On the Source of the Dust Extinction in Type Ia Supernovae and the Discovery of Anomalously Strong Na I Absorption

TL;DR: In this paper, high-dispersion observations of the Na I D 5890, 5896 and K I 7665, 7699 interstellar lines are used as an independent means of probing dust extinction.

Abstract: High-dispersion observations of the Na I D 5890, 5896 and K I 7665, 7699 interstellar lines, and the diffuse interstellar band at 5780 Angstroms in the spectra of 32 Type Ia supernovae are used as an independent means of probing dust extinction. We show that the dust extinction of the objects where the diffuse interstellar band at 5780 Angstroms is detected is consistent with the visual extinction derived from the supernova colors. This strongly suggests that the dust producing the extinction is predominantly located in the interstellar medium of the host galaxies and not in circumstellar material associated with the progenitor system. One quarter of the supernovae display anomalously large Na I column densities in comparison to the amount of dust extinction derived from their colors. Remarkably, all of the cases of unusually strong Na I D absorption correspond to "Blueshifted" profiles in the classification scheme of Sternberg et al. (2011). This coincidence suggests that outflowing circumstellar gas is responsible for at least some of the cases of anomalously large Na I column densities. Two supernovae with unusually strong Na I D absorption showed essentially normal K I column densities for the dust extinction implied by their colors, but this does not appear to be a universal characteristic. Overall, we find the most accurate predictor of individual supernova extinction to be the equivalent width of the diffuse interstellar band at 5780 Angstroms, and provide an empirical relation for its use. Finally, we identify ways of producing significant enhancements of the Na abundance of circumstellar material in both the single-degenerate and double-degenerate scenarios for the progenitor system.

Summary

1. Introduction

• Competition is generally accepted as a positive force in most industries; it is supposed to have a positive impact on an industry’s efficiency, quality of provision, innovation and international competitiveness.
• The issue of competition in banking has always been controversial, as the perceived benefits from increased competition have to be weighted against the risks of potential instability.
• In the European Union, the aim of regulatory developments, which include movements towards the creation of a single market for financial services, was to foster competition in order to improve the productivity, efficiency and profitability of the banking systems and also to increase both national and international competitiveness.
• The pace of domestic consolidation has recently slowed down, whereas the value of cross-border bank M&A has been rising, reaching record levels in 2005.
• The remainder of the paper is structured as follows.

2. Competition and Efficiency in European Banking

• Over the past twenty years, the deregulation and market integration processes, coupled with advances in information technologies, have been a steady feature of EU banking markets and have given way to a profound transformation and restructuring of the banking industry, which materialised in enhanced consolidation and a move away from the traditional intermediation business into more profitable investment services.
• Traditional industrial organisation theory focuses of the Structure-Conduct-Performance (SCP) paradigm.
• If abnormal profits tend to persists from year to year, then there might be barriers to entry or banks might be exploiting monopoly power.
• Here the correlation between profitability and efficiency should be either zero or negative because less efficient banks with more market power will be able to gain extra profits.
• If the authors refer to the extant efficiency literature, banks typically encounter relatively large scale diseconomies once they exceed a certain ‘optimal’ size (Amel et al., 2004).

3.2 Frontier Efficiency Analysis

• The literature on the measurement of efficiency frontiers can be divided in two main streams: parametric techniques, such as the Stochastic Frontier Analysis (SFA) and non-parametric techniques such as Data Envelopment Analysis (DEA).
• The chosen functional form for the cost function is the translog as specified in equation (4) above with the same three inputs, but with two outputs (total loans and total securities) and a time trend.
• On the other hand, DEA is a mathematical linear programming technique developed by Charnes, Cooper and Rhodes in 1978 (CCR) which identifies the efficient frontier from the linear combination of those units/observations that (in a production space) use comparatively less inputs to produce comparatively more outputs.
• The CCR model assumes constant returns to scale (CRS), which is the optimal scale in the long- run.
• The additional convexity constraint ∑ = 1iλ can be included in (8) to allow for variable returns to scale (VRS) (see Banker, Charnes and Cooper (1984) or BCC model.

3.3 Dynamic Panel Data Granger-Type Causality estimation

• Granger testing is a common method of investigating causal relationships (Granger, 1969) by estimating an equation in which y is regressed on lagged values of y and the lagged values of an additional variable x.
• Since the authors expect causality to run in either direction, ity and itx are represented alternatively by a measure of competition (the Lerner Index of monopoly power) and a measure of bank cost efficiency (estimated using parametric and non-parametric methods).
• Moreover, if the sample is unbalanced (as in their case) by increasing the number of lags, the number of observations will be reduced significantly and this may affect the consistency of the results.
• Moreover the authors use an incremental Sargan/Hensen test for the validity of the additional moment restrictions described in (12) required by the SYS-GMM as follows: if S is the Sargan statistics obtained under stronger assumptions and S ′ is the Sargan statistics obtained under weaker assumptions, then the difference SS ′− , is asymptotically distributed as 2χ .

3.4 Data

• The data on EU commercial banks are derived from BankScope, a global database published by Bureau VanDjick.
• The data are collected for an unbalanced sample of 2,701 commercial bank observations operating in France, Germany, Italy, Spain and the United Kingdom between 2000 and 2005.
• The authors restricted the analysis to commercial banks as there are still significant differences in the retail market structure among countries and in some countries the saving banking sector is still partially benefiting from state help9.
• For more details on this test see Bond and Windmeijer (2005).
• The number of banks in the sample is decreasing over time (with the exception of Italy) as the banking sector consolidates further.

4.1 Competition Patterns in European Banking

• Table 2 shows the means of the structural indicators of market concentration across their sample of EU countries over the period 2000-2005.
• In the UK alone, in the six years period from 2000 concentration (measured as the market share of the five largest banks) increased by 28.57%.
• Looking at the separate information for commercial banks, they seem to operate in more concentrated markets and this might be also reflected in their measure for market power.
• Italy and Spain, which display the highest average marginal costs, also display the biggest decrease, possibly because of the reduction of both financial costs and operating costs.
• Comparing their results with averages for the whole banking system (see Fernandez de Guevara and Maudos, 2005), they confirm that, despite being more concentrated, commercial banks enjoy a lower market power compared to saving banks.

4.2 Competition Patterns in European Banking: H-statistic

• Following the empirical literature on competition in banking markets, the authors estimated the reduced form revenue equation specified in (2) using a panel data framework.
• The regression models are estimated using the fixed effect estimators.
• Estimations are carried out at each individual country level.
• The estimated H-statistic indicates monopolistic competition in all countries and ranges from 0.3715 in France to 0.7783 in Germany.
• The impact of the cost of capital seems to be minimal compared to the other input prices (with the exception of UK and Germany).

4. The evolution of bank efficiency: SFA and DEA analysis

• The yearly SFA and DEA results for the countries in their sample, as well as the average efficiency over the period are shown in Table 4.
• Therefore the hypothesis of 1<H<1 (monopolistic competition) holds in countries.
• The equilibrium test can be performed by recalculating the Panzar and Rosse’s H-statistics replacing the dependent variable total revenue over assets with the natural log of return on assets (which is equal to net income over total assets), as shown in equation (2).
• Both methodologies indicate and average inefficiency scores of about 30%, a result that is broadly in line with the main literature on bank efficiency (see Goddard et al., 2007).
• The analysis so far has highlighted that the main EU banking markets are becoming progressively more concentrated and less efficient.

4.2 The Relationship and Causality between Competition and Efficiency

• Tables 5 and 6 report the results of their empirical analysis on the relationship and causality between competition and efficiency.
• This indicates that competition at time t is influenced by previous years’ competition.
• Granger causality is assessed as the joint test of the two lags of efficiency on competition as follows: 021 == ββ .
• The significance of the coefficients for the first and second lags of efficiency seem to suggest that efficiency is affected significantly by previous years’ efficiency (and in some cases the inefficiency as noted by the negative and significant sign of the second lag in panel a).
• This is consistent with the hypothesis that more efficient banks are in a position to exploit market power and therefore there seem to be a trade-off between efficiency and competition.

5. Conclusions

• Competition is generally considered as a positive force, often associated with increased efficiency and enhanced consumers’ welfare.
• The acceleration in the recent consolidation process, however, is raising concerns about increased concentration in the banking sector and its potential implications for public policy deriving from increased market power in the banking sector.
• Furthermore the authors use a dynamic panel data Granger causality test to investigate the dynamics of the relationship between competition and efficiency.
• The authors findings suggest a negative causation between efficiency and competition, whereas the causality running from competition to efficiency, although positive, is relatively weak.
• These results pose further questions for competition policies.

Figures

Table 2 SNe Ia Host Galaxy Information and Light-curve Measurements

Figure 5. (Left) Equivalent width of the DIB feature at 5780 Å in the ISM of the Milky Way plotted vs. AV and the column density of Na i. The open circles in the upper plot correspond to the measurements of Friedman et al. (2011); those in the lower plot are a combination of the EW 5780 Å values of Friedman et al. (2011) with Na i column densities from Welsh et al. (2010). The solid circles in both plots correspond to the objects in our Milky Way sample for which a Galactic component of the 5780 Å feature was detected. The AV values for the latter objects are taken from Schlegel et al. (1998) as rescaled by Schlafly & Finkbeiner (2011). The solid lines are fits to these combined data, with the gray shading corresponding to the 1σ dispersions (see text). (Right) Equivalent width of the host DIB absorption at 5780 Å plotted vs. AV values derived from the SN colors (upper) and the column density of Na i (lower). The fits and 1σ dispersions observed in the Milky Way are reproduced from the left half of the figure. The different symbols used to plot the SNe Ia correspond to the Na i D profile classification scheme of Sternberg et al. (2011).

Figure 7. Na i D profiles of two of the objects in the Milky Way sample (left), and two SNe Ia in the host absorption sample with similar dust extinctions (right). Zero velocity has been arbitrarily set to correspond to the minimum of the D2 line absorption. The anomalous strength of the host Na i D absorption in SNe 2009ig and 2008fp is readily apparent.

Figure 10. Na i column densities are plotted vs. Na i D equivalent widths for both the Milky Way and host absorption samples. Theoretical curve of growth relations for three different values of the Doppler parameter, b, are shown for reference. For each Doppler parameter, we show predictions for a single absorption component as well as for up to 16 components with the same value of b (as indicated in the legend.) In all cases, the single component predictions are the left-most curves with increasing number of components lying progressively to the right.

Figure 9. Na i D equivalent widths of both the Milky Way and host absorption samples are plotted vs. the dust extinction, AV . The Milky Way sample has been augmented by high dispersion measurements of 82 stars (Sembach et al. 1993; Munari & Zwitter 1997) and 30 QSOs (Poznanski et al. 2012). The dashed line corresponds to the Galactic relation derived by Poznanski et al. (2012), with the 1σ dispersion found by these authors indicated by the gray shading. The solid purple line shows the relation for the D1 line found by Munari & Zwitter (1997), multiplied by a factor of 2.53 to match the total equivalent widths plotted here. (A color version of this figure is available in the online journal.)

Table 3 SNe Ia Host Absorption-line Measurements

Figure 4. Host Na i D1 and K i λ7665 absorption in SNe 2006cm and 2008fp. The observations correspond to the black line. The best-fitting VPFIT model is shown as the red line, with the individual Voigt profile components plotted as dashed lines. Zero velocity has been arbitrarily set to correspond to the strongest component of the K i λ7665 profile. The velocities of the 5780 Å DIB in each SN are shown as ±1σ error bars, and were derived assuming a Gaussian profile of 2.1 Å FWHM and a rest wavelength of 5780.55 Å (Tuairisg et al. 2000). The FWHM velocity range covered by the Hα and [N ii] emission in the extracted spectrum of SN 2006cm is also indicated.

Figure 8. (Left) Column densities of Na i and K i in the ISM of the Milky Way, LMC, and SMC plotted vs. AV . Open circles are the Na i D measurements of Sembach et al. (1993) and the K i data of Welty & Hobbs (2001) shown in Figure 3. The fits and 1σ dispersions observed in the Milky Way are also reproduced from Figure 3. The triangles (magenta) are measurements for the LMC (Cox et al. 2006; Welty et al. 2006) and the inverted triangles (cyan) correspond to the SMC (Welty et al. 2006). (Right) Equivalent width of the DIB feature at 5780 Å in the ISM of the Milky Way, LMC, and SMC plotted vs. AV and the column density of Na i. The open circles in the upper plot are the measurements of Friedman et al. (2011) reproduced from Figure 5. Likewise, the EW 5780 Å values of Friedman et al. (2011) are combined with the Na i column densities from Welsh et al. (2010) as in Figure 5. The fits and 1σ dispersions observed in the Milky Way are reproduced from Figure 5. Measurements for the LMC (Cox et al. 2006; Welty et al. 2006) and SMC (Welty et al. 2006) are plotted with the same symbols and colors as in the left-hand side of the figure.

Figure 1. VPFIT models for the Na i D and K i absorption lines for those SNe Ia where both features were detected. The Na i D lines are shown in the left side of the figure. In the right half the K i absorption is displayed. For all but SN 2009le, the stronger component of the K i doublet, λ7665, is plotted. In the case of SN 2009le, the λ7665 line is blended with strong telluric absorption, and so for this object the λ7699 transition is shown. The red lines in each spectrum indicate the velocities of the components used to fit the profiles. The most prominent (deepest) component is used to set the zero point of the velocity scale.

Figure 2. VPFIT models for the Na i D absorption lines for SNe Ia where the K i lines were not observed or detected. The red lines above the D2 line in each spectrum indicate the velocities of the components used to fit the profiles. The most prominent (deepest) component is used to set the zero point of the velocity scale.

Figure 12. Host Na i D1 and K i λ7665 absorption in SN 2009le. The observations correspond to the black line and the best-fitting profile VPFIT model is shown in red. Zero velocity has been arbitrarily set to correspond to the single component observed for of the K i λ7665 line. The FWHM velocity range covered by Hα and [N ii] emission in the extracted spectrum of the SN is indicated. The systematic velocity of the host galaxy is also shown.

Table 4 Goodness of Fit, χ2ν , to the Galactic logNNa i versus AV Relation

Figure 6. Examples of three detections of host DIB absorption at 5780 Å (SNe 2006cm, 2008fp, and 2007sr). Spectral observations are plotted in black, and fits assuming a Gaussian profile of 2.1 Å FWHM are in red. Also shown is the spectrum of SN 2009ig where the 5780 Å is not clearly detected, but where the signal-to-noise ratio of the spectrum allowed a meaningful upper limit to be measured. The fit in this case corresponds to the derived 3σ upper limit on the equivalent width. The scaling of the intensity axis is the same for each SN in order to preserve the relative differences in strength of the 5780 Å feature; the spectra have been offset for illustration purposes.

Figure 3. (Left) Column densities of Na i and K i in the ISM of the Milky Way plotted vs. AV . Open circles are the Na i D measurements of Sembach et al. (1993) and the K i data of Welty & Hobbs (2001); solid circles in both plots show column densities for our Milky Way sample with the AV values taken from Schlegel et al. (1998) as rescaled by Schlafly & Finkbeiner (2011). The solid lines are fits to these combined data (see text), with the gray shading corresponding to the 1σ dispersion. The shaded red area illustrates the uncertainty introduced by the 1σ dispersion in RV values observed in the Milky Way (Fitzpatrick & Massa 2007). (Right) Column densities of Na i and K i for the host absorption sample are plotted vs. the AV values derived from the SN colors. The fits and 1σ dispersions observed in the Milky Way are reproduced from the left half of the figure. The different symbols used to plot the SNe Ia correspond to the Na i D profile classification scheme of Sternberg et al. (2011).

Figure 13. Column densities of Na i and K i plotted against each other. (Left) Measurements for the Milky Way (Welty & Hobbs 2001; Kemp et al. 2002), LMC (Cox et al. 2006; Welty et al. 2006), and SMC (Welty et al. 2006) are compared. The diagonal line shows the average value of log (NNa i/NK i) = 1.9 and rms dispersion of 0.3 dex obtained for the Milky Way measurements. The thick black line is a representation of the logNK i vs. logNNa i trend shown in Figure 2 of Kemp et al. (2002). (Right) Plot of column densities of Na i and K i for our Milky Way and SNe Ia host absorption samples. The fit and rms dispersion of the Milky Way stellar sample from the left half of the figure is duplicated for comparison. Also repeated is the logNK i vs. logNNa i trend from Kemp et al. (2002).

Table 1 Milky Way Na i and K i Column Density Measurements

Figure 11. (Left) Total host absorption column densities of Na i plotted vs. the AV values derived from the SN Ia light curves. The objects are separated into the high-velocity (above) and normal-velocity (below) subclasses of Wang et al. (2013), where high velocity is defined as v(Si ii λ6355) 12,000 km s−1 at B maximum. The fits and 1σ dispersions observed in the Milky Way are reproduced from Figure 3. The different symbols used to plot the SNe Ia correspond to the Na i D profile classification scheme of Sternberg et al. (2011). (Right) Similar plots, but in this case examining subcomponents of the Na i D absorption profiles for the “Blueshifted” sample only. In the top diagram, the column density of the most prominent Na i D profile is plotted, while in the lower diagram the sum of the column densities of the components blueward of the most prominent component are shown.

Full text

Phillips, MM, Simon, JD, Morrell, N, Burns, CR, Cox, NLJ, Foley, RJ, Karakas,
AI, Patat, F, Sternberg, A, Williams, RE, Gal-Yam, A, Hsiao, EY, Leonard, DC,
Persson, SE, Stritzinger, M, Thompson, IB, Campillay, A, Contreras, C,
Folatelli, G, Freedman, WL, Hamuy, M, Roth, M, Shields, GA, Suntzeff, NB,
Chomiuk, L, Ivans, II, Madore, BF, Penprase, BE, Perley, DA, Pignata, G,
Preston, G and Soderberg, AM
On the source of the dust extinction in type Ia supernovae and the discovery
of anomalously strong Na i absorption
http://researchonline.ljmu.ac.uk/6458/
Article
LJMU has developed LJMU Research Online for users to access the research output of the
University more effectively. Copyright © and Moral Rights for the papers on this site are retained by
the individual authors and/or other copyright owners. Users may download and/or print one copy of
any article(s) in LJMU Research Online to facilitate their private study or for non-commercial research.
You may not engage in further distribution of the material or use it for any profit-making activities or
any commercial gain.
The version presented here may differ from the published version or from the version of the record.
Please see the repository URL above for details on accessing the published version and note that
access may require a subscription.
http://researchonline.ljmu.ac.uk/
Citation (please note it is advisable to refer to the publisher’s version if you
intend to cite from this work)
Phillips, MM, Simon, JD, Morrell, N, Burns, CR, Cox, NLJ, Foley, RJ,
Karakas, AI, Patat, F, Sternberg, A, Williams, RE, Gal-Yam, A, Hsiao, EY,
Leonard, DC, Persson, SE, Stritzinger, M, Thompson, IB, Campillay, A,
Contreras, C, Folatelli, G, Freedman, WL, Hamuy, M, Roth, M, Shields, GA,
LJMU Research Online

For more information please contact researchonline@ljmu.ac.uk
http://researchonline.ljmu.ac.uk/

The Astrophysical Journal, 779:38 (21pp), 2013 December 10 doi:10.1088/0004-637X/779/1/38
C

2013. The American Astronomical Society. All rights reserved. Printed in the U.S.A.
ON THE SOURCE OF THE DUST EXTINCTION IN TYPE Ia SUPERNOVAE AND
THE DISCOVERY OF ANOMALOUSLY STRONG Na i ABSORPTION
∗
M. M. Phillips
1
, Joshua D. Simon
2
, Nidia Morrell
1
, Christopher R. Burns
2
,NickL.J.Cox
3
, Ryan J. Foley
4
,
Amanda I. Karakas
5
, F. Patat
6
, A. Sternberg
7,21
, R. E. Williams
8
, A. Gal-Yam
9
,E.Y.Hsiao
1
, D. C. Leonard
10
,
Sven E. Persson
2
, Maximilian Stritzinger
11
, I. B. Thompson
2
, Abdo Campillay
1
, Carlos Contreras
1
,
Gast
´
on Folatelli
12
, Wendy L. Freedman
2
, Mario Hamuy
13
, Miguel Roth
1
, Gregory A. Shields
14
,
Nicholas B. Suntzeff
15
, Laura Chomiuk
4
, Inese I. Ivans
16
, Barry F. Madore
2,17
, B. E. Penprase
18
,
Daniel Perley
19
, G. Pignata
20
, G. Preston
2
, and Alicia M. Soderberg
4
1
Carnegie Observatories, Las Campanas Observatory, Casilla 601, La Serena, Chile;
mmp@lco.cl
2
Observatories of the Carnegie Institution for Science, 813 Santa Barbara St., Pasadena, CA 91101, USA
3
Instituut voor Sterrenkunde, KU Leuven, Celestijnenlaan 200D bus 2401, 3001 Leuven, Belgium
4
Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA
5
Research School of Astronomy and Astrophysics, The Australian National University, Weston, ACT 2611, Australia
6
European Southern Observatory (ESO), Karl Schwarschild Strasse 2, D-85748, Garching bei M
¨
unchen, Germany
7
Max Planck Institute for Astrophysics, Karl Schwarzschild Strasse 1, D-85741 Garching bei M
¨
unchen, Germany
8
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, USA
9
Benoziyo Center for Astrophysics, Faculty of Physics, Weizmann Institute of Science, Rehovot 76100, Israel
10
Department of Astronomy, San Diego State University, San Diego, CA 92182, USA
11
Department of Physics and Astronomy, Aarhus University, Ny Munkegade 120, DK-8000 Aarhus C, Denmark
12
Kavli Institute for the Physics and Mathematics of the Universe, Todai Institutes for Advanced Study,
the University of Tokyo, Kashiwa 277-8583, Japan
13
Universidad de Chile, Departamento de Astronom
´
ıa, Casilla 36-D, Santiago, Chile
14
Department of Astronomy, University of Texas, Austin, TX 78712, USA
15
George P. and Cynthia Woods Mitchell Institute for Fundamental Physics and Astronomy, Texas A&M University,
Department of Physics and Astronomy, College Station, TX 77843, USA
16
Department of Physics and Astronomy, University of Utah, Salt Lake City, UT 84112, USA
17
Infrared Processing and Analysis Center, Caltech/Jet Propulsion Laboratory, Pasadena, CA 91125, USA
18
Department of Physics and Astronomy, Pomona College, 610 N. College Ave., Claremont, CA 91711, USA
19
Cahill Center for Astrophysics, California Institute of Technology, Pasadena, CA 91125, USA
20
Departamento de Ciencias Fisicas, Universidad Andres Bello, Avda. Republica 252, Santiago, Chile
Received 2013 July 11; accepted 2013 November 1; published 2013 November 22
ABSTRACT
High-dispersion observations of the Na i D λλ5890, 5896 and K i λλ7665, 7699 interstellar lines, and the diffuse
interstellar band at 5780 Å in the spectra of 32 Type Ia supernovae are used as an independent means of probing
dust extinction. We show that the dust extinction of the objects where the diffuse interstellar band at 5780 Å is
detected is consistent with the visual extinction derived from the supernova colors. This strongly suggests that the
dust producing the extinction is predominantly located in the interstellar medium of the host galaxies and not in
circumstellar material associated with the progenitor system. One quarter of the supernovae display anomalously
large Na i column densities in comparison to the amount of dust extinction derived from their colors. Remarkably, all
of the cases of unusually strong Na i D absorption correspond to “Blueshifted” profiles in the classification scheme
of Sternberg et al. This coincidence suggests that outflowing circumstellar gas is responsible for at least some of the
cases of anomalously large Na i column densities. Two supernovae with unusually strong Na i D absorption showed
essentially normal K i column densities for the dust extinction implied by their colors, but this does not appear to
be a universal characteristic. Overall, we find the most accurate predictor of individual supernova extinction to be
the equivalent width of the diffuse interstellar band at 5780 Å, and provide an empirical relation for its use. Finally,
we identify ways of producing significant enhancements of the Na abundance of circumstellar material in both the
single-degenerate and double-degenerate scenarios for the progenitor system.
Key words: circumstellar matter – dust, extinction – galaxies: ISM – supernovae: general
Online-only material: color figures
1. INTRODUCTION
Type Ia supernovae (SNe Ia) are one of the most effective
observational tools for measuring the expansion history of
the universe. Their successful use in cosmology is due to the
discovery of empirical relations that dramatically decrease the
dispersion in peak luminosities at optical wavelengths. The first
∗
This paper includes data gathered with the 6.5 m Magellan telescopes at Las
Campanas Observatory, Chile.
21
Minerva Fellow.
of these is the well-known correlation with light curve shape:
intrinsically brighter SNe Ia have broader light curves that
decline more slowly from maximum than do the light curves
of less luminous SNe Ia (Phillips
1993). The second is a strong
dependence of peak luminosity on color that is in the same
sense as dust reddening, but with an average value of the ratio
of total-to-selective extinction, R
V
, that is significantly less than
would be produced by normal interstellar dust in the Milky Way
(Tripp
1998). The latter result has variously been interpreted
as possible evidence that the extinction arises in circumstellar
1

The Astrophysical Journal, 779:38 (21pp), 2013 December 10 Phillips et al.
dust (Wang
2005; Goobar 2008), as the consequence of intrinsic
differences in color between SNe Ia with “normal” and “high”
Si ii expansion velocities (Foley & Kasen
2011), or as a bias due
to a misidentification of the dispersion in the luminosity/color-
corrected Hubble diagram with an intrinsic scatter in luminosity
rather than color (Scolnic et al.
2013).
Our understanding of the progenitors and explosion mech-
anism(s) that produce SNe Ia is still quite limited. Although
there is widespread agreement that these objects correspond to
the thermonuclear disruption of a white dwarf in a binary sys-
tem, it is not yet clear if the companion to the white dwarf
is a main sequence or giant star (“single-degenerate” or “SD”
model) or another white dwarf (“double-degenerate” or “DD”
model). In recent years, observational evidence favoring both
scenarios has been put forward (e.g., see Howell
2011; Maoz &
Mannucci
2012; Patat 2013). In both the SD and DD scenarios,
material ejected from the system prior to the explosion may re-
main as circumstellar material (CSM; Moore & Bildsten
2012;
Raskin & Kasen
2013; Shen et al. 2013). Sternberg et al. (2011)
found a strong statistical preference for blueshifted structures
in the narrow Na i D absorption observed in the line-of-sight to
many SNe Ia, suggestive of gas outflows from the progenitor
systems. In a few such cases, temporal variations of blueshifted
components of the Na i D lines apparently due to changing ion-
ization conditions in the CSM have also been observed (e.g.,
Patatetal.
2007; Simon et al. 2009; Dilday et al. 2012). On
the other hand, radio and X-ray observations of the prototypical
Type Ia SN 2011fe place tight upper limits on the amount of
CSM in the progenitor system before explosion (e.g., Horesh
et al.
2012), and early-time photometry of this event apparently
rules out either a red giant or main sequence companion (Bloom
et al.
2012).
In the Milky Way, the strengths of certain interstellar absorp-
tion features such as the Na i D lines and the diffuse interstellar
bands (DIBs) have been known for many years to correlate with
dust extinction (e.g., Merrill & Wilson
1938; Hobbs 1974). In
this paper, we employ high-dispersion spectroscopy to use these
features as an independent probe of the dust affecting the col-
ors of SNe Ia. As we will show, the data indicate that the dust
extinction for the objects where DIBs are observed is gener-
ally consistent with the extinction derived from the SN colors,
and therefore most likely arises in the interstellar medium of
the host galaxy. However, one-fourth of the SNe Ia, all with
blueshifted structures as per Sternberg et al. (
2011), display
anomalously large Na i column densities that, in the interstellar
medium (ISM) of the Milky Way, would correspond to an order
of magnitude or more greater dust extinction than that implied
by the SN colors.
2. OBSERVATIONS AND ANALYSIS
2.1. Column Densities and Equivalent Widths
Our approach is to first examine the relationship between dust
extinction and interstellar absorption lines in the Milky Way.
These results will then be contrasted with a similar comparison
between the dust extinction in the line-of-sight to the SN Ia as
derived from their optical and near-infrared (NIR) light curves,
and the narrow absorption lines producedby the host galaxy ISM
and/or a pre-existing CSM (hereafter referred to collectively as
“host absorption”).
In the first case, we employ a sample of 46 SNe and
active galactic nuclei (AGNs) as external beacons to study the
absorption lines produced by the ISM of the Milky Way. Echelle
spectra of 22 of these objects are drawn from the observations
of thermonuclear and core-collapse SNe published by Sternberg
et al. (
2011) and available through WISeREP (Yaron & Gal-Yam
2012
22
). The remaining 24 spectra in our data set correspond
to unpublished observations of SNe and AGNs obtained with
the Magellan Inamori Kyocera Echelle (MIKE; Bernstein et al.
2003) on the 6.5 m Clay telescope. Table 1 lists the objects in
this sample. Henceforth, we refer to these objects as the “Milky
Way” sample.
To study the relationship between the SN dust extinction
and the narrow host absorption lines, we have put together a
sample of 32 SNe Ia with both high-dispersion spectra and
well-observed light curves. Spectra for 21 of these SNe Ia were
drawn from the Sternberg et al. (
2011) study (also available
through WISeREP), and an additional 6 are taken from Foley
et al. (
2012b). Results for the remaining 5 SNe are taken from
the literature. Table
2 lists the full sample of 32 SNe Ia along
with host galaxy names, morphologies, and references to the
SN photometry. Table
3 gives the sources and wavelength
resolutions of the high-dispersion spectral observations. These
SNe are referred to as the “host absorption” sample in the
remainder of this paper.
Column densities of neutral sodium and potassium were mea-
sured for both the Milky Way and host absorption components
of the Na i D λλ5890, 5896 and K i λλ7665, 7699 doublets
using the Voigt profile fitting program, VPFIT,
23
developed by
R. F. Carswell, J. K. Webb, and others, in combination with
the VPGUESS
24
interface of J. Liske. Upper limits for non-
detections of both Na i and K i were calculated by first estimating
an upper limit to the equivalent width, and then converting this
to a column density using empirical relations between equiva-
lent width and column density derived from weak, unsaturated
lines in other objects. In cases where the Na i D lines were sig-
nificantly saturated (log N
Na i
 13 cm
−2
), the much weaker K i
lines were used to determine the velocity and Doppler parame-
ter, b, of each visible component, and this information was em-
ployed in fitting the saturated portion of the Na i D profiles. This
procedure was possible for most of the objects observed from
2006 onward. Four of the SNe in the host absorption sample had
extremely strong D lines (log N
Na i
> 13.5cm
−2
). For one of
these—SN 2002bo—the spectral coverage did not include the
K i lines. A model with two absorption components provided a
significantly better fit to the severely saturated profiles of the D
lines in this SN than did one with a single component, and so
we have adopted the results of the two-component model in this
paper. However, without the additional information provided
by the K i lines, the error associated with the measurement of
log N
Na i
is large.
Kemp et al. (
2002) found that Na i column densities measured
from fitting profiles to the D lines for values log N
Na i
>
12.5cm
−2
were systematically underestimated by 0.40–0.70 dex
compared to column densities measured from the much weaker
Na i UV λλ3302, 3303 doublet. The UV lines are not covered
by our echelle spectra so we cannot confirm this, although
as mentioned in Section
4.5, a comparison of the N
Na i
/N
K i
ratio derived for our Milky Way sample with the measurements
of Kemp et al. (
2002) suggests that our N
Na i
values may be
similarly affected. This should be borne in mind when using
the relations involving the Na i column density developed in
22
http://www.weizmann.ac.il/astrophysics/wiserep/
23
http://www.ast.cam.ac.uk/∼rfc/vpfit.html
24
http://www.eso.org/∼jliske/vpguess/
2

The Astrophysical Journal, 779:38 (21pp), 2013 December 10 Phillips et al.
Tab le 1
Milky Way Na i and K i Column Density Measurements
A
V
log N
Na i
log N
K i
Object (mag) (cm
−2
)(cm
−2
) Reference
(1) (2) (3) (4) (5)
2003gd 0.19 ± 0.03 12.775 ± 0.034 ··· 1
2006be 0.08 ± 0.01 12.068 ± 0.038 ··· 2
2006ca 0.64 ± 0.10 13.181 ± 0.052 ··· 2
2006eu 0.52 ± 0.08 12.914 ± 0.039 ··· 2
2007af 0.11 ± 0.02 11.751 ± 0.106 ··· 2
2007hj 0.26 ± 0.04 12.859 ± 0.102 ··· 2
2007kk 0.64 ± 0.10 12.801 ± 0
.122 ··· 2
2007le 0.09 ± 0.02 11.924 ± 0.055 ··· 2
2007on 0.03 ± 0.01 11.178 ± 0.077 ··· 2
2007sr 0.13 ± 0.02 11.734 ± 0.018 ··· 2
2008C 0.23 ± 0.04 12.777 ± 0.467 ··· 2
2008fp 0.54 ± 0.09 13.141 ± 0.061 11.417 ± 0.070 2
2008ge 0.04 ± 0.01 11.307 ± 0.045 ··· 2
2008hv 0.09
± 0.01 12.276 ± 0.016 ··· 2
2008ia 0.62 ± 0.10 13.149 ± 0.010 11.545 ± 0.070 2
2009ds 0.11 ± 0.02 12.489 ± 0.020 ··· 2
2009ev 0.28 ± 0.05 12.737 ± 0.040 ··· 2
2009iw 0.24 ± 0.04 12.543 ± 0.021 ··· 2
2009le 0.05 ± 0.01 11.793 ± 0.011 ··· 2
2009mz 0.08 ± 0.01 11.972 ±
0.030 ··· 2
2009nr 0.07 ± 0.01 11.926 ± 0.031 ··· 2
2010A 0.08 ± 0.01 11.431 ± 0.032 ··· 2
2010ev 0.29 ± 0.05 12.564 ± 0.028 ··· 2
2010jl 0.07 ± 0.01 11.791 ± 0.119 ··· 1
2010ko 0.39 ± 0.06 12.767 ± 0.103 11.210 ± 0.087 1
2011K 0.27 ± 0.04 12.500 ± 0.016 ··· 1
2011di 0.
29 ± 0.05 12.472 ± 0.025 ··· 1
2011dn 0.49 ± 0.08 12.948 ± 0.011 11.614 ± 0.052 1
2011dq 0.31 ± 0.05 12.808 ± 0.022 11.619 ± 0.024 1
2011dy 0.19 ± 0.03 12.960 ± 0.056 11.103 ± 0.041 1
2011ek 0.97 ± 0.15 12.999 ± 0.024 11.867 ± 0.038 1
2011fj 0.47 ± 0.07 13.043 ± 0.
047 ··· 1
2012cg 0.07 ± 0.01 11.218 ± 0.057 ··· 1
PTF11iqb 0.09 ± 0.02 12.004 ± 0.007 ··· 1
3C273 0.06 ± 0.01 12.073 ± 0.014 ··· 1
IC4329A 0.16 ± 0.02 11.999 ± 0.137 ··· 1
Mk509 0.16 ± 0.02 12.142 ± 0.041 ··· 1
NGC 1068 0.09 ± 0.02 12.618 ± 0.156 ··· 1
NGC 2110 1.03 ± 0.16 13.
149 ± 0.071 11.559 ± 0.049 1
NGC 3783 0.33 ± 0.05 12.656 ± 0.013 11.169 ± 0.055 1
PDS456 1.42 ± 0.23 13.309 ± 0.037 11.757 ± 0.022 1
Fairall51 0.30 ± 0.05 12.388 ± 0.036 ··· 1
IRAS06213+0020 1.77 ± 0.28 13.538 ± 0.049 12.167 ± 0.018 1
IRAS08311-2459 0.29 ± 0.05 12.668 ± 0.040 ··· 1
IRAS09149-6206 0.50
± 0.08 12.643 ± 0.100 11.023 ± 0.075 1
IRAS11353-4854 0.53 ± 0.08 12.771 ± 0.075 11.522 ± 0.027 1
Notes. Columns: (1) Object name; (2) Milky Way dust extinction (Schlafly & Finkbeiner 2011);
(3) Logarithm of the total neutral sodium column density; (4) Logarithm of the total neutral potassium
column density; (5) High-dispersion spectroscopy reference [1 = unpublished MIKE spectrum; 2 =
Sternberg et al. (
2011)].
Section
3.1. Nevertheless, since our approach is to compare
the SN host absorption measurements relative to our same
measurements for the Milky Way, this problem should not affect
our conclusions.
Although the Ca ii H & K lines were also present in many of
the SNe spectra, except for a few specific objects, they are not
included in this study since the column density of Ca ii is poorly
correlated with dust extinction in the Milky Way, presumably
due to variations in the large depletion factor of calcium
(Hobbs
1974).
Equivalent widths of the DIB at 5780 Å were calculated
using the IRAF
25
task fitprofs assuming a Gaussian profile
of 2.1 Å FWHM, a typical value in the Milky Way (Tuairisg et al.
2000; Welty et al. 2006; Hobbs et al. 2008).
26
This feature was
25
IRAF is distributed by the National Optical Astronomy Observatory, which
is operated by the Association of Universities for Research in Astronomy
(AURA) under cooperative agreement with the National Science Foundation.
26
Although very high dispersion spectroscopy has shown that the profile of
the 5780 Å feature is not Gaussian (Galazutdinov et al.
2008), the wavelength
resolution and signal-to-noise ratio of our observations do not warrant a more
sophisticated method of determining the equivalent width.
3

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "On the source of the dust extinction in type ia supernovae and the discovery of anomalously strong na i absorption" ?

The authors show that the dust extinction of the objects where the diffuse interstellar band at 5780 Å is detected is consistent with the visual extinction derived from the supernova colors. Overall, the authors find the most accurate predictor of individual supernova extinction to be the equivalent width of the diffuse interstellar band at 5780 Å, and provide an empirical relation for its use. This strongly suggests that the dust producing the extinction is predominantly located in the interstellar medium of the host galaxies and not in circumstellar material associated with the progenitor system. This coincidence suggests that outflowing circumstellar gas is responsible for at least some of the cases of anomalously large Na i column densities. Finally, the authors identify ways of producing significant enhancements of the Na abundance of circumstellar material in both the single-degenerate and double-degenerate scenarios for the progenitor system. 

Q2. What is the effective observational tool for measuring the expansion history of the universe?

Type Ia supernovae (SNe Ia) are one of the most effective observational tools for measuring the expansion history of the universe. 

Q3. What is the average value of RV for the four SNe colors?

An independent test of the values of RV derived from the SN colors is provided by spectropolarimetric measurements, since the wavelength of maximum polarization is well-correlated with RV (Sarkowsky et al. 

Q4. What is the significance of the blueshifted subcomponent of the D lines?

In addition, Maguire et al. (2013) found that the strength of the blueshifted subcomponent of the D lines for those objects with blueshifted profiles was correlated with the (B−V ) colors of the SNe at maximum light.