Evidence for Type Ia Supernova Diversity from Ultraviolet Observations with the Hubble Space Telescope
ABSTRACT We present ultraviolet (UV) spectroscopy and photometry of four Type Ia
supernovae (SNe 2004dt, 2004ef, 2005M, and 2005cf) obtained with the UV prism
of the Advanced Camera for Surveys on the Hubble Space Telescope. This dataset
provides unique spectral time series down to 2000 Angstrom. Significant
diversity is seen in the near maximum-light spectra (~ 2000--3500 Angstrom) for
this small sample. The corresponding photometric data, together with archival
data from Swift Ultraviolet/Optical Telescope observations, provide further
evidence of increased dispersion in the UV emission with respect to the
optical. The peak luminosities measured in uvw1/F250W are found to correlate
with the B-band light-curve shape parameter dm15(B), but with much larger
scatter relative to the correlation in the broad-band B band (e.g., ~0.4 mag
versus ~0.2 mag for those with 0.8 < dm15 < 1.7 mag). SN 2004dt is found as an
outlier of this correlation (at > 3 sigma), being brighter than normal SNe Ia
such as SN 2005cf by ~0.9 mag and ~2.0 mag in the uvw1/F250W and uvm2/F220W
filters, respectively. We show that different progenitor metallicity or
line-expansion velocities alone cannot explain such a large discrepancy.
Viewing-angle effects, such as due to an asymmetric explosion, may have a
significant influence on the flux emitted in the UV region. Detailed modeling
is needed to disentangle and quantify the above effects.
arXiv:1110.5809v2 [astro-ph.HE] 27 Oct 2011
DRAFT VERSION OCTOBER 28, 2011
Preprint typeset using LATEX style emulateapj v. 5/2/11
EVIDENCE FOR TYPE IA SUPERNOVA DIVERSITY
FROM ULTRAVIOLET OBSERVATIONS WITH THE HUBBLE SPACE TELESCOPE
XIAOFENG WANG1,2,3, LIFAN WANG2, ALEXEI V. FILIPPENKO3, GREG ALDERING4, PIERRE ANTILOGUS5, DAVID ARNETT6, DIETRICH
BAADE7, EDDIE BARON8, BRIAN J. BARRIS9, STEFANO BENETTI10, PATRICE BOUCHET11, ADAM S. BURROWS12, RAMON CANAL13,
ENRICO CAPPELLARO10, RAYMOND CARLBERG14, ELISA DI CARLO15, PETER CHALLIS16, ARLIN CROTTS17, JOHN I. DANZIGER18,
MASSIMO DELLA VALLE19,20, DENNIS JACK21, MICHAEL FINK22, RYAN J. FOLEY16, 23, CLAES FRANSSON24, AVISHAY GAL-YAM25,
PETER GARNAVICH26, CHRIS L. GERARDY27, GERSON GOLDHABER4, MARIO HAMUY28, WOLFGANG HILLEBRANDT29, PETER A.
HOEFLICH27, STEPHEN T. HOLLAND30, DANIEL E. HOLZ31, JOHN P. HUGHES32, DAVID J. JEFFERY33, SAURABH W. JHA32, DAN
KASEN34, ALEXEI M. KHOKHLOV31, ROBERT P. KIRSHNER16, ROBERT KNOP35, CECILIA KOZMA24, KEVIN KRISCIUNAS2, MARKUS
KROMER29, BRIAN C. LEE4, BRUNO LEIBUNDGUT36, ERIC J. LENTZ37, DOUGLAS C. LEONARD38, WALTER H. G. LEWIN39, WEIDONG
LI3, MARIO LIVIO40, PETER LUNDQVIST24, DAN MAOZ41, THOMAS MATHESON42, PAOLO MAZZALI10,29, PETER MEIKLE43, GAJUS
MIKNAITIS44, PETER MILNE6, STEFAN MOCHNACKI45, KEN’ICHI NOMOTO46, PETER E. NUGENT4, ELAINE ORAN47, NINO
PANAGIA40, FERDINANDO PATAT37, SAUL PERLMUTTER4, MARK M. PHILLIPS48, PHILIP PINTO6, DOVI POZNANSKI49, CHRISTOPHER
J. PRITCHET50, MARTIN REINECKE29, ADAM RIESS40, PILAR RUIZ-LAPUENTE13, RICHARD SCALZO4, ERIC M. SCHLEGEL51, BRIAN
SCHMIDT52, JAMES SIEGRIST4, ALICIA M. SODERBERG16, JESPER SOLLERMAN24, GEORGE SONNEBORN30, ANTHONY SPADAFORA4,
JASON SPYROMILIO36, RICHARD A. SRAMEK53, SUMNER G. STARRFIELD54, LOUIS G. STROLGER55, NICHOLAS B. SUNTZEFF2,
ROLLIN THOMAS4, JOHN L. TONRY9, AMEDEO TORNAMBE56, JAMES W. TRURAN31, MASSIMO TURATTO18, MICHAEL TURNER31,
SCHUYLER D. VAN DYK57, KURT WEILER47, J. CRAIG WHEELER58, MICHAEL WOOD-VASEY59, STAN WOOSLEY60, HITOSHI
YAMAOKA61, TIANMENG ZHANG62
Draft version October 28, 2011
We present ultraviolet (UV) spectroscopy and photometry of four Type Ia supernovae (SNe 2004dt, 2004ef,
2005M, and 2005cf) obtained with the UV prism of the Advanced Camera for Surveys on the Hubble Space
Telescope. This dataset provides unique spectral time series down to 2000 Å. Significant diversity is seen in
the near maximum-light spectra (∼ 2000–3500Å) for this small sample. The corresponding photometric data,
together with archival data from Swift Ultraviolet/Optical Telescope observations, provide further evidence
of increased dispersion in the UV emission with respect to the optical. The peak luminosities measured in
uvw1/F250W are found to correlate with the B-band light-curve shape parameter ∆m15(B), but with much
larger scatter relative to the correlation in the broad-band B band (e.g., ∼ 0.4 mag versus ∼ 0.2 mag for those
with 0.8 < ∆m15(B) < 1.7 mag). SN 2004dt is found as an outlier of this correlation (at > 3σ), being brighter
than normal SNe Ia such as SN 2005cf by ∼ 0.9 mag and ∼ 2.0 mag in the uvw1/F250W and uvm2/F220W
filters, respectively. We show that different progenitor metallicity or line-expansion velocities alone cannot
explain such a large discrepancy. Viewing-angle effects, such as due to an asymmetric explosion, may have
a significant influence on the flux emitted in the UV region. Detailed modeling is needed to disentangle and
quantify the above effects.
Subject headings: cosmology: observations — distance scale — dust, extinction — supernovae: general –
2Physics and Astronomy Department, Texas A&M University, College
Station, TX 77843, USA.
3Department of Astronomy, University of California, Berkeley, CA
4Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA.
5Laboratoire de Physique Nucleaire des Hautes Energies, Paris,
6Steward Observatory, University of Arizona, Tucson, AZ 85721,
7European Southern Observatory, 85748, Garching bei München,
8Department of Physics, University of Oklahoma, Norman, OK 73019,
9Institute for Astronomy, University of Hawaii, Honolulu, HI 96822,
10Osservatorio Astronomico di Padova, 35122 Padova, Italy.
Yvette Cedex, France.
12Department of Astrophysical Sciences, Princeton University, Prince-
ton, NJ 08544, USA.
13Universidad de Barcelona, Barcelona 8007, Spain.
(THCA), Tsinghua University,
14University of Toronto, Toronto, ON M5S 3J3, Canada.
15INAF, Osservatorio Astronomico di Teramo, 64100 Teramo, Italy.
16Harvard/Smithsonian Center Astrophysics, Cambridge, MA 02138,
17Department of Astronomy, Columbia University, New York, NY
18INAF, Osservatorio Astronomico di Trieste, I-34143, Trieste, Italy.
19Capodimonte Astronomical Observatory, INAF-Napoli, I-80131,
20International Center for Relativistic Astrophysics, I-65122, Pescara,
21Hamburger Sternwarte,Gojenbergsweg 112, 21029 Hamburg,
22Institut für Theoretische Physik und Astrophysik, Universität
Würzburg, Am Hubland, D-97074 Würzburg, Germany.
24Stockholm University, SE-106 91 Stockholm, Sweden.
25Weizmann Institute of Science, Rehovot, 76100, Israel.
26University of Notre Dame, Notre Dame, IN 46556, USA.
27Florida State University, Tallahassee, FL 32306, USA.
28Universidad de Chile, PO Box 10-D, Santiago, Chile.
29Max-Planck-Institut für Astrophysik, Karl-Schwarzschild-Str.1,
2 Wang et al.
The utility of Type Ia supernovae (SNe Ia) as cosmolog-
ical probes depends on the degree of our understanding of
SN Ia physics, and on various systematic effects such as cos-
mic chemical evolution. There is increasing evidence show-
ing that even the so-called “Branch-normal” SNe Ia (Branch
et al. 1993) exhibit diversity in their spectral features and
light-curve shapes that do not correlate with light-curve pa-
rameters such as the decline rate ∆m15(B) (Phillips 1993;
Benetti et al. 2005; Branch et al. 2009; Wang et al. 2009a;
Zhang et al. 2010; Höflich et al. 2010). For example, it
has been recently shown that SNe Ia having the same light-
curve shape parameter such as ∆m15(B) (Phillips 1993) but
faster expanding ejecta are on average ∼ 0.1 mag redder in
B−V colornear maximumlight (Wanget al. 2009a,hereafter
W09a). Possible origins of such a color difference include a
change of the dust obscuring the SN (e.g., circumstellar dust
versus interstellar dust; W09a), an effect of line blanketing
(Foley & Kasen 2011), or a projection effect in an asymmet-
ric explosion (Maeda et al. 2011). Moreover, the peak lu-
minosity of SNe Ia seems to show an additional dependence
85748 Garching, Germany.
30Laboratory for Observational Cosmology, NASA Goddard Space
Flight Center, Code 665, Greenbelt, MD 20771, USA.
31University of Chicago, Chicago, IL 60637, USA.
32Department of Physics and Astronomy, Rutgers, the State University
of New Jersey, Piscataway, NJ 08854, USA.
33Northern Arizona University, Flagstaff, AZ 86011, USA.
34Department of Physics, University of California, Berkeley, CA
35Quest University Canada, Squamish, BC, Canada.
36European Southern Observatory, 85748 Garching bei München,
37Department of Physics and Astronomy, University of Tennessee,
Knoxville, TN 37996, USA.
38San Diego State University, San Diego, CA 92182, USA.
39Massachusetts Institute of Technology, Cambridge, MA 02139,
40Space Telescope Science Institute, Baltimore, MD 21218, USA.
41Tel Aviv University, Wise Observatory, 69978, Tel Aviv, Israel.
42National Optical Astronomy Observatory, Tucson, AZ 85719, USA.
43Imperial College of Science Technology and Medicine, London SW7
44Center for Neighborhood Technology, 2125 W. North Ave., Chicago
45University of Toronto, Toronto, ON M5S 3G4, Canada.
46University of Tokyo, IPMU, Kashiwa, Chiba 277-8583, Japan.
47Naval Research Laboratory, Washington, DC 20375, USA.
48Carnegie Institution of Washington, Washington, DC 20005, USA.
49School of Physics and Astronomy, Tel-Aviv University, Tel Aviv
50University of Victoria, Victoria, BC V8W 2Y2, Canada.
51Physics & Astronomy Department, University of Texas at San
Antonio, San Antonio, TX 78249, USA.
52Australian National University, Canberra ACT 0200, Australia.
53Associated Universities, Inc., Washington, DC 20036, USA.
54Arizona State University, Tempe, AZ 85287, USA.
55Western Kentucky University, Bowling Green, KY 42101, USA.
56INAF, Rome Astronomical Observatory, 00136 Roma, Italy.
57IPAC, California Institute of Technology, Pasadena, CA 91125, USA.
58Department of Astronomy and McDonald Observatory, University of
Texas, Austin, TX 78712, USA.
59Pittsburgh Particle Physics Astrophysics, and Cosmology Center
(Pitt-PACC), University of Pittsburgh, Pittsburgh, PA 15260, USA.
60University of California, Santa Cruz, CA 95060, USA.
61Graduate School of Sciences, Kyushu University, Fukuoka 812-8581,
62National Astronomical Observatory of China, Chinese Academy of
Sciences, Beijing, 100012, China.
on the global characteristics of their host galaxies: events of
the same light-curve shape and color are ∼ 0.1 mag brighter
in massive (presumably metal-rich) host galaxies (Sullivan et
al. 2010). This correlation may also indicate variations in the
explosion physics and/or differences of the progenitors in a
range of environments.
Ultraviolet (UV) observations are of particular importance
in understanding both the diversity of SNe Ia and their phys-
ical evolution from low to high redshifts.
with existing optical and near-infrared data, UV observations
improve the determination of the bolometric light curves of
SNe Ia. The UV spectrum at early times forms mainly in
the outer shells of the explosion ejecta where the unburned
outer layers of the white dwarf play a large role in shaping
the appearance of the spectrum. The dependence of the UV
emission on the progenitor metallicity has been theoretically
studied by several authors (Pauldrach et al. 1996; Höflich et
al. 1998; Mazzali et al. 2000; Lentz et al. 2000; Timmes et al.
2003; Sauer et al. 2008). Both Höflich et al. (1998)and Sauer
et al. (2008)arguedthat increasing the metallicity in the outer
layers would lead to a stronger UV flux, with the latter high-
lighting the contribution from the reverse-fluorescence scat-
tering of photons from red to blue wavelengths. In contrast to
the above studies, Lentz et al. (2000) predicted that increas-
ing the metallicity of the progenitor should cause a decrease
in the level of the UV pseudocontinuum due to a decrease in
the line opacity and cooling. A much larger effect was pro-
posed by Timmes et al. (2003), who argued for a significant
increase of56Ni, and hence of the resulting luminosity, with
increasing metallicity (but see Howell et al. 2009). We note
that current theories do not offer a consensus picture as to the
effects of changing metallicity on the overall UV flux. UV
observations of a large sample of SNe Ia are needed to help
quantify the UV properties and constrain the above models.
The necessity of observing above the Earth’s atmosphere
has resulted in a sparse UV dataset for low-redshift SNe Ia;
see Panagia (2003, 2007) and Foley, Filippenko, & Jha
(2008a) for summaries of International Ultraviolet Explorer
(IUE) and Hubble Space Telescope (HST) spectra. Recently,
the Swift satellite obtained UV photometry of nearby SNe Ia
(Milne et al. 2010; Brown et al. 2010), as well as UV grism
spectra (Bufano et al. 2009; Foley et al. 2011). With the suc-
cessful repair of the Space Telescope Imaging Spectrograph
(STIS) onboard the HST during the 2009 Servicing Mission
4, obtaining UV spectra with STIS has also become possible
(Cooke et al. 2011). Observations of distant SNe Ia (with
ground-based telescopes) provide larger samples that probe
the UV spectral properties at high redshift (e.g., Ellis et al.
2008; Foley et al. 2008b; Sullivan et al. 2009), since the
rest-frame spectrum in the UV shifts to that in the optical re-
gion due to cosmic expansion. All of these datasets suggest
significant dispersion in the UV for SNe Ia at both low and
high redshifts. However, precise evolutionary constraints are
still limited by the absence of high-quality, low-redshift UV
In HST Cycle 13 (year 2004–2005), extensive UV obser-
vations of low-redshift SNe Ia were conducted (program GO-
shortly before the project started, the PR200L prism became
the only available element on HST capable of UV spectro-
ful UV spectra of four SNe Ia (SNe 2004dt, 2004ef, 2005M,
and 2005cf). These four objects represent three different sub-
HST UV Observations of SNe Ia3
classes of SNe Ia. Amongst our sample, SN 2004dt, and per-
haps SN 2004ef, can be put into the high-velocity (W09a)
or the high-velocity-gradient groups (Benetti et al. 2005).
Moreover, SN 2004dt has the highest line polarization ever
observed in a SN Ia (Wang et al. 2006), and is a clear out-
lier in the relationship between the Si II velocity gradient and
the nebular-phase emission-line velocity shift (Maeda et al.
2010). SN 2005M is a spectroscopically peculiar object like
SN 1991T (Filippenko et al. 1992a; Phillips et al. 1992),
with the early spectrum showing weak Ca II H&K lines and
prominent Fe III lines (Thomas et al. 2005). And SN 2005cf
has been regarded as a standard SN Ia (e.g., Wang et al.
2009b, hereafter W09b), belonging to the “Normal" or the
above four events reveal significant diversity of the UV prop-
erties amongSNe Ia. Good understandingof the UV diversity
forms the basis for further improvementsin the application of
SNe Ia as cosmological probes.
This paper is structured as follows. Observations and data
reduction are described in §2, while §3 presents our results
and a comparison with the existing UV data. Discussions are
given in §4, and we summarize our conclusions in §5.
FIG. 1.— Left panels: Direct images (F330W) of SNe 2004dt, 2004ef,
2005M, and 2005cf near maximum light obtained with HSTACS/HRC. Right
panels: Raw images near maximum light obtained in the ACS/HRC PR200L
slitless spectroscopy mode. The dispersed image is only a few times the
point-spread function of the direct images, indicating that the HST prism
spectra have very low resolution.
2. OBSERVATIONS AND DATA REDUCTION
UV observations of SNe 2004dt, 2004ef, 2005M, and
2005cf were performed using the Advanced Camera for Sur-
veys(ACS) high-resolutionchannel(HRC) in both the slitless
prism spectroscopy and the direct imaging modes, as shown
in Figure 1. Observations with the HRC PR200L generally
consist of a “direct" or undispersed image taken through the
F330W filter, used to establish the zero point of the wave-
length scale, and followed by dispersed images with one or
more exposures through the PR200L prism.
imagingobservationswere madethroughtheF220W, F250W,
and F330W filters, with the exposure time split into several
segments.Individual exposures were combined using the
MultiDrizzle task within the Space Telescope Science Data
geometric-distortion corrections. A log of the observations is
given in Table 1, listing the exposure times of the prism and
direct imaging exposures.
The UV spectra were extracted in PyRAF using the “aXe”
slitless spectroscopy reduction package (Kümmel et al. 2004,
2009). The spectral extraction within aXe relies upon the po-
sition, morphology,and photometryof the targets as observed
inthe accompanyingdirectimage. The spectralresolutionob-
tained through the PR200L prism decreases as a function of
increasing wavelength (e.g., from 5.3 Å pixel−1at 1800 Å to
200 Å pixel−1at 4000 Å), resulting in extreme “red pile-up"
pressed into just five pixels). Our analysis is restricted to the
useful wavelength range 1800–3500Å.
photometry on the drizzled images using an aperture radius
of 4 pixels (∼ 0.′′1). The background level was determined
from the median counts in an annulus of radius of 100–130
pixels. The measured magnitudes were further corrected to
an infinite-radius aperture and placed on the Vega magnitude
system (Sirianni et al. 2005). The final HST ACS UV magni-
tudes of SN 2004dt, 2004ef, 2005M, and 2005cf are listed in
The distances to the SNe were computed by using the re-
cession velocity v of their host galaxies (see Table 3), with a
Hubble constant H0= 74 km s−1Mpc−1(Riess et al. 2009).
For objects in the Hubble flow (vhelio? 3000 km s−1) (SNe
mined with respect to the 3 K cosmic microwave background
radiation; for the nearby sample with vhelio< 3000 km s−1
(SN 2005cf), the velocity was corrected to a 220 km s−1Vir-
FIG. 2.— UV light curves of the Hubble-flow SNe 2004dt, 2004ef, 2005M
obtained with the HST ACS/HRC and the F220W, F250W, and F330W filters,
together with the corresponding optical light curves in BVI bands (Gane-
shalingam et al. 2010). Overlaid are the UV-optical template light curves of
the “gold standard” SN Ia 2005cf (W09b).
4Wang et al.
3.1. Ultraviolet Light Curves
Figure 2 shows the UV light curves of SNe 2004dt, 2004ef,
and 2005M obtained with the HST ACS and the F220W,
F250W, and F330W filters. Also plotted are the correspond-
ing optical light curves (in BVI) from Ganeshalingam et al.
(2010). As SN 2005cf was well observed and is judged to be
a normal SN Ia and its light curves have been published else-
where (W09b), we use it as a template SN Ia for comparison
study. To account for effects of the redshift on the observed
SN flux, we applied K-corrections (Oke & Sandage 1968) to
all of the light curves using the observed spectra (as shown in
Figures 3 and 4) and the Hsiao et al. (2007) template.
The light curves have also been corrected for reddening in
the host galaxy as well as in the Milky Way. The total redden-
ing toward a SN Ia can be derived by comparing the observed
color with that predicted by the empirical relation between
the decline rate and color indexes (e.g., W09b, and references
therein). Absorption in the Milky Way (MW) was estimated
byusingthe reddeningmapsof Schlegel,Finkbeiner,& Davis
(1998) and the Cardelli, Clayton, & Mathis (1989) redden-
ing law with RV= 3.1, while extinction in the host galaxy
was corrected applying an extinction law with RV≈ 2.3 (e.g.,
Riess et al. 1996; Reindl et al. 2005; Wang et al. 2006;
Kessler et al. 2009). The difference in RV does not signifi-
cantly affect the light curves shown in Figure 2, since all four
SNe Ia have low host-galaxy reddening, e.g., E(B−V)host?
0.1 mag. Table 3 lists the relevant parameters of these four
FIG. 3.— The UV color curves of the Hubble-flow SNe 2004dt, 2004ef,
2005M. Overlaid are the color-curve templates of the “gold standard” SN Ia
One can see from Figure 2 that the UV light curves of our
SN Ia sample show a large dispersion, especially at shorter
wavelengths,while the optical counterpartsare generallysim-
ilar except in the I band where there is more diversity. Of our
sample, SN 2004ef appears relatively faint in all bands, con-
sistent with the fact that it is a slightly faster decliner with
∆m15(B) ≈ 1.46±0.06 mag. SN 2005M is a spectroscop-
ically peculiar object like SN 1991T. This supernova has a
slower decline rate (∆m15(B) = 0.86±0.05 mag), but it does
not appear to be overluminous either in the UV or optical
bands. However,ourrelativeignoranceregardingthe intrinsic
color of those slow decliners might prevent us from deriving
reliable reddenings for them. A notable feature in the plot is
that the UV emission of SN 2004dt appears unusually strong
imum light. Moreover, SN 2005M seems to peak a few days
earlierthanSN 2005cfin the F220WandF250W bands. Such
a temporal shift seems to exist for SN 2004dt and SN 2004ef,
althoughtheir light curvesin UV are not well sampled around
The color curves of our four SNe Ia, corrected for the MW
and host-galaxy reddening, are presented in Figure 3. One
can clearly see that their UV −V curves show remarkabledif-
ferences despite having similar B−V curves. SN 2004dt is
quite peculiar, being very blue compared with the other three
objects. The colors derived from Swift UVOT photometry of
SN 2008Q were similarly blue (Milne et al. 2010).
Since SN 2004dt has ∆m15(B) ≈ 1.1 mag and a peak B−V
color (∼ 0.0 mag) similar to that of SN 2005cf, it is of inter-
est to perform a detailed comparison of their UV properties.
Around maximum light, SN 2004dt is bluer than SN 2005cf
by ∼ 0.8 mag in F250W −V and by ∼ 2.0 mag in F220W
−V. About 2–3 weeks after maximum, the color difference
becomes less significant as a result of the rapid decline of the
post-maximumUV emission in SN 2004dt(see Fig. 6 and the
discussion in §3.2). In the first 15 days after the maximum,
the decline of the UV light curve is measuredto be ∼1.9 mag
in F220W, ∼2.4mag in F250W, and ∼2.5mag in F330W for
SN 2004dt. These are all largerthan the correspondingvalues
measured for SN 2005cf by ∼ 0.6–0.9 mag (see also Table
10 in W09b). A faster post-maximum decline usually corre-
sponds to a shorter rise time for the light curves of SNe Ia in
the optical bands, as evidenced by the fact that they can be
better normalized through the stretch factor (Perlmutter et al.
1997; Goldhaber et al. 2001). For SN 2004dt, a shorter rise
time in the UV is consistent with the high expansion velocity
of the UV photosphereto higher velocities and hence result in
a larger UV flux. The possible origin of the UV excess in SN
2004dt in the early phase is an interesting issue and is further
discussed in §4.2.
3.2. Ultraviolet Spectroscopy
As shown in Table 1, a total of 34 HST ACS PR200L prism
spectra have been collected for SNe 2004dt, 2004ef, 2005M,
and 2005cf. In column (5) of Table 1, we list the total expo-
sure time for each spectrum, resulting from a series of coad-
ded CR-split exposures. Among the four SNe Ia observed
with the HST ACS, SN 2005cf and SN 2004dt have better
their UV spectra (see Figures 4 and 5). The UV spectra of SN
2004ef have limited S/N, due to the fact that the object is rel-
atively faint and is located at a larger distance (∼ 120 Mpc).
Onlyasingle UV-prismspectrumwas obtainedofSN 2005M.
The UV spectra of SN 2004ef and SN 2005M are shown in
3.2.1. SN 2005cf
Figure 4 shows the evolution of the combined UV-optical
spectrum of SN 2005cf from t = −9 days to t = +24 days
relative to the B-band maximum. The ground-based spectra
are taken from W09a and Garavini et al. (2007). Spectra
obtained within 1–2 days of each other were combined di-
rectly. In cases where no contemporaneous optical spectrum
HST UV Observations of SNe Ia5
was available, an interpolated spectrum was used. After flux
calibration with the corresponding photometry, the UV-prism
spectra agree reasonably well with their optical counterparts
over the spectral range 3300–3500Å.
FIG. 4.— Evolution of the UV-optical spectrum of SN 2005cf. The UV
spectra were obtained with the HST ACS prism (PR200L), and the corre-
sponding optical data are taken from W09b and Garavini et al. (2007). All of
the spectra have been rescaled to match the UV-optical photometry, and were
adjusted to the peak of the spectral flux and arbitrarily shifted for display.
The dashed line marks the position of the ∼ 3250 Å absorption feature.
It is knownthat in the UV spectra ofSNe Ia, almost noindi-
vidual lines can be identified. This is because the vast number
of iron-group element lines in the UV overlap strongly, elim-
inating individual line features. This contrasts with the opti-
cal, where P Cygni lines of neutral or singly ionized ions of
intermediate-masselements(mostlyO,Mg,Si, S,andCa) can
usually be identified. Probably the only individual feature in
SN IaUV spectrabluewardof∼3700Å thatcan beobviously
identified is the P Cygni line (absorptionfeature at ∼ 2650Å)
caused by Mg II λ2800 which was seen in HST spectra of SN
1992A at t = 5 days after B maximum (Kirshner et al. 1993).
In our spectra, we see no clear evidence of this line or any
other line-like feature blueward of ∼ 2800 Å. In most of our
spectra, this is probably caused by the low S/N and low reso-
lution. In the case of SN 2004dt, the effect that led to strong
UV emission (e.g., circumstellar matter interaction)may have
smoothed out line-like features (see the discussion below).
Inspecting the near-maximum UV spectra in the wave-
length range 1800–3500 Å, we identify a prominent absorp-
tion feature near 3250 Å. It is also seen in the Swift UVOT
UV-grism spectrum of SN 2005cf near maximum light (e.g.,
Bufano et al. 2009; see also Fig. 8) and the STIS UV spectra
of the Palomar Transient Factory sample of SNe Ia (Cooke et
al. 2011). However, we note that the relatively weak absorp-
tion featurenear∼3050Å seen in STIS and Swift spectra was
nearly invisible in the HST UV spectrum, indicating the lim-
ited resolution of the HST/ACS prism. The resolution effect
can be demonstrated by degrading the Swift spectrum. The
∼ 3050 Å feature becomes very weak in the Swift spectrum
degraded to a resolution of ∼ 50 Å pixel−1(see the dash-
dotted line in Fig. 8), and this feature almost disappears when
further decreasing the spectral resolution to ∼ 100 Å pixel−1
(see the dashed line in Fig. 8). The absorption feature near
3250 Å also becomes less significant at a resolution ∼ 100 Å
pixel−1, consistent with that observed in the HST prism spec-
The flux ratio RUV, defined as fλ(2770 Å)/fλ(2900 Å), has
been proposedto correlate with the peak luminosity of SNe Ia
(Foley, Filippenko, & Jha 2008). Our HST spectra may not be
appropriate for testing such a correlation due to their limited
resolution (? 40 Å pixel−1at 2800–2900 Å); it could have
significantimpact on the intrinsic UV slope within such a nar-
row wavelength range.
The near-UV features in the ∼ 2700–3500 Å range of the
early-time spectra of SNe Ia were initially studied by Branch
& Venkatakrishna (1986), who suggested that they are pro-
duced by blends of Fe II and Co II lines. In particular, the
Co II absorption (rest wavelength 3350–3500 Å; Branch et
al. 1985), and the absorption feature at ∼ 3050 Å is ascribed
to Fe II absorption. Based on their study of the prominent
example SN 1992A, Kirshner et al. (1993) confirmed that
the near-UV region is largely formed by a complex blend of
iron-group element lines. They found that Cr II, Mn II, and
Fe II contribute significantly to the absorption-like feature at
∼ 3250 Å while the ions Co II and Ni II make no significant
contributionbecause the newly synthesized Ni-Co is confined
to the inner regionsat these early phases. The spectral proper-
ties of SNe Ia in the UV were also investigatedby Sauer et al.
(2008), based on the early-time spectra of SN 2001eh and SN
2001ep obtained with HST STIS. In their study, the contribu-
tion of the Ti II to the ∼3050 Å feature and of doublyionized
species such as Fe III and Co III to the ∼ 3250 Å features of
SN 2001eh are important. The absorption features blueward
of 2700 Å in the spectrum are also thought to originate from
singly and doubly ionized Fe and Co (Kirshner et al. 1993;
Sauer et al. 2008).
3.2.2. SN 2004dt
The evolution of the combined UV-optical spectrum of SN
2004dt is displayed in Figure 5. The UV spectroscopic ob-
servations of SN 2004dt cover the period between t ≈ −2
days to t ≈ +50 days with respect to the B-band maximum.
The optical data shown in the plot are taken from Altavilla
et al. (2007). At t = −2 days and t = +1 days, it is evi-
dent that there is a prominent peak emerging between 2900 Å
and 3400 Å. The absorption-like feature at ∼ 3250 Å, as
seen in SN 2005cf and other normal SNe Ia (Bufano et al.
2009, and this paper), is very weak and nearly invisible in
SN 2004dt. This feature becomes prominent in the t = +15
day and t = +41 day spectra, but it is weak at other epochs
after maximum brightness. This variation may be caused by
6 Wang et al.
FIG. 5.— Evolution of the UV-optical spectrum of SN 2004dt. The UV
spectra were obtained with the HSTACS/HRC PR200L prism, and the optical
data are from Altavilla et al. (2007). All of the spectra have been rescaled
to match the UV-optical photometry, and were normalized to the flux peak
of the spectrum and arbitrarily shifted for display. The vertical dashed line
marks the position of the absorption feature near ∼ 3250 Å .
the poor S/N, the lower resolution of the PR200L spectrum,
possible errors in the data reduction, or a combination of the
above factors. Moreover, the overall spectral flux in the UV
region drops rapidly after maximum light, consistent with the
rapid decline of the UV light curves (see Fig. 2).
The strengthening of the ∼ 3200 Å feature in the post-
maximum spectra is particularly interesting in the case of SN
2004dt. It may reflect an increase in the line blanketing opac-
ity as a result of a retreat of the photosphere into the iron-
group-rich inner region of the ejecta, which is supported by
the fact that the ∼ 5000 Å feature (due to blends of Fe II
and Fe III) gradually gained strength since maximum bright-
ness. The Doppler velocity of Si II λ6355 is found to be ∼
14,000 km s−1for SN 2004dt from the near-maximum spec-
trum, which fits comfortably in the high-velocity category
(W09a). This object is also a member of the high-velocity-
gradient group (Benetti et al. 2005) and the broad-line (BL)
subclass (Branch et al. 2006). It is interesting to examine
whether the distinct UV behavior correlates with the produc-
tion of the high-velocity features in SNe Ia. The origin of the
high-velocityfeatures is still unclear. They could be due to an
increase of the abundance and/or density in the outer layers,
circumstellar material interaction (e.g., Gerardy et al. 2004;
Mazzali et al. 2005), or a clumpy distribution of the outer
ejecta (Leonard et al. 2005; Wang et al. 2006).
3.2.3. SN 2004ef and SN 2005M
FIG. 6.— Evolution of the UV spectra of SN 2004ef (solid curves) and
SN 2005M (the dashed curve). The UV spectra were obtained with the HST
ACS/HRC PR200L prism.
In Figure 6, we show the evolution of the UV spectrum of
SN 2004ef, as well as a single-epoch spectrum of SN 2005M.
The UV spectroscopic observations of SN 2004ef span from
t ≈−2 days tot ≈+29 days with respect to the B-bandmaxi-
mum. A temporalsequenceof the combinedUV-opticalspec-
tra cannot be constructed for SN 2004ef due to the sparse op-
tical data at similar phases.
As in SN 2005cf and SN 2004dt, the UV spectra of SN
2004ef show a broad and shallow absorption trough near
the ionized iron-group elements (Kirshner et al. 1993). Note
that the prominentabsorptionfeaturenear 3250Å, commonly
seen in other SNe Ia, is invisible in the UV spectra of SNe
2004ef and 2005M. This inconsistency reminds us that the
limited S/N and low resolution of the spectra could have sig-
nificant influence on the observed features of the spectra.
3.2.4. Comparison of the UV spectra
In Figure 7, we compare the near-maximum UV spectra of
our four nearby SNe Ia with other mean low-redshift SN Ia
spectra at similar phases. The spectra have been corrected for
reddening in the Milky Way. The mean low-redshift spectra
come from several studies, including the mean spectrum con-
structed primarily with the HST STIS spectra (Cooke et al.
2011), the combined spectrum of a few nearby SNe Ia with
archival HST and IUE data (Foley et al. 2010), and the spec-
tral template established by Hsiao et al. (2007). As Cooke
et al. (2011) did not apply any extinction correction in build-
ing their mean low-redshift spectrum, we apply a reddening
correction to their spectrum, with E(B−V) = 0.10 mag and
RV=3.1, to correctfor the color difference. All of the spectra
shown in the plot are normalized to approximately the same
flux in the rest-frame wavelength range 4000–5500 Å. Note
that the Hsiao et al. (2007) template spectra contain high-
redshift SN Ia data from the Supernova Legacy Survey (Ellis
et al. 2008); hence, they may suffer from evolutionary effects
and may not represent the genuine spectrum of local SNe Ia
in the UV region.
TheFoleyet al. spectrum(meanphaset ≈0.4days)andthe
Hsiao et al. t = −1 day spectrum agree fairly well within the
uncertainties, and the STIS UV spectrum (mean phaset ≈1.5
days)also agreesfairlywell giventhedifferencein epoch. We
see that there is a general shift of UV peak features to the red
between t ≈ −1 day and t ≈ +1 day. Most notably, the large
HST UV Observations of SNe Ia7
FIG. 7.— Comparison of the near-maximum-light UV-optical spectra of SNe 2004dt, 2004ef, 2005M, and 2005cf with the mean spectra of the low-redshift
SNe Ia constructed by Cooke et al. (2011), Foley et al. (2010), and Hsiao et al. (2007). Note that the low resolution of our HST spectra may play a substantial
role in affecting the spectral features below 3500 Å as shown in Fig. 8.
FIG. 8.— Comparison of the UV-optical spectra of SN 2004dt and SN
2005cf at three selected epochs (t = −2, +5, and +14 with respect to B-
band maximum). The dash-dotted (blue) and dashed (green) lines represent
the Swift UV spectra of SN 2005cf degraded to a resolution of ∼ 50 Å and
∼ 100 Å respectively.
peak near 3100 Å at t ≈ −1 day shifts to ∼ 3150 Å at t ≈ +1
day. The underlying cause of the shift is probably just the re-
cession in velocity of the UV photosphere with the expansion
of the ejecta, but detailed modelingwould be needed to verify
Of our sample, SN 2004dt appears to be unusually bright
in the UV. Its UV emission is found to be stronger than
the Foley et al.low-redshift mean spectrum for t ≈ −1
day by ? 75% (or at a confidence level ? 6–7σ) at wave-
lengths 2500–3500 Å. Differences are also present in the op-
tical portion of the spectrum where the absorption features of
intermediate-mass elements (IMEs) are relatively strong and
highly blueshifted (except for the S II lines), although the in-
tegrated flux over this region does not show significant differ-
ences with respect to the templates. The ∼5000 Å absorption
feature due to Fe II and Fe III blends appears to be rather
weak, perhaps suggestive of less iron-peak elements in the
outer ejecta of SN 2004dt.
nent features of both IMEs and Fe-group elements, with the
UV emission being weaker than the local compositespectrum
by about 25%. Another notable aspect of Figure 6 is the defi-
ciency of the UV emission in SN 2005M that is characteristic
of the strong Fe II and Fe III in the earliest spectrum. SN
2005M has a smaller ∆m15; however, the UV flux emitted in
the wavelengthregion2500–3500Å is foundto be evenlower
than the mean value by ∼ 22%. One can see that significant
scatter is observed in these three events at wavelengths below
∼ 3500 Å. In this wavelength region, the continuum and the
spectral features were proposed to be primarily shaped by the
heavy elements such as Fe and Co (Pauldrach et al. 1996).
Thus, the large scatter in our sample might be related to vari-
ations of the abundance of Fe and Co in the outer layers of
8Wang et al.
the exploding white dwarf (Höflich et al. 1998; Lentz et al.
2000; Sauer et al. 2008). On the other hand, the spectrum of
SN 2005cf generally matches well the nearby comparison ex-
cept for the Ca II H&K feature, which exhibits the strongest
difference in the optical region for our sample of four events.
The variations of Ca II and Si II absorptions at higher ve-
locities suggest that additional factors, such as asphericity or
different abundances in the progenitor white dwarf, affect the
outermost layers (Tanaka et al. 2008).
Figure 8 shows a more detailed comparison of the UV-
optical spectra of SN 2004dt and SN 2005cf at three different
epochs (t ≈ −2 d, +5 d, and +14 d). The integrated fluxes
of the spectra have been normalized over the 4000–5500 Å
wavelength range. These two SNe Ia provide us good exam-
ples to examine the generic scatter in the UV region as they
have quite similar photometricpropertiesin the optical bands,
such as the B−V color at maximum and the post-maximum
decline rate ∆m15(B) (see Table 3). At t ≈ −2 day, the nor-
malized flux obtained for SN 2004dt at 2500–3500Å is much
brighter than for SN 2005cf in the same wavelength region,
with the flux ratio F04dt/F05cf≈ 1.9. This flux ratio in the
UV decreases quickly to ∼ 1.2 at t ≈ +5 day, and it becomes
comparable to 1.0 at t ≈ 2 weeks after maximum. Neither
contamination by the background light from the host galaxy
noruncertaintyinthereddeningcorrectionis likely toaccount
for such a peculiar evolution of the UV flux for SN 2004dt
since the influence from the background emission would be-
come more prominent as the object dims and SN 2004dt does
not suffer significant reddening in the host galaxy (see §4).
Such a fast decline of the UV flux is rarely seen in the exist-
ing sample of SNe Ia, possibly an indication of differences in
the explosion physics or progenitor environment with respect
to normal SNe Ia.
3.3. General Properties of the
Ultraviolet Luminosity of SNe Ia
ThepeakluminosityofSNe Iain the opticalbandshas been
extensively studied and found to correlate well with the width
of the light curve around maximum light. Outliers have been
identified as abnormal objects with different explosion mech-
anisms such as the overluminous objects SN 2003fk (Howell
et al. 2006) and SN 2009dc (Silverman et al. 2011; Tauben-
berger et al. 2011), and the underluminous events SN 1991bg
(Filippenko et al. 1992b) and SN 2002cx (Li et al. 2003).
Examining the relationship between the UV luminosity and
∆m15(B) maydisclosea diversity,evenforthe so-called“nor-
mal” SNe Ia, since the UV emission is thought to be more
sensitive to possible variations of the explosion physics or
We have examined a sample of 20 SNe Ia with good pho-
tometry in the UV. Four among this sample are from the HST
observations presented in this paper and the rest are Swift ob-
jects published by Milne et al. (2010). The left panels of
Figure9 show themaximumabsolutemagnitudesofthese ob-
jects in broadbandU, B, and Swift uvw1/HST F250W versus
∆m15(B)63. Following the analysis by Brown et al. (2010),
we applied the red-tail corrections to the uvw1/F250W mag-
nitudes to mitigate the effects of the optical photons on the
63As the instrumental response curve of the HST ACS F250W filter is
similar to that of the Swift uvw1, the magnitudes measured in these two fil-
ters should be comparable. This is demonstrated by the observations of SN
2005cf, forwhich the measurements by the HST ACSand the Swift UVOT are
consistent to within 0.1 mag (W09b). We thus neglect magnitude differences
measured in the F250W and the uvw1 filters in our analysis.
UV flux. The U and uvw1/F250W magnitudes are also K-
corrected for the large variation in the spectral flux at shorter
wavelengths. Note that the UV/blue event SN 2004dt war-
rants a second spectral sequence for the sake of red-tail and
K-corrections. The peak magnitudes in the uvw1/F250W fil-
ters are obtained by fitting the data with a polynomial or the
template light curve of SN 2005cf. The parameters in the op-
tical bands, such as the peak magnitudes and the B-band light
curve decline rate ∆m15(B), are estimated from the published
light curves (see Table 3 and the references).
Assuming an extinction law with RV= 2.3, linear fits to the
subsample in Figure 9 with 0.8 < ∆m15(B) < 1.7 mag yield
a root-mean square scatter (i.e., σ values) of ∼ 0.2 mag in B
and ∼ 0.4 mag in uvw1. Note that SN 2004dt and SN 2006X
were excluded from the fit. The uvw1-band peak luminosity
and larger scatter than that in B and U. Given the fact that
the dust obscuring the SN may have different origins (e.g.,
Wang 2005; Goobar 2008; W09a), the host-galaxy extinction
corrections with RV= 3.1 (Milky Way dust) and RV= 1.8
(LMC dust) are also applied to the absolute magnitudes. In
both cases, the scatter also increases at shorter wavelengths.
The relevant results of the best linear fit with different values
of RVto the Mmax−∆m15(B) relation are shown in Table 4.
The wavelength-dependent scatter can be driven in part by
the uncertainty in the absorption corrections as the UV pho-
tonsare morescatteredbythe dust thanthe optical. Assuming
a mean error ∼ 0.04 mag in E(B−V)host(e.g., Phillips et al.
1999;Wangetal. 2006),theresidualscatter oftheuvw1-band
luminosity can be as large as ∼ 0.3 mag, which is still much
larger than that found in B (∼0.15 mag). A notable feature in
Figure 8 is the UV excess seen in SN 2004dt and perhaps SN
2006X; they are found to be brighter than the corresponding
mean value by ∼ 6σ and ∼ 3.0σ, respectively. The discrep-
ancy of the observed scatter cannot be caused by the error in
distance modulus, which applies equally for the UV and op-
tical bands. This is demonstrated by the correlation between
the peak colors and ∆m15(B), as shown in the right panels
of Figure 9. The color correlation is distance independent,
but it shows significantly large scatter in the UV. In particular,
the uvw1/F250W −V colors of SN 2004dt and SN 2006X are
found to be bluer than the mean value by ∼ 1.3 mag. This
indicates that the uvw1/F250W filter is perhaps more a probe
of SN Ia physics than a cosmological standard candle.
Strong emission in the UV is likely an intrinsic property of
SN 2004dt rather than being due to improper corrections for
the dust reddening, since neither the interstellar Na I D lines
in the spectra nor the observedcolor indices reveal any signif-
icant reddening in the host galaxy (e.g., E(B−V)host? 0.1).
Moreover, SN 2004dt still appears much brighter and bluer
in the UV than the mean values defined by the other SNe Ia
even if it is assumed to be intrinsically red by ∼ 0.1 mag with
respect to the normal objects. On the other hand, SN 2006X
is heavily extinguished by dust having an abnormal extinc-
tion law, perhaps with RV≈ 1.5 (Wang et al. 2008a). Apply-
ing this extinction correction (which is an extrapolation from
the results of Wang et al. 2008a) and the red-tail correction
(which is an extrapolation from the results of Brown et al.
2010) to the UV magnitudes of SN 2006X would make it ap-
pear brighter than the other comparisons by ∼ 1.7±0.6 mag
in uvw1/F250W. We caution, however, that this result may
suffer large uncertainties from the speculated extinction and
HST UV Observations of SNe Ia9
FIG. 9.— Left panels: the uvw1-, U-, and B-band maximum magnitudes versus the B-band decline-rate parameter ∆m15(B) of SNe Ia with UV observations
from the Swift UVOT and HST. Right panels: the peak colors plotted versus ∆m15(B). Thesolid lines represent the linear fit to the SNewith 0.8<∆m15(B)<1.7
mag, and the dashed curves represent the 3σ uncertainties. SN 2004dt and SN 2006X are not included in the fit, while the open symbols show the case assuming
that these two SNe Ia have intrinsically redder color by ∼ 0.1 mag with respect to the other objects.
FIG. 10.— Flux ratio of the UV (1800–3500 Å) portion to the optical flux
(3500–9000 Å), measured from the flux-calibrated UV-optical spectra of SNe
2004dt, 2004ef, 2005M, 2005cf, 2005ke, and 2006X (see text for the sources
of the data). The dotted line is the flux ratio computed from the template
spectra from Hsiao et al. (2007); the dashed curve is the flux ratio estimated
from a few nearby SNe Ia with the UV observations (Stanishev et al. 2007).
4.1. Bolometric Light Curve and Nickel Mass
To examine how the relative flux in the UV evolves since
the SN explosion, we compute the ratio of the UV emission
(1800–3500 Å) to the optical (3500–9000 Å), FUV/Fopt, for
a few SNe Ia, as shown in Figure 10. The flux ratios ob-
tained for SN 2005cf and SN 2005ke are overplotted. The
dotted curve represents the Hsiao et al.
and the dashed curve shows a combined template of SNe
1981B, 1989B, 1990N, 1992A, and 2001el (Stanishev et al.
2007). The flux ratios were calculated by integrating the flux-
calibrated, UV-optical spectra except for SN 2005ke and SN
2006X.Theintegratedfluxes of these two SNe Ia are obtained
approximately by the mean flux multiplied by the effective
width of the passband. It is noteworthy that the flux ratio
shown for SN 2006X might be just a lower limit, as we did
not include the flux contribution below 2500 Å in the anal-
ysis because of the larger uncertainties in the extinction and
red-tail corrections at shorter wavelengths (e.g., Brown et al.
We notice that the FUV/Foptvalue of SN 2006X peaks at
t ≈ 10 days before the B-band maximum, about 5 days earlier
thanforSN2005cfandthetemplates. Thisfeatureis similarly
observed in the fast-decliner SN 2005ke, which exhibited a
possible signature of X-ray emission as well as evidence for
a UV excess in the early nebular phase, perhaps suggestive of
circumstellar interaction (Immler et al. 2006). Owing to the
lack of earlier UV data, we cannot conclude whether such a
feature is present in SN 2004dt. It is apparent that the flux
ratio FUV/Foptexhibits a large scatter at comparable phases
for the selected sample of SNe Ia. Consequently, the spread
of the nickel masses from the light-curve peaks.
The overall properties of the SNe can be represented by
their quasi-bolometric (UV/optical/IR) light curves shown
in Figure 11. The near-infrared (NIR) photometry of SNe
2004ef and 2005M was taken from Contreras et al. (2010).
The NIR emission of SN 2004dt was corrected on the ba-
sis of SN 2005cf (W09b). Similar corrections were applied
to the comparison SNe Ia when the NIR observations were
not available. Overall, the quasi-bolometric light curves of
our SNe Ia are very similar in shape, with the exception of
SN 2004dt, which shows a prominent “bump" feature. This
“bump" is consistent with the secondary maximum visible in
the I band (see Fig. 2), being more prominent and occurring
10Wang et al.
FIG. 11.— TheUV/optical/IR quasi-bolometric light curves of SNe 2004dt,
2004ef, 2005M, and 2005cf. Overplotted are the corresponding light curves
of the comparison SNe Ia.The numbers in parentheses represent the
∆m15(B) values for the SNe Ia.
10 days earlier than that of the other comparisons. This ob-
served behavior suggests that SN 2004dt may have a larger,
cooler iron core, or a higher progenitor metallicity according
to the study of the physical relation between a supernova’s
NIR luminosity and its ionization state (Kasen 2006).
The bolometric luminosity at maximum light of SN 2004dt
is estimated to be Lmax≈ 1.7 ± 0.2 × 1043erg s−1with
RV= 2.3 and E(B−V)host= 0.05 mag. This value is con-
sistent with that obtained for SNe Ia with similar ∆m15such
as SN 2005cf (∼ 1.5±0.2×1043erg s−1) and SN 2007af
(∼ 1.3±0.2×1043erg s−1) within 1–2σ errors. With the de-
rived peak luminosity, we can estimate the56Ni mass ejected
during the explosion (Arnett 1982). According to Stritzinger
& Leibundgut (2005), the56Ni mass (MNi) can be written as
a function of the bolometric luminosity at maximum and the
rise time tr:
∼ 18 d for SN 2004ef, and ∼ 24 d for SN 2005M (e.g., Gane-
shalingam et al. 2011), the corresponding mass of the ejected
56Ni is roughly estimated to be 0.9±0.1 M⊙, 0.8±0.1 M⊙,
0.4±0.1 M⊙, and 0.7±0.1 M⊙, respectively.
We see that the deduced MNiof the four SNe Ia shows sig-
nificant scatter, and the value for SN 2004dt is slightly larger
than the normal values. The net effect of a larger MNiwould
tendto delaythe secondarymaximumof theNIR light curves,
inconsistent with what is seen in SN 2004dt. This inconsis-
tency can be resolved if SN 2004dt has a shorter rise time,
tr≈ 18.0 d. Given the observational evidence that longer tr
usually corresponds to slower post-maximum decline of the
light curve (Ganeshalingam et al. 2011), SN 2004dt may rise
to maximum at a faster pace. A shorter rise time is likely
to be a common feature of the high-velocity SNe Ia (Zhang
et al. 2010; Ganeshalingam et al. 2011). Therefore, the
deduced nickel mass of SN 2004dt may have been overesti-
mated to some extent. On the other hand, the implied MNifor
SN 2005M seems quite normal and is lower than in slowly
declining SNe Ia such as SN 1991T, indicating that the light-
curve widths are not a single-parameter family of the ejected
4.2. Origin of the UV Excess in SN 2004dt
TheUV emissionofSNe Iais thoughtto originatepredomi-
nantly from the outer layers of the ejecta because UV photons
producedin deeperlayers are subject to line blanketingby the
wealth of bound-boundtransitions associated with iron-group
elements. Therefore, UV features are a promising probe to
study the composition of the outer ejecta of SNe Ia.
Previous work has shown that the UV is particularly sen-
sitive to the metal content of the outer ejecta (line-blocking
effect) and their ionization (backwarming effect: Lentz et al.
2000; Sauer et al. 2008). Lentz et al. (2000) suggested a
correlation between the emitted UV flux and the progenitor
On the other hand, Sauer et al. (2008) proposed that the UV
dueto an enhancedreverse-fluorescencescattering of photons
from red to blue wavelengths and a change in the ionization
fraction (backwarming effect). According to the model se-
ries for SN 2001eh and SN 2001ep (Sauer et al. 2008), as
well as the prediction by Lentz et al. (2000), an abundance
change of about ±2.0 dex (e.g., variations from 1/10 normal
metallicity in the C+O layer to 10 times normal metallicity
in the C+O layer) could lead to a change of up to ∼ 0.3 mag
in F250W/uvw1. This is much smaller than that observed in
SN 2004dt. Note that the above studies are based on the one-
dimensional deflagration model W7 (Nomoto et al. 1984).
To illustrate the effect of different progenitor metallicity
on the UV spectra in more detail, we also consider a sub-
Chandrasekhar-mass detonation model proposed by Sim et
al. (2010). In particular, we adopt the model of a 1.06M⊙
CO white dwarf (model 1.06), which was found to give good
agreement with the observed properties of normal SNe Ia in
theoptical. Forthis model,Sim etal. (2010)studiedtheeffect
of progenitor metallicity on the nucleosynthetic yields and, in
turn, the synthetic observables. For that purpose they polluted
the initial CO white dwarf with 7.5%22Ne (corresponding to
a progenitor metallicity of ∼ 3Z⊙). From this they find that
find the B-band peak magnitude to be ∼ 0.5mag dimmer for
the model with Z = 3Z⊙compared to the model with Z = 0).
In Figure 12(a), we show the UV spectra of these models at
the maximum light, which were obtained using Monte Carlo
radiative transfer code ARTIS (Kromer & Sim 2009). To get
better coverage in progenitor metallicity, we added another
model at Z = Z⊙ which was obtained in exactly the same
manner as described by Sim et al. (2010). It is clearly seen
that the UV flux increases significantly with decreasing pro-
genitor metallicity. However, comparing the synthetic spectra
to SN 2004dt,it is also obviousthat the metallicity effect can-
not be the dominant factor responsible for the unusual bright-
ness of SN 2004dt in the UV, as it isn’t possible to enhance
the UV flux farther than for the model with zero metallicity.
SN 2004dt shows line-expansion velocities which are ap-
parently larger than those of normal SNe Ia. To investigate
if these higher ejecta velocities can explain the UV excess
of SN 2004dt, we took the standard deflagration model W7
(Nomoto et al. 1984), which is known to reproduce the ob-
served maximum-light spectra of SNe Ia very well.
t∼0 day models shown in Figure 12(b) were calculated using
the time-dependent PHOENIX code in local thermodynamic
equilibrium (LTE; Jack et al. 2009; 2011). The velocities in
W7 were increased by a uniform factor (20% or 40%). The
densities were adjusted so that the total mass was conserved.
HST UV Observations of SNe Ia 11
FIG. 12.— Left panel: Comparison of the UV spectra of SN 2004dt near maximum light with synthetic UV spectra produced at comparable phase by the
sub-Chandrasekhar detonation model with different metallicity. Right panel: Similar comparison with the synthetic spectra, but obtained with the W7 model and
different expansion velocities.
Figure 12(b) clearly shows enhanced UV emission when the
velocities are larger; the effects of line blending in the UV
serve to increase the radius of the UV pseudo-photosphere.
Normalizing the spectral flux over the 4000–5500 Å region,
we find that the UV flux emitted in the 2500–3500 Å region
can be increased by about 40% if the expansion velocity vexp
is increased by 20% everywhere in W7. Note that increasing
vexpby 40% does not further enhance the UV flux, and the re-
sulting fluxincrease in the 2500–3500Å regiondropsto 34%.
This indicates that an increase of the expansion velocity can
lead to more UV flux, but it cannot reproduce the very large
flux enhancement near 3000 Å in SN 2004dt.
Interestingly, SN 2004dt is the most highly polarized SN Ia
ever observed. Across the Si II line its polarization PSi II
reaches up to 1.6% at ∼1 week before maximum light (Wang
et al. 2006), indicating that its Si II layers substantially de-
part from spherical symmetry. Among the comparison sam-
ple, SN 2006X also has a large degree of polarization in the
early phase, with PSi II≈ 1.1% at 6 d before maximum light
(Patat et al. 2009). Thus, it could well be that viewing-angle
effects play a major role in the observed UV excess of SN
2004dt. Breaking of spherical symmetry in the explosion is
also thoughtto be a criticalfactorresponsibleforthe observed
scatteramongSNe Ia(e.g.,Wang, Baade,&Patat 2007;Kasen
et al. 2009; Maeda et al. 2010; Maund et al. 2011).
Kromer & Sim (2009) recently studied the effect of asym-
metric ejecta on the light curves and spectra of SNe Ia us-
ing an ellipsoidal (prolate) toy model. They found that SNe
observed along the equator-on axis are always brighter than
those observed along the pole-on axis. This effect is strongest
in the bluer bands because the photons at short wavelengths
are more strongly trapped than photons in other bands and
therefore tend to preferentially leak out along the equatorial
plane where the photospheric velocity is smallest. Around
maximum light, the difference ∆M = Mpole−Mequatormea-
sured in the U band could be ∼ 0.4 mag larger than that in
theV band (Kromer & Sim 2009). After maximum light, this
line-of-sight effect becomes weaker with time as the ejecta
become optically thin at these wavelengths. Thus, the UV
excess seen in SN 2004dt could also be due to a geometric
We have described how there are several competing possi-
ble explanationsfor the enhancedUV flux in SN 2004dt. This
is due to the complex nature of line blanketing in the barely
optically thick, but highly scattering dominated differentially
expanding SN Ia atmosphere. This effect was described in
part by Bongard et al. (2008), who showed that the com-
plete spectrum is formed throughout the semi-transparent at-
mosphere and that Fe III lines produce features that are im-
printed on the full spectrum, partially explaining the UV ex-
et al. (2008) and those of Höflich et al. (1998) and Lentz et
al. (2000) are perhaps due to the complex nature of spectral
formation, ionization, and radiative transfer effects.
While asymmetry could play some role, there seems to be
enough variation in spherically symmetric models due to ef-
fects of varying ionization that is likely produced by a den-
sity profile that differs from that of W7. Recall that W7 has
a density bump that occurs when the flame is quenched and
momentum conservation causes material to pile up ahead of
the dying flame. Also, the higher brightness and indications
of a somewhat higher nickel mass in SN 2004dt will affect
the ionization state of the iron-peak elements, which are al-
most certainlyresponsiblein part for the observedUV excess.
More UV observations of nearby SNe Ia are needed to thor-
oughly understand these effects, and whether they are due to
asymmetries or other phenomena in the explosion.
We have presented HST ACS UV photometry and spec-
troscopy of SNe Ia obtained during Cycle 13 (2004–2005).
These data include 34 ACS prism spectra and 110 photomet-
ric observations of four SNe Ia (SNe 2004dt, 2004ef, 2005M,
and 2005cf). The spectral analysis is limited by the low res-
olution and low S/N of the data. However, comparison with
the existing low-redshift mean spectra of SNe Ia clearly in-
dicates that significant dispersion exists at wavelengths be-
low ∼ 3500 Å. In particular, SN 2004dt is found to show
a prominent, broad emission peak at 3000–3500 Å in the
near-maximum spectrum, rather than having an absorption-
like feature at ∼ 3250 Å similar to that in normal SNe Ia such
as SN 2005cf. Another interesting feature of SN 2004dtis the
rapid decline of its UV emission after the peak.
Based on a larger sample of SNe Ia, we studied the prop-
erties of their peak luminosity in the UV region. The lumi-
12Wang et al.
nosity in uvw1/F250W shows a correlation with the light-
curveparameter∆m15(B), but with significantlylargerscatter
than that found in the optical: ∼ 0.4 mag in uvw1/F250W
vs. ∼ 0.2 mag in B. The increased dispersion in the UV
has also been noted in other studies with independent SN Ia
samples of both photometry (e.g., Guy et al. 2007; Brown
et al. 2010) and spectra (Ellis et al.
2010; Cooke et al. 2011), which is likely intrinsic and has
beeninterpretedas compositionaldifferencesbetween events.
However, variations of the abundance based either on the W7
model or the sub-Chandrasekhar-mass detonation model can-
not account for the UV excess seen in SN 2004dt. The W7
model with increased expansion velocities has also been in-
vestigated, but it can explain only part of the large UV flux in
In our study, the comparison object SN 2006X may also
exhibit strong emission in the UV. Some common features
for SN 2004dt and SN 2006X are the distinctly high-velocity
features beyond the photosphere, slower B-band light-curve
evolution in the early nebular phase (Wang et al. 2008a), and
(Patat et al. 2009). The above observational features of these
two objects are perhaps related to an asymmetric explosion of
thehigh-velocitySNeIa(Wangetal. 2011,inprep.). Line-of-
sight effects due to the asymmetric explosioncan have a more
significant effect on the observed scatter in the UV than in
other bands. A new proposal to identify the progenitorsbased
on the symmetrypropertiesof the explosionhas recentlybeen
proposed by Livio & Pringle (2011). However, more detailed
studies are clearly needed to investigate this fully for realistic
The question of whether a UV excess is a more gen-
eral property of the high-velocity subclass merits further
study with a larger sample having earlier observations like
PTF11kly/SN 2011fe (Brown et al. 2011). We note, however,
that the origin of the high-velocity features and large polar-
ization observed in SN 2004dt and SN 2006X probably dif-
fers, since the former does not follow the relation between the
early-phasevelocitygradientand the nebular-phaseemission-
line velocity shift while the latter does (Maeda et al. 2010).
SN 2004dt is also found to be an outlier in the relation be-
tween the line polarization of Si II λ6355 and the nebular ve-
2008; Foley et al.
locity offset (Maund et al. 2010). Moreover, the reddening
of SN 2004dt by its host galaxy is low, while SN 2006X is
heavily extinguished, perhaps by circumstellar dust (Patat et
al. 2007; Wang et al. 2008a). Thus, the physical origin of the
UV excess in these two SNe Ia might not be the same. In SN
2006X,dust scattering may contribute in part to the unusually
brightbehaviorinthe UV, whichis favoredbythedetectionof
surrounding circumstellar material (Patat et al. 2007) and/or
a light echo around this object (Wang et al. 2008b).
We furtheremphasize here the significance of the discovery
of a UV excess in SN 2004dt. It provides not only a new clue
to the study of SN Ia physics and/or the progenitor environ-
ments, but also draws attention to another possible systematic
error that might exist in current cosmological studies: the rel-
atively higher UV flux would result in a bluer U −B color
for some SNe Ia, which could lead to an underestimate of the
host-galaxy reddening and hence an overestimate of the dis-
tance. It is of interest to determine the fractional population
of the SN 2004dt-likeevents or those showing a UV excess in
of such a peculiar subclass of SNe Ia on current cosmological
We thank Mark Sullivan and Andy Howell for their sugges-
tive comments. Financial support for this work has been pro-
vided by the National Science Foundation of China (NSFC
grants 11178003, 11073013, and 10173003) and the Na-
tional Key Basic Research Science Foundation (NKBRSF
TG199075402). A.V.F.’s group at U.C. Berkeley is grate-
ful for NSF grants AST-0607485 and AST-0908886, the
TABASGO Foundation,and US Departmentof Energygrants
DE-FC02-06ER41453 (SciDAC) and DE-FG02-08ER41563.
Substantial financial support for this work was also pro-
vided by NASA through grant GO-10877 from the Space
Telescope Science Institute, which is operated by Associated
Universities for Research in Astronomy, Inc., under NASA
contract NAS 5-26555.The work of L.W. is supported
by NSF grant AST-0708873. J.C.W. is supported by NSF
grant AST-0707769. K.N. is supported by WPI Initiative,
MEXT, Japan. M.T., S.B., and E.C. are supported by grant
ASI-INAF I/009/10/0. P.A.M. is supported by NASA ADP
Altavilla, G., et al. 2007, A&A, 475, 585
Arnett, W. D., 1982, ApJ, 253, 785
Benetti, S., et al. 2004, MNRAS, 348, 261
Benetti, S., et al. 2005, ApJ, 623, 1011
Bongard, S., et al. 2008, ApJ, 2008, 684
Branch, D., Dang, L., & Baron, E. 2009, PASP, 121, 238
Branch, D., Fisher, A., & Nugent, P. 1993, AJ, 106, 2383
Branch D., & Venkatakrishna, K. L. 1986, ApJ, 306, L21
Branch, D., et al. 1985, ApJ, 294, 619
Branch, D., et al. 2006, PASP, 118, 560
Brown, P. J., et al. 2009, AJ, 137, 4517
Brown, P. J., et al. 2010, ApJ, 721, 1608
Brown, P. J., et al. 2011 (arXiv:1110.2538)
Bufano, F., et al. 2009, ApJ, 700, 1456
Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245
Contreras, C., et al. 2010, AJ, 139, 519
Cooke, J., et al. 2011, ApJ, 727, L35
Ellis, R. S., et al. 2008, ApJ, 674, 51
Filippenko, A. V., et al. 1992a, ApJ, 384, L15
Filippenko, A. V., et al. 1992b, AJ, 104, 1543
Foley, R. J., Filippenko, A. V., & Jha, S. W. 2008, ApJ, 686, 117
Foley, R. J., & Kasen, D. 2011, ApJ, 729, 55
Foley, R. J., et al., 2010, ApJ, submitted (arXiv: 1010.2749)
Ganeshalingam, M., et al. 2010, ApJS, 190, 418
Garavini, G., et al. 2007, A&A, 471, 527
Gerardy, C., et al. 2004, ApJ, 607, 391
Goobar, A. 2008, ApJ, 686, L103
Hicken, M., et al. 2009, ApJ, 700, 331
Höflich, P., Wheeler, J. C., & Thielemann, F.-K. 1998, ApJ, 495, 617
Höflich, P., et al. 2010, ApJ, 710, 444
Howell, A., et al. 2006, Nature, 443, 308
Howell, A., et al. 2009, ApJ, 691, 661
Hsiao, E. Y., et al. 2007, ApJ, 663, 1187
Immler, S., et al. 2006, ApJ, 648, L119
Jack, D., Hauschildt, P. H., & Baron, E. 2009, A&A, 502, 1043
Jack, D., Hauschildt, P. H., & Baron, E. 2011, A&A, 528, 141
Kasen, D. 2006, ApJ, 649, 939
Kasen, D., Röpke, F. K., Woosley, S. E. 2009, Nature, 460, 7257
Kirshner, R. P., et al. 1993, ApJ, 415, 589
Kromer, M., & Sim, S. A. 2009, MNRAS, 398, 1809
Kümmel, M., Walsh, J., Larsen, S. & Hook, R. 2004, ST-ECF Newsletter 36,
Kümmel, M., et al. 2009, PASP, 121, 59
Lentz, E. J., Baron, E., Branch, D., Hauschildt, P. H., & Nugent, P. E. 2000,
ApJ, 530, 966
HST UV Observations of SNe Ia13
Leonard, D. C., Li, W., Filippenko, A. V., Foley, R. J., & Chornock, R. 2005,
ApJ, 632, 450
Li, W., et al. 2003, PASP, 115, 453
Livio, M. & Pringle, J. E. 2011, ApJ, 740, L18
Maeda, K., et al. 2010, Nature, 466, 82
Maeda, K., et al. 2011, MNRAS, 413, 3075
Maund, J. R., et al. 2010, ApJ, 725, L167
Mazzali, P. A., et al. 2000, A&A, 363, 705
Mazzali, P. A., et al. 2005, ApJ, 623, L37
Milne, P. A., et al. 2010, ApJ, 721, 1627
Oke, J. B., & Sandage, A. 1968, ApJ, 154, 21
Panagia, N., et al. 2003, in Supernovae and Gamma-Ray Bursters, ed. K.
Weiler (Berlin: Springer), 113
Panagia, N., 2007, in Supernova 1987A: 20 Years After (New York: AIP,
Volume 937), 236
Patat, F., et al. 2007, Science, 317, 924
Patat, F., et al. 2009, A&A, 508, 229
Pauldrach, A. W. A., et al. 1996, A&A, 312, 525
Perlmutter, S., et al. 1997, ApJ, 483, 565
Phillips, M. M. 1993, ApJ, 413, L105
Phillips, M. M., et al. 1992, AJ, 103, 1632
Phillips, M. M., et al. 1999, AJ, 118, 1766
Riess, A. G., et al. 2009, ApJ, 699, 539
Sauer, D., et al. 2008, MNRAS, 391, 1605
Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525
Sim, S. A., et al. 2010, ApJ, 714, L52
Sirianni, M., et al. 2005, PASP, 117, 1049
Stanishev, V., et al. 2007, A&A, 469, 645
Sullivan, M., et al. 2009, ApJ, 693, L76
Silverman, J. M., et al. 2011, MNRAS, 410, 585
Taubenberger, S., et al. 2011, MNRAS, 412, 2735
Thomas, R. C. 2005, IAUC, 8474
Timmes, F. X., Brown, E. F., & Truran, J. W. 2003, ApJ, 590, L83
Wang, L., et al. 2005, ApJ, 635, L33
Wang, L., et al. 2006, ApJ, 653, 490
Wang, L, Baade, D., & Patat, F. 2007, Science, 315, 212
Wang, X., et al. 2008a, ApJ, 675, 626
Wang, X., et al. 2008b, ApJ, 677, 1060
Wang, X., et al. 2009a, ApJ, 699, L139 (W09a)
Wang, X., et al. 2009b, ApJ, 697, 380 (W09b)
Zhang, T., et al. 2010, PASP, 122, 1
14Wang et al.
JOURNAL OF HST ACS PR200L SPECTROSCOPIC
OBSERVATIONS OF SNE IA
SN UT DateJDa
2004 Aug. 20
2004 Aug. 23
2004 Aug. 27
2004 Aug. 31
2004 Sept. 01
2004 Sept. 06
2004 Sept. 07
2004 Sept. 12
2004 Sept. 16
2004 Sept. 21
2004 Sept. 27
2004 Oct. 02
2004 Oct. 10
2004 Sept. 14
2004 Sept. 18
2004 Sept. 22
2004 Sept. 25
2004 Sept. 29
2004 Oct. 02
2004 Oct. 08
2004 Oct. 14
2005 Jan. 31
2005 Jun. 03
2005 Jun. 05
2005 Jun. 07
2005 Jun. 11
2005 Jun. 14
2005 Jun. 16
2005 Jun. 21
2005 Jun. 25
2005 Jun. 26
2005 Jun. 29
2005 Jun. 30
2005 Jul. 05
aJulian Date minus 245,0000.
bDays relative to the epoch of B-band maximum.
HST UV Observations of SNe Ia 15
HST ACS ULTRAVIOLET PHOTOMETRY OF TYPE IA SUPERNOVAE
F220W F250W F330W
2004 Aug. 20
2004 Aug. 23
2004 Aug. 27
2004 Aug. 31
2004 Sept. 01
2004 Sept. 06
2004 Sept. 07
2004 Sept. 12
2004 Sept. 16
2004 Sept. 21
2004 Sept. 27
2004 Oct. 02
2004 Oct. 10
2004 Sept. 13
2004 Sept. 18
2004 Sept. 21
2004 Sept. 24
2004 Sept. 28
2004 Oct. 02
2004 Oct. 08
2004 Oct. 14
2005 Jan. 28
2005 Jan. 31
2005 Feb. 04
2005 Feb. 09
2005 Feb. 13
2005 Feb. 17
aJulian Date minus 245,0000.
bRelative to the epoch of B-band maximum.
Note: Uncertainties, in units of 0.001 mag, are 1σ.
RELEVANT PARAMETERS FOR THE HST UV SAMPLE OF SNE IA.
v3 K,v220(km s−1)
Uncertainty estimates in parentheses are in units of 0.01 mag.
1. Ganeshalingam et al. (2010); 2. Altavilla et al. (2007); 3. Contreras et al. (2010); 4. Silverman et al. (2011, in prep.); 5. Hicken et al. (2009); 6. Wang et
al. (2009b); 7. Garavini et al. (2007).
LINEAR FIT TO THE Mmax−∆m15(B) RELATION
Mmax= M0+α (∆m15(B)−1.1)
Uncertainty estimates in parentheses are in units of 0.01 mag.