The dark recipe generates one single product:

The Aladdin array is operated in three different detector setups: uncorrelated read mode, double correlated read mode at high bias and at low bias.

LW-arm (Aladdin array) dark frame using Uncorr read mode.

LW-arm (Aladdin array) dark frame using DoubleCorr read mode.

As the calibration purpose, the recipe and the pipeline products are the same for both detectors:

Purpose: see reduction of SW-arm (Hawaii array) dark calibrations

Recipe: see reduction of SW-arm (Hawaii array) dark calibrations

LW-arm twilight flats

The twilight recipe is operated to generates three products:

LW-arm twilight flat product H-band	LW-arm twilight flat J+Block filter
LW-arm twilight flat K band	LW-arm twilight flat L-band
LW-arm twilight flat M_NB filter	LW-arm twilight flat NB_3.28 filter

Purpose: see data reduction of calibrations SW-arm.

Recipe. The ISAAC LW-arm can be operated in short wavelength bands (J, H, K) and long wavelength bands (L and M). In the short wavelength bands the telescope tracks and jitters around a position fixed on the sky. The data reduction is identical to the one used in the SW-arm.

For the long-wavelength bands L and M the sky brightness gradient with time during twilight is not large enough to sample a reasonable flux range within the raw frame stack. The telescope instead takes 15 jittered sky images, spread over three fixed zenith distances (5 at low, 5 at intermediate and 5 at high zenith distance). The spatial sky brightness gradient is used sample the required flux range.

The photometric zeropoint recipe generates two products:

The LW-arm photometric zeropoint calibration files are taken in jitter mode (J, H, K bands) or chopping mode (L and M). The data reduction in jitter mode is described in the SW-arm section.

	Combination of 4 chopped raw data cubes.
	Pipeline product frame showing for each of the four raw data cubes a combination of both chopping images.

Purpose. See SW-arm section.

Recipe. The ISAAC pipeline zeropoint recipe can handle all types of zeropoint raw frames: raw images generated by the jitter template (see the SW-arm section) and raw data cubes generated by the chopping template. In chopping mode (L-band and M-band) the recipe reads a stack of usually 4 chopped standard star images. Each of the raw input frames is a data cube, with two planes containing the two chop positions. The layers of the data cubes are subtracted from each other and summed up per raw frame. The zeropoint for each data cube is taken from average difference frame. The default command line parameters are used, meaning the target radius is 30 pixel (=4.4 arcsec) and the background is determined from a ring with 40 < r < 60 pixel (or 5.9 < r < 8.9 arcsec), using a image scale of 0.1478 arcsec/pixel (non-chopping mode) or 0.0709 arcsec/pixel (chopping mode). The recipe provides the median instrumental magnitude. The LW-arm zeropoints are corrected for detector non-linearity.

The data reduction of illumination calibrations is identical to that for the SW-arm. See there.

The detector linearity recipe generates four product images:

LW-arm Linearity A-product	LW-arm Linearity B-product
LW-arm Linearity C-product	LW-arm Linearity Q-product

Purpose: The series of lamp flat frames and dark frames covering a reasonable range of DIT and flat lamp flux are taken to monitor the linearity of the detector in all three read modes. The LW-arm zeropoints are corrected for non-linearity within the data reduction cascade.

Recipe: The recipe subtracts the darks frames from the lamp frames and estimates the the lamp stability. For each pixel a quadratic function is fit to the DIT versus lamp flux relation. The resulting coefficients per pixel are stored in the coefficient product frames.