Welcome at Euler! This page is made to help you with the observation of gravitational lenses for the COSMOGRAIL project, which is led by the Laboratory of Astrophysics of the École Polytechnique Fédérale de Lausanne (EPFL).
During your run at Euler, you will observe some gravitational lenses typically every two nights; these observations take one-quarter of a night. The dates of these “lens-nights” are planned well in advance will be sent to you, as well as the precise target lists, by e-mail. As multiple telescopes are involved in COSMOGRAIL, all observing lenses according to a best possible sampling under the given weather conditions and priorities, we cannot foresee far in advance which lenses will have to be observed in your nights. That’s why the target lists are prepared only relatively shortly before the observations.
About the project (You could skip this…)
In a few words, the COSMOGRAIL project consists of monitoring strong QSO lenses, to measure their “time delays”. We observe distant quasars, gravitationaly lensed by galaxies lying on the line of sight from us to the quasar; as a result of this configuration, we see multiple images (usually 2 or 4) of the same quasar. But the light travel times along the photon paths that form these multiple images are different ! By measuring these differences (aka time delays) between the individual images of the same quasar, we can learn about the absolute geometrical configuration of the gravitational lens. With a model of the mass distribution in the lensing galaxy, these time delays give direct access to the value of Hubble’s constant H0 — at relatively low cost, in a way that is fully independent from other techniques, and without a distance ladder or complicated astrophysics.
In the above light curves of the quad lens RXJ1131 obtained by Euler, the time delay between image D and the other 3 images can be seen by eye ! The light curve of D has been shifted by 90 days towards the left.
The instrument : ECAM
ECAM (or EulerCAM, also called “CAMERA” by the software) is Euler’s imager. It is an E2V 4k x 4k CCD, with a pixel scale of 0.215”, and a field of view of 15′ x 15′. ECAM has two different readout modes (specified by the field “ampname” of an OB) : a 4-quadrant readout called “ALL”, and a single-quadrant readout called “UL” (for “Upper Left”). The lenses are always observed in the UL mode. Hence I will need bias frames and flats also in UL.
We use three filters for our observation: “RG”, “VG” and “IC”.
ECAM’s images are written into /gls/data/raw/[yyyy-mm-dd]. To easily identify them, you could change into this date directory, and type (for instance) :
dfits ECAM.*.fits | fitsort OBJECT FILTER EXPTIME
DS9 and skycat are available to open them (use for instance skycat’s object picker (ctrl-p) to get an estimation of the FWHM of the stars), but please be careful to not modify any files in this directories, or litter them with test files.
There is no guiding available with ECAM, and we do not need guiding for our short (360 second) exposures.
If interested, you’ll find further information about this instrument in the new ECAM Manual (still in development), here !
are easy ! Import the OBs from the small ECAM lenses catalog (“/gls/catalogues/ECAMlenses/ECAM_lenses_UL.rdb, accessible form the epd”), according to the wishlist of targets that I will send you. There is alos two non-lensed quasars in this catalog, namely HE0331 and CTQ252. Once imported, you can still move the lens OBs around your own observations, to plan your night (buttons “swap”, “move”, “copy”, …) and to optimize airmasses. If you’re out of targets you could add one exposure for each targets and therefore help us to increase the photometric precision. To do so click on the OB in the edp and simply duplicate the last exposure
- Each “lens” consists of 5 x 360 s dithering + overheads, this sums up to about 35 minutes per target.
- The non-lensed quasar (CTQ252 and HE0331) consist of one exposure of 360s with the VG filter, one exposure of 240s in RG and one exposure of 240s in IC. In total, it should take about 15 minutes per target
- Number one observational constraint for our project : the seeing !. If the DIMM seeing (as shown on the ESO MeteoMonitor) is worse than 1.5 arcsec, do not observe lenses, but move them to the next night, or even the night thereafter (yes, this is an official agreement).
- In case the MeteoMonitor is not available, you could measure the FWHM of some stars using skycat on the ECAM images (pixel scale 0.215″). If the DIMM seeing is significantly above 1.5″ (or the FWHM of the stars on the actual images is above 2.0″, i.e. FWHM > 9 pixels), the images will be essentially useless.
- There is no need for a perfectly photometric sky. Light cirrus clouds are ok. But if it’s too cloudy, the same rule as for bad seeing applies : shift the lenses to the next night.
- Try to observe the lenses at best possible airmass, at least below 1.5, but not at zenith, as Euler sits on an alt-az mount. So keep elev < 80.0 deg (anyway this applies only to HE2149-274 and Q1355-2257, other lenses don’t come close to the zenith).
- You can start observing in nautical night (the planning if forseen to do so, same applies for the other Euler programs). If you end your night with a lens that is just rising, it is perfectly ok if the last exposure gets shifted out of the nautical night.
- Keep a moon separation of approximatively 30 degree for the full moon. You can go closer when there is less illumination. As rule of thumb : the sky background on the ECAM images should stay below 3000 ADU.
- Finding charts for all the lenses you will observe are available here. It happens regularly that a satellite crosses the field of view (space is full of junk!). Keep an eye on the Real Time Display view of the EDP (bottom right monitor) and if a trail crosses the lens, redo the affected exposure.
- In case of conflicts with other time-critical observations, you can shift a few lenses to the next night. Then, you will still have three nights to observe them.
Finally, to give you a higher flexibility with your other observations, you can, if you want and in case of relatively stable weather, directly shift by your own up to 2 lenses to an adjacent night. Let me know if you have larger clashes with time-critical observations, we can certainly arrange the observing schedule.
- Bias frames: EDP Insert -> CAM_BIAS, then switch the ampname to UL and duplicate the exposure to get 5 images. I need only one such set of 5 UL biases per night. You can take these bias frames for instance after the Coralie calibs, or after the evening flats, waiting for the night to fall.
- Flat fields are very important to us, as we perform differential photometry over sometimes large distances across the image. With ECAM, skyflats have to be taken daily, in all filters and readout modes used during the night. For the lenses, the only filter we use the filter RG, readout mode UL, and 3+ of such flats in the evening and 3+ in the morning would be optimal, at least for the nights containing lenses. For non-lensed quasar, we also use VG/UL and IC/UL in addition to RG/UL. You are always welcome to take flats on any other evenings or mornings of your run : we tend to build masterflats over periods of the order of one week, so any RG, VG, or IC flat of your run is useful, the more the better!
The “skyflat OB” corresponding to the current month is sent into the EDP from : Catalogs: -> Camera -> ECAMutils/flat.rdb. For each twilight, insert the corresponding OB. For each ampname, some desired filters/ampname combinations have now to be specified in the “flatlist” field, separated by spaces or commas. Each combination (like “RG/UL”) written into this flatlist corresponds to 3 exposures with this filter and ampname; the exposure times will be adapted automagically, always aiming at a flux of 20000 ADU. But, you have to respect a given ordering of the filters that you write into the flatlist (e.g., in the evening, we take flats starting with the darkest and ending with the most transparent filter). A list giving this order is displayed on the pinboard behind the printer.
If you lack inspiration, you could try this :
- Evening flats: UG/ALL,BG/ALL,VG/ALL, VG/UL, ZG/ALL,RG/ALL,RG/UL,IC/ALL, IC/UL (note that UG starts already just before sunset !)
- (ALL biases, UL biases, night, and again ALL biases just after the night)
- Morning flats: VG/UL, RG/UL, IC/UL + whatever filters were most used in the night
- Focus sequences are essential to get sharp images, even if Euler’s autofocus will try to compensate for temperature changes during the night. Always leave the autofocus on, even while taking the manual focus sequences! The latter will serve as zero point for the autofocus.
Focus sequences are inserted into the EDP using the menu Catalogs: -> Utils -> focus/focus_for_camera.rdb. Choose any star whose RA is relatively close to the sideral time (visualized at the very bottom of the EDP window, label “ST”), and thus high up in the sky.
Start the night with a focus sequence. This will not cost observing time, as you can launch these observations 7 minutes before the nautical night. There is nothing else to configure for these focus exposures, but the software will almost certainly ask you to pick the focus star by hand. Be ready to do so when observing a focus sequence :
- Minimize all the big windows of the lower right screen (catalogs, ECAM RTD) except for those with the fancy rainbow colors: the manual star selection will be done in the latter.
- At the beginning of the focus sequence, a dialog might appear in the lower left screen, inviting you to select the focus star, by clicking on it in the “Image Xaff” window on the lower right screen.
- If you cannot see stars in that window, click on “Cuts” and then “Redisplay”, or tweak the cuts by yourself. Then just click (left mouse button) on the brightest star of the field. It shouldn’t have a companion, and usually, it should be close to the center of the image.
- One after the other, 5 images of the same star will now appear in the Image Xaff window, with different focus positions. Image #3 should be clearly sharper than images #1 and #5 (both in terms of FWHM and visual aspect), at least in the second focus sequence you’ve inserted at the beginning of the night. The software will determine the optimal focus by a parabolic fit on these 5 measurements. Ideally the best FWHM (given in arcseconds) would be close to what the DIMM gives; however this will not be the case: currently, Euler is not seeing-limited.
Please insert further a focus sequences just before observing a bunch of lenses. No lenses should be observed if a focus hasn’t been performed in the last three hours.