Transcript Document

Introductory X-ray Group Meeting
Spectroscopy and X-rays
23 January 2001
Members of the group:
David Cohen
Prue Schran
Allison Adelman
David Conners
Eric Levy
Geneviève de Messières
Kate Penrose
Project Summaries
Allison – ROSAT X-ray survey of O stars
David – statistics of velocity, density, temperature in hot star
wind computer simulations
Eric – line width analysis of Chandra spectrum of t Sco
Geneviève – line ratio analysis of Chandra spectrum of t Sco
Kate – absorption line analysis of fusion plasmas from OMEGA
laser experiments
contact information
Allison: [email protected]
David: [email protected]
Eric: [email protected]
Geneviève: [email protected]
Kate: [email protected]
Our individual projects are part of a bigger pursuit
understanding the interactions of very strong electro-magnetic
radiation fields (i.e. light) with matter. Contexts are (1) hot star
winds, (2) laser-produced fusion plasmas.
On the stellar wind side (everyone except Kate…for now) the big
specific question we’re trying to answer is:
How do hot stars make their X-rays?
wind shocks vs. a magnetic/coronal mechanism
High intensity light invariably means that things get hot…
Hot things (above about half-a-million degrees) means X-rays
(remember, there’s a relationship between color/wavelength and temperature)
Spectroscopy provides information, through its dependence on
atomic states and the Doppler shift, about the microphysics, as
well as bulk properties, of gases.
A spectrum carries a huge amount of information:
“If a picture is worth a thousand words, then a spectrum is worth
a thousand pictures.” -- Blair Savage
The spectral resolution of x-ray telescopes has improved
many hundred-fold over the past decade, enabling us to
detect and resolve emission lines in hot stars like t Sco.
Spectral resolution,
defined:

E


 E
Where  is the width of the narrowest feature
detectable, and  is the wavelength of that feature.
Thus, the resolution is a (rough) measure of the total
number of independent “channels” or pieces of
information in a spectrum.
Background
Astronomers measure wavelength, , in Angstrom Units (Å)
Remember, frequency, n, and wavelength of any wave phenomenon are related
by:
n=c
Where c is the characteristic wave velocity (speed of light for light, duh…)
Frequency is directly proportional to photon energy: E=hn
And inversely proportional to wavelength: E=hc/
When we talk about single photons, the unit of energy is the electron volt (eV)
12.4 Å = 1000 eV or 1 keV; X-rays are taken to be any photons with energies
greater than 100 eV (0.1 keV) and wavelengths, therefore, below 124 Å
Note that since temperature is proportional to energy, we can relate light of a
given wavelength to a characteristic temperature through Boltzmann’s constant,
k:
E=kT
Where 1 keV corresponds to 11.6 million degrees
ROSAT (1993), resolution E/E ~ 2 - With only a few independent pieces of
information, the shape of this smooth-looking spectrum tells us only about
the relative amount of ‘hard’ to ‘soft’ X-rays (i.e. those with energies above
0.5 keV to those with energies below this level).
ASCA (1997) resolution E/E ~ 40 - Some of the strongest line
complexes are just visible above the pseudo-continuum of blended
weak lines. The number of independent elements of information is
just not great enough to distinguish individual emission lines.
Chandra (2000) resolution E/E ~ 800 - We can now see numerous emission
lines; and not just see them, but resolve them, measuring their intrinsic widths.
The photons of different wavelength, that together constitute a
spectrum, are produced by a handful of specific atomic processes
Much of the power of spectroscopy comes from this macro-micro
connection (observed spectrum vs. movement of electrons inside
atoms) -- What is an atom? To us, it is a central electrostatic
potential plus a bunch of quantum mechanically determined bound
energy levels and a “continuum” of unbound states. Think potential
energy:
an atom
continuum
energy
is like a staircase
ground state
In both these cases, energy is released
Let’s consider some specific atomic processes and see which ones
lead to either the production of photons (emission) or the
destruction of photons (absorption)
key
previous position of an electron
current position of an electron
movement of an electron
g
photon
movement of a photon
e
external electron
simplified 2-level atom
(plus continuum)
key
1. Radiative excitation
previous position of an electron
current position of an electron
movement of an electron
g
photon
movement of a photon
e
external electron
•Photon must be of exactly the right
energy
•Photon is destroyed (absorbed)
g
2. Radiative de-excitation (“spontaneous emission”)
key
previous position of an electron
current position of an electron
movement of an electron
g
photon
movement of a photon
e
external electron
•Photon energy reflects energy
difference of electron bound states
•Photon is produced (emitted)
g
3. Collisional excitation
key
previous position of an electron
current position of an electron
movement of an electron
g
photon
movement of a photon
e
external electron
•Free electron need only have more
energy than the energy difference
between states
•Free electron is not destroyed, but
loses some kinetic energy (note
potential for cooling…)
e
e
4. Collisional de-excitation
key
previous position of an electron
current position of an electron
movement of an electron
g
photon
movement of a photon
e
external electron
•Free electron is not destroyed, but
gains some kinetic energy
e
e
Collisional excitation followed by spontaneous (radiative) emission
is the most important process in X-ray-emitting cosmic plasmas
It is a cooling process (taking kinetic energy from free
electrons -- i.e. heat) and converting it to photons, which
can escape the plasma (and bring information to our
telescopes about the conditions in the emitting plasma).
g
e
e
These basic atomic processes lead to a wealth of
phenomena in X-ray spectra.
An X-ray spectrum reflects the physical properties of the
source (the electron density and temperature, radiation flux
and temperature, bulk motions).
These properties can be used to discriminate among
different physical models of a laboratory or astrophysical
object.
The following slides are supplemental – read at your
own risk
The context of X-ray spectral observations of hot stars
Hot stars are thought not to have outer convection
zones, magnetic fields, or the associated magnetic dynamo and
corona that our sun has. Thus their discovery 20 years ago as
relatively strong soft X-ray sources was a surprise.
Hot stars do have strong radiation-driven winds.
These winds are subject to a line-driving instability which can
lead to shock heating of the wind plasma. Although this
mechanism has been assumed to produce the observed X-rays,
the numerical simulations do not do a very good job of
quantitatively reproducing the observed X-rays.
The new high-resolution X-ray spectra of hot stars
holds out the potential to discriminate between these two
general theories of stellar X-ray emission; or possibly to
inspire the development of a new theory.
Solar-type magnetic heating?
Or massive stellar wind shock heating?
1s2p 1P
1s2p 3P
1s2s 3S
R
I
F
1s2 1S
A partial energy level diagram for helium-like ions, such as Si+12
(see the FIR panel inset on the Chandra spectrum figure). The
resonance transition (R) is strongest, but the intercombination (I)
and forbidden (F) lines can also be strong. Electrons in the longlived, or metastable, 3S state can be collisionally or radiatively
excited to the 3P level (transition energy of order 10 eV), making
the F/I ratio a good diagnostic of density and radiation field
(effectively distance from the star). The forbidden line of the
silicon feature in the t Sco spectrum is strong, indicating that the
X-ray emitting plasma is both far from the star and low
density…but how far and how low?
Line profiles are affected by the hot plasma’s spatial and velocity
distribution, as well as the degree of attenuation by an overlying cold
stellar wind.
In the next panel I show synthetic line profiles for a family of coronal
models (left) and wind-shock models (right). Line profiles as a
function of the scaled wind velocity (x=c/ov) are shown for
different instrumental resolutions in each sub-panel. The panels have
wind attenuation increasing downward. Note that as the wind
attenuation increases less and less of the red side of the profiles are
seen. This is because the redshifted wind necessarily comes from the
back side of the star, which suffers the most absorption. The coronal
models assume an X-ray source that is strongly concentrated at the
base of a slowly accelerating wind, while the wind-shock models
assume that the source is distributed within the supersonic outflow
beyond some minimum radius. The narrow features seen in t Sco
would seem to point to a mostly coronal origin to these X-rays; but
the strong forbidden lines indicate that the hot plasma is not very
close to the surface of the star.
Coronal
Wind-shock
Instrumental broadening
s/v = 0, 0.1, 0.3, 0.5
lx
t*0.01
1
3
5
-1
0
1
10
-1
0
1
x