doing so will cause the beam to be vignetted, and the star's brightness to be underestimated. Trailing bright stars along the array will decrease the flux per pixel without affecting the total flux.

Make several exposures of each standard, moving the source to different parts of the detector. Zenith distance corrections should be made in the standard fashion. Remeber, actual exposures times are slightly longer than indicated by the data-taking system. Consult the users' logbook, the Lirctop log, or a support scientist, if you are not sure of the length of the delay.

Source Frames

Multiple exposures on a single source are the rule. Due to the brightness of the near-infrared sky, broad-band images quickly become background limited, and even observations made through narrow-band filters will become background limited within minutes. Moreover, the array's response becomes non-linear above about 20,000 DN (240,000 e^-). Thus, in the absence of a bright source, readnoise and the detector's linearity limit will combine to constrain the exposure to between about 1,000 and 15,000 DN, that is, well above the readnoise limit, and well below non-linearity.

Lirctop's `Observing' option includes procedures for automating observations.

Sky Frames

Infrared photometric measurements usually require a nodding or chopping technique which can result in half the observing time being spent on sky measurements. These sky frames not only measure the sky brightness, but, in some cases, may also be used for flat-fielding.

The infrared sky, at wavelengths shorter than 2.0 microns, is dominated by emission from the hydroxyl molecule OH. This emission is highly structured--spatially and spectrally--and exhibits intensity variations, in the worst case, of up to 50% on time scales of less than an hour (McCaughrean, 1988). Thus the sky brightness should be measured frequently, and should be recorded at the same bandpass as the source image to which it will be applied. Most observers choose to make skies at intervals of a few minutes.

It should be possible to find a reasonably blank patch of sky relatively near most sources. However, since there is a good chance of faint sources being present in any given piece of `blank' sky, we recommend offsetting by a few arcseconds between sky measurements and then median filtering or averaging the images with a sigma clipping algorithm, in order to remove faint sources. In some cases, the source frames can themselves serve as sky frames, provided the field is sufficiently sparse. Dithering is the prefered method when the source frames can also be used as sky frames.

Lirctop's `Observing' option includes procedures for nodding and other telescope movements.

Nodding, Dithering, and Mosaicing

Nodding refers to the technique of moving the telescope between the target and an adjacent position in the sky, making exposures at each position to create a matched set of source and sky frames. If the field is dense, or the target large compared to the size of the array, the nod may be relatively large. Sparse fields and small targets may only require small movements. In either case, the aim is to create a series of frames which can later be combined in such a way as to remove the sky contribution in the area of the target or targets.

Dithering is a more complex form of nodding. The telescope is moved several times, according to a pattern, and exposures are taken at each point in the pattern. Dithering typically consists of small moves which place the target at several positions on the array. Dithering can be used to make combined sky and target frames--provided the field is relatively sparse and the target small. Usually, frames taken before and after each source frame are combined using a median filter to create a sky frame that is then subtracted from the source frame. There are a number of different techniques commonly used and the best method depends on the source object. A set of images created by dithering can also later be combined to minimize the effects of bad pixels.

Mosaicing allows the imaging of regions of the sky larger than the array's field of view. The goal is to produce adjacent frames with enough overlap to allow them to be assembled into a single image. A variety of approaches are possible, but you must provide for later image registration, either by ensuring the presence of reference stars or by logging precise telescope positions. Sky-subtraction must also be taken into account, keeping in mind that the sky background varies temporally and spatially.

Lirctop's `Observing' option includes procedures for nodding, dithering, and mosiacing.

Guided vs. Unguided Exposures

The Nickel Telescope User's Manual includes a very complete discussion of the autoguider and its operation. At the 3-meter, operation of the autoguider is taken care of by the telescope operator.

However, the short exposures and frequent telescope moves that are typical for near infrared observations, prompt many observers to take at least some of their exposures unguided, relying on the telescope tracking to hold the image steady. Unguided exposures have the advantage of reducing the overhead required to reset the autoguider each time the telescope is moved, but can result in blurred or elongated images if the exposure is too long or the track rate not accurate. The best way to determine whether an observation can be safely made without guiding is to make a few unguided test exposures and carefully examine the shapes of the resulting images.

The nominal value for the 1-meter's right ascension track rate is -0.04. It is entered via the thumbwheel on the Telco panel. The track rate is adjusted according to input from the autoguider. It can be fine tuned by autoguiding for a few minutes on a star in the neighborhood of the target, prior to beginning a series of unguided exposures.