GEPI

Primary Science Drivers

– Chemical labelling and Galactic substructure

It is now widely recognized that the different populations of Galaxies may present distinct signatures of chemical
composition, due to the details of the chemical evolution. The most well known is the distinction in [Mg/Fe] between
thick and thin disc, first highlighted by Fuhrmann [1]. Vanessa Hill introduced the concept of « chemical labelling » :
when one is able to determine abundance ratios with an error of 0.1 dex for the purpose of putting a chemical « label »
on a given stellar population. Even more ambitious is the task of « chemical tagging » [2] which aims at identifying the
members of an individual star-forming cloud which would require accuracies of the order of 0.01 dex.

One of the primary aims of GYES is to provide chemical abundances accurate to ≈0.1 dex for the different
components of the Galaxy : thin disc, thick disc, bulge, halo. The stellar clusters, both open and globular, are primary
targets for GYES.

These requirements set immediately the minimal requirements for the resolution of the spectrograph
and for the field of view of the instrument.
Experiments with both synthetic spectra and real data convinced us that the spectral resolution should not be less
than about 15 000. Multi-element chemical analysis at lower resolution have been conducted [3], [4] these were directed
towards metal-poor stars for which line blending is not a major problem. Below the resolution of R≈15 000 the
chemical analysis relies only on strong lines, which are mainly on the flat part of the curve of growth, whichever the
signal-to-noise ratio of the spectra. This implies that their sensitivity to abundance variations is small.

A minimal requirement on the field of view comes from the lower density environment considered : the halo. To derive this requirement we need to have a crude estimate of what can be the performances of a spectrograph at
the CFHT. A crude, but robust, estimate can be obtained by scaling the performances of Giraffe at the VLT [5], [6].

– Kinematic structure of the Galactic disc

From the study of the stars of the solar neighbourhood we know that there is a significant correlation between stellar
ages and their velocity dispersion. This points towards a scenario in which their Galactic orbits display significant
evolution with time. This implies that their birth places may be very different from their current location.
The complexity of the disc calls for a mapping of the full position and velocity, six dimensional, vector of a sample
of stars well beyond the solar neighbourhood. We should be able to detect stellar streams and trace their origin to
resonances in the interaction between stars and spiral arms or the Galactic bar. We can use sub-giant stars to trace
these streams at distances up to 3 kpc, provided that we can measure radial velocities accurate to 1 km s^-1 for stars
as faint as 16th magnitude.

Other Science Cases

– Massive stars

The hot stars of spectral type O and B are rare objects, however their high luminosity allows to observe them at very
large distances, thus they are important tracers of the young Galactic populations. Because of their short lifetimes
they are always found very near to their birth places, thus they allow to map the sites of star formation. They also
are ideal background sources for the study of the intervening Interstellar Medium. The class of Be stars (B stars with
emission lines) deserves a special mention. These stars are extremely rapid rotators, in fact often close to break-up
velocity, and as a consequence they produce an equatorial excretion disc, where the emission lines are formed. The
fraction of Be stars of a given population and its possible evolution with metallicity is an important diagnostic of the
rotational velocity distribution of massive stars, and, ultimately, on their formation mechanisms. Gaia will be capable
to detect Be stars with strong emission, but, due to its limited spectral range, many Be stars with weak emission will
go undetected. The optical spectra provided by GYES will allow to unambiguously discriminate between B and Be
stars.

– Pulsating stars

In the magnitude range 6 < V < 20 Gaia will detect about 190 000 pulsating stars showing a Period-Luminosity (PL)
relation (35 000 RR Lyrae, 5000 Cepheids and 150 000 δ-Scuti stars [7]).
Once the Gaia parallaxes will be available, these PL relations can be accurately calibrated, including also possible
effects of other quantities, such as metallicity. Therefore the goal of GYES is to obtain chemical compositions for these
stars. RR Lyrae are too faint to be an interesting target for GYES, it should be capable to observe all Cepheids in
the northern hemisphere provided by the Gaia catalogue (in good conditions : SNR better than 50 for a 1h integration
time), as suggested by the histogram of the apparent brightness of known Galactic Cepheids [8]. It should also be
capable to observe 22 500 δ-Scuti stars in good conditions (SNR better than 50). Observations of these pulsating stars
should be scheduled together with the other survey targets.

– Interstellar medium

In the course of the survey, it will be possible to observe also early type stars (earlier than A) which can be used as
background sources to map the Galactic ISM. The most numerous will be hot Horizontal Branch stars.
The warm and cold ISM can be traced by Ca II H and K and Na I D lines observed in absorption against these
background sources. If a spectral resolution as high as 50 000 is available, it may be possible to disentangle the
absorptions due to different ISM clouds, which differ by only a few km s^-1. At this high resolution it may also be
possible to use cooler stars as background sources, provided their radial velocity is high enough to disentangle them from the stellar absorptions. At lower resolutions, like 20 000, it will still be possible to obtain interesting column
densities of Ca II and Na I which may be used as diagnostics of reddening and detect high velocity clouds.