Black-Body Stars [620514]

Black-Body Stars
Nao Suzuki
Kavli Institute for the Physics and Mathematics of the Universe, University of Tokyo,
Kashiwa 277-8583 Japan
Masataka Fukugita
Kavli Institute for the Physics and Mathematics of the Universe, University of Tokyo,
Kashiwa 277-8583 Japan
Institute for Advanced Study, Princeton NJ08540, U.S.A.
ABSTRACT
We report the discovery of stars that show spectra very close to the black-body
radiation. We found 17 such stars out of 798,593 stars in the Sloan Digital Sky Survey
(SDSS) spectroscopic data archives. We discuss the value of these stars for the cali-
bration of photometry, whatever is the physical nature of these stars. This gives us a
chance to examine the accuracy of the zero point of SDSS photometry across various
passbands: we conclude that the zero point of SDSS photometric system is internally
consistent across its five passbands to the level below 0.01 mag. We may also examine
the consistency of the zero-points between UV photometry of Galaxy Evolution Ex-
plorer and SDSS, and IR photometry of Wide-field Infrared Survey Explorer against
SDSS. These stars can be used as not only photometric but spetrophotometric standard
stars. We suggest that these stars showing the featureless black-body like spectrum of
the effective temperature of 100001500 K are consistent with DB white dwarfs with
the temperature too low to develop helium absorption features.
1. Introduction
We report the discovery of stars that have nearly a perfect black-body spectrum without any
features in the wavelength range from ultraviolet to infrared. A search is made for objects with
featureless spectrum in the spectral archives of the Sloan Digital Sky Survey (SDSS)(Abazajian et
al. 2009; Alam et al. 2015), supplemented with Galaxy Evolution Explorer (GALEX)(Morrissey et
al. 2007) and Wide-field Infrared Survey Explorer (WISE)(Wright et al. 2010). We find 17 objects
whose spectral energy distribution is consistent with a black-body radiation to a high accuracy inarXiv:1711.01122v1 [astro-ph.SR] 3 Nov 2017

2
the UV (GALEX; FUV , NUV), the optical (SDSS; u;g;r;i;z ,) and in the infrared (WISE; W1).
These stars all are r >17. The physical identity of these objects is yet to be confirmed, but we
find that they are consistent with DB white dwarfs with a low temperature. Whatever is the nature
of the objects, we emphasize a practical value of these black-body-like objects as photometric or
spectrophotometric standards.
Such an object was found serendipitously while the quasar spectra in Baryon Acoustic Os-
cillation Spectroscopic Survey of SDSS-III (BOSS, Dawson et al. 2013) were visually inspected.
This object was primarily targeted as a quasar candidate (Ross et al. 2012), but it does not resem-
ble any type of quasars, and was eventually rejected from the quasar catalogue (P ˆaris et al. 2014).
It turned out to be a star with a smooth-featureless spectrum. In fact, this object shows a proper
motion, which, together with the close proximity in its spectrum, leads us to conclude that this is a
star having a black-body spectrum.
We have then conducted a systematic search for such stars showing the black-body-like spec-
trum from the archived 4.3 million SDSS spectral objects. We select 22 stellar-like objects showing
smooth black-body-like spectra. We then employ GALEX satellite data for ultraviolet photometry.
We study if a black-body spectrum in the optical wavelengths continues to UV . Particularly, it is
important to find the peak and turnover of the black-body spectrum in the UV region: we drop one
star that does not pass this test. We finally employ WISE satellite data for infrared photometry.
It is amazing to see that the majority of the stars we found for black-body-like star candidates
show spectra that extrapolate smoothly to the infrared as far as 3.4 m. It is known that some stars,
specifically white dwarfs, are occasionally surrounded by a disc or dust debris that exhibits some
IR excess (Debes et al. 2011). We carried out this test: 17 objects survived, and we present these
objects as black-body stars in this paper. We dropped four objects that show an IR excess from our
sample.
We may use these black-body-like stars for the photometric standard for the verification of
the photometric system across various photometric passbands, to the accuracy that they are in-
deed black-body stars. This enables us to examine photometric systems used in the literature, in
particular the accuracy of the zero points, across various photometric bands, which otherwise can-
not easily be carried out. This also allows us to examine possible relative consistency among a
few photometric systems that adopt the ‘AB system’ outside the optical passband used in different
projects. These black-body stars can also be used as standard stars for photometric and spectropho-
tometric observations. Our objects all range in 17 19th mag in the rband, which is appropriate
for standard stars for 8 10 metre telescopes.

3
2. Data and Search for Black-Body Stars
2.1. SDSS
Our primary source is the spectroscopic data archive of SDSS in DR7 (Abazajian et al. 2009)
and DR12 (Alam et al. 2015). There are little overlaps in targets between the two data sets, while
the spectrograph was modified between the two. After DR7, the final data release of the SDSS I and
II projects, the spectrograph has been modified for SDSS-III: the wavelength range has widened,
the fibre aperture size was narrowed, and washers are installed to correct for a tilt against the focal
plane in the spectrograph to enhance the signal-to-noise ratio in blue wavelengths, while sacrificing
the quality of the flux calibration (Dawson et al. 2013).
In the SDSS data reduction spectrophotometric flux is integrated over the filter curve and
calibrated against its broad-band, point-spread-function (PSF) photometry of stars. We confirmed
that spectrophotometric flux agrees with broad-band photometry within a 1% systematic difference
with an rms scatter, however, of 7% in the g,randi-bands. We remark here that the precision
of broadband photometry is typically 1% for the g;r;i;z -bands and 2% for the u-band, as claimed
in the DR7 SDSS document. Spectroscopic flux is then calibrated against the model of Gray et
al. (2001) for spectrophotometric standard stars per spectroscopic plate, typically F subdwarfs, as
described in the SDSS DR6 paper (Adelman-McCarthy et al. 2008).
The observed spectra are automatically classified into object types by the SDSS pipeline
(Bolton et al. 2012). The majority of the objects are correctly classified, but classification into
‘quasars’, often suffers from contaminations of other types of objects, or sometimes redshift mis-
measured, by a few percent probability. Visual inspection is conducted for all spectra classified as
“Quasar”.
Quasars and white dwarfs often have similar colours, resembling the black-body spectrum.
Hence, many white dwarfs and the targets we are looking for are included in the candidate quasar
sample in the target selection of SDSS. We go back to the master data archives. The essential
initial step of our work is to search for featureless spectra with rather close to black-body in the
data archive of SDSS-I/II (Data Release 8, Aihara et al. 2011) and SDSS-III (Data Release 12,
Alam et al. 2015). We use the data reduction version of v5 312 for DR8, and v5 70 for DR12.
We examine 1,843,200 spectra in DR8 and 2,463,000 spectra in DR12. We fit the black-body
spectrum to all spectra with two free parameters, the effective temperature and a flux normalization
factor. We did not introduce reddening or a spectral tilt at this stage. We preselect spectra with
S=N > 20per pixel at the r-band wavelength to avoid noisy spectra that would obscure absorption
features. We collect objects whose reduced 2=dofof the black-body fit is less than 1.05. We find
465 objects from the total of 798,593 star-like objects. We find that the effective temperature is
roughly 10,000K for most of the objects we are looking for. Then, we visually inspected all 465

4
objects and rejected ones that show absorption features. We remark that DC white dwarf is defined
as a star with featureless spectrum with the strength of the absorption lines less than 5 %of its
continuum. We rejected spectra with H or He absorption lines if they are visible, if weak, within
the allowed S/N of the SDSS data. 22 stars survived this selection.
Once we identify candidates for featureless objects with spectra close to the black-body in
the spectral archives, we employ the five broad-band photometry data from SDSS DR8 to enhance
the photometric accuracy in the optical band. We remark that DR8 presents the last release of
photometry of SDSS and includes all SDSS objects1. After confirming that the photometric data
are consistent with spectrophotometry, we refit to obtain the temperature and the flux normalization
from broad-band photometry, whereas we should wait for the introduction of the GALEX UV data
to give more definitive temperatures.
SDSS photometry is designed to be in the AB magnitude system. It is constructed based on
the absolute flux of Lyr (Vega), and used spectrophotometry of BD+174708 as the interme-
diary. Slight offsets, however, are claimed in order to make it closer to the AB system in SDSS
photometry. For example it is suggested in their data release paper (say DR8) that some offset be
added to the values given in the SDSS photometry data release for DR7 and DR8: AB SDSS=
0.042, +0.036, +0.015, +0.013 and 0.0022 3foru;g;r;i;z -bands from SDSS magnitude to
(pseudo)AB magnitude. It is one of our purposes to examine the validity of this possible offset, as
seen in the next section.
2.2. GALEX
We extend our study to include the UV using the all-sky survey of GALEX (Morrissey et
al. 2007). We fit a black-body spectrum to these spectral energy distributions (SED) in addition
to those of SDSS, both taking broad-band photometry, and examine if the fit is consistent with
black-body fitted to the optical band alone.
We collect the photometric data for FUV ( 1350-1780) and NUV ( 1770-2730). GALEX
1The data in DR8 occasionally differ from those in DR7 due to the re-reduction the data by the SDSS team. The
difference between the two data releases is typically of the order of 0.01 mag in either positive and negative way, but
it amounts for some cases to 0.05 mag. The difference does not seem to show any systematic trends
2These offsets are being used in the SDSS pipeline routine named “SDSSFLUX2AB” in their IDL package. It was
derived by David Hogg (2003, private communication).
3IDL package can be found in : https://users.obs.carnegiescience.edu/yshen/IDL/
photoop_doc.html

5
photometry has been calibrated to STScI CALSPEC (Bohlin et al. 2001) in its AB magnitude
system. GALEX photometry is important for our project to find the turnover of black-body spectra
when temperature is <20;000K, which is indicated in our fit to the optical spectrum. One
candidate star, J121886+414800, does not show a turn over in UV , so is dropped from our black-
body star sample. This makes our sample 21 stars. We derive the effective temperature and flux
normalization factor from fit to the photometric data of SDSS and GALEX, which are shown in
Table 1 below. We reserve for a possibility in mind that the zero point of the AB magnitude of
GALEX may deviate from that of the SDSS AB photometric system.
We also check for the proper motion of black-body candidate stars (Roeser et al 2010). We
should reject candidates when they exhibit no proper motions. They might be candidates for BL
Lac objects. All our 22 candidate objects, however, show significant proper motions of the order of
2070 mas/yr per coordinate, which is consistent with their distances tens of parsec. This distance
is consistent with distance moduli inferred from brightness and general luminosity if these stars
are white dwarf like objects. In any case, we conclude that these stars are in the local field.
2.3. WISE
We supplement our study with infrared photometry of WISE, which has observed the sky
at 3.4, 4.6, 12 and 22 m (W1, W2, W3 and W4 filters, respectively) (Wright et al. 2010). We
use the photometric catalogue of Lang et al. (2016), which introduced “forced” photometry,
where the positions of objects are fixed to those in SDSS photometry, and unblurred images are
coadded (unWISE) to deepen the survey depth1. Only the W1 (3.4 m) channel is appropriate to
our work, for photometry matches the depth we require only in this passband. WISE photometry is
presented taking Lyr (Vega) as the zero magnitude. This is transformed to the AB system using
the conversion given in Wright et al. (2010). We reserve here again for a possibility of mismatch
in the zero point of AB photometry between WISE and SDSS. We remark that the original WISE
data release does not reach the depth we need.
It has been known that a significant fraction (say, 20%) of white dwarfs have dust debris
around the star that is manifested as an IR excess (Debes et al 2011). The IR data are essential to
rule out objects with the IR excess. We find that 4 among 22 objects show a significant IR excess,
and drop them from our sample of black-body stars. This makes our final sample to be 17. For this
sample, the black-body fit from the UV to the optical region extrapolates well to 3.4 m.
1We thank Brice M ´enard for his suggestion to use forced photometry to increase the depth to match our study.

6
3. Result
Our initial 22 black body star candidates are presented in Table 1. It includes 4 stars showing
the IR excess and one star that does not fit GALEX UV . We carry out global black-body fits to the
available broad-band photometric data of SDSS and GALEX. The fit parameters are Teand the
flux normalization factor a, as:
f=a2hc2
51
exp(hc=kT e)1(1)
whereis the wavelength, hthe Planck constant and kis the Boltzmann constant. The normal-
izationais a dimensionless factor and given in this table as well as 2/dof. We do not introduce
extinction corrections in this fit. The normalization factor is alternatively represented as the angular
size,
=R
d=s
f(Te)
B(Te)=ra
; (2)
whereB(Te)is the Planck radiation brightness, Ris the radius, and dis the distance to the star.
In Figure 1 we present the spectra of the 22 stars given in Table 1 and the fits to the black-body
spectrum. We include one star with its UV deviated from GALEX. The fit is carried out without
using WISE photometry as a constraint. This then shows that the IR excess is apparent for four
stars beyond 2 sigma. When an IR excess is not apparent, the fit from optical spectra extrapolates
well to IR 3.4 m (within 1.2 sigma), which corroborates the conclusion that the emission from the
stars is close to black-body. To scrutinize the deviation from the black-body spectrum, we show in
the lower panels the residual of the observed flux against the black-body fit given in the fraction.
For WISE W1 we show this separately, as its error is too large to display in the same scale as other
bands. We are left with 17 black-body stars. Note that chi squares are not necessarily good for
some stars, so that one may further reject some depending on their purpose.
We note that the temperature and the normalization are correlated in the fit. We demonstrate
an example for J124535.626 +423824.58 in Figure 2, where thick-hatched region is the 1 allowed
range in the fit to the SDSS data only. This is squeezed to a thicker-hatched region upon the use
of the GALEX data, resulting in the reduced temperature error of 50K. The inclusion of WISE
does not improve the result. This trend differs little for the other 16 stars.
The black-body fit and the residuals are shown in Table 2: the residuals are also plotted in
Figure 3(a). Brightness given in Table 2, assuming that the emission is perfect black-body, may
be used as the photometric standard, till more accurate observation becomes available. We added
our calculation for the J,H, andKbands of Subaru MOIRCS (Suzuki et al. 2008) and W1 for
WISE. NIR is not used as a constraint to the fit. The zero-point is SDSS pseudo-AB system. The
mean residuals over our 17 stars are given in the upper row of Table 3. For the optical bands, they

7
are all consistent with zero within errors of 0.01 mag, smaller than the typical error of uorzband
photometry. This means that the SDSS photometry zero points are consistent across the uto the
zbands: no appreciable deviations are seen in the zero-point of any of the five passbands. We
also see that deviations are of random nature without showing systematic trends. Looking at the
deviations star-by-star, there seems no apparent correlations between colour bands. The same can
be said for GALEX NUV . There seems no offset in the GALEX zero point against the SDSS AB
system beyond the photometric scatter. The deviation is recognizable for GALEX FUV: observed
brightness is fainter by 0.1-0.2 mag than the black-body, while the large photometric scatter does
not allow us to draw a precise conclusion. For W1 of WISE, the offset of mean is 0.452 0.663.
The error is large, while it is consistent with zero. We note that we use forced photometry of WISE
and we do not use WISE W1 as a constraint to fit to the black-body. We cannot derive an accurate
conclusion as to the zero point of WISE photometry.
It has been alleged that SDSS photometry is deviated from the AB system and some offsets
should be added to the bare value of SDSS photometric pipeline output1. We compare SDSS
photometry to that with the suggested offsets (as given in the bottom of Sect. 2.1) added to the
data, and re-minimize the black-body fit that is given in row 3 of Table 3. This is also shown in
panel (b) of Figure 3. We now see the residuals increased significantly, in particular in the uband,
where a rather large offset ( 0:042mag) is added. The residual to the black-body seen in this
figure looks somewhat larger and also wiggly across the five colour bands: the observed is brighter
inuand fainter in g.
We also compare SDSS photometry to photometry with the AB zero point set by CALSPEC
(Holtzman et al. 2008), by re-minimizing again the entire fit, as shown in row 5 in Table 32, and
also in panel (c) of Figure 3. We observe a wiggle similar to the plot with Hogg’s offset above,
indicating brighter uand fainterg. Departures of residuals from zero are larger in the uandz
bands, 0.038 mag, and 0.027 mag, respectively, which are compared to <0:01mag offsets with
the original magnitude system of SDSS. We conclude that original SDSS photometry works better
than adding some offsets or using the CALSPEC standard as the AB zero magnitude.
We see in these Figure and Table that GALEX NUV photometry smoothly matches SDSS
photometry, to the level of a few times 0.01 mag, albeit the scatter of photometry in the NUV
band is as large as 0.08 mag, verifying the consistency between the SDSS photometric system and
GALEX NUV . For FUV photometry a large scatter of the order of 0.1 mag seen in our residual
figure hinders us from deriving an accurate conclusion as to the zero point offset, but the mismatch
1This has been claimed by David Hogg (private communication) and is used in some analyses, in particular of
spectrophotometric data handling, of SDSS.
2The constant added is 0:037(u), 0.024 ( g), 0.005 ( r), 0.018 ( i) and0:016(z)

8
of the AB magnitudes between the two systems is at most of the order of 0.2 mag. Our finding
here indicates that these stars can be used as spectrophotometric standards from UV to IR with
expected errors and fluctuations in mind.
Our fit gives the temperature of our black-body stars 8500 to 12000K. So we may conclude
that these stars are consistent with DB white dwarfs with He lines undeveloped. The white dwarf
spectrum is generically not very far from the black-body spectrum up to absorption features. If
we assume that these black-body stars are white dwarfs, the observed brightness indicates the
distance to be roughly of the order of 50 pc. This is also consistent with the distance indicated
by large proper motions. A typical distance of 50 pc in the Galactic disc indicates extinction
E(BV)0:01orAr0:03mag. Note that we cannot estimate the distance from colours. We
also note that dust reddening is almost parallel to the temperature in colour space. The reddening
correction modifies the resulting temperature, but, as we confirmed, the residuals of the black-body
fits are modified very little, only up to 0.01 mag even for a very large extinction, E(BV) = 0:10.
Therefore, our conclusion as to black-body stars is not modified. The fitted temperature becomes
lower as
T122[E(BV)=0:01]Kupon the inclusion of reddening.
4. Conclusion
We have discovered 17 stars that show spectra very close to the black-body for a wide range
of spectrum from GALEX FUV to the IR = 3:4m band out of 4,300,000 spectra archived
in the SDSS data base. These spectra can give us a unique possibility to examine the magnitude
system, in particular the zero point of the photometric system in AB, across the various colour
bands within SDSS, or even across the different system used by GALEX against that of SDSS, or
WISE versus SDSS. Within the SDSS photometric system, we would indicate that the zero point
of the 5 colour bands is consistent within 0.01 mag. We do not find indication for the offset with
SDSS AB photometry up to the absolute constant common in the 5 bands. We do notneed to
introduce any additional constant to adjust to the ‘AB system’. GALEX NUV photometry zero
point is consistent with SDSS photometric zero point within a few hundredths of mag, although
large scatter of NUV photometry ( 0.07 mag) hinders from a more accurate conclusion. During
this study we also noted a 0.02 0.04 mag error in the optical zero point of the CALSPEC AB
standard.
The black-body stars we found are as faint as 17 19 mag inr, and can be used as photometric,
but also as spectrophotometric standard stars for the work with large aperture telescopes. We
consider that these stars in our sample are consistent with DB white dwarfs with temperature
around 10000 K.

9
Acknowledgment
We thank Brice M ´enard for useful discussion and Ting-Wen Lan for assisting us to collect
the WISE data. NS is partially supported by JST CREST JPMHCR1414 and JSPS Programs for
Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers.
MF thanks Hans B ¨ohringer and Yasuo Tanaka for the hospitality at the Max-Planck-Institut f ¨ur
Extraterrestrische Physik and also Eiichiro Komatsu at Max-Planck-Institut f ¨ur Astrophysik, in
Garching. He also wishes his thanks to Alexander von Humboldt Stiftung for the support during
his stay in Garching, and Monell Foundation in Princeton. He received in Tokyo a Grant-in-Aid
(No. 154300000110) from the Ministry of Education. Kavli IPMU is supported by World Premier
International Research Center Initiative of the Ministry of Education, Japan.
REFERENCES
Abazajian, K. N., Adelman-McCarthy, J. K., Ag ¨ueros, M. A., et al. 2009, ApJS, 182, 543 (DR7)
Adelman-McCarthy, J. K., Ag ¨ueros, M. A., Allam, S. S., et al. 2008, ApJS, 175, 297 (DR6)
Aihara, H., Allende Prieto, C., An, D., et al. 2011, ApJS, 193, 29 (DR8)
Alam, S., Albareti, F. D., Allende Prieto, C., et al. 2015, ApJS, 219, 12 (DR12)
Bohlin, R. C., Dickinson, M. E., Calzetti, D., 2001, AJ, 122, 2118
Bolton, A. S., Schlegel, D. J., Aubourg, ´E., et al. 2012, AJ, 144, 144
Dawson, K. S., Schlegel, D. J., Ahn, C. P., et al. 2013, AJ, 145, 10
Debes, J. H., Hoard, D. W., Wachter, S., Leisawitz, D. T., & Cohen, M. 2011, ApJS, 197, 38
Gray, R. O., Graham, P. W., & Hoyt, S. R. 2001, AJ, 121, 2159
Holtzman, J. A., Marriner, J., Kessler, R. et al. 2008, ApJ, 136, 2320
Lang, D., Hogg, D. W., & Schlegel, D. J. 2016, AJ, 151, 36
Morrissey, P., Conrow, T., Barlow, Tom., et al. 2007, ApJS, 173, 682
Pˆaris, I., Petitjean, P., Aubourg, ´E., et al. 2014, A&A, 563, A54
Roeser, S., Demleitner, M., & Schilbach, E. 2010, AJ, 139, 2440
Ross, N. P., Myers, A. D., Sheldon, E. S., et al. 2012, ApJS, 199, 3

10
Suzuki, R., Tokoku, C., Ichikawa, T., et al. 2008, PASJ, 60, 1347
Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868
This preprint was prepared with the AAS L ATEX macros v5.2.

11Table 1. 22 Black-body Candidates: Coordinates and Photometric Data
SDSS name RA (J2000) DEC (J2000) GALEX FUV GALEX NUV SDSS u SDSSg SDSSr SDSSi SDSSz WISE W1 T (K) a(1023)(”109)2/dof
J002739.497-001741.93 00:27:39.497 -00:17:41.93 21.80 0.06 19.89 0.02 19.05 0.02 18.90 0.03 18.98 0.02 19.12 0.03 19.37 0.05 21.21 0.48 10662 60 0.435 0.010 242.7 2.8 0.84
J004705.818-004820.0900:47:05.818 -00:48:20.09 20.02 0.04 19.29 0.02 19.00 0.03 19.01 0.02 19.24 0.02 19.50 0.02 19.60 0.05 21.07 0.41 14648 94 0.164 0.003 149.0 1.6 8.67
J004830.324+001752.80 00:48:30.324 +00:17:52.80 21.18 0.05 19.14 0.01 18.27 0.03 18.14 0.01 18.23 0.01 18.38 0.01 18.63 0.03 20.46 0.22 10639 40 0.876 0.013 344.5 2.5 5.10
J014618.898-005150.51 01:46:18.898 -00:51:50.51 20.61 0.05 18.80 0.01 18.21 0.02 18.17 0.02 18.24 0.01 18.42 0.01 18.61 0.03 21.19 0.41 11770 48 0.672 0.010 301.8 2.3 8.52
J022936.715-004113.63 02:29:36.715 -00:41:13.63 23.04 0.14 20.73 0.01 19.38 0.03 19.07 0.03 19.09 0.01 19.17 0.02 19.31 0.04 21.00 0.30 8901 31 0.640 0.011 294.5 2.5 1.93
J083226.568+370955.48 08:32:26.568 +37:09:55.48 19.39 0.03 18.98 0.01 18.84 0.03 18.83 0.02 18.93 0.04 20.63 0.24 7952 95 1.137 0.050 392.4 8.5 0.63
J083736.557+542758.64 08:37:36.557 +54:27:58.64 21.18 0.07 19.24 0.03 18.73 0.03 18.56 0.02 18.54 0.02 18.52 0.04 20.34 0.16 7449 61 1.815 0.055 495.8 7.5 3.38
J100449.541+121559.65 10:04:49.541 +12:15:59.65 19.41 0.02 19.18 0.03 19.23 0.02 19.32 0.02 19.43 0.05 22.25 1.07 9773 136 0.440 0.019 244.1 5.2 0.57
J103123.906+093657.8910:31:23.906 +09:36:57.89 18.73 0.03 18.64 0.02 18.80 0.01 18.96 0.02 19.15 0.05 20.63 0.25 11667 224 0.419 0.020 238.2 5.7 0.40
J104523.866+015721.96 10:45:23.866 +01:57:21.96 20.51 0.09 19.32 0.03 19.09 0.01 19.01 0.01 19.07 0.02 19.28 0.07 22.13 1.04 8956 87 0.666 0.021 300.3 4.7 2.56
J111720.801+405954.67 11:17:20.801 +40:59:54.67 20.93 0.34 18.98 0.09 18.26 0.02 18.08 0.01 18.19 0.01 18.34 0.01 18.63 0.04 23.00 1.91 10950 121 0.843 0.026 337.9 5.1 2.56
J114722.608+171325.21 11:47:22.608 +17:13:25.21 19.91 0.06 18.91 0.03 18.65 0.02 18.68 0.02 18.85 0.02 19.00 0.04 22.24 1.06 9962 109 0.669 0.022 301.0 4.9 1.92
J121856.693+414800.29a12:18:56.693 +41:48:00.29 17.59 0.06 17.84 0.04 17.91 0.02 18.18 0.02 18.62 0.02 18.90 0.01 19.22 0.05 22.58 1.26 23357 383 0.133 0.004 134.1 2.0 7.03
J124535.626+423824.58 12:45:35.626 +42:38:24.58 18.30 0.03 17.32 0.02 17.14 0.02 17.18 0.02 17.29 0.01 17.46 0.02 19.60 0.08 10086 67 2.650 0.057 599.1 6.4 0.17
J125507.082+192459.00 12:55:07.082 +19:24:59.00 19.93 0.08 18.83 0.04 18.53 0.02 18.45 0.01 18.50 0.01 18.63 0.03 21.75 0.78 8882 98 1.157 0.039 395.8 6.6 1.46
J134305.302+270623.98 13:43:05.302 +27:06:23.98 20.03 0.14 19.05 0.02 18.93 0.02 19.00 0.02 19.14 0.02 19.34 0.07 22.09 0.85 10678 151 0.427 0.017 240.4 4.8 0.20
J135816.735+144202.2713:58:16.735 +14:42:02.27 21.93 0.12 19.46 0.02 18.34 0.03 18.06 0.03 18.02 0.02 18.07 0.01 18.12 0.03 19.77 0.09 9191 44 1.598 0.030 465.2 4.3 3.39
J141724.329+494127.85 14:17:24.329 +49:41:27.85 20.73 0.29 18.27 0.06 17.36 0.03 17.25 0.03 17.31 0.02 17.43 0.02 17.61 0.03 20.01 0.08 10503 112 2.127 0.066 536.7 8.3 1.01
J151859.717+002839.58 15:18:59.717 +00:28:39.58 19.71 0.03 19.44 0.01 19.37 0.02 19.51 0.03 19.57 0.07 21.57 0.47 9072 131 0.458 0.022 249.1 5.9 0.71
J161605.194+142116.7016:16:05.194 +14:21:16.70 22.32 0.47 19.57 0.07 18.70 0.02 18.61 0.01 18.70 0.01 18.85 0.02 19.02 0.04 19.52 0.07 10918 119 0.533 0.016 268.6 4.1 1.08
J161704.078+181311.96 16:17:04.078 +18:13:11.96 20.60 0.12 19.17 0.03 18.78 0.02 18.73 0.02 18.80 0.02 18.81 0.04 21.56 0.44 8568 88 0.996 0.034 367.2 6.2 2.00
J230240.032-003021.60 23:02:40.032 -00:30:21.60 20.79 0.06 18.88 0.01 17.97 0.02 17.80 0.02 17.90 0.02 18.02 0.02 18.25 0.03 21.19 0.42 10478 42 1.241 0.021 409.9 3.4 1.14
Stars that show IR excess. To be dropped from the list of black-body stars.
aGALEX UV data do not show turn-off. To be removed from the list of black-body stars.

12Table 2. The best fit broad-band flux of the 17 black-body stars (upper rows) and the residuals from the fit (lower rows):
m =
DataBlack-Body in magnitude. Calculations are added for Subaru MOIRCS J, H, K and for W1 of WISE. NIR is not used as a
constraint to the fit
Star name GALEX FUV GALEX NUV SDSS-u SDSS-g SDSS-r SDSS-i SDSS-z 2/dof MOIRCS-J MOIRCS-H MOIRCS-K WISE W1
J002739.497-001741.93 21.705 0.027 19.905 0.010 19.041 0.002 18.911 0.007 18.980 0.010 19.124 0.012 19.306 0.014 19.753 0.015 20.170 0.016 20.647 0.017 21.470 0.018
0.0900.061 -0.017 0.019 0.010 0.022 -0.013 0.025 0.000 0.019 -0.005 0.034 0.060 0.034 0.84 -0.259 0.482
J004830.324+001752.80 20.965 0.019 19.158 0.008 18.289 0.000 18.158 0.003 18.225 0.006 18.368 0.007 18.550 0.008 18.996 0.009 19.413 0.010 19.889 0.010 20.712 0.011
0.2150.049 -0.021 0.013 -0.021 0.027 -0.017 0.014 0.007 0.014 0.007 0.014 0.082 0.014 5.10 -0.252 0.217
J014618.898-005150.51 20.351 0.018 18.839 0.007 18.169 0.001 18.121 0.004 18.250 0.006 18.425 0.007 18.630 0.008 19.108 0.010 19.541 0.010 20.031 0.011 20.867 0.011
0.2630.050 -0.040 0.012 0.045 0.019 0.046 0.019 -0.006 0.013 -0.004 0.015 -0.025 0.015 8.52 0.327 0.411
J022936.715-004113.63 23.126 0.020 20.716 0.007 19.450 0.001 19.140 0.005 19.079 0.008 19.155 0.010 19.289 0.011 19.668 0.012 20.049 0.013 20.498 0.013 21.293 0.014
-0.0860.144 0.009 0.012 -0.066 0.029 -0.065 0.031 0.008 0.015 0.012 0.017 0.025 0.017 1.93 -0.292 0.302
J083226.568+370955.48 23.824 0.097 20.972 0.048 19.420 0.017 18.977 0.002 18.819 0.009 18.844 0.015 18.942 0.019 19.270 0.024 19.624 0.027 20.053 0.030 20.827 0.032
-0.0250.031 0.008 0.015 0.026 0.028 -0.017 0.022 -0.013 0.022 0.63 -0.201 0.240
J083736.557+542758.64 24.149 0.073 21.016 0.037 19.286 0.014 18.757 0.003 18.538 0.005 18.530 0.009 18.604 0.012 18.899 0.016 19.235 0.018 19.651 0.020 20.412 0.022
0.1660.075 -0.045 0.026 -0.023 0.027 0.026 0.017 0.013 0.017 -0.081 0.017 3.38 -0.067 0.162
J100449.541+121559.65 22.541 0.092 20.461 0.046 19.411 0.015 19.199 0.002 19.209 0.008 19.322 0.013 19.482 0.017 19.898 0.022 20.299 0.025 20.763 0.027 21.573 0.029
0.0040.024 -0.019 0.025 0.021 0.023 0.000 0.021 -0.048 0.021 0.57 0.679 1.071
J104523.866+015721.96 23.016 0.071 20.629 0.036 19.377 0.013 19.074 0.002 19.018 0.005 19.096 0.009 19.232 0.012 19.614 0.016 19.996 0.018 20.446 0.020 21.242 0.021
-0.1180.085 -0.053 0.028 0.021 0.010 -0.005 0.011 -0.024 0.016 0.050 0.016 2.56 0.886 1.038
J111720.801+405954.67 20.741 0.066 19.021 0.034 18.211 0.012 18.105 0.002 18.190 0.005 18.343 0.008 18.531 0.011 18.987 0.014 19.409 0.016 19.889 0.018 20.716 0.019
0.1860.336 -0.039 0.086 0.046 0.022 -0.027 0.013 0.002 0.014 -0.001 0.014 0.101 0.014 2.56 2.282 1.910
J114722.608+171325.21 21.893 0.071 19.877 0.036 18.869 0.012 18.676 0.002 18.699 0.006 18.819 0.010 18.985 0.013 19.408 0.017 19.812 0.019 20.279 0.020 21.092 0.022
0.0360.058 0.043 0.031 -0.027 0.016 -0.016 0.017 0.035 0.019 0.017 0.019 1.92 1.152 1.060
J124535.626+423824.58 20.276 0.042 18.300 0.020 17.319 0.006 17.138 0.000 17.170 0.005 17.295 0.007 17.463 0.009 17.891 0.011 18.297 0.013 18.767 0.014 19.581 0.015
-0.0040.032 -0.000 0.017 0.001 0.017 0.015 0.019 -0.007 0.015 -0.004 0.015 0.17 0.022 0.081
J125507.082+192459.00 22.508 0.084 20.090 0.044 18.818 0.017 18.506 0.005 18.443 0.003 18.518 0.008 18.651 0.011 19.030 0.016 19.410 0.018 19.859 0.020 20.653 0.022
-0.1580.081 0.010 0.038 0.021 0.019 0.012 0.015 -0.015 0.015 -0.025 0.015 1.46 1.100 0.781
J134305.302+270623.98 21.712 0.086 19.916 0.043 19.055 0.014 18.927 0.002 18.996 0.007 19.141 0.012 19.323 0.015 19.771 0.019 20.188 0.022 20.665 0.024 21.488 0.026
0.1170.139 -0.008 0.024 -0.002 0.021 0.005 0.016 -0.004 0.020 0.012 0.020 0.20 0.598 0.846
J141724.329+494127.85 20.124 0.066 18.277 0.033 17.382 0.011 17.238 0.001 17.297 0.006 17.436 0.009 17.615 0.012 18.057 0.015 18.471 0.017 18.946 0.019 19.766 0.020
0.6080.288 -0.008 0.056 -0.026 0.026 0.012 0.026 0.010 0.017 -0.005 0.022 -0.006 0.022 1.01 0.242 0.081
J151859.717+002839.58 23.280 0.102 20.940 0.050 19.719 0.017 19.431 0.001 19.384 0.010 19.468 0.016 19.608 0.020 19.995 0.026 20.380 0.029 20.832 0.031 21.630 0.034
-0.0070.028 0.007 0.015 -0.013 0.016 0.040 0.033 -0.042 0.033 0.71 -0.057 0.473
J161704.078+181311.96 23.078 0.080 20.524 0.041 19.164 0.015 18.811 0.004 18.719 0.005 18.778 0.009 18.900 0.013 19.263 0.017 19.635 0.020 20.078 0.021 20.866 0.023
0.0740.123 0.006 0.025 -0.030 0.021 0.012 0.016 0.020 0.017 -0.092 0.017 2.00 0.691 0.436
J230240.032-003021.60 20.731 0.020 18.877 0.007 17.977 0.001 17.831 0.005 17.889 0.007 18.027 0.009 18.205 0.010 18.646 0.011 19.060 0.012 19.534 0.012 20.354 0.013
0.0570.056 0.000 0.013 -0.009 0.021 -0.027 0.017 0.009 0.017 -0.002 0.020 0.041 0.020 1.14 0.838 0.422

13Table 3. Mean of residuals from the black-body fit (
m=data black-body fit): The 17 black-body stars are
used in this calculation. Row 1 is with the raw value of brightness given in SDSS data release DR8; Row 2 gives
the offsets suggested by Hogg to make the SDSS magnitude closer to the AB system, and row 3 is the mean
residuals for this case. Row 4 gives the offsets suggested by Holzman (2009) to make the SDSS magnitude
closer to the AB system with CALSPEC, and Row 5 is the mean residuals.
Data Name GALEX FUV GALEX NUV SDSS-u SDSS-g SDSS-r SDSS-i SDSS-z
Mean Residuals without Offset 0.191 0.202 -0.000 0.081 -0.006 0.031 -0.008 0.025 0.007 0.012 0.002 0.017 0.003 0.052
Hogg’s Offset to SDSS -0.042 0.036 0.015 0.013 -0.002
Mean Residuals with Hogg Offset 0.218 0.215 0.016 0.097 -0.047 0.034 0.020 0.025 0.007 0.014 -0.002 0.016 -0.019 0.054
Holtzman et al Offset (CALSEPC) -0.037 0.024 0.005 0.018 -0.016
Mean Residuals with Holtzman Offset 0.215 0.213 0.018 0.093 -0.038 0.033 0.013 0.025 0.003 0.013 0.009 0.016 -0.027 0.054

14
Fig. 1.— Spectra of 22 black-body star candidates. In the left panels, observed spectra are
indicated with grey, and blue curves show smoothed spectra. Black-body fits are indicated with
red. The right panels show their photometric data from which our fit parameters are derived. In
the bottom panel, the residuals (fractional values) from the fits are shown, with the positions of
hydrogen and helium absorption lines indicated.

15

16

17

18

19

20
Fig. 2.— Examples of the black-body fit, the fit for J124535.626 +423824.58. One sigma outer
contour (thin coloured) is the fit to SDSS five-colour photometric data alone and inner contour
(thicker coloured) is the fit GALEX UV data included. Inclusion of WISE IR data does not change
the contour.

21
Fig. 3.— (a) Residuals of the SDSS five-band photometric data and the GALEX data from the
black body fits in magnitude. (b) The SDSS photometric data are corrected for the offset proposed
by Hogg and re-fitting is made to give the best fit. The figure shows residuals in magnitude. (c)
The offset is applied to SDSS photometry as suggested by Holtzman et al. (2009) based on the
CALSPEC zero point. Another re-fitting is made to give the best black-body fits. The figure shows
residuals in magnitude.