School of Physics amp Astronoy Cataclysmic Variables 10 Breakthroughs in 10 Years P Marenfeld and NOAOAURANSF Christian Knigge University of Southampton University of Southampton School of Physics amp Astronoy ID: 177171
Download Presentation The PPT/PDF document "University of Southampton" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
University of SouthamptonSchool of Physics & Astronoy
Cataclysmic
Variables:10 Breakthroughs in 10 Years
P. Marenfeld and NOAO/AURA/NSF
Christian
Knigge
University of SouthamptonSlide2
University of SouthamptonSchool of Physics & Astronoy
Outline
IntroductionCataclysmic variables: a primer10 breakthroughs in 10 years
(a personal and hugely biased perspective...)
Evolution
Accretion
Outflows
The Role of UV AstronomySummary
35 minutes
9 8 7 6 5 4 3 2 1
XX XXXXX
Links to Other Systems
(BH/NS LMXBs)Slide3
University of SouthamptonSchool of Physics & Astronoy
Cataclysmic Variables: A PrimerThe Physical Structure of CVs
White Dwarf
Accretion Disk
Red Dwarf
White dwarf
primary
UV bright
“Main-sequence”
secondary
75 mins < Porb
< 6 hrs Roche-lobe overflow Accretion usually via a diskUV-bright
Disk accretion is unstable if below critical rate
dwarf novae
Mass transfer and evolution driven by angular momentum
loss
Evolution is (initially) from long to short periods
Credit: Rob HynesSlide4
University of SouthamptonSchool of Physics & Astronoy
Cataclysmic Variables: A Primer
The Orbital Period Distribution and the Standard Model of CV EvolutionClear “Period Gap” between 2-3 hrs
Suggests a change in the dominant angular momentum loss mechanism:
Above the gap:
Magnetic Braking
Fast AML
High
Below the gap:
Gravitational RadiationSlow AML Low Minimum period at Pmin ≈ 80 mindonor transitions from MS
BD
beyond this, Porb increases again This disrupted magnetic braking scenario is the standard model for CV evolution
Knigge 2006Slide5
Breakthrough I: Evolution
Disrupted Angular Momentum Loss at the Period Gap
Standard model predictionThe period gap is caused by a disruption in AML when the donor becomes fully convectiveMagnetic braking drives high above the gap
Donor is slightly out of TE and thus oversizedAt , donor becomes fully convectiveMB ceases (or is severely reduced)
drops
--> d
onor relaxes (shrinks) to TE radius
Donor loses contact with RLCV evolves through gap as detached binaryResidual AML (e.g. GR) shrinks orbit (and RL)
Contact with donor re-established at Observational reality pre-2005
No direct empirical support for this picture (other than the existence of the gap itself)University of SouthamptonSchool of Physics & AstronoyHowell et al. 2001Slide6
University of SouthamptonSchool of Physics & Astronoy
Patterson et al. (2005),
Knigge (2006)
Donors are significantly larger than MS stars both above and below the gap
Clear discontinuity at M
2
= 0.20 M
☼, separating long- and short-period CVs!Direct evidence for disrupted angular momentum loss!
M-R relation based on eclipsing and “
superhumping” CVs Breakthrough I: EvolutionDisrupted Angular Momentum Loss at the Period GapSlide7
University of Southampton
School of Physics & Astronoy
We can even use the donor
relation to quantitatively reconstruct CV evolution
CV Donors are significantly larger than MS stars because they are bloated by mass loss
Higher
Larger
So we can use the
degree of donor bloating at given to infer
Above the gap: slightly reduced “standard” MB recipes work well
Below the gap: need enhanced AML,
significant revision of the standard model!
Breakthrough II: Evolution
Reconstructing CV Evolution Empirically
Knigge
(2006)
Knigge
,
Baraffe
&
Patterson (2011)Slide8
Breakthrough III: Evolution
Period Bouncers with Brown Dwarf Secondaries
Standard model predictions 99% of CVs should be found below the period gapA full 70% should be “period bouncers” with brown dwarf secondaries
Observational reality pre-2006
Not a single definitive period bouncer
Only ~10 candidates out of ~1000 CVs
No secondary with a well-established mass below the H-burning limit
Is this a selection effect or model failure?
University of Southampton
School of Physics & AstronoyHowell et al. 2001Slide9
Breakthrough III: Evolution
Period Bouncers with Brown Dwarf Secondaries
SDSS has yielded a deep new sample of ~200 CVs (Szkody et al. 2002-9)......including a sub-set of faint, WD-dominated systems near P
min (Gaensicke et al. 2009; see later)
A few of these are eclipsing, allowing precise system parameter determinations
At least 3 of these have M
2
< 0.072 M☼
(Littlefair et al. 2006, 2008)
At least some post-period-minimum systems with brown dwarf donors do exist!But one of them is very strange…University of SouthamptonSchool of Physics & AstronoyLittlefair et al. 2006, Science, 314, 1578Slide10
Stehle
et al. (1999)
Breakthrough III: Evolution
Period Bouncers with Brown Dwarf
Secondaries
SDSS J1507 is one of the three eclipsing CVs with sub-stellar donors…
…
but
i
ts
for other CVs
Two ideas:
J1507
is young -- born
with a
sub-stellar donor
(
Littlefair
et al. 2007)
J1507 is a low
metallicity
halo CV
(Patterson et el. 2008)
How can we test which is correct?
UV astronomy to the rescue!
FUV spectroscopy shows that [Fe/H] = -1.2
SDSS J1507 is an eclipsing period bouncer in the Galactic halo!
Rosetta stone for studying e
ffects of
metallicity
on accretion and evolution?
University of Southampton
School of Physics & Astronoy
Littlefair
et al. (2007)
Patterson et al. (2008)
Littlefair
et al. (2007)
Uthas
et al. (2011)Slide11
Breakthrough IV: Evolution
The Period Spike at
PminStandard model prediction
The number of CVs found in a particular Porb range is inversely proportional to the speed with which they evolve through it
So there should be a spike at
P
min
, in the period distribution since
Observational reality pre-2009No convincing spike anywhere near Pmin
in the CV Porb distributionUniversity of SouthamptonSchool of Physics & AstronoyBarker & Kolb 2003Slide12
Previously known CVs
SDSS CVs
Breakthrough IV: Evolution
The Period Spike at
P
min
Boris
Gaensicke
and collaborators have obtained orbital periods for most of the new SDSS CVs
The resulting period distribution does show a spike at Pmin for the first time (Gaensicke et al. 2009)CVs do in fact “bounce” at Pmin!
University of Southampton
School of Physics & AstronoyGaensicke et al. 2009Slide13
Breakthrough V: Evolution
CVs in Globular Clusters
A typical GC should contain ~100 CVs purely based on its stellar mass content
But
b
right X-ray binaries are overabundant in GCs by ~100x
(Clark 1975, Katz 1975)
New
dynamical
formation channels are available in GCstidal capture (Fabian, Pringle & Rees 1976)3- and 4-body interactionsCould CV numbers also be enhanced?Theory says yes, but “only” by a factor of ~2 (di Stefano & Rappaport 1994, Davies 1995/7,
Ivanova et al. 2006)There should be hundreds of accreting WDs in GCs!
Important and useful:Large samples of CVs at known distancesDrivers and tracers of GC dynamical evolution Are GCs SN Ia factories? (Shara & Hurley 2006)
So where are they?
University of Southampton
School of Physics &
Astronoy
CV
space
density:
(e.g
. Pretorius &
Knigge
2007, 2011)
Effective volume of
MW:
Expected # of
CVs in
MW:
Fraction of MW mass in GCs:
# of GCs
in
MW:
→ expected # of CVs per GC:
3-body exchange encounter
White DwarfSlide14
Breakthrough V: Evolution
CVs in Globular ClustersUniversity of Southampton
School of Physics & Astronoy
Shara
et al. (1996)
Difference Imaging of the Core of 47
Tuc
Early searches typically found only a handful per GC
(e.g.
Shara
et al. 1996,
Bailyn
et al. 1996, Cool et al. 1998)
Are CVs not formed or maybe even destroyed in CVs?
Significant implications for GC dynamics!
Selection effects?
Survey depth?
Dwarf nova duty cycle?
X-rays would be a great way to find CVs in GCs
But this used to be really hard!
Chandra has revolutionized the field
Deep X-ray surveys typically find tens per cluster
Numbers scale with collision rate
dynamical formation matters!
GCs do
harbour
significant populations of dynamically-formed CVs!
47
Tuc
with the ROSAT HRI
(
Hasinger
et al. 1994)
Shara
et al. (1996)
Pooley
& Hut (2006)
47
Tuc
with Chandra
(
Grindlay
et al. 2001;
Heinke
et al. 2005)
47
Tuc
with Chandra
(
Grindlay
et al. 2001;
Heinke
et al. 2005) Slide15
UV astronomy has also played a key role
Efficient way of finding new CVs and confirming X-ray-selected candidates
(
Knigge
et al. 2002,
Dieball
et al. 2005, 2009, 2010, Thomson et al. 2012)
Even
slitless
multi-object spectroscopic identification/confirmation is possible!Still many key unsolved questions!Are there enough CVs in GCs?Are they different from field CVs?Where are the double WDs? Are there SN Ia progenitors?
The core of 47
Tuc
: U-band
The core of 47
Tuc
: FUV (~1500A)
University of Southampton
School of Physics &
Astronoy
Knigge
et al (2002, 2003, 2008)
Breakthrough V: Evolution
CVs in Globular ClustersSlide16
Dwarf nova eruption (optical): SS
Cyg
Wheatley et al (2003)
X-ray transient outburst (X-ray): GX 339
Gallo et al (2004)
Adapted from Fender,
Belloni
& Gallo 2004
Breakthroughs VI and VII:
Accretion /
OutflowsOutburst Hysteresis and Jets
University of SouthamptonSchool of Physics & AstronoyGallo et al. 2004
GX339: Gallo et al. (2004)
SS
Cyg
:
Koerding
et al. 2008, Science
Both
CVs (dwarf novae) and XRBs (X-ray transients) exhibit
outbursts
Thermal/viscous disk instability
XRBs
Outbursts trace a q-shape in the X-ray hardness
vs
intensity plane
(Fender,
Belloni
& Gallo 2004)
h
ysteresis
Collimated (radio) jets are seen (almost only) in the hard state
Hard-soft transition produces a powerful jet ejection episode
CVs (pre-2008)
No evidence for collimated jets in any CV
Constraint on theories of jet formation (e.g.
Livio
1999)?
No constraints on outburst hysteresis
Elmar
Koerding
et al. (2008)
Do dwarf novae also execute a q-shaped outburst pattern?
Yes they do!
Best chance to see a powerful jet is during the “hard-to-soft” transition during the rise to a dwarf nova outburst
D
iscovery of the first CV radio jet!Slide17
Both XRBs and CVs often exhibit (quasi-)periodic oscillations on short (~dynamical) time-scales
Origin is poorly understood, but intimately connected to accretion/outflow processes in the innermost disk regions
Key result in XRBs (accreting NSs and BHs):strong correlations between different types of oscillations, especially LKO and HBO
CVs also exhibit two types of oscillations
Is there a direct connection to
LMXBs?s
Yes!
(Warner & Woudt [2002...2010], Mauche [2003])
DNOs : QPOs in CVs ↔ LKOs : HBOs in LMXBsUniversality of accretion physics extends to periodic variabilityModels relying on ultra-strong gravity or B-fields are ruled outUniversity of SouthamptonSchool of Physics & AstronoyBreakthrough
VIII: Accretion
Periodic Variability: Oscillations
Psaltis, Belloni & van der Klis 1999
Warner &
Woudt
2004
NS & BH
LMXBs
26 CVs
DNOs in VW
Hyi
Woudt
& Warner (2002)Slide18
An XRB (
Churazov
et al. 2003)
A CV (Pretorius & Knigge 2007)
University of Southampton
School of Physics & Astronoy
Breakthrough
IX:
Accretion
Non-Periodic Variability: The RMS-Flux Relation
Black Hole XRB (
Uttley & McHardy 2001)Neutron Star XRB (Uttley & McHardy 2001)
AGN (Vaughan et al. 2011)
NGC 4051
(
Seyfert
1)
What about
non-periodic
accretion-induced variability (“flickering”)?
Stochastic variability has been closely studied in XRBs
Key discovery: the
“
rms
-flux relation
”
(
Uttley
&
McHardy
2001)
Rules out “additive” models
(e.g. shot-noise)
What about CVs?
Non-trivial to study: variability time-scales are much longer, so need high-cadence, uninterrupted long-term light curves
-->
Kepler
!
CVs also show the
rms
-flux relation!
(
Scaringi
et al. 2011)
Accretion-induced variability is universal!
Key properties shared by supermassive BHs, stellar-mass BHs, NSs and WDs
MV
Lyr
(
Scaringi
et al. 2011)
MV
Lyr
(
Scaringi
et al. 2011)Slide19
Shara
et al. 2007, Nature 446, 159
We all “know” that CVs burn accreted matter explosively
(Fujimoto,
Iben
,
Starrfield
,
Shaviv
, Shara, Townsley, Bildsten, Yaron...)→ Nova Eruptions (typical recurrence time ~10,000 yrs)But all known novae were actually discovered as suchHow can we establish the general link empirically ?
Ejected nova
shells may be detectable for ~1000 yrs!So Shara et al. (2007) searched for resolved nebulae around ordinary CVs in the GALEX imaging archive.......and
disovered
an ancient nova shell around the proto-typical dwarf nova Z Cam
→ o
rdinary CVs do undergo nova eruptions!
Postscript: Chinese astronomers would have disagreed with the classification of Z Cam as an “ordinary CV”...
University of Southampton
School of Physics & Astronoy
r
Breakthrough X: Evolution /
Accretion
/
Outflows
Do all CVs go nova?Slide20
University of SouthamptonSchool of Physics & Astronoy
Summary
The last decade has seen several breakthroughs in our understanding of CVs, many of which were made possible by ultraviolet observations
Evolution
The basic disrupted-angular-momentum-loss picture of CV evolution is correct !
We know how to reconstruct CV evolution from both primary and secondary properties
CVs do exist in significant numbers in GCs
CVs
not discovered as novae can still have nova shells -->
all CVs experience nova eruptionsAccretion, Outflows and Links to Other SystemsCV outbursts exhibit hysteresis (“turtlehead” diagram) – just like XRBs and AGN
CVs can drive radio jets – just like XRBs and AGN
Accretion-induced oscillations in CVs are… – just like those in XRBsStochastic variability in CVs follows an rms-flux relation –
just like XRBs and AGN
The physics of disk accretion is universal
CVs provide excellent, nearby, bright accretion laboratories