Publications
12-20-2005
The Dust Cloud Around the White Dwarf G29-38 The Dust Cloud Around the White Dwarf G29-38
William T. Reach
California Institute of Technology
Ted von Hippel
University of Texas at Austin
, vonhippt@erau.edu
et al.
Follow this and additional works at: https://commons.erau.edu/publication
Part of the Stars, Interstellar Medium and the Galaxy Commons
Scholarly Commons Citation Scholarly Commons Citation
Reach, W. T., von Hippel, T., & al., e. (2005). The Dust Cloud Around the White Dwarf G29-38.
The
Astrophysical Journal, 635
(2). Retrieved from https://commons.erau.edu/publication/261
This Article is brought to you for free and open access by Scholarly Commons. It has been accepted for inclusion in
Publications by an authorized administrator of Scholarly Commons. For more information, please contact
commons@erau.edu.
L161
The Astrophysical Journal, 635:L161–L164, 2005 December 20
䉷 2005. The American Astronomical Society. All rights reserved. Printed in U.S.A.
THE DUST CLOUD AROUND THE WHITE DWARF G29-38
William T. Reach,
1
Marc J. Kuchner,
2
Ted von Hippel,
3
Adam Burrows,
4
Fergal Mullally,
3
Mukremin Kilic,
3
and D. E. Winget
3
Received 2005 September 17; accepted 2005 November 11; published 2005 December 7
ABSTRACT
We present new observations of the white dwarf G29-38 with the camera (4.5 and 8 mm), photometer (24 mm),
and spectrograph (5.5–14 mm) of the Spitzer Space Telescope. This star has an exceptionally large infrared excess,
amounting to 3% of the bolometric luminosity. The spectral energy distribution (SED) has a continuum peak around
4.5 mm and a 9–11 mm emission feature 1.25 times brighter than the continuum. A mixture of amorphous olivine and
a small amount of forsterite in an emitting region 1–5 from the star can reproduce the shape of the 9–11 mmR
,
feature. The SED also appears to require amorphous carbon to explain the hot continuum. Our new measurements
support the idea that a relatively recent disruption of a comet or asteroid created the cloud.
Subject heading: infrared: stars — stars: individual (G29-38, WD 2326⫹049) — white dwarfs
1. INTRODUCTION
White dwarfs offer a unique view into the properties of
planetary systems. A star can retain much of its planetary sys-
tem as it leaves the main sequence and evolves through the
red giant stage. Although the inner ∼5 AU of a planetarysystem
may evaporate within the red giant atmosphere, and planets
that were in marginally stable orbits may escape during stellar
mass loss, most outer planets and even much of the Oort-Cloud-
like extremities of planetary systems should persist around
white dwarfs (Debes & Sigurdsson 2002).
The white dwarf Giclas 29-38 (WD 2326⫹049; McCook &
Sion 1987; G29-38 hereafter) has garnered intense interestsince
Zuckerman & Becklin (1987) discovered infrared emission
from this object far in excess of the photosphere. The excess
emission, confirmed both by ground-based (Tokunaga et al.
1990) and space-based (Chary et al. 1999) observations, begins
to deviate from a pure photosphere in the near-infrared. While
it was initially speculated that a Jupiter-sized companion could
supply all the infrared excess, follow-up observations (e.g.,
Kuchner et al. 1998; Graham et al. 1990) appear to have ruled
this out. The colors measured with ISOCAM suggested that
the infrared excess is due to particulate matter rather than a
brown dwarf companion (Chary et al. 1999).
Besides strong hydrogen absorption lines, G29-38 has some
photospheric metal lines, including Ca ii which should rapidly
(
!10
4
yr) sink out of the atmosphere due to gravitational sed-
imentation (Alcock & Illarionov 1980; Koester et al. 1997;
Zuckerman & Reid 1998). Possible explanations for the atmo-
spheric metals and the infrared excess include interstellar me-
dium (ISM) accretion (Dupuis et al. 1993; Koester et al. 1997;
Zuckerman & Reid 1998), cometary breakup (Alcock et al.
1986), and asteroid disruption (Jura 2003).
We turned the powerful instruments of the Spitzer Space Tele-
scope (Werner et al. 2004) toward G29-38 to measure its bright-
ness from 4.5–24 mm. These observations probably probe a plan-
etary system in a late stage of evolution. G29-38’smain-sequence
progenitor was probably an A star with mass ∼3.1 (Wei-M
,
1
Spitzer Science Center, MS 220-6, California Institute of Technology, Pas-
adena, CA 91125; reach@ipac.caltech.edu.
2
NASA Goddard Space Flight Center, Greenbelt, MD 20771.
3
Department of Astronomy, University of Texas, 1 University Station
C1400, Austin, TX 78712.
4
Department of Astronomy and Steward Observatory, University of Ari-
zona, 933 North Cherry Avenue, Tucson, AZ 85721.
demann 2000), similar to many stars known to have debris
disks and potential planetary systems (Rieke et al. 2005). The
post-AGB phase age is estimated to be ∼ yr (Bergeron
8
5 # 10
et al. 1995a).
2. OBSERVATIONS
Figure 1 shows the spectral energy distribution (SED) of G29-
38. Observations with the Infrared Array Camera (IRAC; Fazio
et al. 2004) were performed on 2004 November 26 10:54 UT.
At each of five dithers in a Gaussian spatial distribution, a 30 s
frame was taken with the 4.5 and 8 mm arrays. We extracted our
photometry using the method (Reach et al. 2005) used for the
IRAC calibration stars. The quoted 4.5 and 8 mm fluxes, mea-
sured from the basic calibrated data, are and
8.8 Ⳳ 0.3 8.2 Ⳳ
mJy, respectively. We calculated color corrections of 0.995
0.3
and 1.17 in the 4.5 and 8 mm channels, respectively, making the
estimated fluxes 9.2 and 7.0 mJy at the IRAC nominal wave-
lengths of 4.493 and 7.782 mm. Observations with the Multiband
Imaging Photometer for Spitzer (MIPS; Rieke et al. 2004) were
performed on 2004 December 2 04:45 UT. Three cycles of small-
scale photometry dithers were taken with 10 s frames (total
exposure time 420 s); the flux at 24 mmis2.4mJy.
Observations with the Infrared Spectrograph (IRS; Houck et
al. 2004) were performed on 2004 December 8 3:26 UT. The
source was first observed on the blue filter portion of the peak-
up array; the flux measured from the 16 mm peak-up image is
3.7 mJy. The spectrum from 5.2 to 14.4 mm was assembled by
differencing observations at two nods on each of the four spectral
orders. The flux derived from the IRS spectrum convolved with
the IRAC 8 mm bandpass is consistent with the IRAC flux.
The IRS spectrum shows continuum emission and a strong
9–11 mm feature characteristic of silicate minerals. It shows
none of the bands at 6.2, 7.7, 8.6, 11.3, or 12.6 mm characteristic
of polycyclic aromatic hydrocarbons (PAHs) that normally
dominate mid-infrared spectra of the ISM. The excess emission
above photospheric is approximated by two modified black-
bodies with temperatures 890 and 290 K, dilution factors (at
10 mm wavelength) and , and emis-
⫺16 ⫺15
4.2 # 10 6.3 # 10
sivity proportional to n
0.5
. These functions are only intended as
mathematical approximations of the continuum shape.
L162 REACH ET AL. Vol. 635
Fig. 1.—SED of the white dwarf G29-38 from the ultraviolet through in-
frared. The infrared observations are described in § 2, and the ultraviolet
through near-infrared observations and model are described in § 3. The solid
line through the infrared data is a modified blackbody fit to the continuum.
Fig. 2.—Continuum-subtracted spectrum of the silicate feature from G29-
39, compared to similarly subtracted silicate features of the diffuse ISM, the
O-rich star Mira, comet Hale-Bopp, and the zodiacal light.
3. MODELING
We adopt the following basic parameters for G29-38. It is a
DA4-type white dwarf, with atmospheric spectrum dominated by
H lines (Green et al. 1986). It has an assumed mass of 0.69 M
,
,
surface gravity , and temperature 11,800 K (Ber-log g p 8.14
geron et al. 1995b). The parallax is (Har-p p 0⬙.071 Ⳳ 0⬙.004
rington & Dahn 1980), implying a distance of 14 pc.
We assembled the SED from the International Ultraviolet
Explorer (IUE) low-dispersion spectrum covering 0.115–
0.3148 mm (Holberg et al. 2003) and a model atmosphere for
K and covering 0.35–60 mm (D. KoesterT p 12,000 log g p 8
2005, private communication). We used the 2MASS photom-
etry ( , ,J p 13.132 Ⳳ 0.029 H p 13.075 Ⳳ 0.022 K p
s
) and optical spectrophotometry from Palomar12.689 Ⳳ 0.029
(Greenstein & Liebert 1990) for normalization, giving extra
weight to the 2MASS J-band photometry, the longest wave-
length where we think the photosphere dominates. The inte-
grated luminosity of the star is .
⫺3
2 # 10 L
,
The 9–11 mm emission feature indicates silicates. Figure 2
compares the shape of the G29-38 silicate feature to that of
interstellar dust (Kemper et al. 2004), solar system zodiacal
light, the O-rich mass loss from the star Mira (Sloan et al.
2003), and comet Hale-Bopp (C/1995 O1; Crovisier et al.
1997). The G29-38 silicate feature is redder than that of the
ISM, which has a single, rounded peak at 9.7 mm and nearly
linear slopes on either side. This suggests that the material
around G29-38 is not accreted ISM but rather is indigenous to
the star. The G29-38 silicate feature is more compact than the
zodiacal light silicate feature and than the exozodiacal light
feature around b Pic (Reach et al. 2003). The Hale-Bopp spec-
trum has a prominent 11.3 mm peak that is not seen in the G29-
38 spectrum. The Mira silicate feature is somewhat redder and
wider than the ISM feature, roughly intermediate between the
ISM and G29-38 feature shape. The G29-38 silicate feature
shape is most similar to (and approximately intermediate be-
tween) those of Mira and the zodiacal light, but the line-to-
continuum ratio of G29-38 is much greater than that of the
zodiacal light (125% vs. 6%), suggesting a very different par-
ticle size distribution.
We computed theoretical emission spectra for grains (size
–1000 mm, ) of various compositions
⫺3.5
a p 0.01 dn/da ∝ a
(amorphous carbon, crystalline enstatite, amorphous pyroxene
[50/50 Mg/Fe], amorphous olivine [50/50] and crystalline for-
sterite) around G29-38 to further elucidate the properties of the
dust around that star. The temperatures and emissivities over a
range of distances (0.003–3.5 AU p 0.6–750 R
,
) from the white
dwarf were integrated over a spherical, presumed optically thin
shell with a radial profile and an inner cutoff at R
min
.
⫺a
n ∝ r
The temperatures of 0.5 mm radius grains of amorphous carbon
(olivine) are 600 K (920 K) at 3 R
,
and 130 K (110 K) at
1 AU from the white dwarf. Grains hotter than 1500 K are
presumed to be sublimated and are not included in the emission.
Since we assume that the cloud is optically thin, we cannot
further constrain the geometry of the emitting region, which
could range from a spherical shell to a flattened disk.
The G29-38 silicate feature resembles the amorphous olivine
model, with excess on the red side that can be accounted for
by mixing in forsterite. Pyroxene cannot contribute signifi-
cantly to the observed spectrum; its 9–11 mm feature is too
blue. Amorphous olivine and fosterite alone cannot explain the
spectrum, however, because they cannot emit at 3–6 mm without
also producing a 1.6 mm feature that exceeds the near-infrared
H-band flux. The near-infrared continuum shape at these wave-
lengths matches the amorphous carbon model. The contrast of
the silicate feature demands a particle size distribution that
favors small particles, unlike that of interplanetary dust grains
(Gru¨n et al. 1985). On the assumption that all the particles
participate in the same collisional evolution, we assume that
silicate and carbon particles share the same size distribution.
No. 2, 2005 DUST CLOUD AROUND G29-38 L163
Fig. 3.—Infrared excess of G29-38 compared to the best-fit model (solid
line), consisting of three compositions: amorphous carbon (dotted line), amor-
phous olivine (long-dashed line), and forsterite (dot-dashed line).
The distance of the emitting region from the star is con-
strained as follows. In order to produce significant emission at
5 mm, we require . The best-fitting models all re-R
! 5 R
min ,
quire small R
min
, implying that the dust exists nearly up to its
sublimation temperature. The radial profile shape requires high
a, suggesting that the emitting region extends from 1 to 10 R
,
from the star. Figure 3 shows the observed spectrum and the
best-fit model. Experiments with other geometries yielded sim-
ilar results: a face-on slab required , . Aa ∼ 3 R ∼ 2.5 R
min ,
dust cloud this small is consistent with the measurements of
Kuchner et al. (1998), which showed G29-38 to be unresolved
in the K band at 55 mas (160 R
,
) resolution.
4. CONCLUSIONS
A cloud of small grains 1–10 R
,
from G29-38 creates its
mid-infrared emission. The luminosity of the infrared excess
is 3% of the luminosity of the star, high by the standards of
debris disks, indicating nontrivial disk opacity in the ultravi-
olet–visible range.
Our models suggest a dust grain abundance ratio (by number)
of olivine : carbon : forsterite of 5 : 12 : 2. Considering only the
atoms in these grains, assuming material densities of 2.2 and
2.5gcm
⫺3
for the amorphous carbon and silicates, the abun-
dance ratio (by number) of C : O is 3 : 1. The total dust mass
required to generate the observed mid-infrared flux is of the
order of 10
18
g; a larger mass could be present if there are
larger, cooler grains that do not contribute to the observed flux.
This inferred mass corresponds to that of a ∼10 km diameter
asteroid or comet, and is slightly less than the mass of the
interplanetary dust responsible for the zodiacal light in the solar
system (Fixsen & Dwek 2002).
Where could this dust cloud have come from? It was not
generated by the white dwarf (by any known mechanism), nor
is it likely to be planetary system material that was at this
location before the star was a red giant, whose atmosphere
would have extended much further than the current dust cloud.
The silicate feature is suggestive of dust formed in O-rich mass
loss during the red giant or AGB phase, but the presence of
significant and comparable amounts of silicate and carbona-
ceous dust (and no PAH) does not seem compatible with such
dust, which in any event could not have survived at 10 R
,
from the star. The material must have been transported inward.
If the material originates from the planetary system of the
progenitor, then gravitational perturbations would occasionally
send smaller bodies close to the star. If the small body is col-
lisionally fragmented, loses mass by cometary sublimation, or
breaks apart by thermal or gravitational stress, the Poynting-
Robertson effect will cause particles of radius a (mm) and dis-
tance r (R
,
) to spiral into the star on timescales of yr,
2
4ra
i.e., a few years for material contributing to the mid-infrared
emission. The particles would continue toward the star until
they sublimate, suffer mutual collisions, and disrupt, with part
of the material landing in the photosphere, part being blown
out of the system by radiation pressure, and part remaining on
bound orbits. If the entire mid-infrared–emitting mass (i.e.,
particles out to ∼10 R
,
) were lost in the Poynting-Robertson
timescale, the accretion rate is of the order of 10
15
gyr
⫺1
. This
flow could be supplied by collisional comminution of debris
from an asteroidal and cometary cloud like the one in the solar
system (Sykes & Greenberg 1986). This mass-loss rate is com-
parable to the “high” ISM accretion scenario calculated by
Dupuis et al. (1993), which appears inconsistent with G29-38’s
present location in a low-density ISM environment.
A plausible model for the infrared excess around G29-38
involves the tidal disruption of an asteroid near the white dwarf
(Graham et al. 1990; Jura 2003). Since short-period comets are
composed largely of refractory material, we generalize this
model to include the tidal disruption of a comet. Comets are
commonly observed to pass near the Sun, so by analogy a
surviving reservoir of comets (and possibly a surviving massive
outer planet) would inevitably lead to star-grazing cometary
passages. With its low luminosity, the white dwarf would not
drive a high sublimation rate for star-grazing comets, reducing
the related stresses that appear to cause Sun-grazing comets to
split. Comets may then travel rather close to the white dwarf,
where they could be disrupted by tidal forces. Our new ob-
servations support this model, since the emitting region lies
close to the star, where a large solid body would be tidally
disrupted. Based on the high C/O ratio in the dust, it is also
conceivable that the “comet” was condensed out of the AGB
mass loss (Kuchner & Saeger 2005). The details of the tidal
disruption model presented here are different from those in
previous papers, which assume larger grains and an optically
thick disk. If the emitting region is a disk, it is either optically
thin or viewed relatively face-on, with the line of sight from
the white dwarf to the Sun relatively free of dust. For a thin
disk to absorb 3% of the luminosity of the star it would have
to be opaque in the ultraviolet, where there is no evidence of
significant extinction and no 2175 absorption bump in the
˚
A
IUE spectrum.
An alternate and long-standing model for both the photo-
spheric metals and the infrared excess is dredge-up and mass
loss. There are already difficulties with dredge-up models as-
sociated with the need for vertical velocities that are at odds
with G29-38’s g-mode pulsations that are overwhelmingly hor-
izontal. Furthermore, standard white dwarf models are layered
from prior nuclear burning; hypothetical mixing from the pul-
sations would have to reach many orders of magnitude below
the convection zone boundary, well below even the degeneracy
boundary (∼10
⫺6
M
,
) to dredge up material from the C/O layer
(presumably around 10
⫺3
M
,
or deeper). Furthermore, the
dredge-up hypothesis cannot explain the abundant silicates;
invoking these materials just below the He layer is very difficult
to accommodate within our current understanding of nuclear
L164 REACH ET AL. Vol. 635
burning of He in post–main-sequence stars appropriate to white
dwarf stars in the mass range of G29-38. Our observations
therefore argue that the observed material was not formed in
the white dwarf, but formed well outside its current position.
The presence of a dust cloud around G29-38 is likely related
to the anomalous presence of metals in its photosphere. As
discussed by Zuckerman et al. (2003), the photospheric metals
could be accreted from interstellar gas, but then the star would
need to be within a dense cloud (because the lifetime of metals
in the photosphere [
!10
4
yr] is smaller than the crossing time
of a dense cloud [ⲏ10
4
yr]), which is not true for G29-38. The
other major model for white dwarf photospheric metals is com-
etary impact (Zuckerman & Reid 1998; Debes & Sigurdsson
2002). On the basis of the presence of a strong silicate feature
and the overall mid-infrared SED, we believe this latter model
prevails for G29-38.
The abundance of photospheric metals for G29-38 is among
the highest of any known for a white dwarf, and the infrared
excess is also the highest of any known white dwarf. Another
white dwarf, WD 1337⫹705 (G238-44), with even higher
atmospheric Ca abundance (Zuckerman et al. 2003), does not
show infrared excess at the level seen in G29-38 in our Spitzer
photometric survey of white dwarfs at 4.5 and 8 mm. It is
possible that a recent disruption of an asteroid or comet has
occurred in G29-38, repopulating the star’s dust cloud at a
level much higher than the long-term steady state, as indeed
occurs in debris disks around main-sequence stars (Rieke et al.
2005) and the solar system (Sykes & Greenberg 1986).
Are there planets around G39-38? The 9–11 mm spectral
feature proves that there is a cloud of small silicate grains. We
ascribed the 3–6 mm continuum to amorphous carbon dust, and
the carbon-silicate mix matches the SED, but there is no spec-
tral signature for carbon other than its color temperature. If we
ascribe the 890 K continuum to a planet, then it must be not
only very hot but also very large, with radius 0.2 R
,
—an
extremely unusual object. If the dust cloud is indeed due to an
asteroid or comet, perhaps other, cooler planets await discovery
around the fascinating star G29-38.
This work is based in part on observations made with the
Spitzer Space Telescope, which is operated by the Jet Propul-
sion Laboratory, California Institute of Technology under
NASA contract 1407. Support for this work was provided by
NASA through award project NBR 1269551 issued by JPL/
Caltech to the University of Texas.
REFERENCES
Alcock, C., Fristrom, C. C., & Siegelman, R. 1986, ApJ, 302, 462
Alcock, C., & Illarionov, A. 1980, ApJ, 235, 534
Bergeron, P., Wesemael, F., & Beauchamp, A. 1995a, PASP, 107, 1047
Bergeron, P., Wesemael, F., Lamontagne, R., Fontaine, G., Saffer, R. A., &
Allard, N. F. 1995b, ApJ, 449, 258
Chary, R., Zuckerman, B., & Becklin, E. E. 1999, in The Universe as Seen
by ISO, ed. P. Cox & M. F. Kessler (ESA-SP 427; Garching: ESA), 289
Crovisier, J., Leech, K., Bockelee-Morvan, D., Brooke, T. Y., Hanner, M. S.,
Altieri, B., Keller, H. U., & Lellouch, E. 1997, Science, 275, 1904
Debes, J. H., & Sigurdsson, S. 2002, ApJ, 572, 556
Dupuis, J., Fontaine, G., & Wesemael, F. 1993, ApJS, 87, 345
Fazio, G. G., et al. 2004, ApJS, 154, 10
Fixsen, D. J., & Dwek, E. 2002, ApJ, 578, 1009
Graham, J. R., Matthews, K., Neugebauer, G., & Soifer, B. T. 1990, ApJ, 357,
216
Green, R. F., Schmidt, M., & Liebert, J. 1986, ApJS, 61, 305
Greenstein, J. L., & Liebert, J. W. 1990, ApJ, 360, 662
Gru¨n, E., Zook, H. A., Fechtig, H., & Giese, R. H. 1985, Icarus, 62, 244
Harrington, R. S., & Dahn, C. C. 1980, AJ, 85, 454
Holberg, J. B., Barstow, M. A., & Burleigh, M. R. 2003, ApJS, 147, 145
Houck, J. R., et al. 2004, ApJS, 154, 18
Jura, M. 2003, ApJ, 584, L91
Kemper, F., Vriend, W. J., & Tielens, A. G. G. M. 2004, ApJ, 609, 826
Koester, D., Provencal, J., & Shipman, H. L. 1997, A&A, 320, L57
Kuchner, M. J., Koresko, C. D., & Brown, M. E. 1998, ApJ, 508, L81
Kuchner, M. J., & Saeger, S. 2005, ApJ, submitted
McCook, G. P., & Sion, E. M. 1987, ApJS, 65, 603
Reach, W. T., Morris, P., Boulanger, F., & Okumura, K. 2003, Icarus, 164,
384
Reach, W. T., et al. 2005, PASP, 117, 978
Rieke, G. H., et al. 2004, ApJS, 154, 25
———. 2005, ApJ, 620, 1010
Sloan, G. C., Kraemer, K. E., Price, S. D., & Shipman, R. F. 2003, ApJS,
147, 379
Sykes, M. V., & Greenberg, R. 1986, Icarus, 65, 51
Tokunaga, A. T., Becklin, E. E., & Zukerman, B. 1990, ApJ, 358, L21
Weidemann, V. 2000, A&A, 363, 647
Werner, M. W., et al. 2004, ApJS, 154, 1
Zuckerman, B., & Becklin, E. E. 1987, Nature, 330, 138
Zuckerman, B., Koester, D., Reid, I. N., & Hu¨nsch, M. 2003, ApJ, 596, 477
Zuckerman, B., & Reid, I. N. 1998, ApJ, 505, L143
