resolution x - nasa · 2003-05-21 · introduction it has been said that chandra and xmm/newton...
Post on 19-Apr-2020
3 Views
Preview:
TRANSCRIPT
High Resolution X
-ray Spectra
Randall K. Sm
ithChandra X
-ray Center
The Glory and the G
randeur
Introduction
It has been said that Chandra and XM
M/N
ewton have
finally brought high-resolution spectra to X-ray astronom
y,but this is not strictly true.Chandra &
XM
M/N
ewton have m
ade high-resolutionspectra com
mon, but the first high-resolution spectra to be
published was:
A 7 ksec exposure of Capella
with the Einstein Solid State
Spectrometer (SSS) (H
olt et al.1979)
IntroductionO
f course, it could be argued that the Einstein SSS is notreally high-resolution; it is only ~ CCD
resolution. Thefirst published grating spectrum
was:
Wavelength
020
1030
-10-20
-30
Log Counts
A 42 ksec observation
of Capella w
ith theEinstein O
bjectiveG
rating Spectrometer
(OG
S) by Mew
e et al.(1982)
IntroductionClose on their heels w
ith an even higher resolutionspectrom
eter was:
Vedder &
Canizares (1983) with a 59 ksec observation of
(yes!) Capella w
ith the Einstein Focal Plane CrystalSpectrom
eter (FPCS) :
IntroductionA
fter the Einstein mission, EX
OSA
T was launched.
Although it prim
ary claim to fam
e is timing, it did m
anageto get a high resolution spectrum
of:
Capella, using the Transmission G
rating Spectrometer
(TGS) in a 85 ksec observation (Lem
en et al. 1989)
IntroductionCom
ing into the more m
odern era (i.e., when I started in
this business), there was a joint analysis of EU
V/X
-rayem
ission from:
Capella, using a 21 ksec with the A
SCA Silicon Im
agingSpectrom
eter (SIS) and a 120 ksec with the EU
VE
spectrometer (Brickhouse et al. 2000)
Capella
IntroductionO
nce Chandra was
launched, two new
gratingsbecam
e available. Here is
spectrum of Capella
observed for 95 ksec with
the Low Energy
Transmission G
rating(LETG
) using the High-
Resolution Camera (H
RC)as a detector. (Brinkm
an etal. 2000)
Introduction
And, of course, the H
igh-Energy Transmission G
rating (HETG
)also observed Capella, w
ith the Advanced CCD
Imaging
Spectrometer (A
CIS) for 89 ksec (Canizares et al. 2000)Capella
IntroductionW
ith the launch of XM
M/N
ewton, the Reflection G
ratingSpectrom
eter (RGS) w
as finally able to achieve the onetrue goal of all spectroscopists: a 52 ksec observation of
Audard et al. 2001
CAPELLA
!Capella.
Grating Basics
Chandra; notice the position of the removable gratings:
So, where do these gratings go? H
ere is a diagram of
Grating Basics
Here is a the H
ETG, pre-installation. The m
any grating facets areall carefully positioned to catch the light from
the 4 nested mirrors.
Grating Basics
The LETG, pre-installation. N
ote the support structures, which
will later show
up as scattered light in deep LETG observations:
Grating Basics
Canizares et al. (2000)
The two arm
s of the HEG
and MEG
gratings can be clearly seenhere. N
otice the larger effective area of the MEG
; the explodedregion show
s the effect of the HEG
’s higher resolution.
Grating Basics
Canizares et al. (2000)
X-ray gratings w
ork identically to optical gratings; both follow the
grating equation:sin b = m
l/pw
here l is the w
avelength, b is the dispersion angle (measured
from the zero-order im
age), p is the spatial period of the gratingitself, and m
is the order (1st, 2nd, etc).
Grating Basics
High-resolution (grating) spectra on Chandra cover a huge range of
wavelengths: from
1.2-170Å, over tw
o orders of magnitude. N
ote thatalthough “w
avelength” is the natural unit for grating data, many (older) X
-ray astronom
ers use energy (keV) units anyw
ay.
LETG/H
RC-S observation of NG
C6624
If ACIS is used, the CCD
resolution can distinguish between the
different orders; on the HRC, this m
ust be modeled.
The spatial and spectral elements are tightly coupled. If the zero-order
image is slightly displaced (i.e., if the source is piled-up), the +/- order
wavelengths w
ill be offset from each other. If so, it calls for reprocessing.
sin b = ml/p
Grating Basics
The light blue lines show the extraction region in the cross-
dispersion direction. Notice the enhanced background on the
rightmost chips; this can be (largely) elim
inated by using theenergy resolution of the CCD
s. Also, one chip has failed; this
also happened on RGS2, but fortunately at a different position.
The star AB
Dor observed
with the
XM
M RG
S1
Grating Basics
So, when you dow
nload grating data, what do you actually have?
4 spectra; +1,2 from each
of two RG
Ss.X
MM
/RGS
Not recom
mended
HRC/H
ETG
2 spectra; ±∑n i orders
HRC/LETG
6 spectra; ±1,2,3 ordersfor the LEG
ACIS/LETG
12 spectra; ±1,2,3 ordersfor both H
EG and M
EGA
CIS/HETG
For Chandra data, these are stored in one file with m
ultiple spectra,a PH
A Type II file; for X
MM
data, each spectrum is in a different
PHA
Type I file.
Grating Basics
Eff. Area @
1keV (cm
-2)
150
25 34 10
200~0.06
RGS 1st
2500.05
LEG 1st
5400.023
MEG
1st
10000.012
HEG
1st
Resolution@
1keVD
l
(Å)
Grating
The characteristics of the different gratings in theirprim
e arrangement:
Grating Basics
Consider a line at 1.0 keV (12.398Å
), on top of a 0.5 keV blackbody
So, what can w
e do with grating data that w
e can’t with CCD
s?
Grating Basics
Now
consider the same spectrum
observed by each of the four gratings:
If this data were fit assum
ing a pure Gaussian response, the scattering w
ings of theRG
S would m
ean we’d m
easure only 1/2 the true power!
Processing
This talk really isn’t about actual data processing; if you findyourself needing to reduce Chandra and/or XM
M/Newton
grating data, I recomm
end:
http://cxc.harvard.edu/ciao/threads/spectra_letghrc.html
LETG/H
RC
http://cxc.harvard.edu/ciao/threads/spectra_letgacis.html
LETG/A
CIS
http://cxc.harvard.edu/ciao/threads/spectra_hetgacis.html
HETG
/ACIS
http://heasarc.gsfc.nasa.gov/docs/xmm
/abc/node8.html
RGS
The end product will be either m
ultiple Type I or a Type IIPH
A file(s); basically, a histogram
of counts as a function ofw
avelength, along with the instrum
ental response.
Processing
1.Identify the dispersed source(s). In som
e cases, multiple sources m
ay bepresent; both CIA
O and SA
S software can identify and reduce m
ore than onesource, but the user m
ust select which source to extract.
2.Calculate the dispersion distance for each event, and (if available) resolve theorder by com
paring the dispersion wavelength to the m
easured CCD energy.
3.Extract the spectrum
into a PHA
file, suitable for use in Sherpa or XSPEC.
4.Create the detector response for the source. This depends on the position of thesource on the detector, the aspect solution, and the operating m
ode of thetelescope. The variations are not generally large, so using another observation'sresponse is O
K for quick w
ork.
The XM
M/N
ewton RG
S is a reflection grating, while the Chandra LETG
andH
ETG are transm
ission gratings. How
ever, the basic processing steps aresim
ilar:
SAS does steps 1-3 w
ith one comm
and, rgsproc. Chandra breakssteps 1-3 into subprogram
s: tg_create_mask, tg_resolve_events,
and tgextract.
ProcessingK
now your D
ata: RGS event files
unix% dm
list rgs_evt2.fits cols
ProcessingK
now your D
ata: HETG
/ACIS event files
unix% dm
list acis_evt2.fits cols
Processing
unix% dm
list hrc_evt2.fits cols
Know
your Data: LETG
/HRC event files
ProcessingFor bright sources on A
CIS-S, the background is likelynegligible. H
owever, on the H
RC-S or with the RG
S it isn’t:
Source/Background regions for the HR
C/LETG
Proper background subtraction is still a topic of some debate!
Processing
Using a detector w
ith even moderate energy resolution can
substantially reduce the background, because one can cross-correlate the w
avelength measured from
the dispersed positionw
ith the energy measured in the CCD
using the relation:
XM
M RG
S1spectra of thestar A
B Dor
EkeV = 12.398 / l
Å
Processing
What if you w
ant both high-resolution spectra AN
D tim
ing?
1) You’d better have a bright source
2) With Chandra, the A
CIS detectors can be run in continuousclocking m
ode while either of the gratings are in place. The
time resolution is about 3 m
s.3) W
ith XM
M/N
ewton, the RG
S detectors can be run in CCm
ode, but as of this writing (M
ay 2003) this mode is not
available to users for technical reasons.4) The H
RC has good time resolution, but due to an
electronics problem, the tim
e tags for each event suffer froman off-by-one problem
and actually refer to the previousevent. It is possible to run the H
RC in a “timing” m
odew
here this is mitigated, but not w
ith the full grating array.
Processing
Unfortunately, not every X
-ray source is as bright as Capellaor N
GC6624. W
hat can you do to enhance the S/N?
In this case, you have 4 choices:
1)C
o-add plus/minus orders of the sam
e grating ÔCan broaden lines if zero-order is offset.
2)C
o-adding different gratings, such as HEG
andM
EG or R
GS1 and R
GS2 Ô
Complicates line shape
function.3)
Co-add separate observations Ô
Instrumental
background can vary, plus the same issues of zero-
order offsets.4)
Co-add separate observations and instrum
ents ÔAll of the above
Grating A
nalysisunix%
sherpasherpa> data acis_pha2.fitssherpa> param
prompt off
sherpa> rsp[hm1]
sherpa> rsp[hp1]sherpa> rsp[m
m1]
sherpa> rsp[mp1]
sherpa> hm1.rm
f = acisheg1D1999-07-22rm
fN0004.fits
sherpa> hm1.arf = acisf01318H
EG_-1_garf.fits
sherpa> hp1.rmf = acisheg1D
1999-07-22rmfN
0004.fitssherpa> hp1.arf = acisf01318H
EG_1_garf.fits
sherpa> mm
1.rmf = acism
eg1D1999-07-22rm
fN0004.fits
sherpa> mm
1.arf = acisf01318MEG
_-1_garf.fitssherpa> m
p1.rmf = acism
eg1D1999-07-22rm
fN0004.fits
sherpa> mp1.arf = acisf01318M
EG_1_garf.fits
sherpa> instrument 3 = hm
1sherpa> instrum
ent 4 = hp1sherpa> instrum
ent 9 = mm
1sherpa> instrum
ent 10= mp1
sherpa> ignore allsets all sherpa> notice allsets w
ave 14.9:15.4 sherpa> source 3,4,9,10 = poly[b1] + delta1d[l1] + delta1d[l2] + delta1d[l3]sherpa> l1.pos = 15.014sherpa> l2.pos = 15.079sherpa> l3.pos = 15.2610 sherpa> freeze l1.possherpa> freeze l2.possherpa> freeze l3.possherpa> fitsherpa> lp 4 fit 3 fit 4 fit 9 fit 10sherpa> im
port(``guide'')sherpa> m
dl2latex\begin{tabular}{lllllll}M
odelNam
e & Line M
odel & Position &
Flux & Flux Error &
Fit Data &
Label \\ &
& A
ngstrom &
ph/cm$^2$/s &
ph/cm$^2$/s &
\\l1 &
delta1d & 15.014 &
0.00308923 & 6.7101e-05 &
3,4,9,10 & \\
l2 & delta1d &
15.079 & 0.000270431 &
2.81612e-05 & 3,4,9,10 &
\\l3 &
delta1d & 15.261 &
0.00125857 & 4.79625e-05 &
3,4,9,10 & \\
Load data
Define
instrument
response
Define source
Display results
Grating A
nalysis
Results from Sherpa
fits; notice the lowbackground in theH
EG data, and the
higher resolution.
Guide
GU
IDE is a collection of scripts w
hich access the atomic
database ATO
MD
B. GU
IDE provides a num
ber of informational
functions:
Output ionization balance fractions for
a given ionionbal
Convert fit parameters into a latex table
mdl2latex
Describe atom
ic parameters of a line
describe
List strong lines at a given temperature
strong
Print finding chart of wavelengths
identify
GU
IDE w
orks in Sherpa or Chips; initialize it with im
port(“guide”)
WebG
UID
E
WebG
UID
E
WebG
UID
E
Grating A
nalysis
When faced w
ith the task of understanding a high-resolution spectrum, especially a line-
dominated one, it is likely that sim
ple equilibrium m
odels will not w
ork adequately, despitepossibly giving low
reduced c2 values--usually because the counts/bin is low
and so theerrors are overestim
ated, not because the fits are good.
In this case, some new
strategies are needed. Here are som
e suggestions for starting points.
First: Determ
ine if your plasma is dom
inated by photoionization or collisional ionization.For exam
ple, the initial analysis of ASCA
\ data of Cygnus X-3 used a collisional m
odel,even though the em
ission is due to photoionization (see Liedahl & Paerels 1996)
Make a list of all processes that could be affecting line em
ission. For example, in helium
-like ions, the three dom
inant lines are the resonance, the forbidden, and theintercom
bination lines. These can be excited by direct excitation, radiative recombination,
dielectronic recombination of hydrogen-like ions, innershell ionization of lithium
-like ions,cascades from
higher levels, or by photoexcitation or photoionization. Once created, the
lines can be absorbed or scattered by like ions or by different ions. In many cases, sim
plephysical argum
ents can be used to limit or exclude various processes, reducing the
parameter space that m
ust be searched.
Grating A
nalysisSecond: A
ttempt to determ
ine the true continuum level w
ith confidence. Thecontinuum
in a hot plasma is not necessarily dom
inated by bremsstrahlung:
Weak em
ission lines will also blend in to m
ake the continuum seem
larger, which
will lead to a system
atic underestimate of line fluxes, and therefore elem
ental andionic abundances. A
fter you have found an acceptable spectral model, search for
regions in the model w
ith no or few lines, and com
pare the model to the data in this
region. If the model continuum
overestimates the continuum
here, it is likely itoverestim
ates it everywhere due to unresolved line em
ission
AstroA
tom
Conclusions
• Reprocessing grating data is no longer absolutely required, but hasgotten far easier and provides a sense of confidence about the data.• Co-A
dding and/or binning grating data should be avoided when
possible. Remem
ber that, statistically, nothing is gained by it,although it m
ay be much faster to fit it and easier to see the results.
• A num
ber of new facilities for atom
ic data analysis have beencreated for Sherpa and X
SPEC How
ever, remem
ber to check thecaveats on this data before trusting it totally! For the A
TOM
DB,
they are at http://asc.harvard.edu/atomdb/doc/caveats.htm
l• G
lobal fitting of generic equilibrium m
ay be useful for guiding theanalysis, but any project should begin w
ith a physics-basedapproach, follow
ed ideally by a line-based analysis and finally bychecking regions w
hich should well-understood (such as line-free
areas or those dominated by a single line).
top related