CFHT, Instruments, Spectroscopy, ESPaDOnS

ESPaDOnS: an Echelle SpectroPolarimetric Device
for the Observation of Stars at CFHT


NEW! Added June 26, 2007:

List of all questions available

What spectral resolution can I get?
The spectral resolution is about 80,000 for the spectroscopic mode 'star only', and 68,000 for the spectroscopic 'star + sky' mode and the polarimetric mode.

What is the relationship between the 2.6 km/s bin mentioned in the Exposure Time Calculator and a CCD pixel?
The 2.6 km/s bin corresponds to one CCD pixel. The spectra are actually recovered with spectral bins of 1.8 km/s (in other words, the spectral sampling is 1.8 km/s, or 0.6923 CCD pixel).

What is the wavelength coverage? Where are the gaps?
Each exposure covers the 40 orders, from 370nm to 1050nm, with 3 very small gaps: 922.4-923.4 nm, 960.8-963.6 nm, 1002.6-1007.4 nm.

What is the size of the aperture hole?
The aperture hole through which the light is gathered is 1.6 arcsec across.

What is the Stokes parameter that you call 'W'?
In this circular polarization mode, the top rhomb rotates at about 0.1Hz, eliminating the cross-talk from linear to circular polarization by averaging the linear polarization. This is useful if your target is highly polarized in linear polarization.

What is the wavelength at the center of each order?

61 372 47 482 34 666
60 377 46 492 33 686
59 384 45 503 32 708
58 390 44 515 31 730
57 397 43 527 30 755
56 404 42 539 29 781
55 412 41 552 28 809
54 419 40 566 27 839
53 427 39 581 26 871
52 435 38 596 25 906
51 444 37 612 24 944
50 453 36 629 23 985
49 462 35 647 22 1029
48 472        

What is field of view of the guiding camera?
The total field of view is about 100 arcsec. A subraster of 5-15 arcsec on a side is usually used to reduce the readout time and speed up the guiding corrections.

Is it possible to guide on my target?
Yes, and this what is done by default. The guiding algorithm takes into account that it can only use the ring of light outside the star hole to guide.

Is it possible to guide on a nearby star?
Yes. It is not possible to do this with ESPaDOnS right now (this option does not work at the moment) but it is possible to use the Cassegrain guiding instead.

What if it is not possible to guide on my target and there are no nearby stars to guide on?
It is possible to use the Cassegrain guiding camera instead.

What is the magnitude of the faintest target observed so far?
About magnitude 14.

How can I get an estimate of the seeing?
Put your star close to the aperture hole, but not in it. Make the guiding zone big enough to contain the star. Start the mode "Calculate guiding corrections" (without sending the corrections to the telescope): seeing and magnitude estimates will be given.

When should I use the Atmospheric Dispersion Corrector?
An Atmospheric Dispersion Corrector is an optical device usually comprising two thin prisms which can rotate to compensate for the elongation of a star image caused by the wavelength dependence of the refractive index of air. At high airmasses, it is very easily seen on the guider that the star image is not round and small. Using the ADC will make the star look round and small again. The ADC can be left in the beam at all times since it does not absorb a lot of light and does not introduce polarization. It should be noted that the telescope focus will change dramatically when the ADC is moved in or out of the beam; the star image will look like a big ugly donut. The telescope has to be refocussed if the ADC is moved in or out of the beam, by about 100 'old' focus units, or about 1.0mm (new units).

What is a Hartmann mask?
A Hartmann mask blocks one half of the beam at a time. In ESPaDOnS, the mask can block the UP part or the BOTTOM part of the beam of light. This is used to focus the spectrograph. If the spectrograph is not focussed, each position of the Hartmann mask will produce spectral lines at different locations on the detector. If the spectrograph is focussed, the spectral lines are superimposed. This is a figure to explain how the Hartmann mask is used.

ESPaDOnS has a fiber agitator; what's that?
Like Gecko, ESPaDOnS has a fiber agitator, which agitates (shakes) the fiber just before it enters the spectrograph. This device is used to remove modal noise present in optical fibers, and ensures a S/N ratio compatible with photon noise.

How many exposures are taken to measure each Stokes parameters?

Is it possible to take only 2 exposures to measure a Stokes parameter?
No. Four exposures are required.

How much time does it take to move the rhombs from one position to the next?
The rhombs are moved while the CCD is read out, and this is what takes the longest time (40 sec in Normal mode).

How much time does it take to reduce data in polarimetric mode?
After the geometric and wavelength calibrations, which take about 5 minutes total, 4 exposures taken in the polarimetric mode can be completely reduced in about 3 minutes.

What are the telescope limits?
The CFHT Observatory Manual presents the telescope's range of movement, although from experience, the following limits are more practical (especially for the hour angle):

Can I do CONTINUUM polarization with ESPaDOnS?

Generally speaking, ESPaDOnS is not a good instrument for continuum polarization. There are mainly 2 reasons:

  1. The entrance hole of the instrument is a 1.6 arcsec hole drilled in a metallic (reflective) surface. As any metallic surface, it produces continuum polarization. It has been shown that a non-polarized star well centered in the hole will not show continuum polarization (as expected), but if the guiding keeps the star at the edge of the hole, a few percents of continuum polarization are introduced (and there's no way to correct for that).

    This problem will not affect extended objects that overfill the aperture hole (because the whole circular edge will produce a polarization that will cancel itself by symmetry).

    One might also imagine that long exposures with random guiding errors will result in an averaged out continuum polarization, but this has not been tested.

  2. The amount of polarization is the difference in intensity of the 2 beams produced by the Wollaston prism and recorded simultaneously. Each beam is injected in a fiber (that then goes to the spectrograph), without the help of Fabry lenses. Therefore, seeing variations or guiding errors that make the image of the star wander in the aperture hole (and the rest of the optics) will make the image also wander at the entrance of the 2 fibers, but not necessarily by the same amount. One fiber might lose a bit more (or less) light than the other beam. This difference in intensity will be calculated as continuum polarization even if the target is not polarized.

    The exact extent of that effect has not been measured, but it might be minimal for extended objects.

For more details on spectropolarimeters and continuum polarization, you may consult Eversberg et al. (PASP, 1998, 110, 1356 - jump to page 1364), or Moffat & Eversberg (A&A, 2000, 358, 572). These papers present an instrumental scatter of about 1% due to the 2 effects mentioned above.

What is the stability of the spectrograph throughout a night?
As the night goes on, the spectra move with respect to the CCD (or the CCD moves with respect to the spectra, or both). When looking at the corrections provided by Libre-Esprit (position of telluric lines), the corrections are increasing towards negative values (for an initial value around 0.0). At the end of the night, the correction ends up around -0.3 km/s with respect to the beginning of the night. This is only a small fraction of a pixel, about 0.1 pixel (one CCD pixel = 2.6 km/s), but it is a clear and systematic effect.

Note that the latest version of Libre-Esprit incorporates analysis of telluric lines to provide a better wavelength calibration. Therefore, it should not be necessary to take calibration exposures during the night to track that effect.

Can I take daytime flats?
Observers usually take about 25 flats in the evening, and 25 more in the morning. If your project requires an exquisite flat field correction, you may take additional flats during the day. You may start those additional flats in the morning before you leave the summit, or in the afternoon from Hale Pohaku. It has been verified that if lights are turned ON in the dome, the flats are not affected (the polarimeter is light-tight and the dome lights are not very bright compared to the internal flat field lamps).

What is the maximum number of flats that Libre-Esprit can process?
Libre-Esprit can process up to 100 flats. If you absolutely need more than that, there is a (somewhat involved) work-around:

  1. Use Iraf or write a script/routine to average all the desired flats
  2. Modify one line in a Libre-Esprit file (the file that tells the program which file to use for the flat)
  3. Execute Libre-Esprit, with the option to NOT delete the calibration files but delete the previously reduced spectra (-r option)

What are the units in the reduced spectra?
The units of the polarization column of the reduced spectra depend on the use or not of the option -c:

Guiding (with either the ESPaDOnS guider or the Cass guider) does not keep the star centered in the hole, I'm losing flux, what can I do?
At some telescope positions (especially at high airmass or high Hour Angle), the telescope can start to "oscillate", usually in Declination. If too much flux is lost or the star really moves around too much, it is always possible to stop all guiding, and just use the Guider Control button "Center star" to adjust manually the position of the star, every 30-60 seconds.

Back to ESPaDOnS's page
This CFHT Web page is maintained by Nadine Manset (manset -=AT=-