Transcript Presentation

Recombination line spectroscopy
- theory and applications
Robert Bastin and Peter Storey
UCL
Mike Barlow (UCL) and Xiaowei Liu (Peking University)
with particular thanks to XWL for use of some figures.
Collisionally-excited lines (CELs) and optical recombination lines (ORLs) in the
spectrum of a planetary nebula
• Nebula formed from
stellar ejecta
• Ionization is maintained
by UV radiation from
central star
• Classical temperature and
density diagnostics give
electron temperatures
near 10000K and electron
densities 100-100000/cm3
• Balmer jump temperatures
tend to be lower than
those derived from CELs
OIII forbidden and infra-red lines
provide a measure of
a) electron temperature due to different
excitation energies (eg 4363 and 5007Å)
b) O2+ abundance by comparison with
hydrogen recombination lines
Other ions provide density diagnostics
The OII Grotrian diagram
• prominent optical recombination
lines shown in green – no significant
collisional excitation.
• until recently all calculations of
the line intensities were carried out
in LS-coupling.
• 4f-3d transitions might be expected
to be well enough represented by
hydrogenic approximations to use
these lines to determine abundances
of O2+ relative to H+.
• Expect good reliability since hydrogen
metal recombination lines have very
similar dependence on temperature.
• The O/H abundance ratio derived
from ORL observations of OII is
systematically larger than that
derived from CEL observations
of O2+ for a range of
photoionized nebulae.
• The most extreme example (so far)
is Hf2-2 where the ORL abundance
is 70x the CEL abundance.
• Similar results are obtained for C/H
and N/H.
• O/H from CELs spans one order of
magnitude – O/H from ORLs two.
Oxygen abundances relative to hydrogen
from ORLs and CELs
By contrast, Mg/H from MgII
shows no such enhancement
And relative abundances of C, N, O
are approximately the same from CELs
And ORLs
(Resembles an inverse FIP effect?)
Images and spectra of Hf2-2
• the Balmer jump is
unusually large and the
recombination lines (eg
CII at 4267Å) are
exceptionally strong
• the Balmer jump indicates a
temperature of 900K while
the CELs give 8820K
A possible explanation for the ORL/CEL abundance discrepancy is that the
ORLs arise from a different physical region within the nebula.
The ORL emitting region would have enhanced CNO abundances compared
to the background material and would therefore be at a much lower temperature
due to increased cooling from forbidden and infra-red fine-structure transitions.
The colder metal-rich material would not be visible at all in CELs.
Hydrogen recombination line emission would come from both regions
The nebula A30 might be a model,
with cold (500K) knots of material
emitting very strong ORLs and
very little CELs and containing
only 1% hydrogen by number.
Unfortunately such knots are not
visible in any of the other nebulae
with high ORL abundance
discrepancies.
Recombination line diagnostics - measuring temperature
If the cold, metal-rich knot model is correct it becomes important to try to
determine the conditions in which the recombination lines are emitted.
ie to find diagnostics of the temperature and density that use only ORLs
and not CELs.
The ratios of the intensities of ORLs from states of different orbital angular
momentum shows some temperature dependence
eg CII 4f-3d (4267Å)/4s-3p(3920Å).
CII 4f-3d (4267Å)/4s-3p(3920Å)
against electron temperature
Emissivity of higher l states is
enhanced at low temperature
and also at higher temperatures
due to high-temperature
dielectronic recombination.
The OII recombination spectrum
The recombination spectrum of
OII holds out the prospect of
measuring electron density as
well as temperature since the relative
populations of the 3PJ levels of
O2+ are sensitive to density in
about the right density range for
nebulae (assuming the CEL
densities are appropriate).
ORL temperatures may be very
low, possibly below 500K.
Since the OII recombination lines
are observed between low-lying
states a complete model of the
atom is needed. The maximum
bound state principal quantum
number is n=62 in the 3P1 series
Relative populations of the
3P states of O2+ as a function
J
of electron density.
Are these changes of population
reflected in the intensities of the
recombination lines between
low-lying states, providing a way
to measure density?
• Below the collision limit,
populations are assumed to
be determined by radiative processes
only.
• The appropriate coupling scheme
varies from near pure LS-coupling for
low l to near j-j coupling for l>3 and
high n for any l.
• In the low-n, low-l region we perform an
R-matrix calculation of all radiative
data (bound-bound and bound-free).
• Note that some two-body magnetic
terms are not implemented in the
R-matrix code.
• States are described by J, parity and
energy.
• For the remainder j-j coupling is assumed
and various approximate methods used.
So for low n and l bound-bound
O+(EJπ) --> O+(E'J' π') + hν
and bound-free and free-bound
O2+(Jπ) + e <--> O+(EJ'π') + hv
while for the remainder angular core
momentum, Jc, is conserved
O+(3PJcnlj;EJπ) -> O+(3PJcn'l'j';E'J' π’) + hv
and
O2+(3PJc) + e <--> O+(3PJcnlj;EJπ) +hv
• Above the collision limit, l-changing and energy changing collisions must
also be included.
O+(3PJcnlj;EJπ) + e --> O+(3PJcn'l'j';E'J'π') + e'
• These states are assumed hydrogenic in terms of energy, radiative and
collisional properties
• Collisions with protons and heavier ions may also be important and
are included.
• Above the O+ ionization limit
dielectronic capture and
autoionization can also occur
O+ (3PJc’) + e <--> O+ (3PJc ) nlj (EJπ)
followed by radiative decay
O+(3PJcnlj;EJπ) -> O+(3PJcn'l'j';E'J' π’) + hv
via the outer electron.
• The rates for both these processes
are relatively small at high n, so
n and l changing collisions also
compete and distribute the
population to states with small
dielectronic capture rates..
• More complex than the usual di-electronic mechanism where there
is a rapid core decay that dominates.
n>63
Autoionization probabilities are computed
in intermediate coupling with Nigel Badnell’s
AUTOSTRUCTURE for n.le.1000 and l.le.40.
For l.le.3, radiation damping is neglected
Results for the OII transition
array
3p (4D) --> 3s (4Po)
• The intensity of each component
relative to the total is shown as
a function of electron density and
at 104 K.
• The strongest component at the
higher densities is
3p (4D7/2) --> 3s (4Po5/2)
whose upper state can only be
formed from the 3P2 state of O2+.
• The intensity variation mirrors
the variation of population of the
3P state
2
• Also shown are observed values
for two HII regions
Densities from CELs are 5500/cm3 for M42 and
310/cm3 for 30Dor.
• The same figure with observed
values from two planetary
nebulae.
• CEL densities are 390/cm3 for
S311 and 4000/cm3 for
NGC5882.
The variaton of the intensity
of the strongest component
of multiplet V1 with electron
density
• the HII regions have densities
down to 100/cm3 and
temperatures close to 104 K
• theoretical results in red with
fine-structure dielectronic
processes included show
improved agreement with
observation.
Line ratios from different
multiplets as a function of
density and temperature
• At “high” temperatures, both
density and temperature can
be determined from line ratios
• At low temperatures, all
density information is lost
• The figure seems to confirm
that for some nebulae the OII
lines are being emitted at
temperatures below 1000K
At low temperatures, the average free
electron energy is comparable to the
fine-structure energy separation.
Rydberg states with a 3P1 or 3P2 core are
populated by dielectronic capture from
the 3P0 + e continuum, even at low
density when the populations of the
O2+ 3P1 and 3P2 are negligible.
Relative intensities are almost
independent of density
Conclusions
Atomic:
• Despite the complexity of the atomic model and the approximate
treatments used, the observed spectral features can be reproduced.
• Fine-structure dielectronic processes are important in modelling ionic
recombination spectra, particularly at low temperatures (comparable to
the fine-structure energy separations)
• The absence of some two-body terms from the current R-matrix codes
does not lead to serious errors in this case
Astrophysical:
• Various workers have suggested that the line ratios that we have interpreted
as arising from very low temperature material are actually caused by
variations in the density of the material. These results show that low
temperatures are obtained irrespective of the density.
• The physical cause of these observed phenomena is still uncertain.
The continuation of this work forms part of the application to extend
the funding of the UK members of the Iron Project, along with continued
work on ions of importance in solar physics