Alex Gianninas
April 16 th, 2014
UT Arlington Physics Colloquium
Collaborators

University of Oklahoma
 Mukremin Kilic, Sara Barber, Paul Canton

Smithsonian Astrophysical Observatory
 Warren Brown, Scott Kenyon

University of Warwick
 JJ Hermes

Université de Montréal
 Patrick Dufour, Pierre Bergeron
2
White dwarfs are the dense stellar
remnants of 97% of all stars
in the Galaxy

M ~ 0.6 M

R ~ 0.01 R ~ 1 R

 ~ 106 

log g ~ 8.0
(log g = 4.4)

Teff ~ 150,000 
3000 K

No H burning 
WDs cool over
several Gyrs
3
Gravitational settling quickly
creates a stratified internal
structure
H
He
C/O
He
C/O
Non-DA (DO, DB, DQ, DZ)
(20%)
C/O
DA
(80%)
Hot DQ
4
WDs with M < 0.45 M must be
formed in binary systems
Binary Evolution
Gianninas et al. (2011)
5



Binary companions strip a significant amount of
material from the progenitor, preventing the ignition of
He burning, giving birth to an extremely low mass (ELM)
WD
ELM WDs are found almost exclusively in short-period
binaries (P < 1 day)
Many will merge in less than a Hubble time!
6
ELM WDs should have
He cores
H
He
C/O
DA
H
He
ELM DA
7
ELM WDs are the potential progenitors of many
types of astrophysical objects
Progenitors of Type Ia supernovae
(Iben & Tutukov 1984)
 Progenitors of underluminous .Ia supernovae (Bildsten et al.
2007)
 Progenitors of AM CVn systems (Kilic et al., 2014)
 Progenitors of R CrB stars (Clayton 2013)
 Pulsar companions (Kaplan et al. 2013, Ransom et al.
2014)
 Shortest period systems are important sources of
gravitational waves


The ELM Survey is a targeted search for Extremely Low
Mass WDs (M < 0.30 M)
8
Candidates are selected mostly
via SDSS colors
 Hypervelocity Survey
(Warren Brown,
SAO): search for B type stars leaving the
Galaxy, colors similar to ELM WDs 
15% of targets are in fact ELM WDs
 Spectroscopy from the Sloan Digital Sky
Survey (SDSS)
 SDSS colors (u-g, g-r)
9
ELM WDs straddle the color-color space between
normal WDs and A stars
Brown et al. (2012)
10
We have an ongoing multi-site campaign to
confirm the nature of
our ELM WD candidates

Observatories
 Mount Hopkins, AZ:
○ MMT (6.5m)
○ FLWO (1.5m)
 Kitt Peak, AZ: KPNO 4m

Once confirmed, we seek to
improve our orbital solution by
better sampling all phases of
the orbit
11
The ELM Survey has been
extremely successful thus far
Paper
ELM WD
Merger Systems
Brown et al.
(2010)
I
12
6
Kilic et al. (2011)
II
4
2
Brown et al.
(2012)
III
7
6
Kilic et al. (2012)
IV
7
5
Brown et al.
(2013)
V
17
6
• Before ELM Survey:
• Only 6 known merger systems
• Shortest known period is P = 1.5 hr
• ELM Survey has found 8 systems with
hr and 3 with P < 1 hr
P < 1.5
12
Orbital solutions yield the period (P) and
velocity semi-amplitude (K)
Brown et al. (2013)
13
The orbital parameters put limits on
M2 and the merger time

Mass Function  Lower limit on M2

Merger time  Upper limit on 
14
Some ELM WDs may be Type Ia
progenitors




Super-Chandrasekhar
mass systems not
necessarily Type Ia
progenitors
Mass ratio is
important
0.2 + 1.2 M systems
→ Stable mass
transfer system (AM
CVn, .Ia SN)
Can be WD+NS
binaries
Brown et al. (2013)
15
We have identified the first unambiguous
progenitors of AM CVn (cataclysmic variables,
novae)
Have massive
Companions
 Pulsars? No
detections with
Chandra
 Will undergo stable
mass transfer

Kilic et al. (2013)
16
The orbital parameters put limits on
M2 and the merger time

Mass Function  Lower limit on M2

Merger time  Upper limit on 
17
There are model dependent and
model independent methods to
determine M

Model dependent: two ingredients
 Precise measurements of the atmospheric
parameters (Teff and log g)
 Evolutionary models

Model independent:
 Eclipses
 Ellipsoidal variations (tidal distortion)
18
The Hydrogen Balmer lines are very
sensitive to Teff and log g
19
Model fits allow us to measure the atmospheric
parameters
20
New grids of models were
required for ELM WDs
New regime in surface gravity (log g < 7.0)
 New grid of models (down to log g = 4.5)
computed
 Need to include higher Balmer lines (up to H12)
in the models and the fits since lines are still
present in WDs with log g < 7.0

21
22
ELM WDs undergo a series of Hshell flashes as they evolve
Althaus et al. (2013)
23
A statistical approach is used to
calculate masses and ages
Mean weighted by
the time spent at
each point in Teff - log
g plane
 Still much
uncertainty

Althaus et al. (2013)
24
ELM WDs with log g < 6.0 all
have metals
25
26
27
J0745 has more than
just Ca
28
J0745 is the most metal rich WD from
the ELM Survey
Gianninas et al. (2014)
29
J0745 is unique for its Teff
 Teff = 8380 K
 log g = 6.21
Abundances
log Ca/H = -5.8
log Mg/H = -3.9
log Cr/H = -6.1
log Ti/H = -5.6
log Fe/H = -4.5
All nearly solar!

Gianninas et al. (2014)
30
For more massive WDs, metals
originate from circumstellar disks
Circumbinary disks for ELM WDs?
Optically thick
Dotted
i = 30°
Dashed-dotted
i = 60°
Rin = 1.3 R
Rout = 1.4 – 5.0 R
Radiative levitation?
 Recent H-shell flash?

31
HST
(COS)
Yes!
10 orbits
Keck
(HIRES)
TBD…
32
33
J0651 is the poster child for ELM
WDs
Shortest period ELM
WD binary

P = 12.75 min!
 Eclipsing!
 After initial discovery,
photometric follow-up
at McDonald, APO,
Gemini North and
GTC

34
The midpoint of the eclipses reveals that
the orbit is decaying !
Hermes et al. (2013)
35
The rate of decay agrees with the
prediction of General Relativity!
Hermes et al. (2013)
36
The Hulse-Taylor binary pulsar take 30
years to display the same period shift!
Weisberg & Taylor (2005)
37
eLISA

Replaces LISA
 Two beams instead of three
 Reduced sensitivity
 Frequency range spans four decades (~0.1
mHz – ~1 Hz)

“The Gravitational Universe” approved as
science theme for ESA L3 project ( launch
in ~2034)
38
J0651 will be a verification
source for eLISA
Galactic foreground
eLISA, after 2 yrs
Gianninas et al. (2014, submitted)
39
J0651 presents a unique opportunity to
measure rotation rate
Isolated WDs : slow rotators ( < 10’s of
km/s), measured with Ca line profile
 Evolution in compact binaries and
interactions with companions  could
ELM WDs be fast rotators?
 Tidal forces  synchronized orbits
 For J0651 = 200 km/s

Berger et al. (2005)
40
The Rossiter-McLaughlin Effect
produces an anomaly in the RV curve
Winn et al. (2005)
41
We can predict the magnitude of
the R-M effect using models

Follow recipe from Winn et al. (2005)
 1. Rotationally broaden line profile of primary
to simulate integrated spectrum (S)
 2. Unbroadened line profile, Doppler shifted to
the red/blue (Sp)
 3. Str = S - Sp
42
The R-M effect in J0651 would
produce a ~40 km/s shift in RV
43
We were awarded ½ night on
Keck I with LRIS
Only telescope where we could obtain
necessary S/N
 5 - 7 spectra (t ~ 25s) per exposure to reduce
number of times we read the CCD and ensure
the best possible temporal sampling
 Needed to synchronize blue and red sides of
instruments (different read times)
 6 hours of observing = 313 spectra

44
The folded RV curve does not show
any obvious signature of the R-M
effect
45
Conclusions
ELM Survey: targeted search for extremely
low-mass WDs
 Success: >60 systems avec P < 1 day
 Observed Phenomena:

 Pulsations
 Metals
 Eclipses
 Orbital Decay
 Ellipsoidal Variations
46
There is still plenty of
work to be done!

ELM Survey
 Upcoming observing runs: MMT, KPNO 4m
 LAMOST?
 Southern Hemisphere?
Spitzer
 J0745

 HST data to be analyzed
 Keck: TBD

J0651
 Finish data reduction and analysis
47
Is this the fate of J0651?
48