The Astrophysical Journal, 745:7 (9pp), 2012 January 20 doi:10.1088/0004-637X/745/1/7
C
2012. The American Astronomical Society. All rights reserved. Printed in the U.S.A.
NEOWISE STUDIES OF ASTEROIDS WITH SLOAN PHOTOMETRY: PRELIMINARY RESULTS
A. Mainzer
1
, J. Masiero
1
,T.Grav
2
, J. Bauer
1,3
, D. J. Tholen
4
, R. S. McMillan
5
,E.Wright
6
, T. Spahr
7
,
R. M. Cutri
3
, R. Walker
8
,W.Mo
2
, J. Watkins
6
,E.Hand
1
, and C. Maleszewski
5
1
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109, USA
2
Department of Physics and Astronomy, Johns Hopkins University, Baltimore, MD 21218, USA
3
Infrared Processing and Analysis Center, California Institute of Technology, Pasadena, CA 91125, USA
4
Institute for Astronomy, University of Hawaii, 2680 Woodlawn Drive, Honolulu, HI 96822-1839, USA
5
Lunar and Planetary Laboratory, University of Arizona, 1629 East University Blvd., Kuiper Space Science Bldg. #92, Tucson, AZ 85721-0092, USA
6
UCLA Division of Astronomy and Astrophysics, P.O. Box 91547, Los Angeles, CA 90095-1547, USA
7
Minor Planet Center, Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA
8
Monterey Institute for Research in Astronomy, Monterey, CA, USA
Received 2011 July 14; accepted 2011 October 3; published 2011 December 27
ABSTRACT
We have combined the NEOWISE and Sloan Digital Sky Survey data to study the albedos of 24,353 asteroids
with candidate taxonomic classifications derived using Sloan photometry. We find a wide range of moderate
to high albedos for candidate S-type asteroids that are analogous to the S complex defined by previous
spectrophotometrically based taxonomic systems. The candidate C-type asteroids, while generally very dark,
have a tail of higher albedos that overlaps the S types. The albedo distribution for asteroids with a photometrically
derived Q classification is extremely similar to those of the S types. Asteroids with similar colors to (4) Vesta have
higher albedos than the S types, and most have orbital elements similar to known Vesta family members. Finally, we
show that the relative reflectance at 3.4 and 4.6 μm is higher for D-type asteroids and suggest that their red visible
and near-infrared spectral slope extends out to these wavelengths. Understanding the relationship between size,
albedo, and taxonomic classification is complicated by the fact that the objects with classifications were selected
from the visible/near-infrared Sloan Moving Object Catalog, which is biased against fainter asteroids, including
those with lower albedos.
Key words: atlases – catalogs – infrared: general – minor planets, asteroids: general – surveys
Online-only material: color figures
1. INTRODUCTION
Asteroids and comets represent the leftovers from the forma-
tion of our solar system. By studying their compositional vari-
ation, we can begin to better understand the conditions present
at the earliest stages of planet formation as well as their subse-
quent evolution and processing. Asteroids are grouped into three
main categories: C type or carbonaceous asteroids, thought to
be the most common type in the main belt; the S type or stony
asteroids, a spectrally diverse group, and the X types, another
diverse group of asteroids that have relatively featureless spec-
tra but a wide range of albedos, probably representing a broad
range of mineralogies and thermal histories. Gaffey et al. (1993)
give an overview of some of the earlier taxonomic systems
(e.g., Chapman et al. 1975;Bowelletal.1978; Barucci et al.
1987; Tedesco et al. 1989a, 1989b;Howelletal.1994). Tholen
(1984) created 14 taxonomic classes based on the combina-
tions of colors available from the Eight-Color Asteroid Survey
(ECAS; Zellner et al. 1985; Tholen & Barucci 1989). ECAS
used filter passbands with wavelengths ranging from 0.34 to
1.04 μm, since UV wavelengths were detectable by the photo-
multiplier available at the time. The Tholen and Tedesco sys-
tems also used albedo to differentiate asteroids; the X-group
of asteroids have similar ECAS colors but markedly different
albedos.
The Tholen and Tedesco systems offer a powerful means of
distinguishing unique asteroid groups. However, when CCDs
replaced the photomultipliers at most observatories, they ush-
ered in a factor of several improvement in quantum effi-
ciency at red wavelengths compared to blue; while the ECAS
photomultipliers extended out to 1.1 μm, most CCDs’ responses
were diminished at this wavelength. These effects combined to
make it difficult for observers to collect the full set of UV and
near-infrared bands used by ECAS. Groups such as Xu et al.
(1995), Lazzaro et al. (2004), and Bus & Binzel (2002a) obtained
visible spectra of asteroids and used them for classification with-
out albedos. Bus & Binzel (2002b) created a new taxonomic
scheme relying solely on visible spectroscopy with 26 taxo-
nomic types, and DeMeo et al. (2009) revised that scheme to
24 taxonomic types based on visible and near-infrared (VNIR)
spectral signatures. Neither the Bus & Binzel (2002b) nor
DeMeo et al. (2009) systems rely upon having albedo mea-
surements for classification. Nevertheless, systems based solely
on VNIR spectroscopy and photometry are widely used because
observations at these wavelengths have generally been far more
widely available than albedo measurements for the ∼500,000
asteroids known today. However, as discussed in Gaffey et al.
(2002), the ability to link taxonomic classification to asteroid
mineralogy is complicated by the unknown surface properties
such as particle size, the fact that some minerals have limited or
no spectral features in the wavelengths used for classification,
overlapping features, etc. It is therefore useful to try to under-
stand how t he various taxonomic types are linked to physical
properties such as albedo and density.
The fourth release of the Sloan Digital Sky Survey
(SDSS) Moving Object Catalog (MOC; Stoughton et al. 2002;
Abazajian et al. 2003) provided near-simultaneous observations
of ∼100,000 known asteroids in five bands (u, g, r, i, and z)
(Ivezi
´
cetal.2001), and these have been used to study the distri-
bution of colors t hroughout the main belt (cf. Parker et al. 2008;
1
The Astrophysical Journal, 745:7 (9pp), 2012 January 20 Mainzer et al.
Figure 1. Visible albedo as a function of size for the nine individual classes defined by Carvano et al. (2010). The classes have been separated into two sub-panels for
clarity. The bias of the SDSS survey against small, low-albedo objects is evident.
(A color version of this figure is available in the online journal.)
Nesvorn
´
yetal.2005). These studies have been conducted
largely without reference to albedo, simply because the num-
ber of asteroids in the SDSS MOC vastly outnumbers those
with well-measured albedos. To date, the Infrared Astronomical
Satellite (Tedesco et al. 2002) has been the largest source of
radiometrically measured asteroid albedos, providing measure-
ments for ∼2000 asteroids.
With the Wide-field Infrared Survey Explorer’s (WISE)
NEOWISE project (Wright et al. 2010; Mainzer et al. 2011a),
thermal observations of >157,000 asteroids are now in hand. In
Mainzer et al. (2011e), we compared the albedos derived from
NEOWISE observations of ∼1900 asteroids with taxonomic
types derived from VNIR spectroscopy. Here, we use the SDSS
MOC photometry to obtain approximate taxonomic classifica-
tions for asteroids with NEOWISE observations. We initially
focus on the system of Carvano et al. (2010), who define a clas-
sification algorithm based on SDSS colors that is compatible
with previous taxonomic systems.
2. OBSERVATIONS
WISE surveyed the entire sky in four infrared wavelengths,
3.4, 4.6, 12, and 22 μm ( denoted W 1, W 2, W 3, and W 4,
respectively). Descriptions of the pre-launch mission design and
testing can be found in Liu et al. (2008) and Mainzer et al. (2005),
and the post-launch description is given in Wright et al. (2010).
A series of enhancements to the WISE data processing pipeline,
known as NEOWISE, have enabled the detection of >157,000
asteroids and comets throughout the solar system, including the
discovery of ∼34,000 new minor planets (Mainzer et al. 2011a).
A total of 24,353 objects, including nine NEOs and ∼24,275
main belt asteroids (MBAs), were detected during the fully
cryogenic portion of the NEOWISE survey and had matches
with SDSS MOC observations of sufficient quality to enable
classification according to the method described in Carvano
et al. (2010). The observations of these objects were extracted
from the WISE archive using the First Pass version of the
2
The Astrophysical Journal, 745:7 (9pp), 2012 January 20 Mainzer et al.
Figure 2. NEOWISE-derived albedos of asteroids observed and classified by Carvano et al. (2010). Only classes with more than ∼100 asteroids are plotted, and these
include some objects that have multiple SDSS observations that produced different classifications. The median p
V
value is shown as a vertical red line.
(A color version of this figure is available in the online journal.)
WISE data processing pipeline (Cutri et al. 2011) following the
methods and parameters given in Mainzer et al. (2011c, hereafter
M2), Mainzer et al. (2011d, M3), and Mainzer et al. (2011e,
M4).
3. PRELIMINARY THERMAL MODELING
We have created preliminary thermal models for each asteroid
using the First-Pass Data Processing Pipeline described above.
As described in references M2, M3, and M4, we employ the
near-Earth asteroid thermal model (NEATM) of Harris (1998).
The NEATM model uses the so-called beaming parameter η
to account for cases intermediate between zero thermal inertia
(Lebofsky & Spencer 1989, the Standard Thermal Model) and
high thermal inertia (Lebofsky et al. 1978, the Fast Rotating
Model; Veeder et al. 1989, the Fast Rotating Model; Lebofsky
& Spencer 1989, the Fast Rotating Model). With NEATM, η
is a free parameter that can be fit when two or more infrared
bands are available. Bands W 1 and W 2 typically contain a
mix of reflected sunlight and thermal emission. The flux from
reflected sunlight was computed for each WISE band using
the methods described in M2, M3, and M4; when sufficient
reflected sunlight was present in bands W 1 and W 2, it was
possible to compute the reflectivity at these wavelengths, p
IR
,
where we make the assumption that p
IR
= p
3.4
= p
4.6
.The
validity of this assumption and the meaning of p
IR
is discussed
in M4 and will be the subject of future work. As described in
M2 and M3, the minimum diameter error that can be achieved
using WISE observations is ∼10%, and the minimum relative
albedo error is ∼20% for objects with more than one WISE
thermal band for which η can be fitted. For objects with
large amplitude light curves, poor H measurements, or poor
signal-to-noise measurements in the WISE bands, the errors
will be higher.
3
The Astrophysical Journal, 745:7 (9pp), 2012 January 20 Mainzer et al.
Tab le 1
Median Values of p
V
and p
IR
/p
V
using NEOWISE Cryogenic Observations of Asteroids with Taxonomic Types Derived from SDSS Colors
According to the Method of Carvano et al. (2010) and Hasselmann et al. (2011)
Class N Median SD Min Max N Median SD Min Max N Median SD N 50 Median 50 SD
p
V
p
V
p
IR
/p
V
p
IR
/p
V
SD Min Max p
V
50 p
V
50 50 p
IR
/p
V
p
IR
/p
V
50
Sp 4880 0.262 ± 0.001 0.084 0.034 0.779 474 1.452 ± 0.028 0.603 0.607 4.945 2725 0.267 ± 0.002 0.080 362 1.449 ± 0.030 0.570
Cp 9779 0.064 ± 0.001 0.055 0.013 1.000 566 1.147 ± 0.032 0.765 0.260 4.971 3069 0.067 ± 0.001 0.054 374 1.095 ± 0 .035 0.680
Xp 1773 0.106 ± 0.003 0.118 0.020 1.000 201 1.318 ± 0.044 0.625 0.507 4.801 502 0.109 ± 0.006 0.127 138 1.322 ± 0.056 0.658
Lp 1838 0.202 ± 0.002 0.085 0.030 0.638 155 1.263 ± 0.047 0.580 0.577 4.036 768 0.204 ± 0.003 0.087 109 1.252 ± 0.050 0.518
Dp 984 0
.080 ± 0.002 0.073 0.014 0.516 104 2.079 ± 0.073 0.744 1.034 4.543 235 0.073 ± 0.003 0.041 67 2.165 ± 0.094 0.766
Ap 85 0.248 ± 0.009 0.087 0.081 0.488 9 2.526 ± 0.212 0.637 1.307 3.166 20 0.273 ± 0.018 0.082 4 2.187 ± 0.262 0.523
Qp 424 0.253 ± 0.005 0.110 0.038 0.613 14 1.456 ± 0.106 0.398 0.939 2.144 118 0.295 ± 0.009 0.093 7 1.100 ± 0.172 0.455
Op 16 0.076 ± 0.039 0.157 0.035 0.436 0 0.000 ± 0.000 0.000 0.000 0.000 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
Vp 650 0.
343 ± 0.004 0.105 0.047 0.771 47 1.470 ± 0.092 0.631 0.886 4.046 288 0.352 ± 0.006 0.100 29 1.560 ± 0.099 0.532
CSp 1 0.130 ± 0.000 0.000 0.130 0.130 0 0.000 ± 0.000 0.000 0.000 0.000 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
XSp 36 0.203 ± 0.017 0.100 0.044 0.507 2 0.885 ± 0.040 0.056 0.829 0.941 1 0.507 ± 0.000 0.000 1 0.829 ± 0.000 0.000
LSp 1488 0.240 ± 0.002 0.083 0.042 0.623 124 1.476 ± 0.043 0.479 0.673 3.472 396 0.247 ± 0.004 0.083 79 1.449 ± 0.054 0.480
SQp 453 0.252
± 0.004 0.092 0.062 0.617 17 1.561 ± 0.117 0.484 1.069 3.085 93 0.262 ± 0.009 0.084 8 1.444 ± 0.098 0.276
SAp 31 0.273 ± 0.016 0.091 0.135 0.451 3 1.672 ± 0.125 0.216 1.299 1.811 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
SVp 44 0.316 ± 0.015 0.097 0.111 0.549 2 1.874 ± 0.407 0.575 1.299 2.450 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
CXp 1144 0.067 ± 0.003 0.087 0.020 0.999 67 1.243 ± 0.104 0.851 0.643 5.031 104 0.069 ± 0.011 0.110 20 1.122 ± 0.062 0.276
CDp 22 0.064 ±
0.012 0.057 0.030 0.240 0 0.000 ± 0.000 0.000 0.000 0.000 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
CQp 6 0.139 ± 0.017 0.042 0.083 0.192 0 0.000 ± 0.000 0.000 0.000 0.000 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
XLp 170 0.159 ± 0.005 0.060 0.027 0.452 10 1.116 ± 0.122 0.387 0.491 2.015 26 0.160 ± 0.016 0.081 5 1.059 ± 0.099 0.221
XDp 209 0.076 ± 0.006 0.091 0.018 0.715 11 1.628 ± 0 .172 0.569 0.996 2.689 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
DLp 170 0.178 ± 0
.007 0.091 0.044 0.531 9 1.610 ± 0.148 0.445 1.062 2.456 2 0.214 ± 0.085 0.120 0 0.000 ± 0.000 0.000
QVp 40 0.352 ± 0.015 0.097 0.088 0.575 1 1.488 ± 0.000 0.000 1.488 1.488 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
QLp 5 0.249 ± 0.031 0.069 0.145 0.361 0 0.000 ± 0.000 0.000 0.000 0.000 0 0.000 ± 0.000 0.000 0 0.000 ± 0.000 0.000
Notes. The number of objects, median, standard deviation of the mean, standard deviation (SD), minimum and maximum p
V
and p
IR
/p
V
are given for all the objects with a particular classification. Spectral classes
with median p
IR
/p
V
= 0.000 did not have enough asteroids with measurements in W 1andW 2 to compute p
IR
/p
V
. Columns with a “50” in the heading indicate the statistical properties of only those asteroids that
had taxonomic classifications assigned by Hasselmann et al. (2011) with a probability of 50% or greater.
4
The Astrophysical Journal, 745:7 (9pp), 2012 January 20 Mainzer et al.
Figure 3. NEOWISE-derived ratio p
IR
/p
V
for asteroids observed and classified according to the system of Carvano et al. (2010). Only asteroids for which p
IR
/p
V
could be fitted are included in this plot. While the C
p
and D
p
classes have similar p
V
values, p
IR
/p
V
is distinctly different. The median p
IR
/p
V
value is shown as a
vertical red line.
(A color version of this figure is available in the online journal.)
3.1. High-albedo Objects
We note that among the 24,353 asteroids considered here,
there are ∼55 that have p
V
> 0.65. Of these, 48 have W 3 peak-
to-peak variations >0.3 mag, indicating that they are likely to
be highly elongated. Almost all of the extremely high-albedo
objects have orbital elements consistent with membership in
either the Vesta family or the Hungarias. Harris & Young (1988)
and Harris et al. (1989) noted that E- and V-type asteroids
can have slope values as high as G ∼ 0.5. The assumption
that we have used of G = 0.15 for an object like this would
cause an error in the computed H for observations at 20
◦
phase angle of ∼0.3 mag; this would drive the albedo derived
using such an H value up by 0.3. These objects would greatly
benefit from an improved determination of their H and G
values.
4. DISCUSSION
The Carvano scheme uses the SDSS colors as well as
their measurement uncertainties to define nine spectral classes:
V
p
, O
p
, S
p
, A
p
, L
p
, D
p
, X
p
, Q
p
, and C
p
. These are roughly
analogous to the spectroscopically defined systems of Bus &
Binzel (2002b) and DeMeo et al. (2009); the “p” indicates
that the classification was derived photometrically. The nine
Carvano classes were defined based on the ability of the SDSS
colors to represent the system of Bus & Binzel (2002b). The
exception to this is the
L
p
class, which is a conglomeration of
the Bus K, L, and Ld classes. For each SDSS observation, the
probability that an object could be associated with a particular
class was computed using the five SDSS magnitudes and their
associated uncertainties; classifications and probabilities for
∼63,000 asteroids were taken from Hasselmann et al. (2011).
5
