[Skip to Content]


General Links
Professional Links
Australian Human Resources Institute Diversity Awards 2013 Finalist

UCLES ThAr Atlas with the MITLL3 and EEV2 2K x 4K CCDs

NOTE: A recent article by Murphy et al. (astro-ph/0703623) contains a much more accurate ThAr line list than was available when this atlas was originally prepared, and UCLES users are encouraged to consult these documents when wavelength precision is paramount.

With the advent of larger-format array detectors at the AAO (e.g., the MIT/LL and EEV 4K x 2K CCDs), the process of arc line identification and spectrograph setup has in some ways become more complicated than before, owing to the bewildering number of arc lines and spectral orders captured in one image. Although a variety of arc line atlases are available for the ThAr lamp (e.g., D'Odorico et al. 1987, Bessell & Pettini 1991, Meurer et al. 1991, and Lipman et al. 1993), most of them are available in hardcopy only, every arc lamp produces slightly different line ratios, and even the line ratios from any given lamp will change with time. Indeed, the lamp used to produce this atlas is no longer in use, so the line intensities you observe will be different again! This on-line atlas is not intended to replace these useful references, but rather to serve as a complement, particularly when using UCLES with the MITLL and EEV CCDs.

The data was obtained on 1999 October 22, using the MITLL2 CCD. This detector suffered a catastrophic failure in January 2000, but was re-born as the MITLL2A detector in March 2000, by switching to the opposite readout amplifier. The only significant difference between MITLL2 and MITLL2A images is that they are reversed laterally relative to one another. The detector was binned x2 in the spatial direction, and 2 settings of the 31.6 grooves/mm grating were used: one centred on 3946.0 Å covering 3522-4653 Å (orders 160-123), and the other on 6182.6 Å covering 4824-9386 Å (orders 61-117). Note that the detector was not centred on the peak of any one order, but since the extent of each order recorded in one MITLL/EEV CCD image extends way beyond the free spectral range, there is still almost complete wavelength coverage throughout. The exposure time in the blue was 10 seconds, and 2 seconds in the red (which unavoidably results in some extremely saturated lines beyond 7000 Å). The line identification and fitting was carried out using ecidentify within IRAF, relying mainly on the line identifications in Lipman et al. (1993) and the IRAF line list in linelists$thar.dat. Generally a 4th order fit in x should be sufficient to remove any trend in the residuals with pixel number, and a 5th order fit in y should remove any trend with order number.

The arc atlas is presented as a series of Encapsulated Postscript images. To view them, just click on the order number (the wavelength ranges shown are those covered by the detector, which is generally much greater than the FSR listed in Appendix A.2 of the UCLES manual). If this does not launch a Postscript viewer such as ghostview, then check that the application is set properly (for Mozilla, go to Edit -> Preferences -> Helper Applications and set "application/postscript" to "ghostview %s"). The plots are small, but you can change the "Magstep" setting, or use the Zoom (middle mouse) button to enlarge them. Alternatively, a complete hardcopy of the atlas is available in gzipped Postscript format.

The image on the right shows the gaia/ds9 display for the MITLL3 CCD when the grating is set to 6182.6 A. The MITLL2A display has an identical vertical orientation to MITLL3, but appears flipped left to right. In this case, wavelength increases (order number decreases) going from left to right, and redder wavelengths are at the top of each order. Note that this is the reverse of the Tek and the EEV2 CCD orientation, which has redder orders to the left, and redder wavelengths at the bottom. Thus, it is not necessary to flip MITLL2A/MITLL3 CCD images about the x-axis prior to reduction (as would usually be done with Tek or EEV CCD images) if you wish to have wavelength increasing with pixel number along each order. Note that the EEV pixels are 10% smaller than the MITLL pixels, and hence cover 10% less of the echellogram.

The leftmost order which does not run off the edge of this image is order 117; the rightmost complete order is 61. The most heavily saturated line (which has another saturated line just above it, and three more in a right-angle triangle just below and to the left) is at 8115 Å in order 70.

MITLL3 6182.6A

Red Arc

The image on the right shows the gaia/ds9 display for the MITLL3 CCD, when the grating is set to 3946.0 Å. Once again, wavelength increases (order number decreases) going from left to right, and redder wavelengths are at the top of each order. The leftmost order which does not run off the edge of this image is order 160; the rightmost complete order is 123.

MITLL3 3946.0 A
order 123 (4581-4653 Å)
order 124 (4544-4616 Å)
order 125 (4508-4578 Å)
order 126 (4472-4542 Å)
order 127 (4437-4507 Å)
order 128 (4402-4472 Å)
order 129 (4368-4438 Å)
order 130 (4335-4402 Å)
order 131 (4302-4369 Å)
order 132 (4269-4336 Å)
order 133 (4237-4304 Å)
order 134 (4206-4273 Å)
order 135 (4175-4240 Å)
order 136 (4144-4209 Å)
order 137 (4114-4179 Å)
order 138 (4084-4149 Å)
order 139 (4055-4120 Å)
order 140 (4026-4088 Å)
order 141 (3997-4059 Å)
order 142 (3969-4031 Å)
order 143 (3941-4003 Å)
order 144 (3914-3976 Å)
order 145 (3887-3949 Å)
order 146 (3861-3921 Å)
order 147 (3834-3894 Å)
order 148 (3808-3868 Å)
order 149 (3783-3843 Å)
order 150 (3758-3818 Å)
order 151 (3733-3793 Å)
order 152 (3708-3768 Å)
order 153 (3684-3741 Å)
order 154 (3660-3717 Å)
order 155 (3636-3693 Å)
order 156 (3613-3670 Å)
order 157 (3590-3647 Å)
order 158 (3567-3624 Å)
order 159 (3544-3599 Å)
order 160 (3522-3577 Å)

Blue Arc


Stuart Ryder, sdr -@- aao.gov.au