Cataclysmic Variables
Mark Cropper
Mullard Space Science Laboratory
University College London
1
Course Structure
1. Brief historical introduction and context
2. Types and classes.
3. Non-magnetic systems
4. Magnetic systems: Polars and Intermediate Polars
5. Short and ultra-short period systems
6. Formation, evolution
7. The role of CVs within the larger scheme of astrophysical
studies.
2
Cataclysmic Variables: what are they?
3
CVs: what are they?
• Cataclysmic Variables are
– semi-detached binaries accreting
– from a red dwarf main-sequence-like secondary star
– to a more massive white dwarf primary star
• Roche potential: the gravitational potential around two
orbiting point masses – resultant force on a test mass:
Ftot  FM 1 (grav)  FM 2 (grav)  FCoM (centrifugal)
Centre of Mass

2
credit: csep10.phys.utk.edu
4
Roche Lobe Overflow
•
Semi-detached  secondary star fills its Roche lobe so that it is
distorted into a pear shape.
2
credit:
csep10.phys.utk.edu
•
At Lagrangian 1 (L1) point, gravitational and centrifugal forces cancel
and material is lost from the secondary star into the primary Roche
lobe.
•
Material falls towards the white dwarf in a stream
•
The 4 other stationary points
L2 – L5 are important for
orbit theory
credit:
www.genesismission.org
5
CVs: main types
non-magnetic
Intermediate
Polar
Polar
6
CVs: Role of the Magnetic Field (1)
• magnetic field on primary <106 G (100T)  non-magnetic CV
• accretion takes place through a disk
• via boundary layer
on white dwarf
7
CVs: Role of the Magnetic Field (2)
8
•
magnetic field > 107 G (1000 T)  polar/AM Her system
•
NO DISK: accretion takes place via a stream and accretion column
directly onto white dwarf
•
the magnetic field controls the flow from
some threading region
CVs: Role of the Magnetic Field (3)
9
•
magnetic field ~106 G  intermediate polar/DQ Her system
•
accretion takes place through a hollowed-out disk and then via
accretion columns
onto the white dwarf
•
magnetic field controls the flow in the final stages
CVs: Some background
by
Hevelius
• First European observation of a CV,
Nova Vulpecula, were made in
1670 by a monk Pére Dom
Anthelme (2nd magnitude)
• Another nova, Nova Oph 1848
discovered by John Russell Hind
• The first Dwarf Nova, U Gem, also
discovered by J. R. Hind in 1855:
noted that it was blue – most
unusual for variable stars
(mv~13.5); substantial body of
observations accrued and early
harmonic analyses (Whittaker 1911)
Shara, Moffat & Webbink (1985)
Shara, Moffat & Webbink (1985)
• Next Dwarf Nova to be discovered
was SS Cyg, in 1896, by Louisa
Wells (Harvard College Plates)
10
credit: Warner (1995)
Historical light
curves: SS Cyg
• SS Cyg observed almost
continuously since 1896
(AAVSO)
• Brightness history
shows a variation on a
timescale of ~50 days,
• Unequal length maxima,
no strict periodicity but
remarkably regular
timescale
• varying between
mv~8 and mv~12
(a factor of 40)
• Less than the amplitude
seen in novae, hence
known as Dwarf Novae
Warner (1995) from J. Cannizzo
11
Light Curves: Novae
• Typical amplitudes of Nova outbursts are larger, perhaps 10–
15 magnitudes (factor 104–106) and recur on at least very
long timescales  different mechanism is operating
• This is now known to be thermonuclear burning of the
accreted material on the white dwarf
Hellier (2001) from AAVSO data
12
Light Curves: Z-Cam stars
Warner (1995) from J. Mattei/AAVSO
• Some CVs show
variability which is
different from that of
other Dwarf Novae:
sometimes the
brightness remains at
a constant level before
outbursts resume
• Mean level during
outburst phases is
similar to that during
constant phases
• Called Z Cam stars
after prototype
13
CV Optical Spectra
14
•
Spectroscopy of CVs started in
1860s (Huggins), mostly Novae
•
First Dwarf Novae U Gem and SS
Cyg observed in 1891 and 1897.
•
Spectral characteristics vary
during outburst
•
Characterised by strong broad
Balmer lines in absorption or
emission, with Helium I and II
•
Other stars with similar spectra,
but no outbursts known as Novalike variables.
•
It was recognised that these
might be explained by stars that
were stuck in outburst, as in the
non-outbursting episodes of the Z
Cam stars
Hessman et al (1984)
The CV Zoo: subtypes
• Cataclysmic Variables (non-magnetic)
– Novae
large eruptions 6–9 magnitudes
– Recurrent Novae
previous novae seen to repeat
– Dwarf Novae
regular outbursts 2–5 magnitudes
› SU UMa stars
occasional Superoutbursts
› Z Cam stars
show protracted standstills
› U Gem stars
all other DN
– Nova-like variables
› VY Scl stars
show occasional drops in brightness
› UX UMa stars
all other non-eruptive variables
• Intermediate Polars/DQ Her stars
• Polars/AM Her stars
15
magnetic systems
CVs: How do we know they are binaries?
16
•
Joy (1940) found that RU Peg (Dwarf Nova) had a G3 absorption spectrum,
as well as an emission spectrum, suggesting it was double
•
Joy (1956) found that SS Cyg had composite spectrum,
and that radial velocity variations occurred on a timescale
of 6h 38m again suggesting it was double
•
Walker (1954) found DQ Her to be an eclipsing binary
•
Kraft (1962) suggested all CVs might be binaries
•
Crawford & Kraft (1956) found that the secondary of AE Aqr occupied its
“zero velocity surface” (Roche Lobe) so that some gas might be lost at the
L1 point
•
Kuiper (1941) had suggested for other binaries that turbulent gas would
have angular momentum and swirl around the primary
•
Greenstein & Kraft (1959) found line profile changes through eclipse of DQ
Her which result from the eclipse first of one side of the disk then the other
 confirmed this was the eclipse of a prograde rotating disk
•
Kraft (1961) found a spectroscopic variation on the orbital period resulting
from impact region of the stream on the disk
•
Large survey by Kraft (1962) on Palomar 200” telescope found orbital
motion in almost all CVs, indicating they are close binaries with a white
dwarf primary and a low mass main sequence secondary
Eclipsing CVs: light curves
• If the system is seen edge-on then the secondary star will
cover the disk and primary star, causing an eclipse
• This occurs on the orbital period
Patterson et al (2000)
IY UMa
• Also evident is the bright hump resulting from the accretion
stream impact region
17
Eclipsing CVs: light curves
• Successive orbital phases cover different parts of the disk and
primary star obscuring it from view from the observer
Horne (1985)
18
Eclipsing CVs: light curves
• This permits the different components of the system to be
deconvolved
bright spot
disk
white dwarf primary
Wood et al (1986)
Z Cha
19
Disk changes during outburst
• At quiescence the
contributions from the
white dwarf and accretion
spot are clearly evident
• As outburst starts the disk
component becomes more
important
• At maximum the light
from the system is mostly
from the disk, as evident
in the strong U shaped
eclipse
20
OY Car
Warner (1995)
from Vogt (1983)
Disk brightness
profiles from
eclipses
• From the eclipse shadows
the brightness of each
part of the disk can be
determined
• Behaviour can be followed
through an outburst
• Changes can be seen in
the structure of the disk
• In bright states the disk
temperature is T  r –3/4
• In faint states the disk
temperature has a flatter
relation and the effect of
the hot spot from the
stream impact is evident
21
EX Dra
Baptista & Catalàn (2000)
Disk Models
• In the simplest sense, disks can be modelled as a sum of
annular blackbodies, each with the appropriate weighting for
its area: this gives T  r –3/4 dependence
• Temperature set by local dissipation, for example through
Shakura & Sunyaev  parameter for thin disks
• More detailed models can be sums of atmospheres etc.
La Dous (1989)
22
CV Disks through outburst
Baptista & Catalàn (2000)
• By following the X-ray and
optical brightness of the disk
during an outburst (needs
coordinated observations) it
can be seen that the X-ray
brightness lags the optical
brightness
• This is because the outburst
starts in the middle/outer
parts of the disk, then
propagates inwards
• This is the result of an
instability, or because of
increased mass accretion
rates from the secondary
(Bath/Osaki debate)
Mason et al (1978)
23
Disks: X-rays in
outburst
Wheatley, Mauche & Mattei (2003)
optical
• During outburst, outer
regions of disk brighten
first giving optical
emission
UV
• When region of high
dissipation reaches inner
parts of disk X-ray
emission starts to
increase
• As accretion rate
increases, inner region
becomes optically thick
and able to radiate very
efficiently: temperature
drops so that emission is
mostly in the Ultra-Violet
24
X-ray
Disks in quiescence:
the boundary layer
• X-rays are emitted in region at
the inner edge of the disk, called
the boundary layer
optical
• Can be explored through eclipse
studies in the X-rays and
UV/optical
UV
• Non-magnetic CVs are relatively
faint in X-rays so only recently
have observations achieved
sufficient count rates to resolve
the emission region
• Eclipses are sharp, indicating
that the emission is from a
region only the width of the
white dwarf, and not extended
into the disk
X-ray
• Some indication that emission is
from the polar regions: weak
magnetic fields or obscuration
25
Ramsay et al (2001)
Non-magnetic CVs: X-ray Spectra
• Only recently has it
been possible to obtain
X-ray spectra of high
resolution of nonmagnetic systems in
quiescence
• Spectra show strong
emission lines
characteristic of
optically thin emission
from collisionally
ionised plasmas
• Boundary layer model
is consistent with this
• Rotational broadening
is appropriate for
white dwarf spin but
not the inner Keplarian
orbit of the disk
26
Pandel, Còrdova & Howell (2003)
VW Hyi
Horne & Marsh (1986)
Lines from disks
• Each part of the disk has a particular
velocity and brightness
• Line profiles can be constructed
from adding contributions from each
part
• Conversely, the brightness from
each velocity zone can be
determined from the profile:
Doppler Tomography
• If there is a relationship between the
velocity and the location in the disk,
(such as Keplerian motion, or freefall within the Roche lobe, then can
map further from
velocity to
spatial coordinates
27
Doppler Tomograms of Disks
• Using phase-resolved spectra (corresponding to different
views of the system) the emission from the different emission
regions can be mapped (similar to medical tomograms)
• The location of the impact region can also be seen in Doppler
tomograms
• Since highest velocities in disks are at the centre and lower
velocities outwards, the maps need to be “inverted” in the
transformation to spatial coordinates for the disk component
Spruit & Rutten (1998)
28
Superhumps
• Some non-magnetic CVs, the SU UMa class of Dwarf Novae
have occasional large amplitude outbursts, followed by more
normal outbursts
• During these large outbursts, a hump appears in the light
curve: called superhumps
• These humps evolve with time as outbursts decay.
• Superhump period generally slightly longer than orbital period
V1159 Ori
adapted fom Patterson et al (1995) by Hellier (2001)
29
Superhumps
•
30
Superhumps are thought to be
caused by tidal interactions in the
outer disk
•
Disk is larger during outburst,
reaching towards the edge of the
Roche Lobe
•
There fluid elements in the disk
experience the attraction of the
secondary star
•
Individual orbits are perturbed
•
Global disk precession set up
Hellier (2001)
Spiral Waves
• Emission line
structures can
sometimes show
strange structures at
outburst
Harlaftis, et al (2000)
IP Peg
• Doppler Tomograms
indicate arc-shaped
patterns, indicative
of spiral structures
• Can be reproduced
using spiral
structures in model
data
• Thought to be a
spiral shock induced
by the secondary as
a result of noncircular orbits in disk
Steeghs & Stehle (1999)
31
Polars
32
Magnetic systems: history
• AM Her is a V~13.5 variable star which Berg & Duthie (1977)
suggested is the optical counterpart of a source in the Uhuru
catalog (also a SAS-3 source) with a period of ~3.1 hr
• Lightcurve was unlike that of other CVs: too variable for Dwarf
Novae and Nova-like CVs, no eclipses
• In 1976,
polarimetry
obtained by
Tapia, showed
large circular
polarisation
variations on
orbital period.
• Largest circular
polarisation
seen in any
celestial object:
hence ‘Polar’
33
AM Her
Tapia (1977)
Polars: Evidence for magnetically confined
emission
• To achieve large levels of circular polarisation, radiation
process must be largely cyclotron emission
• For emission to be in the optical, require high values of
magnetic field strength, ~30 MGauss (~3000 Tesla)
• Expect this to disrupt the disk
• Eclipse lightcurves
show very sharp
(seconds) drops and
rises in brightness,
consistent with a
small spot on the
white dwarf
• Evidence for
magnetically
controlled accretion
regions
34
HU Aqr
Bridge et al (2002)
Polars: magnetically controlled accretion
• Two accretion
regions evident in
some systems
(UZ For)
• No evidence of
white dwarf
• Weak accretion
stream
Perryman et al (2001)
UZ For
35
Polars: magnetically controlled accretion
• During eclipse, different parts of the system are successively
eclipsed and uncovered
• In some systems, accretion stream between stars can be very
bright
accretion spots
+ accretion stream
+ secondary
accretion stream
+ secondary
secondary only
HU Aqr
36
Bridge et al (2002)
Polars
• Magnetic field is too strong
for a disk to form
• material falls directly from
secondary to primary
• At some point
material in stream
threads onto
magnetic field
37
• Subsequent
accretion is quasiradial onto white
dwarf
Rothschild et al (1981)
Polars: X-ray emission
• Polars/AM Her stars were
found to be strong soft X-ray
emitters (~1033 erg/s) in
early surveys
• X-ray emission characterised
by thermalised free-fall
velocities from a white dwarf
so emission was from a hot
region close to the white
dwarf surface: post-shock
• Cyclotron emission
must also be from a
hot region (otherwise
narrow cyclotron
emission lines rather
than continuum)
38
AM Her
Polars: Spectral Energy Distribution
• Most of the energy from these systems is a result of accretion
• 3 main components:
cyclotron
radiation from
accretion column
39
soft X-ray emission,
from heated surface
of primary
Beuermann (1998)
hard X-ray emission,
also from accretion
column
Polars: Radial Accretion
• Infalling material is forced
to follow the magnetic field
lines
• Gas is initially in free-fall
but then it encounters a
shock front
• Shock converts kinetic
energy into thermal energy
(bulk motion into random
motion)  temperature
increases to ~50 keV
• Velocity drops by 1/4 and
density increases by 4
• Material radiates by
cyclotron and
bremstrahlung and
gradually settles on white
dwarf
40
cold
supersonic
flow
optical/IR
cyclotron
radiation
shock
hot
postshock
flow
white dwarf
hard X-rays
soft X-rays/
extreme UV
Polars: accretion
flow hydrodynamics
• Equations of
density
velocity
pressure
– mass continuity
– conservation of momentum
height
– conservation of energy
• 1-dimensional accretion
• An analytical solution
can be found to
generate solutions
in a step-wise
scheme
41
cooling term
bremsstrahlung
cooling
cyclotron
cooling
Polars: accretion
region hydrodynamics
• Solutions to equations produce
run of hydrodynamic variables
(Temperature, Pressure etc) from
which emissivity as a function of
height can be calculated…
42
Cropper, Wu & Ramsay (2000)
Polars: Emission from
post-shock flow
• Given the run of temperature and
density, and assuming collisional
ionisation (for this density regime)
it is possible to show that the
emission region is optically thin to
X-rays
• Also possible to calculate the
ionisation fraction of any ion
species, and therefore the
emissivity, as a function of height in
the post-shock flow as well as
parameters such as
– mass of the white dwarf
– accretion rate
• Predictions can be matched to
spectra and continuum emission to
derive fundamental parameters –
X-ray calorimetry
43
Wu, Cropper &
Ramsay (2001)
Polars: Energy Balance
• The form of the spectral energy distribution, and particularly the
relationship between the direct component of emission (X-ray
thermal bremsstrahlung + cyclotron) and the reprocessed
component been the topic of much debate.
• Until recently measurements have indicated that the soft X-rays are
much stronger than the direct emission, in contradiction to the basic
model for magnetically confined accretion
• One possibility has been that the accretion flow is not smooth,
known as “blobby” accretion:
– gives rise to flares in light curve (c.f. VV Pup)
– blobs bury themselves deep in the white dwarf, so no visible
post-shock flow for direct emission  soft X-ray excess
AM Her
Frank, King &
Raine (1992)
Beuermann (1998)
44
King (1995)
Polars: energy balance revisited
• New measurements have been made of ~40 polars (60% of
known systems) to examine the energy balance using
XMM-Newton which covers both hard and soft X-ray bands
and also the optical/UV
• Generally coverage of most of orbital; good fits achieved using
stratified accretion column model
GG Leo
Ramsay et al (2003)
45
EU UMa
Polars: Energy Balance
• Recent results find that most Polars have a direct
emission/reprocessing balance (hard X-ray+cyclotron/soft X-rays)
which is consistent with the standard view of magnetic accretion
• A minority of systems have a soft excess
• Still not clear what causes this
(magnetic field?)
• Difference to past studies
ascribed to
– better instrumentation
– better wavelength coverage
(simultaneous)
– better calibration
– better models
46
Ramsay & Cropper (2003)
Polars:
Spectral Characteristics
• Optical spectra are
dominated by strong
emission lines of
Hydrogen and Helium
• General slope influenced
by Balmer and Paschen
continua
• UV spectra show strong
CIV, SiIV and NV lines
(plus others)
• Strong lines indicative of
substantial emission from
the accretion stream
• In high state generally no
signature of the
secondary or primary
47
Tanzi et al (1980)
Schacter et al (1991)
Feigelson, Dexter & Liller (1978)
Polars: Long-term
light curves
• Unlike disk-dominated CVs,
polars tend to emit at an
approximately constant level,
with occasional drops to fainter
levels
• There is a continuum of levels,
but the states tend to be called
“high”, “intermediate” and
“low”
• In low states the accretion rate
drops, no reservoir in a disk, so
the underlying stars can
become visible: important for
measuring the system
parameters
48
Polars: Shorter timescale variability
• Light curves are strongly variable on orbital timescales
• If accretion is occurring mainly near one magnetic pole, then,
depending on the latitude of the accretion region, this can come
into view and disappear over the limb as the synchronised binary
rotates: “bright” and “faint” phases
VV Pup
Cropper &
Warner (1986)
• Many systems show strong variability due to flaring on timescales
of tens of seconds
• Some systems show quasiperiodic oscillations on a
timescale of seconds: this is
due to an intrinsic instability
in the cooling gas accreting
onto the white dwarf
49
Larsson (1992)
V834 Cen
Polars: Synchronisation
•
All of the variability in Polars occurs at a single period: the orbital
period
– radial velocity curves of the secondary
– X-ray light curves from the primary
– polarisation variations
 the white dwarf/red dwarf are locked into the same orientation:
synchronised rotation
50
•
The mechanism for synchronisation is the dissipation due to the
magnetic field of the primary being dragged through the secondary
•
As relative spin rate of primary decreases, locking can occur due to
the dipole-dipole magnetostatic interaction between primary and
(weaker) secondary magnetic field
•
Some Polars not quite in synchronism; in these systems it typically
takes 5–50 days for white dwarf orientation to repeat itself
•
Very useful systems to study the effect of orientation of magnetic field
on the accretion process
Polars: Asynchronous Systems
• Changes from night to night on EUVE2115 (7-day slip period)
1
X-ray
2
3
4
5
6
7
51
Cropper, Ramsay & Marsh (2003)
UV
Polars: optical spectra orbital variation
orbital phase
Hb
velocity
velocity
5
orbital phase
orbital phase
52
velocity
3
orbital phase
1
• Doppler
Tomograms
showing
resultant maps
from changes in
line emission
night 5
orbital phase
• Time sequence
of spectra on
different nights
folded on orbital
period for
EUVE2115
night 3
orbital phase
night 1
HeII
4686
Cropper, Ramsay & Marsh (in prep)
velocity
velocity
velocity
Polars: measuring the magnetic field
• The magnetic field can be measured in two main ways
– cyclotron harmonics: measures the magnetic field in the
cyclotron emitting region
– Zeeman splitting: measures the field over the whole visible
white dwarf (flux weighted)
Thomas et al (2000)
3
RX1313–32
4
2
53
Polars: measuring the magnetic field
• Zeeman split lines can sometimes be seen in low-state spectra
(see also the signature of the secondary star)
AM Her
AM Her
Schmidt, Stockman & Margon (1981)
54
Latham, Liebert & Steiner (1981)
Polars: measuring the magnetic field
•
Sometimes both techniques can be used on the same star
MR Ser
Schwope et al (1993)
55
•
And often two separate field strengths can be determined
 field strengths at the two poles are different;
 decentred dipole, or more complex field
(superposition of multipoles)
•
Generally accretion takes place preferentially near lower-field pole
Polars: Polarised Optical Emission
• Cyclotron emission is elliptically polarised
– linearly polarised viewed perpendicular to field
– circularly polarised viewed along the field
– position angle traces magnetic field line projected on sky
• Diagnostic power very strong
Harrop-Allin, Potter & Cropper (2001)
ST LMi
56
Cropper (1996)
HU Aqr
Polars: modelling the
cyclotron radiation
• The cyclotron radiation pattern
(intensity and polarisations) has
been calculated by several
groups as a function of viewing
angle, wavelength: various
levels of sophistication
• From this the total polarised
emission from any surface
element can be calculated
• It is possible to iterate the map
of emission points to gradually
converge to the right map for
the observed polarisations
(Stokes Imaging, Potter,
Hakala)
• Produces a map of cyclotron
emission
Potter, Cropper & Hakala (2000)
57
V347 Pav
blue
red
Polars: Eclipse mapping studies
• It is possible to use eclipse mapping to calculate the stream
brightness as a function of position
HU Aqr
• Then can use the brightness to
generate the lightcurve from
the stream only, and isolate
cyclotron emitting region
Harrop-Allin, Potter & Cropper (2001)
HU Aqr
58
Polars: Stream mapping
• More recent work even allows the emission regions not to
follow a prescribed trajectory – allows trajectory to be
determined freely
• Emitting regions are defined and allowed to be located
anywhere in Roche lobe of primary
• Emission “swarm” evolved to produce a good fit
• Fits which follow some linear configuration
are preferred (self-organising maps – SOM)
EP Dra
59
Bridge et al (2003)
Intermediate Polars
60
Intermediate
Polars
•
After Polars were
identified, Charles et al
(1979) found an
X-ray emitting V~13
star, AO Psc, with an
optical spectrum like
that of Polars, but
without any identifiable
polarisation
•
They also showed
variability on three
different timescales
now known to be
– the orbital,
– the spin period of
the white dwarf &
– the mixture of the
two (beat/synodic
period)
61
Cropper et al (2002)
AO Psc
Intermediate Polars: another example
Cropper et al (2002)
X-Ray
62
FO Aqr
Intermediate Polars: Power Spectra
• By performing a Fourier
Transform of the previous
data, the main periodicities
can be identified
– orbital period
Evans et al (2003)
UV
– white dwarf spin
– beat (very faint in this
system)
• Also evident are harmonics
when the variations are
non-sinusoidal (2, 3, 2)
• The variation of X-rays at a
higher frequency suggested
that due to magnetically
controlled accretion – but
with a lower field than
Polars/AM Her systems
63
X-ray
Intermediate Polars: folded light curves
• Can now fold the data on the main periods, to derive the
phase relationship between different wavebands
FO Aqr
UV
UV
FO Aqr
X-ray
X-ray
64
Evans et al (2003)
Intermediate Polars
• Since the magnetic field is not as strong as in Polars, a disk can
form; field hollows out central parts of the disk
• From the inner part of the disk, accretion occurs down field lines
similar to that in Polars – so get radial accretion around both
magnetic poles
• Because field is lower, cyclotron radiation is less strong
 unpolarised (generally)
65
Intermediate Polars: more evidence
• Classic observations by Nather (1978) and
Patterson, Robinson & Nather (1978)
found that DQ Her had a 72 sec oscillation
that went away during eclipse
• Also found that the phase of the oscillation
changed through the centre of the eclipse,
as a result of first one side of the disk
being eclipsed then the other
• Interpreted as resulting from the
reprocessing of an X-ray beam sweeping
the disk like a “light house”
• Only recently seen in X-rays (scattered)
Nather (1978)
66
Patterson, Robinson & Nather (1978)
Intermediate Polars: models
• Intermediate Polars spin variability can be explained in several
ways (much debate on this over the years)
– visibility of the accretion region on the white dwarf
– visibility of the accretion “curtains”
– reprocessing of flux on the disk (optical/UV)
stream
Adapted from Hellier (2001)
• From studies of the relative phasing in different wavelength
bands and including to the absorption effects now known to
be a combination of the above models leading to the complex
behaviour in Intermediate Polar light curves
67
Intermediate Polars: X-ray Spectra
• X-ray spectra of Intermediate Polars generally show just the
multi-temperature thermal bremsstrahlung component from
the hot radial accretion flow – no soft reprocessed component
from the white dwarf
• Main explanation is likely to be the larger area over which
accretion takes place, but also photoelectric absorption is
important
AO Psc
Cropper et al (2002)
68
FO Aqr
Intermediate Polars: radial accretion flow
• Very strong evidence for
radial accretion can be
obtained by using the
same accretion flow
model (Stratified
Accretion Column)
shown earlier for Polars,
and applying it to
Intermediate Polars
• This provides a detailed
fit to high resolution Xray spectra of systems
such as EX Hya
• Single or even add-hoc
3-temperature thermal
bremsstrahlung models
do not provide
sufficiently good fits
69
Cropper et al (2002)
EX Hya
Intermediate Polars: Fast-spinning systems
• Some Intermediate Polars have very short spin periods, for
example DQ Her: 71 sec, AE Aqr: 33 sec
• In this case the magnetic field is thought to be small, so that
the hollowed-out part of the disk is also small and the
Keplarian velocities in the field coupling region large
• AE Aqr is a special case: the white dwarf spin period is
decreasing rapidly, and it can be calculated that almost all of
the luminosity of the system arises as a result of the spindown energy
• Spinning white dwarf kicks
the gas from the stream
out of the system –
propellor
• Much of the optical and
radio emission from the
system arises in colliding
shocks in the ejecta
70
Wynn, King & Horne (1997)
Cataclysmic Variable Evolution
• The white dwarf in CVs is the relic of the more massive star in
the binary, already past a giant phase
• The secondary star will have spent some time in the envelope
of the primary red giant perhaps accreting material from it
• Dynamical friction reduces the separation of the remnant core
of the primary and the secondary, causing the secondary to
spiral inwards and perhaps contributing to the ejection of the
envelope
• After this phase have a detached hot white dwarf with a main
sequence secondary; such stars are known (eg BE UMa) and
show strong reflection effects from the hot primary
illuminating the secondary
• Further angular momentum losses can shrink the binary
separation until the secondary comes into contact with its
Roche lobe, and mass transfer starts, giving rise to a CV
• Alternatively in longer period systems the secondary can
evolve, increasing in size and filling its Roche lobe
71
Evolution (ctd)
• Transfer of material from a lower mass star to a higher mass
star (as in a CV) causes the orbital separation to increase,
lengthening the period and causing the star to fall away from
contact with its Roche Lobe and accretion to stop
• Hence some other mechanism has to been in place for stable
mass transfer
• Main candidate for more than two decades has been the
magnetic braking caused by wind material from the secondary
being threaded by field lines to large distances (Zwaan &
Verbunt)
• This magnetic braking robs the binary of angular momentum,
so the two stars move closer together, maintaining contact
with the Roche Lobe and quasi-stable mass transfer
• From this an expected population of CVs can be generated as
a function of orbital period, given ages and mass transfer
rates
72
CV Evolution (ctd)
• Any prediction from population models needs to confront the
statistics of the observed period distribution
Ritter & Kolb (1998)
• From 10 down to ~3 hrs the distribution is approximately
correct, but then there is a lack of systems in the 2-3hr range,
known as the “Period Gap”
• More shorter period systems
73
CV Evolution (ctd)
• Zwaan & Verbunt suggested that at 3hr the magnetic braking
switches off, because secondaries with this mass are fully
convective, so stellar dynamo is quenched
• Mass transfer rate then is reduced, so secondary falls away
from Roche Lobe, and mass transfer stops, so these systems
are not seen as CVs
• Then gravitational radiation would continue to operate, slowly
driving the two stars together, until contact with the Roche
Lobe was reestablished and accretion restarted: the origin of
the <2hr binaries
• While this has been the “standard” explanation for some time,
questions have always remained about the real effect of fully
convective secondaries on the generation of magnetic field by
the dynamo
• Some evidence for this mechanism clear from the absence of
a significant gap in the magnetic systems, in which braking is
affected by the field from the primary (in addition to that from
the secondary)
74
Spin-orbit coupling in magnetic systems
• The two stars in Polar systems rotate synchronously with the
orbit by the magnetic interactions, except in the case of 4
systems which deviate slightly from synchronism
• Intermediate Polar systems are not synchronised
• Many IPs are
found to have
Pspin~0.1Porb
due to
field/stream
interactions
• As the binary
evolves and the
period decreases
it is likely that
some IPs evolve
into Polars
75
Norton, Somerscales & Wynn (2003)
CVs: Short Period Systems
• CVs with main sequence secondaries have minimum orbital period of
80 minutes
• However, some accreting binaries are seen to have shorter periods

the secondaries must be white dwarfs, or perhaps the
degenerate core of a giant which has been stripped away
• The optical spectra of these systems show only Helium (no Hydrogen)
• The disk behaviour is
different because of the
different ionisation of He
compared to He;
interesting astrophysical
laboratories
• Orbital periods from ~65
min to ~15 min
• Known as the AM CVn
systems
Marsh, Horne & Rosen (1991)
76
GP Com
Ultra-short period Binaries
• 3 systems have recently been discovered with binary periods
<10 min (600 sec)! Small enough to fit within Jupiter
RXJ1914.4+2457
(V407 Vul)
9.5 min Cropper et al (1998)
RXJ0806.3+1527
5.4 min Ramsay, Hakala & Cropper (2002)
KUV01584–0939
(ES Cet)
77
Israel et al (2002)
10.3 min Warner & Woudt (2002)
Observational
Characteristics
Cropper et al (1998)
• X-ray light curve
shows sharp rise
to maximum and
slower decline
• X-rays essentially
“off” for ~50% of
the cycle  small
emission region
• RXJ1914.4+2457
and
RXJ0806.3+1527
X-ray light curves
are almost
identical
78
RXJ1914.4+2457
Burwitz & Reinsch (2000)
RXJ0806.2+1527
Observational Characteristics
Only one period in power spectra; if this is the orbital period
then the secondary has to be degenerate (a white dwarf)
RXJ1914.4+2457
X-ray
optical
window
window
window
prewhitened
79
prewhitened
ROSAT HRI data
NOT optical data
Cropper et al (1998)
Ramsay et al (2000)
prewhitened
Observational Characteristics (3)
3) Optical/Infrared Colours
RXJ1914.4+2457
Ramsay et al (2000)
• No evidence for a late main-sequence secondary star
• Spectral flux distribution in optical is that of a hot (~50000 Kelvin)
black body
80
Observational Characteristics
• Source is not constant in X-ray or optical/IR
 most likely some sort of accretion or similar mechanism
Ramsay et al (2000)
Long-term light-curve: RXJ1914.4+2457
81
Observational
Characteristics (6)
• Phasing of X-ray and optical/IR light
curves argues against a single star
• simplest explanation is a reprocessing
source in a binary system where the
X-rays from the primary are
reprocessed from the secondary
reprocessed
X-ray
RXJ1914.4+2457
Ramsay et al (2002)
82
Observational Characteristics (8)
• Spectra of RXJ0806.3+1527 show no H emission – all
emission is He, C and N and weak
RXJ0806.2+1527
83
Israel et al (2002)
Models
• Four models exist for these systems
a) Intermediate Polar
Motch et al (1996)
Norton et al (2002)
b) Double Degenerate Polar
Cropper et al (1998)
c) Double Degenerate Algol
Marsh & Steeghs (2002)
Ramsay et al (2002)
d) Electric Star
Wu et al (2002)
• Of these models, three involve magnetic fields
• Neutron star models are ruled out because of the
absence of a hard tail of X-ray emission; also they
need to be extremely close
(too close for the observed level of reddening)
84
Model 4: Electric Star
•
Model introduced by Wu et al (2002) to deal with absence
of polarisation and emission lines in RXJ1914.4+2457
•
Scaled-up version of Jupiter-moon interactions heating
Jupiter’s atmosphere at footpoints of magnetic field lines
threading moon, BUT:
a) energy release boosted by stronger magnetic field
b) shorter period
c) secondary larger than a moon
•
Up to 1036 ergs/s available (depending on degree of
asynchronism)
•
Star shines by electrical
resistive heating:
new mechanism after
1. nuclear fusion
2. accretion
3. magnetic (magnetars)
HST image (NASA)
85
Model 4: Electric Star (ctd)
• Conductor (secondary) moving in magnetic field of primary has
EMF induced across it (as in any generator)
• EMF drives a current in 2 circuits if low density plasma in system
• Emission takes place
from two small
hotspots at the base of
the field lines on
primary
Wu et al (2002)
86
Model Comparison
IP
DD
Polar
DD Algol
Electric Star
(ES)
Absence of Hard X-rays
x
√
√
√
Only one modulation period
x
√
√
√
Optical-IR colours
x
√
√
√
Phasing of X-ray and Optical
~
√
~
√
Shape of X-ray modulation
√
√
√
√
Absence of Polarisation
√
~
√
√
Absence of H
x
√
√
√
Absence of strong emission
lines
x
~
~
√
Period decreasing
~
x
x
√
Long-term variability
√
√
√
√
• On current knowledge, only strongly ruled out model is
the IP, mainly because of lack of photometric evidence for
secondary
87
Further thoughts: ES Model
• ES model suffers from being ‘unconventional’ but nevertheless:
a) provides the best interpretation of the data
b) evidently works in astrophysical systems (Jupiter)
• Predicts the right shape of X-ray light curve (retrograde spin on
primary WD)
• Lifetime of the system is ~106 years, short but not particularly
short – NOTE: the 103 years quoted by Marsh & Steeghs (2002)
is a misinterpretation of the text in Wu et al (2002)
• It is an expected end point for double WD pairs if one has a
modest (<1MG) field; secondary is not necessarily in contact
with Roche Lobe – expect many of these systems?
• Predicts decreasing orbital period, as observed for both
systems
88
Gravitational Radiation
•
Ultra-short period binaries are expected to be strong
sources of gravitational radiation. These will be the
calibrators for LISA
RXJ0806.3+1527
RXJ1914.4+2457
KUV01584–0939
89
Figure from NASA/JPL; strain from Nelemans et al (2001)
CVs in the grander scheme of things
• Cataclysmic variables are fairly common systems, the later
stages of much of binary evolution
• They produce the low-level background of discrete sources in
galactic X-ray emission – fainter but much more numerous than
neutron-star or Black hole X-ray binaries
• They are highly important laboratories for studies of accretion
– disk behaviour
› instabilities,
› stream impacts,
› warps,
› tidal resonances,
› spiral waves etc.
– magnetically dominated accretion
› accretion columns,
› emission from post-shock flow,
› shocks, instabilities etc.
• Multi-wavelength emission (polarised in many cases) allows a
multi-wavelength approach, providing very strong observational
constraints on the interpretation of data
90
CVs in the grander scheme of things (ctd)
91
•
Important for investigations on how material interacts with a magnetic
field:
– threading region in Polars,
– inner region of disk in Intermediate Polars,
– Dwarf Nova oscillations in non-magnetic CVs
•
In general, the balance of:
– visibility of underlying system (to provide the context) &
– the emission (X-ray, optical) has been fundamental to making
enormous progress in understanding a wide range of astrophysics
•
It is a field which incorporates fluid dynamics, MHD, a full range of
emission processes, stellar evolution, gravitational radiation etc.
•
A large number of important observational techniques have been
developed in the context of CVs and then used elsewhere:
– Doppler tomography,
– eclipse mapping of disks and streams,
– Stokes imaging,
– timing analyses
•
Many well-known astronomers/astrophysicists have worked in this field,
developing their theoretical understanding and observational &
interpretational skills before carrying these into other fields
CVs: Open Issues
• This has been an extremely active field in the last 3 decades.
Many scientists addressing the issues, large amount of time
devoted on observational facilities, large and small, on ground
and in-orbit
• Much is now understood about these systems, and many of
the fundamental issues in these systems have been addressed
(energy balance in Polars, disk-instabilities in non-magnetics,
accretion curtains in IPs)
• Mature field, which is developing around its boundaries:
ultra-compact systems are an important and exciting
new development
• However: the effect of new facilities, particularly in the X-rays
(XMM-Newton and Chandra) and UV (FUSE, XMM-OM) is only
now beginning to be felt, also greater access to 8m telescopes
on the ground
 we can still expect to learn a great deal
92
CVs: Open Issues
• Many aspects deserve further investigation: here are some
– boundary layer in non-magnetics
– the base of the post-shock accretion flow in magnetics and
the way this diffuses into the white dwarf
– heating of the atmosphere around the accretion region in
magnetics, and effect on overall energy distribution
– low accretion rate regimes in magnetics, whether this
results in a bombardment solution (no shock)
– disk-magnetosphere interaction in IPs: important in a
number of contexts
– disk-stream interactions in non-magnetics
– magnetosphere-stream interactions in Polars
– irradiation of the stream and secondary by X-ray flux
– more astrophysics in the post-shock flow models (such as
the separation of electron and ion fluids)
• Combinations of high quality data (eg. eclipse mapping of
spectra) and new astrophysical fluid computations will transform
the field and allow ever more intricate understandings of
accretion phenomena to be achieved
93
Resources
• Books:
Brian Warner: Cataclysmic Variable Stars
(Cambridge University Press, 1995, ISBN: 0521412315)
Coel Hellier: CVs - How and Why They Vary
(Praxis Publishing, 2001, ISBN: 1852332115)
Frank, King & Raine, Accretion Power in Astrophysics
(Cambridge University Press, 3rd edition)
• North American Workshops on Cataclysmic Variables
• Magnetic Cataclysmic Variable Workshops
Acknowledgements: These lectures drew partly from material in Warner
(1995) and Hellier (2001)
94