GALPROP code for CRGALPROP code for CRGALPROP code for CR propagation and diffuse
emissions
GALPROP code for CR propagation and diffuse
emissionsemissionsemissions
Igor V. MoskalenkoIgor V. MoskalenkoStanford & KIPACStanford & KIPAC
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 1
ContentsContents
Introduction to propagation of CRsIntroduction to propagation of CRs
Building a model of the diffuse emission
Effect of anisotropic inverse Compton scattering
Building a model of the diffuse emission
Effect of anisotropic inverse Compton scattering
A view from far far away: multi-wavelength luminosity of the Milky WayA view from far far away: multi-wavelength luminosity of the Milky Wayy y
Bayesian analysis of propagation parameters
y y
Bayesian analysis of propagation parameters
GALPROP WebRunGALPROP WebRun
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 2
CR Propagation: Milky Way GalaxyCR Propagation: Milky Way Galaxy1 kpc ~ 3×1021 cm1 kpc ~ 3×1021 cm1 kpc 3 10 cm1 kpc 3 10 cm
Optical image: Cheng et al. 1992, Brinkman et al. 1993Radio contours: Condon et al. 1998 AJ 115, 1693
Optical image: Cheng et al. 1992, Brinkman et al. 1993Radio contours: Condon et al. 1998 AJ 115, 1693
Halo0.1-0.01/ccm
NGC891
SunSun
Intergalactic spaceR Band image of NGC891R Band image of NGC891
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 3
R Band image of NGC891R Band image of NGC8911.4 GHz continuum (NVSS), 1,2,…64 1.4 GHz continuum (NVSS), 1,2,…64 mJymJy/ beam/ beam “Flat halo” model (Ginzburg & Ptuskin 1976)“Flat halo” model (Ginzburg & Ptuskin 1976)
CRs in the Interstellar MediumCRs in the Interstellar Medium
ISMISMSNR RX J1713SNR RX J1713--39463946SNR RX J1713SNR RX J1713--39463946SNR RX J1713SNR RX J1713--39463946
ee±±e±
X,γX,γISMISM
42 sigma (2003+2004 data)SNR RX J1713SNR RX J1713--39463946
42 sigma (2003+2004 data)42 sigma (2003+2004 data)SNR RX J1713SNR RX J1713--39463946SNR RX J1713SNR RX J1713--39463946
ChandraChandraChandra
eee
PPHeHePP
HeHeISRFISRF•diffusion•diffusion
•energy losses •energy losses •diffusive reacceleration•diffusive reacceleration
•diffusion•diffusion•energy losses •energy losses
•diffusive reacceleration•diffusive reacceleration
ICICHESSPSF
HESSHESSPSFPSF
BBBHESSHESSHESS
CNOCNOCNOCNO gasgasgasee++--e +-
ππ++--π+-
•diffusive reacceleration•diffusive reacceleration•convection•convection
•production of •production of secondariessecondaries
•diffusive reacceleration•diffusive reacceleration•convection•convection
•production of •production of secondariessecondaries
ππ 00π 0 FermiFermiFermiWIMPWIMPaannihilnnihil..WIMPWIMPaannihilnnihil..
PP__PP__
P,P,P,P, X,γX,γ
ee±±e± gasgasgas
PP__PP__
LiBeBLiBeBLiBeBLiBeBpppp
HeHeCNOCNOHeHe
CNOCNO
Flux
Fluxππ
++--π+-
ee±±e±
eee solar modulation
ACEACE
CNOCNOCNOCNO
20 GeV/n20 GeV/n
CR species:O l 1 l ti
CR species:O l 1 l ti
BESSBESS
eee
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 4
ACEACEhelio-modulationhelio-modulation
Only 1 locationmodulationOnly 1 locationmodulationPAMELAPAMELA
Components of realistic CR propagation modelComponents of realistic CR propagation model
Detailed gas distribution (energy losses, π0, bremsstrahlung)
I t t ll di ti fi ld (IC ± l )
Detailed gas distribution (energy losses, π0, bremsstrahlung)
I t t ll di ti fi ld (IC ± l )Interstellar radiation field (IC, e± energy losses)
Nuclear & particle production cross sections & the reaction
Interstellar radiation field (IC, e± energy losses)
Nuclear & particle production cross sections & the reaction
network
G d i b hl C 0
network
G d i b hl C 0Gamma-ray production: bremsstrahlung, IC, π0
Energy losses: ionization, Coulomb, brems, IC, synchrotron
Gamma-ray production: bremsstrahlung, IC, π0
Energy losses: ionization, Coulomb, brems, IC, synchrotrongy , , , , y
Solve transport equations for all CR species
gy , , , , y
Solve transport equations for all CR species
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 5
Derive the propagation parametersDerive the propagation parameters
Transport Equations ~90 (no. of CR species)Transport Equations ~90 (no. of CR species)
t∂ r )(
VD
prqttprψ
∇∇
=∂
∂
rrr
r
][
),(),,( sources (SNR, nuclear reactions…)sources (SNR, nuclear reactions…)sources (SNR, nuclear reactions…)sources (SNR, nuclear reactions…)
diff idiff idiff idiff i VxxD
ψ
ψψ
⎤⎡ ∂∂
−∇⋅∇+
2
][
convectionconvectionconvectionconvection
diffusiondiffusiondiffusiondiffusion
ppppDp
pψ
⎥⎥⎦
⎤
⎢⎢⎣
⎡
∂∂
∂∂
+ 22 convection convection
(Galactic wind)convection convection
(Galactic wind)diffusive reaccelerationdiffusive reacceleration(diffusion in the momentum space)
diffusive reaccelerationdiffusive reacceleration(diffusion in the momentum space)
Vpdtdp
pψψ ⎥⎦⎤
⎢⎣⎡ ⋅∇−
∂∂
−rr
31EE--losslossEE--lossloss
ψψ((rr p tp t)) –– densityψψ((rr p tp t)) –– densitydf τψ
τψ
−−fragmentationfragmentationfragmentationfragmentation radioactive decayradioactive decayradioactive decayradioactive decay
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 6
ψψ((rr,p,t,p,t) ) density per total momentumψψ((rr,p,t,p,t) ) density per total momentum+ boundary conditions+ boundary conditions+ boundary conditions+ boundary conditions
Secondary/primary nuclei ratio & CR propagationSecondary/primary nuclei ratio & CR propagationTypical parameters (model-dependent):
D 1028 ( /1 GV)α 2/Typical parameters (model-dependent):
D 1028 ( /1 GV)α 2/D ~ 1028 (ρ/1 GV)α cm2/sα ≈ 0.3-0.6Zh ~ 4-6 kpc; VA ~ 30 km/s
D ~ 1028 (ρ/1 GV)α cm2/sα ≈ 0.3-0.6Zh ~ 4-6 kpc; VA ~ 30 km/s
Be10/Be9Be10/Be9
Zh increaseZh increase
Using secondary/primary nuclei ratio (B/C) & radioactive isotopes (e.g. Be10):Diffusion coefficient and its indexGalactic halo size ZhPropagation mode and its parameters (e g reacceleration V convection V )
Using secondary/primary nuclei ratio (B/C) & radioactive isotopes (e.g. Be10):Diffusion coefficient and its indexGalactic halo size ZhPropagation mode and its parameters (e g reacceleration V convection V )
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 7
Propagation mode and its parameters (e.g., reacceleration VA, convection Vz)Propagation parameters are model-dependentPropagation mode and its parameters (e.g., reacceleration VA, convection Vz)Propagation parameters are model-dependent
Relevant CR and gamma-ray (also CR!) instrumentsRelevant CR and gamma-ray (also CR!) instruments
PAMELApbar
đ,α
pbar
đ,αBESS-Polar
i-mat
ter
i-mat
ter
−
e+
e-
e+
e- Fermi/LATHESSMagic
Integral
AM
S-I HEATWMAP
CAPRICE
anti
anti
p
He
p
He
BESS-PolarAMS-I
gMilagroVeritasCOMPTEL
EGRET ATICC At
erter
Z≤8
8
Fermi-LAT 1-year Gamma-Ray SkymapFermi-LAT 1-year Gamma-Ray Skymap
Galactic plane
Buiding a modelBuiding a model
of the diffuse emissionof the diffuse emission
Sources
Dominating source of gamma rays on the sky. Produced by CR interactions with interstellar Dominating source of gamma rays on the sky. Produced by CR interactions with interstellar
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 9
gas and radiation field – the so-called diffuse emission. The diffuse Galactic gamma rays are the tracers of CR proton and electron spectra throughout the Galaxy.gas and radiation field – the so-called diffuse emission. The diffuse Galactic gamma rays are the tracers of CR proton and electron spectra throughout the Galaxy.
Diffuse Galactic GammaDiffuse Galactic Gamma--ray ray EmissionEmissionDiffuse Galactic GammaDiffuse Galactic Gamma--ray ray EmissionEmission
Origin and propagation of cosmic rays– Nature and distribution of CR
sources
Origin and propagation of cosmic rays– Nature and distribution of CR
sources
Foreground against which sources are detected− Point sources: limitation on
Foreground against which sources are detected− Point sources: limitation on
– Abundances of primary species– Production of secondary species– The propagation mode and its
– Abundances of primary species– Production of secondary species– The propagation mode and its
sensitivity− Extended sources: disentanglement
Indirect dark matter detection
sensitivity− Extended sources: disentanglement
Indirect dark matter detectionp p grelationship to magnetic turbulence in the ISM
Interstellar Medium
p p grelationship to magnetic turbulence in the ISM
Interstellar Medium
− Predicted γ-ray/CR signals− relies on accurate treatment of
standard astrophysical sources
− Predicted γ-ray/CR signals− relies on accurate treatment of
standard astrophysical sources– Distribution of HI, H2, HII gas– Nature of XCO relation in Galaxy– Distribution and intensity of
i ll di i fi ld d
– Distribution of HI, H2, HII gas– Nature of XCO relation in Galaxy– Distribution and intensity of
i ll di i fi ld d
Foreground for isotropic diffuse background− Whatever its nature
Foreground for isotropic diffuse background− Whatever its nature
interstellar radiation field and formation of H2interstellar radiation field and formation of H2
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 10
Fermi-LAT: diffuse gammasFermi-LAT: diffuse gammasC ti l GALPROP d l i i t ith thC ti l GALPROP d l i i t ith thConventional GALPROP model is in agreement with the Fermi-LAT data at mid-latitudes (mostly local emission)The EGRET “GeV excess” is not confirmedThis means that we understand the basics of cosmic ray
Conventional GALPROP model is in agreement with the Fermi-LAT data at mid-latitudes (mostly local emission)The EGRET “GeV excess” is not confirmedThis means that we understand the basics of cosmic rayThis means that we understand the basics of cosmic ray propagation and calculate correctly interstellar gas and radiation field, at least, locally
This means that we understand the basics of cosmic ray propagation and calculate correctly interstellar gas and radiation field, at least, locally
model
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 11
Abdo+’09
Diffuse emission at low to high Galactic latitudes
Diffuse emission at low to high Galactic latitudes
Pion-decay and inverse Compton emission are two dominant components – allow us to probe the average CR proton and electron spectra
Pion-decay and inverse Compton emission are two dominant components – allow us to probe the average CR proton and electron spectra
High latitudesThe GALPROP predictions agree wellThe GALPROP predictions agree well
along the line of sightalong the line of sight
Low latitudes
predictions agree well with the LAT datapredictions agree well with the LAT data
Abdo+’10
Mid-latitudes
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 12
Spectrum of isotropic diffuse γ-ray emissionSpectrum of isotropic diffuse γ-ray emission
index=2.41±0.05
2010
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 13
Accessed every 6 years (1998, 2004, 2010, 2016,…)Time-dependent: increasingly steepens with time…Accessed every 6 years (1998, 2004, 2010, 2016,…)Time-dependent: increasingly steepens with time…
Diffuse Gammas – Local SpectrumDiffuse Gammas – Local Spectrum
Mid-latitudes: Mid-latitudes: (200°
Example: latitude profile of the inner Galaxy Example: latitude profile of the inner Galaxy
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 15
Fractional count residuals: (theory-data)/dataFractional count residuals: (theory-data)/data
Agreement for models is overall goodAgreement for models is overall good
srs=SNR, Zh=4kpc, Rmax=20kpc, TS=150K, mag=5srs=SNR, Zh=4kpc, Rmax=20kpc, TS=150K, mag=5
Agreement for models is overall good but features in residuals are at ~ % levelDifference between illustrative models due to variations of model parameters
Agreement for models is overall good but features in residuals are at ~ % levelDifference between illustrative models due to variations of model parametersA)A) pp
difference between residualsdifference between residualsA)A)
A-B=A-B=srs=Lorimer,
Zh=6kpc, Rmax=20kpc,
mag=5,
srs=Lorimer, Zh=6kpc,
Rmax=20kpc, mag=5, g
optically thin HIg
optically thin HIB)B)
A-C=A-C=srs=OB,
Zh=8kpc, Rmax=30kpc,
2
srs=OB, Zh=8kpc,
Rmax=30kpc, 2
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 16
mag=2, optically thin HI
mag=2, optically thin HIC)C)
Anisotropic IC scattering in the GalaxyAnisotropic IC scattering in the Galaxy
The formalism and effect of anisotropic IC The formalism and effect of anisotropic IC p Cscattering in the Galaxy: IM & Strong’00Varies over the sky
p Cscattering in the Galaxy: IM & Strong’00Varies over the skyyGives ×(0.7-2) more gammas vs isotropic approximation
yGives ×(0.7-2) more gammas vs isotropic approximationppHas to be taken into account when modeling the diffuse emission
ppHas to be taken into account when modeling the diffuse emission
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 17
IM & Strong’00
Standard ISRF (10 kpc halo)Standard ISRF (10 kpc halo)
i t i /i t i i ii t i /i t i i ianisotropic/isotropic aniso sum iso sum anisotropic/isotropic aniso sum iso sum
50 MeV
35 GeV
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 18Porter+, in preparation
Bulge×10 model (10 kpc halo)Bulge×10 model (10 kpc halo)i t i /i t i i ii t i /i t i i ianisotropic/isotropic aniso sum iso sum anisotropic/isotropic aniso sum iso sum
50 MeV
35 GeV
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 19
Ratio Bulge×10/Standard (10 kpc halo)Ratio Bulge×10/Standard (10 kpc halo)i t i k i ii t i k i ianisotropic skymap aniso sum iso sum anisotropic skymap aniso sum iso sum
50 MeV
35 GeV
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 20
Milky Way from far far away…Milky Way from far far away…Milky Way from far far away…Milky Way from far far away…
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 21
Multi-wavelength luminosity of the Milky WayMulti-wavelength luminosity of the Milky Way
“Proximity” of the MW plus direct
t ll
“Proximity” of the MW plus direct
t ll
opticalopticalpp
measurements allow us to construct its detailed “microscopic” model (GALPROP)
measurements allow us to construct its detailed “microscopic” model (GALPROP)
IRIR
HeHe( )
Integral properties of the MW – a useful template
( )
Integral properties of the MW – a useful template e−e−for studies of the normal galaxies
S l 2010 A JL
for studies of the normal galaxies
S l 2010 A JL
synchsynchπ0π0
ICICγ totalγ total
ee
Strong et al. 2010, ApJLin press Strong et al. 2010, ApJLin press
bremssbremss
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 22
Milky Way as electron calorimeterMilky Way as electron calorimeterCalculations for Zhalo= 4 kpcCalculations for Zhalo= 4 kpchalo pLeptons lose ~60% of their energyγ-rays: 50-50 by nucleons and by leptons
halo pLeptons lose ~60% of their energyγ-rays: 50-50 by nucleons and by leptons
Cosmic rays7.90×1040 erg/sCosmic rays7.90×1040 erg/s
SynchrotronSynchrotron
Primary nucleons98.6%
Primary nucleons98.6%
Synchrotron0.35%
Synchrotron0.35%
BremsstrahlungBremsstrahlung PrimaryPrimarydd Bremsstrahlung0.15%Bremsstrahlung
0.15%
Inverse ComptonInverse Compton
yelectrons
1.41%
yelectrons
1.41%
Secondaryleptons
e+: 0.33%e−: 0.10%
Secondaryleptons
e+: 0.33%e−: 0.10%
Total gamma raysTotal gamma raysNeutral pionsNeutral pions
p0.58%
p0.58%
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 23
g y1.6%g y1.6%
p0.85%
p0.85%
* The percentages in brackets show the values relative to the luminosity of their respective lepton populations
* The percentages in brackets show the values relative to the luminosity of their respective lepton populations
Constrains on CR propagation d l f l b l
Constrains on CR propagation d l f l b lmodels from global scansmodels from global scans
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 24
Bayesian approach (Θ vector of parameters D available observations):Bayesian approach (Θ vector of parameters D available observations):
MethodologyMethodologyBayesian approach (Θ – vector of parameters, D – available observations):Bayesian approach (Θ – vector of parameters, D – available observations):
P(Θ | D) =P(D |Θ)P(Θ)
P(D)P(Θ|D) – posterior distribution of parameters P(D|Θ) – likelihood function (for fixed D)P(Θ|D) – posterior distribution of parameters P(D|Θ) – likelihood function (for fixed D)
P(D)
( | ) ( )P(Θ) – prior distributionP(D) – Bayesian evidenceBayesian approach allows for a more sophisticated treatment of systematics
( | ) ( )P(Θ) – prior distributionP(D) – Bayesian evidenceBayesian approach allows for a more sophisticated treatment of systematicsBayesian approach allows for a more sophisticated treatment of systematicsNested sampling algorithm by John Skilling’04,06 and MultiNest by Feroz&Hobson’08 (SuperBayeS code by Ruiz de Austri+’06, Trotta+’08)T t l 1 4×105 l 13 CPU
Bayesian approach allows for a more sophisticated treatment of systematicsNested sampling algorithm by John Skilling’04,06 and MultiNest by Feroz&Hobson’08 (SuperBayeS code by Ruiz de Austri+’06, Trotta+’08)T t l 1 4×105 l 13 CPUTotal 1.4×105 samples, ~13 CPU yearsThe posterior distribution is in excellent agreement with one from MCMC Representative data from ACE, ISOMAX, HEAO-3, ATIC-2, CREAM
Total 1.4×105 samples, ~13 CPU yearsThe posterior distribution is in excellent agreement with one from MCMC Representative data from ACE, ISOMAX, HEAO-3, ATIC-2, CREAM
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 25
Fit a total of 17 parameters, best fit chi-squared/dof ~ 1 More details in the paper (to be submitted in a few days)Fit a total of 17 parameters, best fit chi-squared/dof ~ 1 More details in the paper (to be submitted in a few days)
Posterior probability distributionsPosterior probability distributionsNormalization of the diffusion coeff. and its index
- the best fit
ti l li t i
- the best fit
ti l li t ivertical line – posterior mean
vertical line – posterior mean
Alfven speed Halo sizecolor bar – 68%, 95% error
rangescolor bar – 68%, 95% error
ranges
Reacceleration modelReacceleration modelInjection index below/above the break @ 1 GV
Parameters are very close to those derived by the eye-fitting
Parameters are very close to those derived by the eye-fitting
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 26
y y gmethod
y y gmethod
2D posterior probability distributions2D posterior probability distributions
Contours enclose 68% and Contours enclose 68% and 95% probability regions
- best fit
95% probability regions
- best fit
● posterior mean● posterior mean● - posterior mean● - posterior mean
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 27
Examples of the fits: B/C and Be ratiosExamples of the fits: B/C and Be ratios
B/C ratio
Be10/Be9 ratioBe /Be ratio
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 28
Examples: Carbon and pbarsExamples: Carbon and pbars
Carbon
pbars
(not fitted!)
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 29
(not fitted!)
GALPROP WebRunGALPROP WebRunGALPROP WebRun
galprop.stanford.edu
GALPROP WebRun
galprop.stanford.edugalprop.stanford.edugalprop.stanford.edu
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 30
Igor Moskalenko
for the GALPROP development team
Igor Moskalenko
for the GALPROP development team
GALPROP WebRunGALPROP WebRun
GALPROP WebRun is a service that allows to run GALPROP via the WWWGALPROP WebRun is a service that allows to run GALPROP via the WWW
No local installation of the code or related libraries is necessary, only a web browser is requiredNo local installation of the code or related libraries is necessary, only a web browser is required
Available at http://galprop.stanford.edu/webrun
Calculations can be performed on a new cluster at Stanford University, using
Available at http://galprop.stanford.edu/webrun
Calculations can be performed on a new cluster at Stanford University, using the most recent GALPROP v54, older versions are also available
The service is free and open to the community. Registration is required
the most recent GALPROP v54, older versions are also available
The service is free and open to the community. Registration is required
Acknowledge by citing the web-page and the introduction paper http://arxiv.org/abs/1008.3642 (submitted to Computer Physics Acknowledge by citing the web-page and the introduction paper http://arxiv.org/abs/1008.3642 (submitted to Computer Physics
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 31
Communications)Communications)
Configuring GALPROP via WebRunConfiguring GALPROP via WebRun
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 32
Interactive interface for parameter entryParameters are validated to avoid misconfgured runsInteractive interface for parameter entryParameters are validated to avoid misconfgured runs
Running calculationsRunning calculations
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 33
Multiple simultaneous runs are allowedSharing resources fairly between usersMultiple simultaneous runs are allowedSharing resources fairly between users
Evaluating and obtaining resultsEvaluating and obtaining results
Archive with FITS files (output) available to the ownerArchive with FITS files (output) available to the owner
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 34
Archive with FITS files (output) available to the ownerURL to the archive may be shared with anyoneQuick peek possible with built-in plot templates
Archive with FITS files (output) available to the ownerURL to the archive may be shared with anyoneQuick peek possible with built-in plot templates
Computing cluster (galprop.stanford.edu)Computing cluster (galprop.stanford.edu)
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 35
Thank you !y
You are hereYou are here
SciNeGHE 2010, Trieste – September 6, 2010 :: IVM/Stanford-KIPAC 36
Top Related