SN 1991T: Re Ections of Past Glory1

SN T Reections of Past Glory

astro-ph/9407097 29 Jul 94

2 2

Brian P Schmidt and Rob ert P Kirshner

HarvardSmithsonian Center for Astrophysics Garden St Cambridge MA

Bruno Leibundgut

Europ ean Southern Observatory KarlSchwarzschildStrasse D Garching b ei M unchen Germany

Lisa A Wells and Alain C Porter

Kitt Peak National Observatory PO Box Tucson AZ

Pilar RuizLapuente and Peter Challis

HarvardSmithsonian Center for Astrophysics Garden St Cambridge MA

Alexei V Filipp enko

Department of Astronomy and Center for Particle Astrophysics University of California Berkeley California

Submitted to The Astrophysical Journal Letters

Abstract

We have obtained photometry and sp ectra of SN T which extend more than days past

maximum light by far the longest a SN Ia has b een followed Although SN T exhibited nearly

normal photometric b ehavior in the rst days following maximum by days its decline had

slowed and by days the sup ernova brightness was consistent with a constant apparent magnitude

of m Sp ectra near maximum showed minor variations on the SN Ia theme which grew less

conspicuous during the exp onential decline At days the nebular sp ectrum was comp osed of Fe

and Co lines common to SNe Ia However by days past maximum light these lines had shifted in

wavelength and were sup erimp osed on a strong blue continuum The luminosity of SN T at

38 2 1

days is more than D Mp c ergs s with a rate of decline of less than mags days

We show that this emission is likely to b e light that was emitted by SN T near maximum light which

has reected from foreground dust much like the light echos observed around SN A

Based in part on observations made at the Multiple Mirror Telescope jointly op erated by the Smithsonian Institution

and the University of Arizona

Guest Observer at Kitt Peak National Observatory op erated by AURA Inc under contract to the National Science

Foundation

Deceased Octob er

Kitt Peak National Observatory National Optical Astronomy Observatories op erated by the Asso ciation of Universities for Research in Astronomy Inc AURA under co op erative agreement with the National Science Foundation

Introduction

SN T was discovered in NGC by a host of observers Waagen et al on April

and reached a maximum of m mag days later Phillips et al Ford et al

So on after its discovery this ob ject was noted as an unusual member of the SN Ia o ck In a class

known for exceptionally uniform sp ectral and photometric evolution SN T b efore maximum showed

abnormally weak sp ectral features of intermediate mass elements Filipp enko et al RuizLapuente

et al Phillips et al Jeery et al and an unusually broad maximum in its B and V

light curves Phillips et al Ford et al Furthermore this black sheep of the SN Ia family

app eared to b e more luminous at maximum than standard SNe Ia Filipp enko et al Phillips

although other views on the absolute magnitude are tenable Phillips et al However in the weeks

after maximum light SN T b egan to lose its unique sp ectroscopic and photometric features and

so on app eared indistinguishable from other wellobserved SNe Ia Filipp enko et al RuizLapuente

et al Phillips et al such as SN E Kirshner et al SN B Branch et al

SN B Barb on et al Wells et al and SN N Leibundgut et al In

this letter we present photometry and sp ectra of SN T which extend days past maximum the

longest time for which a SN Ia has b een followed to date These observations show that ab out days

after maximum SN T stopp ed the exp onential fading characteristic for SNe Ia Although SN T

has once again strayed from its sibling SNe Ia we b elieve that this ob jects p eculiar late time b ehavior

has nothing to with the eccentricities seen in its youth but instead is due to a light echo formed in

intervening dust b etween the sup ernova and the Earth

Observations

Our optical light curve of SN T was determined from CCD images obtained using the Kitt Peak

National Observatorys KPNO m telescop e and the Fred Lawrence Whipple Observatory FLWO

m telescop e A sequence of standard stars Figure Plate near the SN was calibrated on two

photometric nights with the FLWO m and one photometric night on the m by observing Landolt

standard stars Transformation co ecients for each detector were derived using the metho d

describ ed by Harris Fitzgerald Reed and used to determine the BVR I magnitudes of the

c c

sequence stars as given in Table The magnitude of the SN was determined by comparing the magnitude

of the SN relative to the sequence stars using the p oint spread tting program DOPHOT Schechter

Mateo Saha correcting for the dierence in the color b etween the standard stars and SN using

the color terms determined on the photometric nights To estimate the error in each measurement we

have used DAOPHOT Stetson to place several articial stars of known brightness onto regions

of the galaxy similar to that where the SN lies The error estimate for each observation is then derived

from the mean deviation b etween the articial stars known brightnesses and that output by DOPHOT

The derived photometry and error estimates are presented in Table and Figure

Our sp ectra Figure were obtained using the Multiple Mirror Telescope MMT red channel sp ec

trograph Schmidt Weymann Foltz To obtain our last three sp ectra it was necessary to p osition

h m s 0 00

the telescop e by osetting to SN Ts p osition B from

a nearby star Star Table The sp ectrograph has slit allowing us to subtract the background

galaxy by tting it from the region immediately adjacent to the SN Although this is an eective metho d

to remove the background contamination it is prone to errors if the background emission varies signif

icantly over the scale of the spatial resolution After careful insp ection of the two dimensional sp ectra

we b elieve that the features seen in these sp ectra are intrinsic to the SN and are not the result of con

tamination from any underlying source The B V colors synthesized from the sp ectra using the lter

functions of Bessell and zero p oints derived from the Vega sp ectrum of Hayes Latham Hayes

agrees to magnitudes with our photometry providing further supp ort that the shap e of the

sp ectrum is correct The last three sp ectra app ear identical within the noise and have b een combined to

increase the signal to noise ratio The sp ectra are presented in Figure

Discussion of Observations

Few SNe Ia have b een well observed at late times however SN Ts photometric b ehavior is

extraordinary Although the rate of decline b etween and days past maximum

d d

mag mag is typical for SNe Ia Turatto et al Figure shows that by

days after maximum SN T had stopp ed the exp onential fading characteristic of SNe Ia Assuming

they had the same luminosity at maximum SN T was intrinsically more than two magnitudes

brighter than SN E Kirshner Oke and SN N Phillips et al at days If

SN T was intrinsically brighter than SN E at maximum light as suggested by Filipp enko et al

and Phillips then its excess luminosity at late times would b e even more pronounced Over

our last three photometric observations b etween and days after maximum light SN Ts rate

of decline was consistent with zero mag The total luminosity of the SN on this

38 04A 1

late time plateau is substantial D Mp c ergs s b eing emitted in the optical

B V RI alone

At days past maximum SN Ts sp ectrum was similar to other SNe Ia at this ep o ch Ruiz

Lapuente Filipp enko Correcting for the redshift of NGC lines primarily made up of FeII I

multiplets are observed at A A A and A multiplets

obs obs obs obs

of FeII are contributing at A A A and A and lines

obs obs obs obs

from CoI I I are apparent at A and A The width of these lines indicate an

obs obs

expansion velocity of the Fe p eak ejecta of v km s which is approximately faster than

either SN E Kirshner Oke or SN N Phillips et al

More than a year and a half later days past maximum the SNs sp ectrum was dierent SN T

app ears to b e emitting a larger fraction of its ux blueward of A and the feature attributed to

several Fe I I I lines at A at days is narrower and substantially blue shifted to

obs

A Emission features are also apparent at A and A It is not

obs obs obs

proven however that SN T is dierent from other SNe Ia sp ectra at this age SN E is the only

other SN Ia with a sp ectrum at such an age Kirshner Oke and there is very little information

in this observation Between and days past maximum no further evolution in SN Ts

sp ectrum was detected

Understanding SN Ts Late Time Luminosity

There are several p ossible explanations for SN Ts large late time luminosity Before embarking

on explanations for this observation a skeptical reader might wonder if we are seeing the SN at all We

can b e condent that the ob ject we are seeing is indeed SN T and not some other source for several

00 00

reasons The observed ob ject is a p oint source FWHM lo cated within of the p osition of

SN T The absolute magnitude of the ob ject M logD Mp c is equivalent to

O Ia stars Allen and could b e a luminous OB asso ciation however no ob ject is present at this

p osition on a photographic plate taken a decade b efore the SN explosion Figure Plate Sandage

Binggeli Tammann even though ob jects as faint as SN T are clearly visible Furthermore

the ob jects color B V V R shows a large R deciency which is not consistent

with the thermal sp ectrum of hot stars or the H rich sp ectrum of an H II region but is consistent

with SNe Ia colors we also do not see the narrow emission lines asso ciated with H I I regions and OB

asso ciations Most convincingly the strongest sp ectral features observed in the ob jects sp ectrum after

days are broad lines like those of a sup ernova

The physics of burning material to nuclear statistical equilibrium is well understo o d and the resulting

isotopic ratios are reasonably insensitive to the exp ected variations in SNe of neutron fraction density

and temp erature Woosley Therefore although an unusual abundance of Ti for example is

dicult to rule out observationally it is implausible that SN T overproduced this isotop e by an

order of magnitude more than M would b e required Other isotop es such as Co can b e ruled

out by the extremely slow rate of decline Strong circumstellar interaction p owering the light curve is

also unlikely b ecause such a mechanism would probably pro duce strong lines of hydrogen among others

which are not observed Fransson Houck Kosma have run timedep endent mo dels of SNe Ia

through late times Their calculations show that mo dels of SNe Ia such as DD Woosley b egin

to undergo ionization freeze out cf Fransson Kosma for SN A at approximately

days several hundred days later than necessary to explain SN Ts late time luminosity A more

general problem with these mo dels however is encountered b etween to days after explosion

when they undergo an IRcatastrophe as describ ed by Axelro d causing the optical band light

curves to fall sharply from their constant rate of exp onential decline This scenario while dicult to test

with SN T is in disagreement with more typical SNe Ia such as SN E and SN N

We b elieve that SN Ts late time luminosity can b e attributed to a light echo The p ossibility

of detecting light echos the reection of a SNs light from dust in the surrounding ISM has long

b een discussed Zwicky van den Bergh Chevalier Schaefer but it was not until

SN A Suntze et al Crotts Gouies et al that this phenomenon was unam

biguously seen in asso ciation with a sup ernova The total integrated magnitude of SN As light

echo was m and remained essentially unchanged over the p erio d b etween and Crotts

p ersonal communication This ratio of the SNs maximum light brightness and the light echo brightness

mag and the constant ux is consistent with what we see in the case of SN T Furthermore

we know that there was a signicant amount of dust in front of SN T from interstellar absorption

lines seen at the recession velocity of the host galaxy Meyer Roth

The light echo hypothesis makes a strong prediction the sp ectrum observed in reected light is

exp ected to consist of light emitted at early times We should see the intensityweighted timeintegrated

sp ectrum of SN T mo died by preferential scattering of blue light by the dust We have computed the

timeintegrated sp ectrum of SN T for the days following discovery by weighting each sp ectrum of

SN T with its V magnitude Phillips et al and interpolating b etween sp ectra to give uniform

time coverage We have parameterized the eects of scattering as a S / scattering law Suntze

et al The results are compared compared to SN Ts late time sp ectrum in Figure The

match is very go o d using with the ma jor features in the sp ectrum all represented and at the

correct wavelength Because we could only integrate the sp ectrum of SN T for the days following

discovery for lack of observations it is not surprising that some of the lines eg that at A do

not app ear suciently strong in the integrated sp ectrum It is worth noting that the timeintegrated

sp ectrum of SN T is dierent from its maximum light sp ectrum and simply comparing the echo

sp ectrum with the maximum light sp ectrum would b e misleading

At late times SN T is relatively bright and constant in luminosity These observations lead us to

b elieve that the light echo is b eing formed in a cloud of dust in NGC which is tens or hundreds of

light years in front of SN T and not with dust asso ciated with the sup ernova If the reecting dust

were close to the sup ernova the scattering angle for the light would b e large very inecient compared

Schaeer claimed to see evidence for a light echo around SN G from a slowing in its light curve by days

However Phillips p ersonal communication has reanalyzed these data and concludes that SN G continued to fall at its previous rate and did not slow in its decline

to forward scattering and would change signicantly over a year thereby changing the brightness of

the echo If the light echo is b eing formed in a foreground dust cloud it probably has an apparent

size greater than and should b e resolvable using the Hubble Space Telescope Such observations

would determine the p osition of the intervening cloud Unfortunately it is unlikely that a direct geometric

distance determination as discussed by Sparks will b e p ossible for this ob ject His metho d exploits

the fact that the p olarization of scattered light has a sharp maximum at scattering angles of He

demonstrates with this geometry that assuming there is sucient dust lo cated in the appropriate p osition

in space the light echo in p olarized light will have a radius r ct Unfortunately the dust pro ducing

SN Ts light echo is most likely lo cated well in front of the sup ernova and therefore the dust will

not b e lo cated close enough to the SN to pro duce the ring of p olarized light

We b elieve this is after SN A the second conclusive detection of a sup ernova light echo However

there is no reason to think that SN Ts light echo has anything to do with its early idiosyncrasies

Careful insp ection of other bright sup ernovae of all types which have some extinction from their host

galaxy SN J in M and SN C in M come to mind might yield similar ndings to those

presented here

This work is dedicated to the memory of Alain Porter whose hard work and determination will not b e

forgotten whose contributions we valued and whose companionship in scientic adventures we will miss

We would also like to thank Bruno Bingelli for providing us with the preexplosion image of NGC

as well as Mark Phillips for providing his earlytime SN T data and for his helpful comments We

are grateful for valuable discussions with Philip Pinto and Peter Hoich This work is supp orted in part by NSF grants AST and AST as well as NASA grants NAG and NGT

References

Allen C W Astrophysical Quantities rd ed London Athlone Press

Barb on R Benetti S Rosino L Capp ellaro E Turatto M AA

Bessell M S PASP

Branch D Lacy C McCall ML Uomoto A Wheeler C Wills B J Sutherland PG

ApJ

Chevalier R A ApJ

Crotts A P S ApJ L

Crotts A P S Kunkel W E McCarthy P J ApJ L

Filipp enko A V et al ApJ L

Ford C H Herbst W Richmond M W Baker M L Filipp enko A V Treers R R Paik Y

Benson P J AJ

Fransson C Kozma C ApJ L

Fransson C Houck J Kozma C IAU Coll No Sup ernovae and Sup ernova Remnants

Xian China ed R McCray Cambridge Cambridge University Press in press

Harris W E Fitzgerald M P Reed B C PASP

Hayes D S Latham D W Hayes S H ApJ

Gouies C Melnick J Remy M Rosa M Danziger I J AA L

Kirshner R P Oke J B ApJ

Landolt A U AJ

Leibundgut B Kirshner R P Filipp enko AV Shields JC Foltz CB Phillips MM Son

neb orn G ApJ L

Meyer D M Roth K C ApJ L

Phillips M M ApJ L

Phillips M M Wells L A Suntze N B Hamuy M Leibundgut B Kirshner R P Foltz C

B AJ

Phillips M M et al in preparation

Sandage A Binggeli B G A Tammann AJ

RuizLapuente P et al ApJ L

RuizLapuente P Filipp enko A V in Origin and Evolution of the Elements ed N

Prantzos E VangioniFlam M CasseCambridge University Press p

Schechter P L Mateo M Saha A PASP

Schmidt G D Weymann R J Foltz C B PASP

Schaefer B E ApJ L

Sparks W B ApJ Sep issue

Suntze N B Heathcote S Weller W G Caldwell N Huchra J P Nature

Stetson P B PASP

Turatto M Capp ellaro E Barb on R Della Valle M Ortolani S Rosino L AJ

van den Bergh S ApSpaceSci

Waagen E et al IAU Circ

Wells L A et al AJ in press

Woosley S E in Nucleosynthesis and Chemical Evolution ed B Hauck A Maeder Geneva

Observatory CH Sauverny Switzerland

Woosley S E GammaRay Line Astrophysics ed P Durouchoux N Prantzos New York

American Institute of Physics

Zwicky F Rev of Mo d Physics

TABLE Comparison Star Magnitudes

Star B V R I

h m s 0 00

Uncertainties in hundredths of magnitudes are listed in parentheses

TABLE LateTime Photometry of SN T

Date JD B V R I BV VR VI Telescope Observer

Dec m Schmidt

Mar m Leibundgut

Apr m Porter

Jun m Porter

Jan m Wells

Mar m Wells

May m Wells

Dec m Wells Schmidt Uncertainties in hundredths of magnitudes are listed in parentheses

Figure B CCD image of SN T taken with the KPNO m on Jan East is up and north

is to the right The lo cations of the SN and the comparison stars are shown Precise p ositions of the SN

and comparison stars are given in Table Also shown is a plate of NGC taken by Sandage et al

in The SN p osition is marked as well as a star which was of a comparable brightness to the sup ernova on Jan 10

B V R 15 I 15 m R V , m , m B I m 20 20

SN 1972E "B"

25 25 8400 8600 8800 9000 9200

Julian Date (2440000+)

Figure B V RI photometry of SN T Also shown is the early BV photometry of SN T Phillips

et al and a template of B photometry derived from SN E Kirshner Oke The V

and I scales are oset by one magnitude from the B and R scales

Figure Late time sp ectral observations of SN T are displayed for and days

after maximum light corrected for the recession velocity of NGC The last three sp ectra app ear

identical and have b een combined to increase the signal to noise ratio Suggestions as to the principal contributing ions for each line are included

Figure A comparison of the timeintegrated intensityweighted sp ectrum with the observed latetime

sp ectrum for SN T The integrated sp ectrum has b een corrected for the eect of scattering using a

law The go o d agreement b etween these two sp ectra suggests that we are seeing light emitted by

SN T around maximum light that has b een reected o intervening dust in the sup ernovas host galaxy NGC