DOI: 10.1051/0004-6361:20054102
Astronomy&
Astrophysics
Exploring the lower mass function in the young
open cluster IC 4665
W. J. de Wit, J. Bouvier, F. Palla, J.-C. Cuillandre, D. J. James,T.R.Kendall,
N. Lodieu, M. J. McCaughrean E. Moraux,S.Randich, and L. Testi
1. Introduction
Formation theories envisage brown dwarfs (BDs) evolving
from the gravitational collapse of the smallest unities in the
opacity limited fragmentation of giant molecular clouds like
stars, e.g. Shu et al. (1987). Otherwise they may form from the
condensation of solids within circumstellar (or circumbinary)
disks like planets (Lin et al. 1998; Papaloizou & Terquem 2001;
Jiang et al. 2004). Innovative ideas let BDs undergo a starlike
 gravitational collapse initially, but they terminate the subsequent
 mass accretion on the formed hydrostatic core. This
could be due to the removal of this stellar seed from the material
 reservoir after a dynamical interaction with one of its
siblings (Reipurth & Clarke 2001). Conversely the growth of
the stellar seed can be halted by the removal of the material
reservoir itself either due to the intense radiation pressure from
nearby young high-mass stars (Kroupa 2001; Whitworth &
Zinnecker 2004) or due to the actual collision of protostars, by
which the impact causes a large portion of the envelope to be
removed. As such these various ideas supply testable observational
 predictions like "BD halos" of young clusters, a relative
proximity to massive stars or a regular mixture with low mass

stars. They will lead to an increased understanding of BDs and
the star formation (SF) process in general.
Deriving the underlying initial mass function (IMF) is generally
 considered as an outstanding tool for revealing the physical
 mechanisms actually at play. However, it is only for a handful
 regions that the mass distribution has been derived crossing
the sub-stellar boundary at 0.072 Mcircledot, (see Bouvier et al. 2003,
and references therein). The mass function of young open clusters
 with an age of~100 Myr seems best described by a lognormal
 law over the full mass range in a dN/dlog M presentation,
regardless of their exact age, metallicity or richness. On the
contrary, young regions with active SF (<~3 Myr) may indicate
a mutually discordant number of BDs per star (Luhman et al.
2003). Additionally the binary frequency among field BDs indicates
 that the formation of BD binaries cannot be a mere extension
 of the formation of binary low mass stars (Close et al.
2003). As argued by Kroupa & Bouvier (2003), it would indicate
 that BDs do not follow the same general rules for formation
 as stars do. On the other hand, BDs are known to be
accompanied by CS disks (Natta et al. 2002, 2004). Mohanty
et al. (2005) specifically demonstrate that young BDs do experience
 a T Tauri-like phase. These findings combined with
the recent detection of a wide (>200 AU) BD binary (Luhman
2004), and the similar velocity distribution for stars and BDs
in synthetic small-N clusters (Delgado-Donate et al. 2004) all
argue in favor of BD formation being very much similar to lowmass
 star formation. It is clear that the dominant process that
determines the final mass of a low mass star or a BD is not
settled upon.
Young open clusters that occupy the age interval
of 10-50 Myr (hereafter called pre-main sequence clusters or
PMS clusters) can provide the clues to the issues listed above.
There are a number of reasons why this age interval is critical
in deriving the sub-stellar IMF. Unlike SF regions, PMS clusters
 provide the sampling of a complete stellar population immediately
 after the end of the active star forming phase, thus
revealing the full stellar product. Secondly BDs in PMS clusters
 are still bright and easily detected and in addition they do
not suffer from prodigious and inhomogeneous extinction as is
generally the case for SF regions. From a methodological point
of view, the IMF is obviously not what is observed, but is what
is obtained after transformation of the luminosity function. The
derived IMF thus depends on the theoretical predictions for the
dependence of magnitude on mass and age (and metallicity).
Baraffe et al. (2002) emphasized that no current BD model can
be relied upon for ages less than 1 Myr, due to uncertainties
in the treatment of convection, mass accretion rate, and atmospheric
 parameters (H2 abundance). It is therefore that mass
functions derived for somewhat older regions, like PMS clusters
 that will play a more accurate role than the MF derived
for young SF regions. Clusters older than ~150 Myr start to
be significantly affected by dynamical evolution, leading to serious
 problems in deriving the present-day MF. An additional
complicating factor for these somewhat older clusters is the intrinsically
 lower luminosities of BDs. Clusters in the PMS age
interval are therefore well-suited to explore in a relatively unbiased
 way the MF and thus to constrain BD formation theories,
provided that a large enough area is covered. Unfortunately,
only a few PMS clusters do actually exist in the solar neighbourhood
 (<500 pc).
IC 4665 is one of the few young open clusters within the
solar neighbourhood, at a distance of ~350 pc (DHipp = 385 ±
40 pc, Hoogerwerf et al. 2001). Previous work was undertaken
primarily by Prosser and collaborators, scrutinizing candidate
members on spectral type, Li-abundance, proper motion, radial
velocity, Halpha-emission and X-rays with ROSAT (Prosser 1993;
Prosser & Giampapa 1994; Martin & Montes 1997; Giampapa
et al. 1998). An age estimate of IC 4665 by Mermilliod (1981)
indicated an isochronal age of 30-40 Myr based on an analysis
of the upper main-sequence stars. However Prosser (1993) argued
 that due to membership uncertainties, photometric errors
and a colour effect of high rotational velocities, the upper main
sequence stars are consistent with the age of alpha Per and may
even indicate an age similar to the Pleiades. The age of IC 4665
is therefore not firmly established, but should lie somewhere
between 30 Myr and 100 Myr. By definition IC 4665 is possibly
not a PMS cluster. The cluster sequence is well fitted both by a
50 Myr and 100 Myr isochrone (see Fig. 1). Although located
out of the Galactic plane at a latitude of b =+17 a significant
background stellar population does exist in that specific direction.
 We refer to Prosser (1993) for an overview of the work by
various authors prior to 1990. In Fig. 1, we summarize current
membership of IC 4665 prior to the work presented here in the
form of a colour-magnitude diagram. The listed characteristics
of IC 4665 combined with its rather well-known intermediate
mass and solar-mass members render the cluster an excellent
target for probing the IMF into the regime of very low mass
stars and brown dwarfs.

This paper is organized as follows. We give a brief overview
of the observations, data reduction and calibration in Sect. 2.
Section 3 describes the procedure for selecting new candidate
members reaching well below the hydrogen burning limit. The
fractional contamination by non-member objects is estimated
in Sect. 4 using control fields, K-band photometry, and proper
motion. We derive and discuss the radial distribution, the mass
function and the total mass of the cluster in Sect. 5. A summary
of the work presented is given in Sect. 6.
2. Observations and data reduction
2.1. Optical photometry
As part of a CFHT large program, surveying star forming regions
 and young open clusters, IC 4665 was observed in the
light of the I (Mould) and z filters. The telescope at that time
was equipped with the CFH12K mosaic camera, the predecessor
 of the wide-field imager MegaCam. We refer to Moraux
et al. (2003) for more details regarding observations, data reduction
 and astrometric calibration of the CFH12K. In what
follows we describe some specific calibration steps applied to
the IC 4665 set of observations.
The CFH12K camera consists of 12 individual 2048 x
4096 CCDs each with angular scale of 0.206 pix arranged
in a 2 x 6 mosaic (Cuillandre et al. 2001). In total 13 CFH12K
pointings (named field A to M) centred on IC 4655 were observed
 for a total of 3.82 square degrees. Two control fields
(C1 and C2), located at three degrees from IC 4665 at the same
Galactic latitude were observed in order to estimate the foreground
 and background contamination of field stars. The central
 coordinate and the observation dates of each field are given
in overview in Table 1. The coordinates of the fields were chosen
 such that each mosaic would overlap another. The overlaps
are used to determine the external error on the photometry and
allow the creation of an internal photometric system by calibrating
 the fields to one "master" field. The projection on the
sky of the covered area relative to the brightest members of
IC 4665 is shown in Fig. 2. Each of the 13 fields was observed
for three different exposure times: 2 s, 30 s and 300 s or 360 s
in I and z band. It allows the sampling of member stars for a
range in masses between approximately 1 and 0.01 Mcircledot.
Objects were extracted from the CCD frames using the
Sextractor software package (Bertin & Arnouts 1996). To increase
 the accuracy on the photometric measurements, we used
sextractor in conjunction with a point spread function (PSF)
source extraction procedure. The PSF is determined for each
individual CCD of the CFH12K mosaic for each field by the
most well-behaved stars using the PSFex software (Bertin,
priv. comm.; see also Kalirai et al. 2001). The procedure obviously
 requires a constant PSF profile across a CCD. A changing
 PSF profile however was observed on the I-band images
 for each CCD of the CFH12K mosaic and for each of
the three exposures. The PSF profile was constant as function
 of X-coordinate, but changed as function of Y-coordinate:
at Y-coordinates less than ~500 the PSF profile turns into
an asymmetric, "teardrop"-like shape. This instrumental effect
 becomes especially apparent when comparing magnitudes
obtained with aperture photometry and PSFex photometry.
Such a comparison reveals an underestimation of the I-band
magnitude and reaches a maximum 0.2 for PSF extracted
stars at Y-coordinates less 50. This is due to the loss of flux
by the difference in shape of the actual stellar PSF and the applied
 model-PSF. Surprisingly, images obtained in the z-band
do not suffer from variable PSF profile, hinting at an I-band

filter problem of the CFH telescope. As far as we know, this
is the first time this problem is reported for CFH12K I-band
images. The immediate consequence of the variable PSF is
that objects at low Y-coordinates on the CCD appear redder
than they actually are. The PSF effect introduces therefore a
higher fractional contamination of non-members to our photometrically
 selected candidate member catalogues, but does
not lead to detection failure of IC 4665 members. On the other
hand, magnitudes determined from PSF fitting are more accurate
 than aperture photometry. We therefore chose to continue
using the PSF magnitudes, however flagging candidate member
stars when found at low Y-coordinates. The completeness of
the survey is exemplified in Fig. 3, after applying nominal zeropoints
 as determined by the CFHT's Elixir pipe-line (Magnier
& Cuillandre 2004). These limits have been determined from
the stars detected in the overlapping region of the 300 s exposure
 between fields D and K. The figure shows that I and z band
are 90% complete down to a magnitude of ~23;theI-band is
about 0.5 deeper than the z-band.
Some residual zeropoint offsets between the CCDs of the
CFH12K mosaic were found to be present after cross correlating
 stars located in the overlapping regions (see Fig. 2). In
Fig. 4 we show an example of the difference in instrumental
magnitudes for the same stars observed in the overlapping region
 between CCD 06 of field I and CCD 11 of field J (30 s ex-
posure). The upper part of the figure indicates the difference
in I and z magnitude (first and second panel) and the considerable
 difference in the resulting I-(I - z) CMD (third panel).
For every overlap the photometric shift in I and z-band was determined
 on a field-by-field and a CCD-by-CCD basis. In addition,
 an instrumental magnitude difference between the various
 exposures times was also found to occur. In this case we
cross-correlated the stars taken at different exposure times to
obtain the photometric shifts between 2 s, 30 s and 300 s. In
this way an internal calibration of the full dataset was achieved.
The lower part of Fig. 4 exemplifies the improved photometric
 correspondence between CCD 06 of field I and CCD 11 of
field J after the internal calibration. Figure 5 shows the median
difference in I-band and z-band for all the 33 overlapping regions
 that were used in the internal photometric calibration. It
shows the decreased differences between the photometry of the
various overlaps.

2.2. Infrared photometry
The next step in the derivation of IC 4665 properties after the
optical selection of the new candidate members using the calibrated
 I, z colour-magnitude diagram (see Sect. 3) is the step
of weeding out interloper contaminants to the selected member
 dataset. K-band imaging of 101 optically selected IC 4665
candidate members was obtained with the 1kx1k CFHT
IR camera (Starr et al. 2000) on July 10-12, 2003, and on
May 30-June 3, 2004. The targets stars were chosen such as
to probe the radial contamination factor, i.e. the stars observed
with CFHT-IR were selected from either field A, E or L. All the
observed fields with CFHT-IR are indicated in Fig. 2. The exposure
 time varied from 4 to 14 min for candidates with I magnitudes
 ranging from 17 to 22. Each object was dithered on 5
to 7 positions on the detector, so that the median of the images
would provide an estimate of the sky background during the
exposure. Individual images were dome flat fielded, sky subtracted,
 then registered and added to yield the final K-band
image. K-band photometric standards from Hunt et al. (1998)
were observed every couple of hours each night. Aperture photometry
 was performed on candidates and photometric standards.
 For a few close visual binaries a large aperture was
first used to estimate the system's total flux, and PSF photometry
 was performed to derive the flux ratio of the system
components. The derived K-band photometric zero point is
22.86 ± 0.03 for both runs, assuming an average extinction
coefficient of 0.07 mag/airmass.
For the brighter optically selected candidate members, infrared
 data were obtained from cross-correlating all selected
candidate members with the Two-Micron All-Sky Survey
(2MASS) catalogues. The 2MASS catalogues are complete
down to Ks ~ 14.3 with an uncertainty of 0.05.Thecom-
pleteness limit corresponds to a mass of M ~ 0.20 Mcircledot at the
estimated age and distance of IC 4665.
3. IC 4665 cluster member selection
3.1. Binary sequence
The photometric selection of new members relies on the accurate
 positioning of the isochrone corresponding to the age and
distance of IC 4665 in the optical I, z CMD. Throughout the
paper we use evolutionary tracks calculated by Baraffeetal.
(1998) and Chabrier et al. (2000) (the NEXTGEN and DUSTY
models) unless otherwise stated. In this subsection we use a
total of 10 known members (see Fig. 1) that have reliable I
and z photometry in order to provide the actual positioning of
the isochrone for an efficient member selection; the remaining
IC 4665 members are saturated in the I and z images. A close
look at the 10 faintest IC 4665 members in a NIR CMD (Fig. 6)
using the Two Micron All-Sky Survey data indicate that their
distribution is split in two groups. One half of the ten member
stars is thus suspected to be binary objects. The reality of the
binary sequence is supported by the star P 166 (filled square), a
double-line spectroscopic binary (Prosser & Giampapa 1994).
Surprisingly, a binary sequence is not obvious from the optical
CMD presented in Fig. 1, where we have indicated the names
of the 10 members in question. The cluster isochrone for the
selection of new members which is presented in the next subsection
 is thus positioned in the I, z CMD on the presumably
five non-binary members.
3.2. Newly selected candidates
In Fig. 7 we present the isochronal selection of new candidate
members of IC 4665. For reference, brown dwarfs have an expected
 I magnitude of 18.8 or dimmer at the estimated age
and distance of the cluster. Candidate members are selected
when their I, z photometry locates them to the right of the
100 Myr isochrone (full line) down to a magnitude of I = 22,
if their PSF profile was found to be stellar. This age is a conservative
 estimate and in accordance with the isochronal age of
the more massive members as presented in Fig. 1. The upper
magnitude limit ensures the completeness of selected members
down to an equivalent mass of~30 MJup. The selection strategy
therefore ensures the inclusion of all members, given the uncertainties.
 The ten member stars described in the previous paragraph
 are depicted by the large symbols at the top of the figure.
The 100 Myr isochrone is positioned to go through the bluest
confirmed members of the single star sequence. The isochrone
actually envelopes also quite well the probable members which
are indicated by the small crosses. The presence of a binary
sequence is also confirmed by the CFH12K I, z photometry.
Figure 7 in fact shows the 100 Myr isochrones (full lines)
of the NEXTGEN models (Baraffe et al. 1998) and DUSTY
models (Chabrier et al. 2000). For reference the 50 Myr (dotted
lines) isochrones are drawn as well. DUSTY model calculations
 include the settling of dust in the upper atmospheres of
cool objects, which is not included in the NEXTGEN models.
 The transition between the dusty atmosphere and gaseous

atmosphere is thought to occur at (I - z) similarequal 0.85 corresponding
 to a NEXTGEN Teff of 3000 K. This is the reason for the
artificial "hook" in the selected objects at this colour. We especially
 note that had we chosen only NEXTGEN models for
selecting candidate members of the lowest mass and not the
DUSTY models, then only a very small amount of BDs candidates
 would have been found.
At the distance of IC 4665, the 100 Myr isochrone intersects
 the foreground and background main sequence and giant
branch stars in the CMD at about 0.8 Mcircledot, rendering the selection
 procedure at these masses highly inefficient. For clarity,
Fig. 7 does not show all the stars extracted from the three exposures,
 but for two fields only (A and E). This translates to
about one sixth of the total number of extracted stars in the surveyed
 area. On the contrary, all the selected candidate members
 are plotted. The selected objects were further checked by
eye to eliminate any spurious detection due to hot pixels and/or
bad CCD columns. The selection procedure delivered a total
of 94 brown dwarf candidate members (18.8 < I < 22)
from 300 s measurements, 529 low mass stars (15.8 < I <
18.8) measured in the 30 s exposure and 163 low mass stars
(14.8 < I < 15.8) measured in the 2 s exposures. A list of
these objects is found in Table 3 (available only at the CDS).
4. Contamination by foreground and background
stars
The contamination in the foreground by faint M-dwarfs and by
highly reddened bright objects behind the cluster in the direction
 of IC 4665 is anticipated to be quite substantial. The large
numbers of newly selected candidate members makes this obvious.
 In this section we try to quantify the statistical contamination
 fraction and constrain memberships using K-band photometry
 and public proper motion data.
4.1. Statistical contamination
An estimate for the statistical contamination is derived from
the two control fields. These fields are located well beyond
the cluster boundary of IC 4665. The same selection procedure
 to identify candidate members as explained in the previous
 section was applied to the two control fields. In the low
mass stellar range down to the hydrogen burning limit (14.8 <
I < 18.8), we select 108 (48+60) objects in the two control
fields. This translates to an absolute contamination of 164±13
stars per square degree. The surveyed area centred on IC 4665
equals effectively 3.82 square degrees, and we thus expect

626 contaminating stars. We selected 691 IC 4665 candidate
members and we thus find a statistical contamination fraction
of~85%. That is only 1 in 6 photometrically selected candidate
member is expected to be an actual cluster member. We note
that contamination is expected to increase for the brighter stars,
as IC 4665 sequence intersects the Galactic field giant branch
at I < 15. For example in the range (14.8 < I < 15.8),
i.e. the candidate members selected from the short exposure,
we calculate 90% contamination. In the substellar domain,
corresponding here to 18.8 < I < 22, we estimated the
background contamination from the 300 s exposures. In the
control fields (9+2) 11 objects were selected from the 300 s
exposure, i.e. 17±5 per square degree. Therefore we expect to
find ~64 contaminants. In the surveyed area a total of 94 substellar
 candidates are found in the 300 s exposures. We conclude
 that the statistical contamination of our photometric selection
 of BD candidate members is on the order of 70%. A
more sophisticated estimate of the fractional contamination as
function of I-band magnitude bins is not reliable. The cluster
signal does not reach a high enough S/N per bin.
The relatively low efficiency in selecting new member object
 compared to the efficiency reported by Moraux et al.
(2003) for a similar procedure of photometric selection for
new Pleiades members can be attributed to a smaller (absolute)
Galactic latitude (b =-23 for Pleiades, b =+17 for IC 4665).
A second factor yielding a much higher fractional contamination
 is the larger distance to IC 4665 (350 pc against 125 pc) resulting
 in 2.2 difference in apparent magnitude. The IC 4665
stellar and substellar sequences are therefore much closer in
magnitude to the bulk of the foreground and background stars
in a CMD. This difference is only partially alleviated by the
younger age of IC 4665, as can be judged from the location of
the isochrones corresponding to 50 Myr and 100 Myr in Fig. 7.
4.2. Contamination estimate from K-band photometry
A first step in actually removing contaminant objects is by using
 K-band photometry. Although from first principles cluster
 BDs and stars are expected to be only intrinsically more
luminous than field stars of the same effective temperature
(Zapatero Osorio et al. 1997), IR photometry does effectively
separate out a fraction of non-member objects as shown in the
case of alphaPer (Barrado y Navascués et al. 2001). In part the explanation
 may be sought in photospheric differences, not large
enough to produce a significantly different (I - z) colour, but
that do become significant in the (I - K) colour. This is exemplified
 in the optical-NIR CMD of Fig. 8. The small dots and
open/filled asterisks represent all the selected stars based on
the (I - z)CMD,forwhichK-band photometry is available,
observed by either 2MASS or with CFHT-IR as described in
Sect. 2.2. A fraction of the (I - z) colour selected objects are
separated out in the diagram, not being located near the appropriate
 isochrone. The K-band contamination as function of
I-band magnitude displays two peaks of ~35% contamination
at I-band magnitude of 14 and 18,i.e.thelociwherethe
isochrones intersect the bulk of the field population.
The correction step using K-band on the optically selection
 procedure has a relatively low efficiency for stars brighter
than I ~ 18. This can be judged from the 2MASS data in
the NIR CMD of Fig. 8. At these magnitudes the I - K colour
allows at maximum the elimination of 35% per I-magnitude
bin. A better estimate of the efficiency is to perform the same
analysis using the control fields. Cross-correlating the optically
 selected objects in the two control fields (surely nonmembers)
 with the 2MASS dataset then we find that 80% of
these objects are to the right of the cluster isochrone down to
the I ~ 18 in a I,(I - K) CMD, and thus consistent with
membership. For magnitudes fainter than I ~ 18 using the
CFHT-IR data, Fig. 8 shows that K-band photometry has an efficiency
 of ~35%. We see thus that K-band photometry does
a reasonable job in separating out probable non-member stars
from members. However the lower fractional contamination as
determined from the K-band compared with the statistical estimate
 of the previous subsection directly implies that K-band
is not the definitive answer to construct a clean sample of cluster
 members, at least not for the age, latitude and distance of
IC 4665. Additional NIR photometry at other wavelength bands
does certainly allow to constrain membership of potential candidate
 members that have initially been selected by making use
of isochronal colours and magnitudes, as we'll see in the following
 section.
4.3. Membership constraints from proper motion
surveys
The proper motion of IC 4665 is small and does not stand
out significantly from the bulk field motion to be a strong
constraint on membership. Hipparcos measured the astrometry
of 13 IC 4665 members and obtained the following proper motion
 in mas yr: µl* =-7.2±0.3 µb =-3.0±0.3orconverted
to equatorial coordinates (µalpha,µdelta) = (-0.4,-7.5) mas yr

(Hoogerwerf et al. 2001). Nevertheless, here we explore two
public proper motion catalogues to evaluate membership for
the optically selected candidate members, viz. the Tycho-2 catalogue
 (Høg et al. 2000), and the USNO UCAC2 (Zacharias
et al. 2004). The Tycho-2 catalogue is complete down to
Ks = 8, and provides proper motion with an accuracy
of 2.5 mas yr. UCAC2 on the other hand is not complete for
stars brighter than R similarequal 8 due to saturation problems. For
the stars in the R-band magnitude interval 13-16, UCAC2
provides proper motion with an error of 6 mas yr. Crosscorrelation
 with 2MASS All Sky Catalogue learns that 97%
of the 2MASS entries are in the UCAC2 catalogues for
K-magnitude between 8-11.
We find 35 IC 4665 members in the UCAC2 catalogue for
a mean proper motion of (µalpha,µdelta) = (0.0,-7.6) mas yr with
a 0.8 error in the mean value and a ~5masyr standard deviation
 in both the right ascension and declination proper motion
 distributions. The UCAC2 proper motion for IC 4665 is in
good correspondence with the Hipparcos result of Hoogerwerf
et al. The spread in proper motion space, i.e. the internal motion
of IC 4665 can be judged from Fig. 9, in which the 35 member
 stars are indicated by the small open circles. Some 14 of
these members are also found in Tycho-2. Their Tycho proper
motions deviate from the UCAC2 results with (deltaµalpha,deltaµdelta) =
(3,3) mas yr, consistent within the uncertainties. The only
discrepant result between the two catalogues we find for the
visual binary K 67, whose UCAC2 proper motion is not consistent
 with membership, but its Tycho-2 is.
Using UCAC2 to identify proper motion members to
IC 4665, we require for selection that a star has a proper motion
that lies within 5 mas yr of the cluster proper motion. This
threshold is indicated by the large circle in Fig. 9. It equals the
standard deviation in the UCAC2 proper motion of the member
 stars and this criteria will therefore exclude 32% of genuine
cluster members. It is a compromise between optimising member
 selection and excluding background objects. We stress that
proper motions deviant from the internal motion of the cluster
as defined above is not a strong enough constraint to exclude
membership.
In Fig. 10 we present the result of the proper motion selection
 in a CMD. The tiny dots are the proper motion selected
candidate stars that are within the encircled area of the vector
 point diagram in Fig. 9. The cluster sequence is clearly recovered,
 and all member stars (open circles) have been plotted
 to indicate this. The small filled circles indicate all the
rejected objects from the P93 catalogues. Note that on and
near the cluster sequence for stars with a K-magnitude brighter
than 10.5 (corresponding to about 1.2 Mcircledot) objects are located
that do not have a counterpart in the P93 catalogues within a
2 cross-correlation radius, neither as a rejected nor as a confirmed
 member. Some among them are located within 1 of the
cluster centre, i.e. within the tidal radius of the cluster. These
objects are clearly candidate IC 4665 members. The crosses in
Fig. 10 identify the optically selected candidate members for
which a UCAC2 proper motion exists. Figure 10 shows that
currently public proper motion information for IC 4665 is limited
 to members brighter than K similarequal 12 or equivalently to
about 0.6 Mcircledot.The(J - K) colour provides a strong constraint
for cluster membership in this case. Deep NIR observations
in addition to the K-band are therefore strongly warranted.
The new generation of NIR wide field cameras (WFCAM,
WIRCAM) are especially designed to perform large surveys
of young open cluster, e.g. like the Galactic Cluster Survey as
part of the UKIDSS survey. For now we consider the stars that
are too red in the upper of the diagram at color J-K > 0.6and
K < 10.5 as likely non-members, and that the highest mass
cluster stars are still on the main sequence. All eleven red giants
 except one that are brighter than K < 5.5 and lie outside
the plotted range in Fig. 10 are confirmed non-members.

5. Discussion
The radial and the mass distribution for members and candidate
members are derived and discussed.
5.1. The radial distribution for members and candidate
members
The radial dependence of the surface density in square degrees
is derived for the high-mass member stars and the optically
selected low-mass candidate members. For the latter we distinguish
 between stellar and BD candidates. To obtain these distributions,
 we first determine the mass-weighted central coordinates
 of IC 4665 using the high-mass members only. Adopting
weights equaling the third power of the flux in the V-band magnitude,
 the cluster is found centred on RA(2000) = 174604,
Dec(2000) =+053853.
Three radial distributions for high-mass members are presented
 in Fig. 11. High-mass stars in this case are stars with a
K-band luminosity smaller than K = 10.5 or equivalently a
mass larger than ~1.2 Mcircledot. This mass threshold corresponds to
the location in a CMD where the cluster sequence intersects
the background population (e.g. Fig. 7). We estimate the completeness
 of the cluster census as function of radius and the
total radial extent of IC 4665 using both the P93 members and
the unbiased 2MASS catalogues. From 2MASS we select highmass
 stars located within 1.5 of the centre that have a colour
bluer than J - K = 0.6. The colour requirement excludes the
foreground red GB and AGB stars, a population that makes up
a large contaminant fraction at K-band luminosities K < 10.5
(see Fig. 10). The plain histogram with error bars in Fig. 11 depicts
 the 2MASS selected high-mass stars. The radial extent
of IC 4665 is easily read off as the distribution is seen to attain
 background levels at radii larger than ~1. The dashed line
denotes in fact the average surface density determined for a
similar sized area centred on the C1 and C2 control fields. The
control field surface density clearly marks the level of the stellar
 background level at the coordinates of IC 4665.
For comparison we added to Fig. 11 two radial distributions
of P93 member stars: a single-hatched and a double-hatched
histogram. They represent the surface density profile for all
candidate members with K < 10.5 and J - K < 0.6andthe
census of confirmed high-mass members with the same colour
and K-band magnitude constraints (29 objects), respectively.
The candidate high-mass members have been investigated for
cluster membership using various criteria in various studies.
Among them there exist confirmed members, confirmed nonmembers
 and stars without definite classification (see Sect. 1).
The bin at 1 of the single hatched histogram contains zero
confirmed candidates. This confirms the maximum extent of
the cluster derived from 2MASS selected stars.
We note the ratios written above the bins of the plain and
single-hatched histograms. The ratios of the plain histogram
indicate the statistical fraction of cluster stars as function of radius,
 given the background level (dashed horizontal line). The
ratios on the single-hatched histogram indicate the number of
confirmed high-mass members to high-mass candidates. The
two ratios at each bin are in good agreement, indicating that the
current knowledge of the relative number of high-mass member
 stars of each bin is a good representation of the actual relative
 number. If the radial distribution of cluster members is
best-described by a tidally truncated King profile (King 1962),
we can thus fit and compare the profile of high-mass members
 with that of the low-mass constituency adopting a tidal
radius of 1. The bottom histogram of the left panel in Fig. 12
shows again the distribution for the high-mass members, i.e.
the double-hatched histogram of Fig. 11. A by-eye fit to the
first four radial bins points to a core radius rc equaling ~0.25
(full line). In order to illustrate the uncertainty on this core radius,
 a King profile for rc = 0.5 is also indicated by a dashed
line. This profile still fits the observed distribution within the
error bars. We consider it as the upper limit to the high-mass
star core radius for IC 4665.
The upper histogram of the left panel in Fig. 12 shows the
radial dependence of the optically selected candidate stellar
members in the mass range of 0.07 and 0.6 Mcircledot. Error bars
are equivalent to the statistical uncertainty for each bin. We
note first that the surface densities derived for the control fields
(shaded areas) are in good agreement with surface densities
at radii larger than rt = 1.0. This is the case for both the
low-mass stars and the BD candadite members. It shows that
IC 4665 cannot be much more extended on the sky, which
gives confidence that the area covered with the CFH12K camera
 has been sufficient (see Fig. 2). The King profiles derived
for the high-mass stars are scaled and plotted using the same
line styles. A comparison between the best-fit high-mass King
profile (full line) with the radial distribution of low-mass stars
reveals that the two are not consistent. On the other hand a
King profile with rc = 0.5 (dashed line) is in reasonable correspondence.
 We thus see that the low-mass stellar population

of IC 4665 has a rc >~ 0.5. One can thus conclude that for
rc = 0.5 the surface densities of high-mass and low-mass stars
are consistent within the statistical uncertainties. Nevertheless,
it is probable that the low-mass stars have a broader distribution
 on the sky than the high-mass stars. This is illustrated by
the dotted King profile in Fig. 12. It corresponds to the distribution
 for the low-mass stars of the Pleiades. Scaled to the
distance of IC 4665 this profile is described by rc = 0.71,and
rt = 1.97 and is not extremely far off the observed distribution
of the low-mass candidate members of IC 4665.
The right panel of Fig. 12 displays the BD candidate distribution
 for which only a profile with rc = 0.5 is overlaid.
This profile corresponds best to the observed distribution, the
uncertainties are however large. Nevertheless a King profile fit
with rc of 0.25 (high-mass stars) or 0.71 (low mass stars in
Pleiades) are not consistent with the observed distribution (not
plotted). We conclude that the radial distribution of BD candidates
 is less certain than the one of the low-mass stars, however
it seems more reminiscent to the low-mass star than the highmass
 star distribution. It would indicate a mixed BD and lowmass
 stellar population in IC 4665. On the contrary the highmass
 stars may occupy a smaller area on the sky, which may
constitute the first indication of present-day mass segregation
in IC 4665. It is worth mentioning that mass segregation between
 stars of masses >3 Mcircledot and lower mass stars has been
reported for the open cluster NGC 2547 (Littlefair et al. 2003;
Jeffries et al. 2004). This cluster has an age that is probable
only slightly less than the one in IC 4665, and the clusters are
of comparable mass. Littlefair et al. and Jeffries et al. provide
evidence in favour of a primordial origin for the mass segregation
 of NGC 2547, adding to other instances of inferred primordial
 mass segregations in case of more massive clusters (see
Bonnell & Davies 1998; de Grijs et al. 2002). Therefore the reality
 of mass segregation in IC 4665 between stars of >~1.2 Mcircledot
and lower mass stars is of considerable importance, but in need
of confirmation.
Finally, using the King profiles for the low-mass stars and
BDs candidates we can estimate their total number in the cluster.
 IC 4665 consists of 30±10 BDs and 105±20 low-mass stellar
 members, where the uncertainties are estimated from trying
King profiles for various normalization factors. A background
contamination within the tidal radius then proves to be 80±5%
and 60±10% for low-mass stars and BDs respectively; within a
radius of 0.55 contamination factors of 65±5% and 40±15%
are expected. These estimates also take into account the statistical
 uncertainties in the background level. From the radial
distribution we estimate the fraction of BDs to total members
in IC 4665. At the upper mass end there are 35 members with
masses in the range 1 < M/Mcircledot < 5 and an estimated 39 stars
between 1 and 0.6 Mcircledot, based on a Salpeter IMF. Taking into
account the uncertainties we find a range of 10-19% for the
fraction of IC 4665 BDs (down to a mass of ~0.03 Mcircledot)tototal
members. These factions are consistent with fractions found
for other young open clusters for which such data are available,
 viz. 10-15% for the Pleiades, NGC 2516, and Blanco 1
(see Moraux et al. 2005, and references therein). In this respect
IC 4665 thus seems to be a "normal" young open cluster.
5.2. The mass function
The age of IC 4665 is not better determined than within a
margin of nearly 40 Myr. We derive therefore in this section
 the mass function of IC 4665 assuming ages of 50 Myr
and 100 Myr and the concomitant interstellar extinctions.
Masses are assigned to low-mass objects by converting the
I-band magnitudes using the NEXTGEN and DUSTY models.
At overlapping magnitudes these two models do not assign the
same mass for given magnitude. To overcome this, we made
an interpolation between the two models at the expected transition
 effective temperature of 3000 K (I - z similarequal 0.85). For the
high-mass bins we adopt the models without overshooting from
the Padua group (Girardi et al. 2000) and convert the K-band

magnitude to mass. We stress that in the derivation of the mass
function presented here we do not incorporate any correction
for the cluster depth (4% uncertainty on magnitude), undersampling
 of the cluster population due to various effects (hiding,
bad pixels etc.) nor binarity (see e.g. Jeffries et al. 2004, for a
discussion of these various effects).
The two mass functions for the two adopted ages of IC 4665
are presented in Fig. 13 in dN/dM (Salpeter slope equals-2.35)
and dN/dlog M (Salpeter slope equals -1.35) representations.
The 50 Myr mass function in numbers is presented in Table 2.
Both mass functions correspond to candidate members located
within the tidal radius of 1.0: high-mass stars (black squares,
histogram) from 2MASS all sky catalogues applying the same
colour and magnitude selection as in the previous subsection,
and low-mass stars/BD candidate members from our CFHT
survey. For the latter we use asterisk symbols, open circles
and filled circles for candidate members found in 2 s, 30 s
and 300 s exposures respectively. All mass bins have been statistically
 corrected for background contamination. The three
high-mass bins from 2MASS are multiplied by 15%. This factor
 has been derived using a 30 square degree region centred on
control field C2. The low-mass bins are corrected using factors
that are valid inside the tidal radius that have been estimated
in the previous subsection for the low-mass and BDs candidates,
 viz. 20% and 40%. Error bars correspond to statistical
uncertainties and the uncertainties in the above correction factors.
 Additional correction was made for the high-mass candidate
 members by applying the criteria reducing the nonmember
 objects that have been discussed in Sects. 4.2 and 4.3.
We specifically note the vertical dotted lines in all four panels

of Fig. 13. They indicate the masses for which the cluster sequence
 intersects the background stellar population. At these
masses therefore a large relative number of non-member contaminants
 is expected, for which no specific correction is made.
The two mass functions in dN/dM representation (left column
 of Fig. 13) are compared to a Salpeter mass function for
mass bins >~1 Mcircledot, and find reasonable correspondence. More
crucial is the shape of the mass function for bins <~1 Mcircledot.For
this mass range we performed an actualchi-fit to the mass function,
 taking into account the 10 low-mass bins that represent
the investigated mass range in this paper. For the 50 Myr case
(upper panel) we find a best linear fit for a slope of-0.61±0.13
for a chi = 3.6. We have indicated this fit by a full line, and the
1sigma uncertainties on the fit-parameters by the two dotted lines.
For the 100 Myr case (lower panel) the slope is less steep and
the fit less good as can be judged from Fig. 13. We note that
a power law relation with a slope of -0.6 is approximately the
slope of the mass function for masses between 0.4 and~0.1 Mcircledot
found in other young open clusters (see e.g. Bouvier et al.
2003). We thus conclude that down to ~0.1 Mcircledot, IC 4665 confirms
 this trend of the mass function.
For the three mass bins corresponding to the lowest masses
for which the mass function crosses the H-burning limit the
situation is more ambiguous. Either the mass function shows
a constant -0.6 slope well into the substellar regime rendering
 two mass bins discordant, e.g. a dip at the stellar-substellar
boundary as indicated by the linear fit; or on the other hand, the
mass function flattens, rendering the last mass bin discordant.
For both interpretations there exists supporting arguments.
In favour of the first interpretation is that such a dip in the
mass function is a recurrent feature of young open cluster. It
is discussed in detail by Dobbie et al. (2002) and suggested
to be due to dust formation different from the dust formation
recipe incorporated in the current generation of cool model
atmospheres. It occurs at Teff similarequal 2700 K, and reduces the ob-
ject's luminosity at the investigated wavelengths. For the age
of IC 4665 a dip in the mass function due to this effect is expected
 to be at ~0.07 Mcircledot, i.e. corresponding to the observed
dip in the left columns of Fig. 13.
The argument in favour of a flattening of the mass function
in dN/dM, can be derived from the right-hand side column of
Fig. 13. Here we present the mass function in the dN/dlog M
representation and we have fitted the mass function with lognormal
 functions. Lognormal functions are defined by an average
 quantity (mass in this case) and the width (sigma)ofthe
lognormal distribution. A lognormal prescription is found to
be a good representation for the mass function in the Pleiades
(Moraux et al. 2003), Blanco 1 and NGC 2516. Moraux et al.
(2005) reports for these three young open clusters a similar
average mass of around 0.3 Mcircledot and a similar sigma of ~0.5, producing
 quite a homogeneous picture of clusters in the age
range 100-150 Myr. The case for IC 4665 is shown in the righthand
 column of Fig. 13. The top right panel corresponds again
to a cluster age of 50 Myr and the mass distribution is well fitted
 by a lognormal function with an average mass of 0.32 Mcircledot
(full line), except for the lowest mass bin. If one thus assumes
that the complete mass function is described by a lognormal
relation, then one expects the mass function to actually flatten
in a dN/dM representation, as is suggested by the mass bins at
the H-burning limit on the left-hand side column of Fig. 13.
In Fig. 13, we also compare the lognormal fit of IC 4665 to
the one of the Pleiades (with an ¯M = 0.27 Mcircledot, dashed line), and
we note that they are quite similar. In the case of the somewhat
younger cluster Blanco 1, Bouvier et al. (2005) actually find
the same average mass as we deduce here for IC 4665. In fact
the average mass of the Pleiades is suggestively bracketed by
a higher average mass in younger clusters and lower average
mass in the field (viz. 0.25 Mcircledot, Chabrier 2003).
Finally we note, that despite the similar shape of the mass
function in the dN/dM representation (left column) for both
adopted ages, the dN/dlog M representations (right column)
show a marked difference between the two ages. If IC 4665
were to be 100 Myr old then a lognormal fit to the mass
distribution tells us that the average mass of the cluster is
significantly larger than found for clusters of such an age,
viz. 0.45 Mcircledot. This would indicate that IC 4665 for some reason
 might have lost a substantial fraction of its lower mass constituent.
 In light of this we point out the remarkable coincidence
in space (few degrees distance, where 1 degree is about 6 pc),
age, and proper motion of IC 4665 with the young open cluster
Collinder 359 for which we refer to Lodieu et al. (2005). We
speculate that tidal fields between the two cluster could have
stripped off preferentially the low-mass members of IC 4665.
Be this as it may, and assuming the younger age for
IC 4665 we summarize that the mass function of IC 4665 be
well approximated by a power law with slope -0.6 between 1
and ~0.1 Mcircledot. The shape of the mass function of IC 4665 below
 the H-burning limit can either be explained in terms of the
"M 7-8" gap as first described by Dobbie et al. (2002) producing
 a dip with respect to a linear relation, or alternatively, by
assuming that mass functions in general are lognormal functions,
 in which case the mass bin corresponding to the lowest
mass in Fig. 13 would be erroneous for a yet unknown reason.
At the moment we cannot exclude either interpretation for the
shape of the substellar mass function of IC 4665.
5.3. The total mass and tidal radius
In the previous subsection, we arrived at a tidal radius for
IC 4665 of ~1 along various lines of analysis. The tidal radius
depends amongst others on the total cluster mass and Galactic
orbital radius of the cluster, a relation that was cast in the form
of an equation for the Solar neighbourhood by Pinfield et al.
(1998). We note that, as the authors point out, this equation is
based on a simplified model. In the following we estimate the
total mass of IC 4665 and confront this with the tidal radius.
Under the assumptions that (1) the cluster census is complete
for the high-mass stars down to V < 10 (which is likely to
be an underestimate, see Fig. 1); and (2) a Salpeter-law holds
between 5 and 0.7 Mcircledot and a power law with slope -0.6down
to 0.01 Mcircledot, we find the total cluster mass of~300 Mcircledot.Herewe
accounted for an estimated 50% binarity with a flat q distribution.
 An upper limit of<440 Mcircledot to the cluster mass is furnished
by assuming that all high-mass candidate member stars are actual
 members. An other and independent mass estimate can be

found using 2MASS catalogues. Selecting blue (J - K < 0.6)
high-mass stars within the tidal radius of IC 4665 and correct
 them for the background per K-band bin, one counts between
 4.5 and 1.15 Mcircledot a total of 56 stars. This translates to
a total cluster mass of 350 Mcircledot, within the above upper limit.
IC 4665 has thus a mass which is probably about half the mass
of the Pleiades (Pinfield et al. 1998). Comparing now the cluster
 mass estimate with the tidal radius we find that IC 4665 does
not follow the relation between cluster mass and tidal radius of
Pinfield et al. (1998), their Eq. (12). In fact IC 4665 seems too
small for the lowest of the above mass estimates by 0.5.The
mass and tidal radius can be reconciled in terms of the Pinfield
relation if the distance of IC 4665 would have been underestimated
 by 200 pc. It would render the distance modulus 1 mag
larger, which seems unlikely (see Fig. 1). Alternatively tidal
stripping of cluster members may also depend on the local
environment of the cluster. We again point, as in the previous
paragraph to the fact that IC 4665 is located in the Galactic
field suspiciously close to the young open cluster Collinder 359
(Lodieu et al. 2005). We speculate that a small tidal radius for
the total mass and a relatively large average mass for a cluster
 age of 100 Myr could possibly be both the manifestation of
ongoing tidal interactions and subsequent stripping of cluster
members of IC 4665 and Collinder 359. Whether this process
is really in effect certainly warrants a deeper investigation into
the environment of these two young open clusters.
6. Summary
We have presented the photometric selection of 691 low-mass
stellar and 94 BD candidate members to the young open cluster
IC 4665. This extends the census of cluster (candidate) members
 well into the BD regime, going down to a few tens of
Jupiter masses. We make an estimate of the contamination due
to foreground and background objects in the field of IC 4665
on a statistical basis using control fields located at the same
Galactic latitude. An additional selection procedure of contaminant
 objects for low-mass stellar and BD candidates is reported
on the basis of public (2MASS) and new K-band photometry.
 Public proper motion survey data (Tycho-2, UCAC2) are
also used to weed out candidate members for the somewhat
brighter objects. Despite the substantial background contamination
 (up to 85% taken over the full surveyed area) the cluster
does clearly stand out on the sky from the background population.
 This is the case for the brighter objects, but also the lower
mass stars and BDs clearly show an increased number surface
density towards the centre of the cluster. We study the radial
distribution of various cluster populations and find a consistent
tidal radius of 1. There could be an indication of mass segregation,
 but this needs confirmation.
We find IC 4665 typical for a cluster in this age range with
respect to (1) the fraction of BD (down to ~0.03 Mcircledot)tototal
members, viz 10-19%; (2) the slope of the mass function that
it is well described by a power law with a power of-0.6forthe
low-mass objects downto~0.1 Mcircledot; (3) a cusp in the mass function
 at about the Hydrogen burning limit, than can be explained
in terms of the "missing" M 7-8 dwarfs, alternatively the cusp
may be the product of an actual flattening of the mass function
and a spurious value for the lowest mass bin, in concordance
with the mass function being lognormal; (4) an average mass
of 0.32 Mcircledot when assuming an age of 50 Myr; (5) the width of
sigma = 0.5 for the lognormal function fitted to the mass function
 in a logarithmic representation. However in two respects
IC 4665 may stand out. We estimate the total cluster mass and
find that it does not agree with the expected mass derived from
the relation between tidal radius and total cluster mass (Pinfield
et al. 1998). Secondly, we find that the average mass of IC 4665
is relatively large when assuming an age of 100 Myr. We speculate
 that increased stripping of member stars due to the interaction
 with the close-by young open cluster Collinder 359
could be the cause of these two effects. Clearly a better age
determination for IC 4665 may resolve this issue. The analysis
presented here adds in creating a consistent picture of the mass
distribution in young open clusters. A rigorous evaluation of
the mass function of IC 4665 using a large body of spectroscopic
 data will be the subject of paper II in our series on the
young open cluster IC 4665.
References


______________________________________________________________
ccopyright ESO 2006
1 Laboratoire d'Astrophysique, Observatoire de Grenoble, BP 53, 38041 Grenoble, Cedex 9, France
e-mail: dewit@obs.ujf-grenoble.fr
2 INAF, Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, 50125 Florence, Italy
3 Canada-France-Hawaii Telescope Corp., Kamuela, HI 96743, USA
4 Physics and Astronomy Department, Vanderbilt University, 1807 Station B, Nashville, TN 37235, USA
5 Centre for Astrophysics Research, Science & Technology Research Institute, Department of Physics, Astronomy & Mathematics,
University of Hertfordshire, College Lane, Hatfield AL10 9AB, UK
6 University of Leicester, Department of Physics & Astronomy, University Road, Leicester LE1 7RH, UK
7 University of Exeter, School of Physics, Stocker Road, Exeter EX4 4QL, UK
8 Astrophysikalisches Institut Potsdam, An der Sternwarte 16, 1448 Potsdam, Germany
9 Institute of Astronomy, Cambridge, CB3 0HA, UK
Received 25 August 2005 / Accepted 28 October 2005
ABSTRACT
We present a study of the young (30-100 Myr) open cluster IC 4665 with the aim to determine the shape of the mass function well into the
brown dwarf regime. We photometrically select 691 low-mass stellar and 94 brown dwarf candidate members over an area of 3.82 square
degrees centred on the cluster. K-band follow-up photometry and Two-Micron All-Sky Survey data allow a first filtering of contaminant objects
from our catalogues. A second filtering is performed for the brightest stars using proper motion data provided by the Tycho-2 and UCAC2
public catalogues. Contamination by the field population for the lowest mass objects is estimated using same latitude control fields. We fit
observed surface densities of various cluster populations with King profiles and find a consistent tidal radius of 1.0. The presence of possible
mass segregation is discussed. In most respects investigated, IC 4665 is similar to other young open clusters at this age: (1) a power law fit to
the mass function between 1 and 0.04 Mcircledot results in best fit for a slope of -0.6; (2) a cusp in the mass function is noticed at about the substellar
boundary with respect to the power law description, the interpretation of which is discussed; (3) a fraction between 10-19% for BDs with
M >~ 0.03 Mcircledot to total members; (4) a best-fit lognormal function to the full mass distribution shows an average member mass of 0.32 Mcircledot,
if IC 4665 has an age of 50 Myr.
Key words. Galaxy: open clusters and associations: individual: IC 4665 - stars: low mass, brown dwarfs -
stars: luminosity function, mass function - techniques: photometric
star Based on observations obtained at the Canada-France-Hawaii
Telescope (CFHT) which is operated by the National Research
Council of Canada, the Institut National des Science de l'Univers of
the Centre National de la Recherche Scientifique of France, and the
University of Hawaii.
starstar Table 3 is only available in electronic form at the CDS via
anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via
http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/448/189
00.5 11.5
14
12
10
8
6
Fig. 1. Known member stars of IC 4665. Two isochrones for 50 Myr
and 100 Myr with indicated reddening (for Rv = 3.1) from the Padua
group (Girardi et al. 2000). Binary and candidate binary systems
have specific symbols. Some members are named (see also Fig. 6).
The plain histogram represents relative numbers in each magnitude
bin of candidate members (Prosser 1993); the hatched part the confirmed
 members. The masses on the y-axis correspond to the 50 Myr
isochrone.
Table 1. Central coordinates of IC 4665 pointings (fields) with CFHT
using CFH12K camera. Fields C1 and C2 are control fields.
Field RA (2000) Dec (2000) Date
(h,m,s) (,,)
A 17:46:18 +05:43:00 19-5-2002
B 17:43:38 +05:43:00 20-5-2002
C 17:48:58 +05:43:00 20-5-2002
D 17:44:58 +06:10:00 20-5-2002
E 17:48:06 +06:07:00 20-5-2002
F 17:48:00 +05:17:00 09-6-2002
G 17:44:58 +05:13:00 09-6-2002
H 17:46:18 +06:31:00 10-6-2002
I 17:47:25 +04:53:00 10-6-2002
J 17:45:00 +04:51:00 09-6-2002
K 17:43:38 +06:31:00 09-6-2002
L 17:49:25 +06:31:00 10-6-2002
M 17:42:23 +05:17:00 09-6-2002
C1 17:51:03 +08:05:54 12-6-2002
C2 17:40:53 +02:50:14 12-6-2002
Fig. 2. Thirteen CFH12K fields (large boxes) covering 3.82 square degrees
 of IC 4665 were observed. Note the overlaps between the various
 fields. The small boxes correspond to the fields observed with
CFHT-IR. IC 4665 members are represented by the large filled dots,
whose sizes are indicative of their flux in V-band. The large dashed
circle corresponds to a tidal radius for the best fit King profile (see
Sect. 5).
Fig. 3. Histogram of objects in field D located in the overlap with
field K. The hatched histogram presents objects detected in both fields,
and measures the completeness at each magnitude bin.
16 18 20 22
-0.4
-0.2
0
0.2
0.4
16 18 20 22
-0.4
-0.2
0
0.2
0.4
I-magntude
I-magnitude z-magnitude (I-z)
I-magnitude
Delta magnitude
Delta magnitude
Fig. 4. Exemplifying the systematic shift between observations of the
same part of the sky but with different CCDs of the CFH12K mosaic.
 The left (I-band) and middle (z-band) columns show CCD06 of
field I minus CCD10 of field J in a 30 s exposure, before (upper row)
and after (lower row) applying a zeropoint shift. The final effect in a
CMD is shown in the right column.
Fig. 5. The resulting median difference in I-band (top panel)and
z-band (bottom panel) between the 34 overlapping regions of IC 4665.
Thefulllineindicatesthedifference in photometry before the internal
calibration ("raw"), the dotted line after applying a photometric shift.
This shift has been derived from the overlapping regions, putting the
photometry on one internal system.
0.4 0.6 0.8 1
13
12
11
10
Fig. 6. The NIR CMD for IC 4665 members with confident I, z photometric
 measurements (large symbols), based on 2MASS photometry.
 Small symbols are member stars without accurate I,z magnitudes.
Star P 166 is a binary. The figure shows that 5 out of 10 objects are
candidate binary stars.
0 0.5 1 1.5
25
20
15
Fig. 7. Colour magnitude diagram of CFHT I, z observations for IC 4665. Some 94 Brown Dwarf candidate members and 691 low mass stellar
candidate members have been photometrically selected using the theoretical isochrone corresponding to 100 Myr, down to a magnitude of
I = 22. Member stars are given as circles or as a square (the confirmed binary P 166). Probable member stars identified by P92 are given
as small crosses. Error bars have been determined from the overlap regions, with Y(pixel) > 300. Note that for clarity the Galactic field is
plotted corresponding to two CFH12K fields, whereas all new candidate members are represented. The mass scale corresponds to the 50 Myr
NEXTGEN model for >0.2 Mcircledot and to the 50 Myr DUSTY model for lower masses.
Fig. 8. Separating contaminant stars from optically selected objects
using K-band photometry, either from the 2MASS survey (small dots)
or observations with CFHT-IR instrument.
Fig. 9. A vector point diagram for stars within 1 degree from IC 4665,
based on UCAC2 (Zacherias et al. 2004). The estimated bulk proper
motion (circle) for IC 4665 based on the brightest stars using Tycho-2
data.
Fig. 10. The colours and magnitude for stars with proper motion consistent
 with IC 4665 (small dots), i.e. that are inside the circle of
Fig. 9). Confirmed member stars (small circles) are added for reference.
 Small filled symbols are P93 non-members. Crosses are objects
with UCAC2 proper motion that have been selected as candidate members
 in the optical survey.
Nb of high-mass candidate members per sq. deg.
Fig. 11. Histogram for stars with K < 10.5 and J - K < 0.6. Plain
histogram 2MASS All Sky Catalogue selected stars; single-hatched
histogram all selected stars from P93 catalogues; double-hatched histogram
 confirmed members. Dashed line is the average background
determined from control fields C1 and C2. Upper percentages: expected
 fraction of members, given the background level. Lower ratios:
fraction of confirmed members to P93 selected candidates. See text
for explanation.
Fig. 12. Left: upper histogram presents the radial distribution for candidate members with 14.8 < I < 18.8. Lower histogram equals the
double-hatched histogram of Fig. 11. The shaded areas are the stellar background estimates derived from the control fields. Full lines are King
profiles with rc = 0.25 and rt = 1.0; dashed lines King profiles with rc = 0.5 and rt = 1.0; dotted line is the scaled Pleiades low-mass profile
with rc = 0.71 and rt = 1.97. Right: the radial distribution for the BD candidate members.
Fig. 13. The IC 4665 mass function in the dN/dM (left column)anddN/dlog (M)(right column); top row assuming an age of 50 Myr, bottom
row for an age of 100 Myr. Symbols and lines in each panel have the same meaning. Squares denote candidate members within the tidal radius
of 1.0 selected from 2MASS, asterisks, open circles and filled circles from the 2 s, 30 s and 300 s exposures. The vertical dotted lines indicate
the masses for which the cluster sequence intersects the background stellar population. Full lines in left column are Salpeter-law mass functions
and power law functions (fits) with slopes of respectively -0.6and-0.54, and their associated 1sigma fitting uncertainties (dotted). In the right
columns lognormal fits to the observed mass functions (full line) and the scaled Pleiades lognormal fit (dashed) are indicated (see text for
details). Note the differences between the logarithmic mass functions for the two adopted ages.
Table 2. The mass function for IC 4665 for an age of 50 Myr. First
column gives the central I-band magnitude of each bin. The column
indicated with " f " is the correction factor applied to account for the
contamination by non-member objects.
I DELTAMNobs fNcorr dN/dM dN/dlog M
(Mcircledot)%
13.30 0.92-0.77 111 15 16.7 109.6 213.0
14.30 0.77-0.66 57 15 8.6 75.0 122.9
15.30 0.66-0.47 115 15 17.3 93.4 120.2
16.18 0.47-0.31 107 20 21.4 134.5 119.8
16.93 0.31-0.20 114 20 22.8 203.7 118.5
17.68 0.20-0.135 96 20 19.2 291.2 111.1
18.43 0.135-0.093 62 20 12.4 299.2 77.8
19.23 0.093-0.059 25 40 10.0 287.6 49.5
20.08 0.059-0.044 11 40 4.4 293.2 34.3
20.93 0.044-0.035 17 40 6.8 799.6 72.3
Acknowledgements. We would like to thank E. Bertin for providing
 PSFex, I. Baraffe for computing isochrones for CFH12K filters,
 and E. Magnier for help in the derivation of photometric ZP.
The research benefitted from financial assistance from the European
Union Research Training Network "The Formation and Evolution
of Young Stellar Clusters" (RTN1-1999-00436). This research has
made use of the Simbad database, operated at the Centre de Donnees
Astronomiques de Strasbourg (CDS), and of NASA's Astrophysics
Data System Bibliographic Services (ADS). This publication makes
use of data products from the Two Micron All Sky Survey, which is
a joint project of the University of Massachusetts and the Infra-red
Processing and Analysis Center/California Institute of Technology,
funded by the National Aeronautics and Space Administration and the
National Science Foundation. W.J.D.W. would like to thank the staff
and students at the Osservatorio di Arcetri for their hospitality.
Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 1998, A&A,
337, 403
Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 2002, A&A,
382, 563
Barrado y Navascués, D., Stauffer, J. R., Bouvier, J., & Martín, E. L.
2001, ApJ, 546, 1006
Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393
Bonnell, I. A., & Davies, M. B. 1998, MNRAS, 295, 691
Bouvier, J., Moraux, E., Stauffer, J. R., Barrado y Navascués, D., &
Cuillandre, J. 2003, in IAU Symp., 147
Chabrier, G. 2003, PASP, 115, 763
Chabrier, G., Baraffe, I., Allard, F., & Hauschildt, P. 2000, ApJ, 542,
464
Close, L. M., Siegler, N., Freed, M., & Biller, B. 2003, ApJ, 587, 407
Cuillandre, J., Luppino, G., Starr, B., & Isani, S. 2001, in SF2A-2001:
Semaine de l'Astrophysique Française, 605
de Grijs, R., Gilmore, G. F., Johnson, R. A., & Mackey, A. D. 2002,
MNRAS, 331, 245
Delgado-Donate, E. J., Clarke, C. J., & Bate, M. R. 2004, MNRAS,
347, 759
Dobbie, P. D., Pinfield, D. J., Jameson, R. F., & Hodgkin, S. T. 2002,
MNRAS, 335, L79
Giampapa, M. S., Prosser, C. F., & Fleming, T. A. 1998, ApJ, 501,
624
Girardi, L., Bressan, A., Bertelli, G., & Chiosi, C. 2000, A&AS, 141,
371
Høg, E., Fabricius, C., Makarov, V. V., et al. 2000, A&A, 355, L27
Hoogerwerf, R., de Bruijne, J. H. J., & de Zeeuw, P. T. 2001, A&A,
365, 49
Hunt, L. K., Mannucci, F., Testi, L., et al. 1998, AJ, 115, 2594
Jeffries, R. D., Naylor, T., Devey, C. R., & Totten, E. J. 2004, MNRAS,
351, 1401
Jiang, I., Laughlin, G., & Lin, D. N. C. 2004, AJ, 127, 455
Kalirai, J. S., Richer, H. B., Fahlman, G. G., et al. 2001, AJ, 122, 257
King, I. 1962, AJ, 67, 471
Kroupa, P. 2001, MNRAS, 322, 231
Kroupa, P., & Bouvier, J. 2003, MNRAS, 346, 369
Lin, D. N. C., Laughlin, G., Bodenheimer, P., & Rozyczka, M. 1998,
Science, 281, 2025
Littlefair, S. P., Naylor, T., Jeffries, R. D., Devey, C. R., & Vine, S.
2003, MNRAS, 345, 1205
Lodieu, N., Bouvier, J., James, D. J., et al. 2005, A&A
Luhman, K. L. 2004, ApJ, 614, 398
Luhman, K. L., Stauffer, J. R., Muench, A. A., et al. 2003, ApJ, 593,
1093
Magnier, E. A., & Cuillandre, J.-C. 2004, PASP, 116, 449
Martin, E. L., & Montes, D. 1997, A&A, 318, 805
Mermilliod, J. C. 1981, A&A, 97, 235
Mohanty, S., Jayawardhana, R., & Basri, G. 2005, ApJ, 626, 498
Moraux, E., Bouvier, J., & Clarke, C. 2005, Mem. Soc. Astron. It., 76,
265
Moraux, E., Bouvier, J., Stauffer, J. R., & Cuillandre, J.-C. 2003,
A&A, 400, 891
Natta, A., Testi, L., Comerón, F., et al. 2002, A&A, 393, 597
Natta, A., Testi, L., Muzerolle, J., et al. 2004, A&A, 424, 603
Papaloizou, J. C. B., & Terquem, C. 2001, MNRAS, 325, 221
Pinfield, D. J., Jameson, R. F., & Hodgkin, S. T. 1998, MNRAS, 299,
955
Prosser, C. F. 1993, AJ, 105, 1441
Prosser, C. F., & Giampapa, M. S. 1994, AJ, 108, 964
Reipurth, B., & Clarke, C. 2001, AJ, 122, 432
Shu, F. H., Adams, F. C., & Lizano, S. 1987, ARA&A, 25, 23
Starr, B. M., Doyon, R., Beuzit, J., et al. 2000, in Optical and
IR Telescope Instrumentation and Detectors, ed. M. Iye, & A. F.
Moorwood, Proc. SPIE, 4008, 999
Whitworth, A. P., & Zinnecker, H. 2004, A&A, 427, 299
Zacharias, N., Urban, S. E., Zacharias, M. I., et al. 2004, AJ, 127, 3043
Zapatero Osorio, M. R., Martin, E. L., & Rebolo, R. 1997, A&A, 323,
105