The Interstellar Medium and Modeling Galactic Diffuse Emission

Transcription

1 Gamma-ray Large Area Space Telescope The Interstellar Medium and Modeling Galactic Diffuse Emission S. W. Digel SLAC EGRET >100 MeV LAT Dark Matter & Exotic Physics Group Workshop, July

2 Diffuse emission from the Milky Way Cosmic-ray interactions with interstellar gas and photons Nothing to it: all you have to know is the distributions of cosmic rays and of interstellar gas across the Galaxy, and the production functions for gamma rays So the topic is basically one of details The emphasis of this talk will be issues related to the interstellar medium We ll start with some background information and then get more specific about current issues LAT Dark Matter & Exotic Physics Group Workshop, July

3 Milky Way as a Galaxy A large Sbc barred spiral galaxy with ongoing star formation (Pop II) Makes (and keeps) its own cosmic rays and has a continuing supply of massive stars Massive (~ M _ ) stars (OB- WR) have short lives, tremendous luminosities and winds, and explode when they are done, leaving remnants that can accelerate cosmic rays Idealized diagram of Milky Way - don t take this too literally From in-plane perspective tracing arms is hard 8.5 kpc NGC 1187 (Adam Block/NOAO/AURA/NSF, Sbc M109 (AURA/NOAO/NSF), SBc Sun Powell LAT Dark Matter & Exotic Physics Group Workshop, July

4 Historical note The spiral structure of the Milky Way was difficult to elucidate before spectral line radio astronomy; even the direction of the center of the Galaxy was not known particularly well The principal difficulty was dust LAT Dark Matter & Exotic Physics Group Workshop, July

5 More notes By the way, stars themselves are not gamma-ray sources from cosmic-ray interactions. They are too thick absorbing any gamma rays that are produced. Another important aspect of the interstellar medium is that although it is vast it is quite diffuse, with typical densities orders of magnitude less than can be achieved in the lab. A large column density in a given line of sight might be H atoms cm -2, but the line of sight through the disk is easily is cm long. The ISM is vast, diffuse, cold, and not particularly intuitive. The chemistry is odd and magnetic fields and cosmic rays are surprisingly important. LAT Dark Matter & Exotic Physics Group Workshop, July

6 Components of an interstellar emission model Although radiative transfer is simple (MW is transparent at these energies), good models are not easy. Main components Interstellar gas (molecular, atomic, and ionized) Molecular gas (H 2 ) in ISM is indirectly measured via CO J = 1_0 line; even N(H I) can be tricky Distance ambiguity, velocity dispersion, & non-circular velocity field of Milky Way Interstellar radiation field Dependence on position, wavelength, and direction Cosmic rays Spatial and energy distribution Cosmic-ray spectra are measured only within solar system, with low-energy range affected by solar modulation Difficult or impossible to get unique answers for any of the above LAT Dark Matter & Exotic Physics Group Workshop, July

7 Tracers of interstellar gas The 21-cm line of H I is a spin flip transition not hard to collisionally excite and even H I as cold as in the ISM can be detected from this radio emission Importantly, the frequency of the emission can be measured very accurately spectra are routinely measured with fraction of km/s resolution First detected by Ewen & Purcell (1951) Very early on it was recognized that the Doppler shifts caused by differential rotation of the galaxy could be used to estimate distances; large-scale surveys were made using converted radar antennas 1959 LAT Dark Matter & Exotic Physics Group Workshop, July

8 Another historical note CO line commonly used tracer of molecular hydrogen densest parts of the ISM H I _ H 2 apparently catalyzed by (and at least partially protected by) interstellar dust. H 2 and does not have any low-lying transitions CO, the 2nd most abundant molecule, has a dipole moment, and a spectrum of allowed rotational transitions. The lowest-level transition J = 1 _ 0, has a frequency of 115 GHz (2.6 mm). CO is collisionally excited by H 2. The spontaneous transition downward is forbidden, and the excited state has a lifetime of about a year, but under interstellar conditions, a given CO molecule has at least that long between collisions. Interstellar CO was detected in 1970 Large-scale surveys were undertaken starting in the early 1980s, including with very small telescopes. First Observation of Interstellar CO LAT Dark Matter & Exotic Physics Group Workshop, July

9 Distribution of interstellar gas Neutral interstellar medium 21-cm H I & 2.6-mm CO (standin for H 2 ) CO (25, 0 ) ~25 km s Dame et al. (1987) H I G.C. W. Keel Hartmann & Burton (1997) Near-far distance ambiguity, rotation curve V(R), optical depth effects, LAT Dark Matter & Exotic Physics Group Workshop, July

10 Spiral structure in CO CO clouds molecular clouds in particular turned out to be good tracers of the spiral structure of the Milky Way, because they form in spiral density waves and have relatively low random velocities Certainly some handwaving goes into how the intensity of the CO line is related to the column density of H I [more later on Isabelle s results] With painstaking work to resolve the distance ambiguity and account for noncircular (streaming) motions the spiral structure of the inner Milky Way has been mapped out in pretty good detail but there s still wiggle room. Dame et al. (1986) LAT Dark Matter & Exotic Physics Group Workshop, July

11 Spiral structure (2) Observations along tangent directions of arms indicate that the arm-interarm contrast is high, at least a factor of several in molecular gas, but less so for atomic. So you can t assume that all gas is in arms. Having a basic idea of the spiral structure does not mean that you have solved the question of the distribution. In terms of the 3-dimensional, large-scale distribution of the interstellar medium ~1 kpc resolution is about all that is reasonable. Longitude Profile of CO Intensity Norma tangent Crux tangent Grabelsky et al. (1987) LAT Dark Matter & Exotic Physics Group Workshop, July

12 Local interstellar medium At higher latitudes, certainly by 10 degrees from the plane, the interstellar medium can be considered local. The scale height of CO emission is ~80 pc and that of H I is ~200 pc, although H I is detectable in every direction. LAT Dark Matter & Exotic Physics Group Workshop, July

13 Ionized interstellar medium Detected indirectly via pulsar dispersion measures and fit to model D = n ds e This is hard to get a handle on complicated, carefully tuned model but not much mass relatively speaking 1143 pulsars Cordes & Lazio (2003) LAT Dark Matter & Exotic Physics Group Workshop, July

14 Components examples Distribution of interstellar gas (cont) Hunter et al. (1997) Two studies that started with essentially the same data disagree in many details Sun Galactic Center Pohl & Esposito (1998) LAT Dark Matter & Exotic Physics Group Workshop, July Strong 2005 & Moskalenko 14

15 Interstellar Radiation in the MW Inverse Compton characteristic gamma-ray energy ~ (E e E ph ) For ~GeV gamma rays, optical-uv photons are most relevant Not much reprocessing has to happen for IC to be important to Galactic diffuse emission Band Radio IR Optical X-ray >100 MeV Luminosity (10 38 erg s -1, aka 2.5 _ 10 4 L _ ) 3 3 _ _ 10 5! Zombeck, M. V. 1990, Handbook of Astronomy and Astrophysics, Second Edition (Cambridge, UK: Cambridge University Press). LAT Dark Matter & Exotic Physics Group Workshop, July

16 Interstellar radiation field Interstellar radiation field: ε from COBE/DIRBE emissivities + detailed stellar model, k from extinction curves, grain albedo ISRF ( r, ν ) ε = e 2 r r MW ( r, ν ) k( r, ν ) ds dv See recent update by Porter & Strong (astro-ph/ ) Wavelength (mm) Strong, Moskalenko, & Reimer (2000) LAT Dark Matter & Exotic Physics Group Workshop, July Strong 2005 & Moskalenko 16

17 So how did the models work for EGRET? Three approaches for modeling Build in assumption of coupling to gas and derive 3-dim distribution of gas (e.g., Hunter et al. 1997) Fit the cosmic-ray density profile (e.g., Strong & Mattox 1996, and studies of individual clouds) Calculate the density using source distribution and codes for cosmic-ray propagation (e.g., Pohl & Esposito 1998; Strong et al. 2000) All reproduce large-scale structure observed in the diffuse emission differences generally are in contrast and at lower intensities Relative contributions depend on energy & latitude Comparison with EGRET Data Total π 0 Inverse Compton Bremsstrahlung Strong, Moskalenko, & Reimer (2004)* *Galprop, cf. talk by Igor Moskalenko LAT Dark Matter & Exotic Physics Group Workshop, July

18 Models for EGRET (cont) Pion bump and the well-known GeV excess π 0 shoulder M π 0/2 point sources subtracted l < 60 b < 10 Hunter et al. (1997) LAT Dark Matter & Exotic Physics Group Workshop, July

19 Recent work and open issues Special regions of the sky Cygnus in particular, and other spiral arm tangents source crowding, need detailed ISRF of OB associations? Cygnus Sagittarius 3 kpc arm Scutum NormaCrux Carina Galactic center loss of CO as a mass tracer Center and anticenter loss of kinematic distances Metallicity gradient in Milky Way Strong et al. (2005) Abundances of C and O relative to H are not uniform across the Galaxy Relation between N(H 2 ) and W CO should depend on abundances; although the relationship is complicated, N(H 2 )/W CO is certainly inversely related to metallicity LAT Dark Matter & Exotic Physics Group Workshop, July

20 Open issues (2) High latitude clouds (& variations of X) Torres et al. (2005) Dame & Thaddeus (2004) Real radiative transfer H I is not really optically thin - self-absorption of H I HVCs Dark interstellar gas Cosmic ray production and propagation LAT Dark Matter & Exotic Physics Group Workshop, July

21 Galactic Center (3EG J ) Recent re-analyses of EGRET data Source not coincident with the Galactic center itself Variable, too, although systematics are significant (Nolan et al. 2003) Many plausible candidates exist for the point source and its nature is of intense interest Many complications affect modeling the diffuse emision of this region Hooper & Dingus used the same diffuse emission model as EGRET team (Hunter et al. 1997) Complications go beyond the loss of kinematic distance resolution 3EG source confidence region CS (2-1) line Arches cluster (~150 O stars) Yusef-Zadeh (2002) Sgr A* Sgr A East >5 GeV γ-rays Hooper & Dingus (2002) 20 cm radio continuum LAT Dark Matter & Exotic Physics Group Workshop, July

22 Modeling diffuse emission in the Galactic center (cont.) Failure of CO as a tracer of molecular hydrogen in the Galactic center region Molecular clouds there have extremely broad velocity dispersions, and this makes the clouds much brighter in the (optically thick) CO line Also, higher-density tracers (e.g., CS 2-1, HC 3 N) CO (J = 1 0) C 18 O (J = 1 0) Galactic Center Bitran (1987) Dahmen et al. (1998) LAT Dark Matter & Exotic Physics Group Workshop, July

23 Dark interstellar gas The idea of cold gas too cold to see in CO or H I has been around since about 1993 (Pfenniger & Combes) You can have it 2 ways clumpy or diffuse In principle it is not a surprise that could have some molecular gas too diffuse or too cold to shine in CO. Unfortunately it is hard to get a direct handle on this component of the interstellar medium. Dense was suggested (e.g., Lequeux et al. 1993, Dixon et al. 1998, Walker et al. 2002) as a way to have a lot of the dark matter in the Milky Way be baryonic; the idea was that dense, cold gas would not shine in CO. A problem is how it doesn t collapse to form stars. Diffuse (e.g., Grenier, Casandjian, & Terrier 2005) doesn t have to add up to a lot of mass GCT s investigation of it is mostly as a residual left over when gamma-ray intensity (at high latitudes) is modeled as a linear combination of CO + H I + IC. High latitude part is important because that way you can assume that the CR density is uniform, pretty much. Implications of GCT result still need to be explored, e.g., for effect on isotropic emission and (as they point out) on EGRET point sources LAT Dark Matter & Exotic Physics Group Workshop, July

24 Cold, Dense Dark Clouds (BIG) Assumptions: Galactic dark matter is cold gas (i.e., not seen in emission or absorption somehow and stable against collapse) CDM-type clustering model clustering of the dark matter into mini-halos Consequences: Clumps will be gamma-ray sources (although not necessarily optically thin to cosmic rays) Cold Dark Gamma-Ray Sources Walker et al. (astro-ph/ ) Many would be EGRET point sources (i.e., detected but not resolved) Sources would be steady & without counterparts (although might be detectable in thermal microwave emission) Not strongly concentrated in the plane LAT Dark Matter & Exotic Physics Group Workshop, July

25 N γ _ diffus = (qhi N ( HI ) + qh 2W(CO ) + qdust I dust + qic I IC + I bkg )SA ΔT ΔΩ [ ] + = Grenier, Casandjian, & Terrier 2005 Page Number From: Jean-Marc Casandjian, Catalogs & Diffuse WG Meeting, May 2005

26 Metallicity and γ-ray models Metallicity gradient may be as steep as kpc -1 in log Z Z dependence of N(H 2 )/W CO may be as steep as Z -1 or Z -2.5 This gradient could explain the long-standing discrepancy between the cosmic-ray gradient inferred from modeling and that expected from the SN distribution Ref. Strong et al. (astroph/ ) N(H 2 )/W CO Inferred Distribution of CR Sources Lorimer (2004) from pulsars Gradient used in earlier work Case & Battacharya (1998) Inferred Variation of N(H 2 )/W CO LAT Dark Matter & Exotic Physics Group Workshop, July

27 Confusion with point sources At low latitudes especially, an inaccurate model will cause spurious detections or at least errors in positions of sources Historical example in Ophiuchus 2CG was a COS-B source, 90% conf. diam. 3 Flux ~10-6 cm -2 s -1 (>100 MeV) required an extreme enhancement of cosmic rays in in Ophiuchus (~10 4 M _, ~100 pc distant) EGRET marginally resolved Oph and found most of the flux to be from a blazar, with accurate position [& variability] CO (115 GHz) EGRET (>100 MeV) Diffuse Emission in Ophiuchus 2CG Swanenburg et al. (1981) Dame et al. (2001) PKS ! LAT Dark Matter & Exotic Physics Group Workshop, July Hunter et al. (1994)

28 Summary The diffuse emission of the Milky Way is bright and pervasive In high-energy γ-rays most of the luminosity of the MW is diffuse Owing to the limited area and angular resolution of the LAT, a good model of the diffuse emission will be required for accurate characterization of point sources It relates to the density of cosmic rays and the distribution of interstellar gas and radiation, which are themselves of interest to astronomers Models have been used since COS-B; refinements are being investigated for the LAT to maximize the scientific return EGRET Phases 1-5 LAT Sim. 1-yr LAT Dark Matter & Exotic Physics Group Workshop, July

29 Backup slides follow LAT Dark Matter & Exotic Physics Group Workshop, July

30 Angular resolution of the model What angular resolution is required for the LAT model of the interstellar emission? The consideration is that you don t want the angular resolution of the model to be the limiting factor for the determining the positions of point sources (or of detecting them). So one question is for a bright source observed at low latitude, how large does it have to be to be resolvable? Also, on what angular scale does the diffuse emission vary significantly? [The size of the PSF may not be directly relevant because we can localize sources to better than the size of the PSF] The question is probably on what angular scale does the diffuse emission vary by something of the order of the flux of a detectable point source (integrated over that angular scale)? Of course it also depends on the spectrum of the source and of the diffuse emission LAT Dark Matter & Exotic Physics Group Workshop, July

31 Isotropic background The updated model of interstellar emission by Strong et al. (2004) predicts a significantly lower isotropic intensity than the original analysis by the EGRET team (Sreekumar et al. 1998) Most of the difference is attributable to inverse Compton emission (reevaluation of ISRF including its anisotropy and electron spectrum) Spectrum of the Isotropic Emission Sreekumar et al. (1998) Strong, Moskalenko, & Reimer (2004) LAT Dark Matter & Exotic Physics Group Workshop, July

32 Production mechanisms for high-energy cosmic γ-rays Bremsstrahlung scattering of CR electrons by nucleons Inverse Compton scattering of lowenergy photons by CR electrons e π 0 decay secondaries from CR protonnucleon collisions with dn/de ~ A E -Γ Z e Bremsstrahlung Γ γ Γ e E γ <~E e e e Inverse Compton (Γ e + 1)/2 <<E e π 0 decay ~Γ p * * with peak at m π 0/2 = 68 MeV << E p p π 0 p LAT Dark Matter & Exotic Physics Group Workshop, July

33 GLAST Large Area Telescope (LAT) Spectrum Astro Within its first few weeks, the LAT will double the number of celestial gamma rays ever detected 5-year design life, goal of 10 years 1.8 m Tracker ACD EGRET AGILE AMS GLAST LAT Years ? 2007 Ang. Res. (100 MeV) Ang. Res. (10 GeV) Eng. Rng. (GeV) A eff _ (cm 2 sr) 1,500 1,600 25, _ _ 10 6 /yr 7 _ 10 5 /yr 1 _ 10 8 /yr e + e Calorimeter LAT Dark Matter & Exotic Physics Group Workshop, July # γ rays 3000 kg

34 GeV excess (2) Evident in Hunter et al. (1997) spectrum of the inner Galaxy Not obviously a problem with calibration or with calculated production functions (Mori 1997) The excess is clearly associated with the interstellar medium; the isotropic spectrum does not harden (Sreekumar et al. 1998) Is it a local phenomenon? Digel et al. (1996 & 2001) LAT Dark Matter & Exotic Physics Group Workshop, July

35 Galactic Diffuse Radiation & ISRF Inverse Compton emission - ISRF - recent progress ε ( ) ( r, ν ) k( r, ν ) ds ISRF r, ν = e dv 2 r r MW ε from COBE/DIRBE emissivities + detailed stellar model, k from extinction curves, grain albedo Application: Simulated SNR sources of cosmic rays Inverse Compton γ- ray longitude distributions for 3 SNR/century TeV GeV MeV Strong & Moskalenko LAT Dark Matter & Exotic Physics Group Workshop, July

36 Cosmic rays Measured locally Corrected for solar modulation Electrons Protons Barwick et al. (1998) Menn et al. (2000) LAT Dark Matter & Exotic Physics Group Workshop, July

37 Derived LAT Capabilities Point Source Sensitivity (5_, >100 MeV) Source Location Determination Splitting cm -2 s -1 sources Resolving cm -2 s -1 extended sources EGRET ~ cm -2 s min (7.5 max) cm -2 s -1 (at high gal. latitude for 1-year sky survey, for photon index of -2) 0.4 (1σ radius, flux 10-7 cm -2 s -1 >100 MeV, 1-year sky survey, high b) 6 5 LAT For flaring or impulsive sources the relative effective areas (~6x greater for LAT), FOV (>4x greater for LAT), and deadtimes (>3 orders of magnitude shorter for LAT) are relevant as well More fine print: E -2 sources, EGRET: 2-week pointed obs. on axis, LAT: 1-year sky survey, flat highlatitude diffuse background LAT Dark Matter & Exotic Physics Group Workshop, July