Photometric properties of low-mass stars and brown dwarfs
The colour-magnitude relations plotted here have been constructed by combining data
from Mike Bessell's catalogue of BVRI photometry of stars from the CNS1 and CNS2
catalogues (we refer to these as the Gliese stars),
together with data for stars within 8 parsecs of the Sun and observations of
a small number of ultracool dwarfs. Photometry for the 8-parsec stars is taken mainly
from Leggett (2), while the data on ultracool dwarfs are from Dahn et al (3).
Both the Gliese stars and the 8-parsec sample have been cross-referenced against the
Hipparcos catalogue, and against the USNO parallax program. The CMDs plotted below
include only stars with Hipparcos parallaxes accurate to (sigma_pi/pi) < 15%:
the USNO measurements, primarily of late-type dwarfs in the 8-parsec and
ultracool dwarf samples, are of much higher precision.
Cross-referencing Bessell's catalogue against the Hipparcos catalogue gives 831 matches.
All stars known to have suspect photometry (i.e. binaries with joint colours, or primaries
with companions of similar luminosities (delta V < 3 mag) at small separations)
have been excluded from the datasets - but it is clear that some still remain.
That leaves 739 stars in the Gliese dataset, and 120 (out of 150) in 8-parsec sample.
Dave Monet has cross-referenced the Gliese catalogue against the 2MASS database (Oct 22, 1999):
224 stars correlated, and ~150 have JHKS data with acceptable accuracy (most of
the remainder are too bright at near-infrared wavelengths for 2MASS!).
The sample includes a higher proportion than usual of southern stars (from Bessell
and Cousins) and a fair number of bright stars - G dwarfs within 25 parsecs are
kind of bright. This sample will be extended in the near future.
There are about 2000 stars in the CNS3 that have Hipparcos parallaxes measured to a
precision of better than 20%, including faint companions whose distances can be inferred
from the measurement of the primary. A large number of those stars have at least BV
photometry, so it will be possible to obtain (V-J), (V-K) etc colours once those
2MASS data for those sources.
(1) Bessell, M.S., 1991, AAS 83, 357
(2) Leggett, S.K., 1992, ApJS 82, 351
(3) Dahn, C.C. et al, 2001, in prep.
1. Nearby stars - disk dwarfs
In the following figures, CNS3 stars are plotted as green crosses; stars
from the 8-parsec sample as blue crosses; M dwarfs from the USNO ultracool sample (Dahn et al, 2002)
as red crosses; L dwarfs (from Dahn et al) as solid magenta points; and T dwarfs as 5-point stars.
The (MV, (V-I)) relation can be represented as
MV = 3.98 + 1.437(V-I) + 1.073 (V-I)2 - 0.192 (V-I)3, 0.85<(V-I)<2.85
MV = 3.66 + 4.66(V-I) - 0.517(V-I)2 + 0.0448 (V-I)3, 2.96<(V-I)<3.45
The discontinuity at (V-I)~2.9 is real, and is evident in several other colour-magnitude diagrams, notably
(MJ, (I-J)). There is no theoretical explanation at present, but a change in luminosity at a fixed
colour (temperature) suggests a change in radius, and the break happens to match the mass where M dwarfs are
predicted to become fully convective.
- (MV, (V-K))
- (MI, (V-I))
- (MI, (R-I))
- (MJ, (I-J))
- (MJ, (J-K))
- (MK, (V-K))
- (MJ, (R-J))
- (MR, (R-K))
- (MK, (I-K))
- (MI, (I-K))
- (B-V), (V-I)
- (B-V), (R-I)
- (V-I), (R-I) Note very small dispersion in these data
- (V-R), (R-I)
- (V-I), (I-J)
- (V-J), (I-J)
- (R-I), (I-K)
- (R-J), (I-J)
- (R-J), (J-K)
- (V-I), (I-K) note a possible offset between the Gliese
stars and the 8 pc sample amongst GK dwarfs: this may reflect a systematic
offset between KS from 2MASS and KCIT from Leggett's
- (J-H), (H-K) with the Leggett et al early-type T dwarfs
- (J-H), (H-K) with main-sequence stars (blue), L dwarfs (magenta),
T dwarfs (cyan), M subdwarfs (yellow open circles) and red giants (yello crosses).
- mid-IR data
(K-L') and (L'-M') as a function of spectral type
(from Reid & Cruz, 2001). Magenta points mark data from Reid & Cruz; red points are
from Berriman & Reid (1987); green points from Leggett et al (2001); and the cyan point
is Gl 229B.
A couple of the above diagrams appeared in Sky & Telescope's August (September?)
article on L & T dwarfs. Slightly modified versions of those diagrams are given below:
The correlations between broadband colours and spectral types are illustrated on
2. Halo subdwarfs
Metal-poor halo main-sequence stars appear to lie below the main-sequent outlined by
nearby (near-solar abundance) disk dwarfs. In fact, reduced metal-line blanketting moves the
stars to higher luminosities and higher temperatures (bluer colours), so a metal-poor star of
a given mass actually lies up and to the left of its disk counterpart in the HR diagram.
Multicolour broadband photometry (at least, reliable photometry) exists for relatively few FGK
dwarfs; most of the available data are included in the datasets listed by Reid et al (2001).
Later-type M subdwarfs are classified under two heading by Gizis (1997): sdM, intermediate
abundance subdwarfs, probably with -0.5 > [m/H] > -1.5; and esdM, extreme subdwarfs,
[m/H] < -1.5. These stars have substantially bluer colours than their main-sequence counterpart -
for example, the hydrogen burning limit for esdM subdwarfs lies at MV~15, (V-I)~3,
rather than MV~20, (V-I)~6. The following colour-magnitude diagrams illustrate
the general behaviour:
- MV, (B-V): blue points are nearby disk dwarfs;
magenta points and red points, FGK subdwarfs (red for dwarfs from Carney et al sample); green
points are sdM; yellow points are esdM. Fiducial main-sequence for 47 Tuc (red), M5 (yellow),
NGC 6397 (green). Note that the late-type subdwarfs are actually redder than disk dwarfs in (B-V).
- MV, (V-I): symbols as above - except
cyan dots are the NGC 6397 main-sequence (from Cool et al, 1996 - see
Reid & Gizis, 1998, AJ 116) and the cyan 5-point stars are esdM dwarfs.;
- MR, (R-I): symbols as in MV, (V-I): . Note that the FGK
subdwarfs are barely subluminous in this CMD - there aren't many strong features at
the relevant wavelengths in those stars.
- (B-V)/(V-I): symbols in MV, (B-V):.
- (V-I)/(R-I): symbols in MV, (B-V):.
Bolometric corrections are vital in comparing observations
of low-mass stars and brown dwarfs to theoretical predictions of fundamental properties.
- V, I and K bolometric corrections
The individual relations are:
BCV = 0.27 - 0.604(V-I) - 0.125(V-I)2, rms=0.059 mag.
BCI = 0.02 + 0.575(V-I) - 0.155(V-I)2, rms=0.061 mag.
BCK = 0.42 + 1.486(I-K) - 0.220(I-K)2, rms=0.048 mag.
from Reid & Hawley (New Light on Dark Stars, 2000, chapter 2), using data from Leggett et al (1996, ApJS 104, 117).
- M-band bolometric corrections from Reid & Cruz, 2001
Photometric data for low-mass stars and brown dwarfs
The above plots are based on the following datafiles. In each case the
N, MU, MB, MV, MR, MI, MJ,
MH, MK, distance, uncertainty in distance, +dm, -dm (the
corresponding uncertainty in absolute magnitude), name
As noted above, all distances are based on trig parallax measurements, primarily Hipparcos or USNO;
RI magnitudes are on Cousins system; JHK on the CIT or 2MASS system. Negative numbers mean (obviously)
no measurements available.
- CNS3/Hipparcos stars all 831 stars from the Bessell BVRI
sample that also have Hipparcos astrometry.
- CNS3 stars Stars classed as single or resolved primaries in
the Bessell dataset; as the CMDs plotted above illustrate, some are obviously unrecognised
- CNS3 stars beyond 8 pc There are ~70 stars from the Bessell dataset
that lie within 8 parsecs and are included in the 8-parsec sample listed below. Some stars have
photometry by Leggett, others have more accurate ground-based parallax data, and those data are included
in the 8-parsec file. To simplify combining datasets, we omit those stars from this file.
- CNS3 stars Subset of the Bessell sample that have JHK data
from 2MASS; also some stars from Leggett's sample (2).
- 8 parsec sample Leggett/Bessell data for stars within 8 parsecs
of the Sun; again, single or resolved primaries/secondaries. If duplicate photometry exists, preference
is given to the Leggett compilation, and ground-based parallax measurements are used in preference to
Hipparcos if the former have higher formal accuracy.
- Ultracool dwarfs Data from USNO (see
Dahn et al, 2002, AJ (in press).
Note that most of the JHK data are on the CIT system. Transformations
between that system and the 2MASS system are given by
- Gl 229B, Gl 570D, 2M0559-14, SDSS1254-16 and SDSS1624+00,
the only T dwarfs with known distances
at the present time (July, 2002). The last three targets are from Dahn et al (op. cit.).
- Updated photometry file August 2004.
CNS3 and 8-parsec data for single starscross-referenced against the final 2MASS data release,
format is NN (PMSU no.), Hip, U, B, V, R, I, J, H, K (apparent mags),
TiO5, CaH1, CaH2, CaH3, CaOH, parallax, sig(pi) (arcsec), mu)RA), mu(dec),
RA, Dec, numerical sp. type, name, sp. type, 2MASS designation, 2MASS photometric uncertainties.
Metal-poor subdwarfs: UBV photometry is available for a large sample of metal-poor stars,
notably from the surveys of Giclas proper motion stars by
Sandage & Fouts, 1986, AJ 91, 1189 and by
Carney et al, 1994, AJ 107, 2240. RI data are much less readily available. The following files
include most of the reliable photometry currently available for metal-poor stars with
reasonably accurate trigonometric parallaxes. All of the data are on the Johnson/Cousins system:
- FGK subdwarf data: literature data plus SAAO photometry of
known subdwarfs from Reid et al. (2001) - MNRAS 325, 931.
- sdM subdwarfs: intermediate abundance subdwarfs
with measured parallaxes from Gizis (1997) AJ 113, 806
- esdM subdwarfs: extreme ([m/H] < -1.5) subdwarfs
with measured parallaxes from Gizis (1997) AJ 113, 806
NStars home page
INR home page
page by Neill Reid, last updated 15/07/02