Gamma-ray emission from molecular clouds: A probe of cosmic-ray origin and propagation

Progress in Particle and Nuclear Physics 66 (2011) 681–685

Contents lists available at ScienceDirect

Progress in Particle and Nuclear Physics journal homepage: www.elsevier.com/locate/ppnp

Review

Gamma-ray emission from molecular clouds: A probe of cosmic-ray origin and propagation Sabrina Casanova ∗ Max Planck Institut für Kernphysik, Saupfercheckweg 1, 69117 Heidelberg, Germany Ruhr Universität Bochum, Universitätsstrasse 150, 44801 Bochum, Germany

article

info

Keywords: Cosmic rays Gamma-ray emission Cosmic-ray flux

abstract Cosmic rays up to at least 1015 eV (PeV) are believed to be emitted by Galactic sources, such as supernova remnants. However, no conclusive evidence of their acceleration has been found yet. A trace of ongoing cosmic-ray acceleration is the gamma-ray emission produced by these highly energetic particles when they scatter off the interstellar medium gas, mainly atomic and molecular hydrogen. Whereas the atomic hydrogen is uniformly distributed in the Galaxy, the molecular hydrogen is usually aggregated in dense clouds, and the gamma-ray emission from such clouds is particularly intense and localised. A multifrequency approach, which combines the data from the upcoming and future gamma-ray emissions with the data from the submillimeter and millimeter surveys of the molecular hydrogen, is therefore crucial to probe the Galactic cosmic-ray flux. In order to fully exploit this multi-frequency approach, one needs to develop predictions of the expected emission. Here we will discuss the GeV to TeV emission from runaway CRs penetrating molecular clouds close to the young supernova remnant RX J1713-3946 and in molecular clouds illuminated by the background cosmic-ray flux. © 2011 Published by Elsevier B.V.

1. The standard model of cosmic rays Cosmic rays (CRs) are the highly energetic protons and nuclei which fill the Galaxy and carry, at least in the vicinity of the Sun, as much energy per unit volume as the energy density of starlight or of the interstellar magnetic fields or the kinetic energy density of the interstellar gas. One hundred years after their discovery by the Austrian physicist Victor Hess, the origin of cosmic rays is still unclear. Diffusive shock acceleration in supernova remnants (SNRs) is the most widely invoked paradigm to explain the Galactic cosmic-ray spectrum. Galactic SNRs provide, in fact, the necessary power to sustain the Galactic cosmic-ray population. One expects about one supernova event every 30 years, and, in order to account for the energy density of cosmic rays (about 1 eV/cm3 ) and the cosmic-ray confinement time deduced from spallation, the typical non-thermal energy release per supernova has to be about 1050 ergs, which is about ten percent of the total energy released in supernova explosion [1]. This is in good agreement with the typical amount of energy predicted to be produced during the acceleration of relativistic particles in SNR shocks. If the bulk of Galactic CRs up to at least PeV energies are indeed accelerated in SNRs, then TeV γ -rays are expected to be emitted during the acceleration process CRs undergo within SNRs [2]. Indeed TeV γ -rays have been detected from the shells of SNRs, such as RX J1713.7-3946 [3]. However, such observations do not constitute a definitive proof that CRs are accelerated in SNRs, since the observed emission could be produced by energetic electrons up scattering low energy photon fields. The direct observation of cosmic rays from the candidate injection sites such as supernova remnants is not possible since CRs escape the acceleration sites and eventually propagate into the Galactic magnetic fields. CR secondary data suggest that

∗

Corresponding address: Max Planck Institut für Kernphysik, Saupfercheckweg 1, 69117 Heidelberg, Germany. E-mail address: [email protected].

0146-6410/$ – see front matter © 2011 Published by Elsevier B.V. doi:10.1016/j.ppnp.2011.01.029

682

S. Casanova / Progress in Particle and Nuclear Physics 66 (2011) 681–685

E cosmic protons and nuclei diffuse in the magnetic fields for timescales of the order of about tescape ≈ 107 ( 10
2. On the origin of cosmic rays Cosmic rays escaping supernova remnants diffuse in the interstellar medium and collide with the ambient atomic and molecular gas [6–9]. From such collisions gamma rays are created, which can possibly provide the first evidence of a parent population of runaway cosmic rays. Before being isotropised by the Galactic magnetic fields, the injected CRs produce, in fact, γ -ray emission, which can significantly differ from the emission of the SNR itself, as well as from the diffuse emission contributed by the background CRs and electrons, because of the hardness of the runaway CR spectrum, which is not yet steepened by diffusion. These diffuse sources are often correlated with dense molecular clouds (MCs), which act as a target for the production of γ -rays due to the enhanced local CR injection spectrum. SNRs are located in star forming regions, which are rich in molecular hydrogen. In other words, CR sources and MCs are often associated and target–accelerator systems are not unusual within the Milky Way [10,6]. γ -rays in association with dense MCs close to candidate CR sources, have been detected both at GeV energies [11–14] and TeV energies [15–19]. The γ -ray radiation from hadronic interactions from regions close to CR sources depends not only on the total power emitted in CRs by the sources and on the distance of the source to us, but also on the ambient interstellar gas density, the local diffusion coefficient and the injection history of the CR source. It is therefore difficult to definitely recognize the sites of CR acceleration from γ -ray observations alone, since very often only qualitative predictions are provided, rather than robust quantitative predictions, especially from a morphological point of view. In order to fully exploit the present and future experimental facilities and to test the standard scenario for CR injection in SNRs and propagation, we computed the expected γ -ray emissivity from hadronic interactions of runaway CRs in a region around the SNR RX J1713.7-3946, by constructing as quantitative a model as possible. In particular, we have conveyed all information concerning the environment, the source age, the acceleration rate and history, which all play a role in the physical process of CR injection and propagation [20]. Fig. 1 shows the γ -ray spectra from hadronic interactions of background and runaway CRs in four regions around the SNR RX J1713.7-394. In all four locations the hadronic γ -ray emission is enhanced with respect to the hadronic emission due only to background CRs at energies above few TeVs as a consequence of the fact that the CR fluxes are enhanced above 100 TeV. CRs of about 100 TeV are, in fact, thought to be released now by RX J1713.7-394 [21]. The hadronic γ -ray emission from the region d is particularly enhanced with respect to the hadronic emission from background CRs. This is due to two factors, the enhanced CR flux nearby the injection source and the particularly high ambient gas density. In the three regions a, b and d the high energy emission is clearly dominated by the radiation produced along the line of sight distance between 900 and 1100 parsecs, plotted in dashed lines in Fig. 1. The γ -ray emission from high latitude regions, such as region c, is instead

S. Casanova / Progress in Particle and Nuclear Physics 66 (2011) 681–685

a

b

1e-11

1e-12

c

1e-11

d

1e-11

1e-12

1e-11

1e-12

1e-15

1e-16 0.001

1e-15

0.01

0.1

1

Energy (TeV)

10

100

-1

-2

-2

1e-13

1e-14

1e-16 0.001

1e-15

0.01

0.1

1

Energy (TeV)

10

100

1e-13

1e-14

2

-1

1e-14

2

1e-14

1e-13

2

1e-13

E dN/dE ( TeV s cm )

E dN/dE ( TeV s cm )

-2 -1

E dN/dE ( TeV s cm

E 2 dN/dE ( TeV s-1 cm -2 )

)

1e-12

683

1e-16 0.001

1e-15

0.01

0.1

1

Energy (TeV)

10

100

1e-16 0.001

0.01

0.1

1

10

100

Energy (TeV)

Fig. 1. The γ -ray energy flux in four different regions of 0.2° × 0.2° around the positions a = (346.8, −0.4), b = (346.9, −1.4), c = (347.1, −3.0) and d = (346.2, 0.2). The emission produced along the line of sight distance between 900 and 1100 parsecs is plotted in dashed lines, while the emission obtained by summing the radiation contributions over the whole line of sight distance, from 50 parsecs to 30 000 parsecs, is shown in solid lines. The emission in the panels a, b, c and d, corresponding to the different locations, is plotted for different diffusion coefficients D0 . The emission from background CRs is also shown in each panel for comparison. The contribution to the emission from inverse Compton scattering of background electrons is indicated with a dashed light blue line. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. Ratio of the emission due to the sum of background CRs and runaway CRs and background CRs only. The SNR is supposed to have exploded at 1 kpc distance from the Sun at 347.3° longitude and −0.5° latitude 1600 years ago and to have started injecting the most energetic protons 100 years after the explosion. The diffusion coefficient assumed within the region 340° < l < 350° and −5° < b < 5° is 1026 cm2 /s in the left panel, 1027 cm2 /s in the middle panel and 1028 cm2 /s in the right panel. In the three panels the ratio of the emission is shown for different energies from 1 GeV to 10 TeV.

dominated by the contribution from IC scattering of background electrons, almost at all energies. In regions closer to the Galactic Plane the emission from inverse Compton scattering of background electrons is subdominant at TeV energies, where runaway cosmic rays produce the enhanced emission. Therefore the regions where to look for the emission from runaway particles are low latitude regions of higher gas density. The γ -ray spectra show a peculiar concave shape, being soft at low energies and hard at high energies, which, as discussed in [9], might be important for the studies of the spectral compatibility of GeV and TeV gamma-ray sources. The peculiar spectral and morphological features of the γ -ray due to runaway CRs can be therefore revealed by combining the spectra and γ -ray images provided by the Fermi and Agile telescopes at GeV and by present and future ground based detectors at TeV energies. As shown by the surveys of the Galaxy, published by Fermi at MeV–GeV energies, by HESS at TeV energies and at very high energies by the Milagro Collaboration, the various extended Galactic sources differ in spectra, flux and morphology. However, there is growing evidence for the correlation of GeV and TeV energy sources. These sources appear often spectrally and morphologically different at different energies, possibly due not only to the better angular resolution obtained by the instruments at TeV energies, but also to the energy dependence of physical processes, such as CR injection and CR diffusion. For this reason it is important to properly model what we expect to observe at different energies by conveying in a quantitative way all information by recognizing that the environment, the source age, the acceleration rate and history, all play a role in the physical process of injection and all have to be taken into account for the predictions. Fig. 2 shows the ratio of the hadronic gamma-ray emission due to total CR spectrum to that of the background CRs for the entire region under consideration. In our modeling only CRs with energies above about 100 TeV have left the acceleration site and the morphology of the emission depends upon the energy at which one observes the hadronic gamma-ray emission. The different spatial distribution of the emission is also due to the different energydependent diffusion coefficients, assumed in the three different panels, making it into a useful tool to investigate the highly unknown CR diffusion coefficient [20,22].

684

S. Casanova / Progress in Particle and Nuclear Physics 66 (2011) 681–685

3. On the level of the CR background The cosmic-ray flux measured close to the Earth, with its characteristic soft −2.7 spectrum, is usually assumed to be representative of the average cosmic-ray flux throughout the Galaxy. By combining the locally observed spectrum and the propagation effects, one would conclude that the injection spectra from individual sources have a power-law slope close to −2, in accordance with the predictions of the diffusive shock acceleration theory. However, on the basis of the state of the art knowledge of CR propagation processes, a somehow steeper source spectrum up to −2.4 cannot be excluded assuming that CRs might re-accelerate in the Galaxy [2]. Also, for protons and nuclei a good working hypothesis to interpret their locally measured energy spectrum is to assume that the CR sources are distributed uniformly and continuously in the Galaxy and all contribute to the locally observed cosmic-ray flux. However, in principle, one cannot exclude that the cosmic-ray spectrum in the immediate vicinity of the Sun is due to a single local source [23], and that the level of the cosmic-ray sea in other locations of the Galaxy might be different from the one observed locally. Direct measurements provide the CR spectrum and flux only in the vicinity of the Solar System and the level of the CR spectrum and flux in other regions of the Galaxy is unknown. Interestingly, based on EGRET data [24] concluded that the γ -ray emission from the inner Galaxy at energies greater than about 1 GeV exceeds, by some 60%, the intensity predicted by model calculations, in which the average Galactic CR flux is equal to the one measured locally [25,24]. Recent observations, performed by the LAT instrument on board, the Fermi satellite, show that the spectra of the Galactic diffuse emission at MeV–GeV energies, at least at intermediate latitudes and in the Outer Galaxy [26,11], can be explained by cosmic-ray propagation models based on local observations of cosmic-ray electron and nuclei spectra. On the other hand, at TeV energies the spectral features of the γ -ray emission detected by HESS from the Galactic centre (GC) region [3], and the emission measured by Milagro from the Cygnus Region [27,28], suggest that the CR flux might significantly vary in the different locations of the Galaxy. The emissivity of molecular clouds located far away from CR sources, so called passive molecular clouds, i.e. clouds which are illuminated by the supposedly existing CR background, can be used to probe the level of the CR sea [7]. Given that the γ -ray-emission from the molecular cloud depends only upon the total mass of the cloud, M, and its distance from the Earth, d, the CR flux, ΦCR , in the cloud is uniquely determined as

ΦCR ∝

Fγ d2 M

(1)

where Fγ is the integral γ -ray flux from the cloud. Under the assumption that the CR flux in the cloud is equal to the locally observed CR flux, the calculated γ -ray flux from the cloud can be compared to the observed γ -ray flux in order to probe the CR spectrum in distant regions of the Galaxy. The detection of under-luminous clouds with the respect to predictions based on the CR flux at Earth would suggest that the local CR density is enhanced with respect to the Galactic average density. This would cast doubts on the assumption that the local CRs are produced only by distant sources, and that the CR flux and spectrum measured locally is representative of the typical CR flux and spectrum present throughout the Galaxy. In what follows we will describe a methodology to test the CR flux in distant molecular clouds. The longitudinal profile of the γ -ray emission due to protons scattering off the atomic [29] and molecular hydrogen [30,31] from the region which spans Galactic longitude 340° < l < 350° and Galactic latitude −5° < b < 5° is shown in Fig. 3 (Left). A peak in the emission at longitude of about 345.7° close to the Galactic Plane is clearly visible, next to a dip in the longitudinal profile. While the atomic gas is generally broadly distributed along the Galactic Plane, the molecular hydrogen is less uniformly distributed and the peaks in the γ -ray longitudinal profiles correspond to the locations of highest molecular gas column density. The peaks in the longitudinal profile reveal the directions in the Galaxy where massive clouds associated with spiral arms are aligned along the line of sight. Fig. 3(Right) shows that the peak in the γ -ray emission from the direction 345.7° close to the Galactic plane is mostly produced within 0.5 and 3 kpc distance from the Sun. The contribution to the emission from the atomic hydrogen in the 0.5 kpc < d < 3 kpc amounts to about 10% of the total contribution from atomic and molecular hydrogen. Thus the γ -ray emission from this direction provides a unique probe of the CR spectrum in 0.5–3 kpc [32]. 4. Conclusions Here we have presented a methodology to test the cosmic-ray flux in discrete distant regions of the Galaxy by combining the most recent high resolution γ -ray and gas data. Through the knowledge of the mass and distance of distant molecular clouds we have predicted their γ -ray emissivity under the assumption that the CR flux in the clouds is equal to that measured at Earth. By comparing the predicted and the measured γ -ray flux one can obtain a test of the level of the CR flux in distant regions of the Galaxy. The emission from the regions surrounding young SNR shells can provide crucial information on the history of the SNR acting as a CR source and important constraints on the highly unknown diffusion coefficient. Detailed modeling of the energy spectra and of the spatial distribution of the gamma-ray emission in the environment surrounding RX J1713.7-3946 have been presented by using the data from atomic and molecular hydrogen surveys. These predictions show that the age and acceleration history of the SNR, the particle diffusion regime and the distribution of the ambient gas are all paramount. Also, depending on the time and energy at which one observes the remnant and the surrounds, one

S. Casanova / Progress in Particle and Nuclear Physics 66 (2011) 681–685

2

0.045 0.04 0.035 0.03 0.025 0.02

0.45

× 10

0.4

HI H2

0.3 0.25 0.2 0.15 0.1 0.05 340

HI

342 344 346 348 350 Galactic Longitude [deg]

-8

10

H2

0.14

sum

0.12

HI

0.1

H2

0.08 0.06 0.04 0.02 340

1 TeV

-6

sum

0.35

sum

0.16

342 344 346 348 350 Galactic Longitude [deg]

-1

-1

2 Gamma–Ray Profile [ cm s -1 sr -1 TeV ]

100 GeV

340

2 Gamma–Ray Profile [ cm s -1 sr -1 TeV ]

0.015

-3

-1

H2

× 10

0.18

(photons cm-2 s-1 kpc )

HI

0.05

Gamma–Ray Profile [ cm s -1 sr -1 TeV -1 ]

10 GeV sum

0.055

2

-1

Gamma–Ray Profile [ cm s -1 sr -1 TeV ]

1 GeV

685

× 10

342 344 346 348 350 Galactic Longitude [deg]

-9

1

sum HI

0.8

H2

-9

10

-10

10

0.6 0.4 0.2 -1

10 340

342 344 346 348 350 Galactic Longitude [deg]

1 Distance (kpc)

10

Fig. 3. (Left) The longitudinal profile of the γ -ray emission from the region 340° < l < 350°, integrated over the latitude range −5° < b < 5°. The emission is convolved with the energy-dependent PSF of Fermi at 1 and 10 GeV, and of a future Cherenkov telescope such as CTA at 100 GeV and 1 TeV. The dotted green lines are for the emission arising from CRs scattering off molecular hydrogen, the dashed red ones for the emission arising from CRs scattering off atomic hydrogen and the solid black ones for the sum. (Right) Flux above 1 GeV as a function of the line of sight distance, produced by CRs interacting with the atomic and molecular gas [32]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

will observe different spectra and morphologies. This has important implications for the current and future generations of gamma-ray observatories. The high sensitivity and high resolution, which will be reached by future detectors, such as AGIS, CTA and HAWC (e.g. see reviews [4,5]), makes the detection of the predicted emission possible. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32]

V.L. Ginzburg, S.I. Syrovatskii, Origin of Cosmic Rays, Pergamon Press, London, 1964. L.O’C. Drury, F.A. Aharonian, H.J. Völk, Astronomy and Astrophysics 287 (1994) 959. F.A. Aharonian, et al., [H.E.S.S. Collaboration], Nature 432 (2004) 75. F. Aharonian, J. Buckley, T. Kifune, G. Sinnis, Reports on Progress in Physics 71 (2008) 9. J.A. Hinton, W. Hofmann, Annual Review of Astronomy and Astrophysics 47 (2009) 523. T. Montmerle, Astrophysical Journal 231 (1979) 95. M.R. Issa, A.W. Wolfendale, Nature 292 (1981) 430. F.A. Aharonian, A.M. Atoyan, Astronomy and Astrophysics 309 (1996) 917. S. Gabici, F.A. Aharonian, S. Casanova, Monthly Notices of the Royal Astronomical Society 396 (2009) 1629. J. Paul, M. Casse, C.J. Cesarsky, Astrophysical Journal 207 (1976) 62. A.A. Abdo, et al., [Fermi Collaboration], Astrophysical Journal 703 (2009) 1249A. A.A. Abdo, et al., [Fermi Collaboration], Science 327 (2010) 1103. M. Tavani, et al., Astrophysical Journal 710 (2010) L151. D. Castro, P. Slane, Astrophysical Journal 717 (2010) 372C. Albert, et al., Astrophysical Journal 664 (2007) 87. F.A. Aharonian, et al., Astronomy and Astrophysics 481 (2008) 401. F.A. Aharonian, et al., Astronomy and Astrophysics 483 (2008) 509. F.A. Aharonian, et al., Astronomy and Astrophysics 490 (2008) 685. F.A. Aharonian, et al., [H.E.S.S. Collaboration], Nature 439 (2006) 695. S. Casanova, et al., Publications of the Astronomical Society of Japan 62 (2010) 1127. V.N. Zirakashvili, F.A. Aharonian, Astrophysical Journal 708 (2010) 965Z. S. Gabici, S. Casanova, F.A. Aharonian, G. Rowell, Preprint arXiv:1009.5291. A.D. Erlykin, A.W. Wolfendale, Journnal of Physics G (Nuclear and Particle Physics) 23 (1997) 979. S.D. Hunter, et al., Astrophysical Journal 481 (1997) 205. D.L. Bertsch, et al., Astrophysical Journal 416 (1993) 587. A.A. Abdo, et al., [Fermi Collaboration], Physical Review Letter 103y1101A (2009). A.A. Abdo, et al., [Milagro Collaboration], Astrophysical Journal L658 (2007) 33. A.A. Abdo, et al. [Milagro Collaboration], Astrophysical Journal, 688, 1078A. P.M.W. Kalberla, et al., Astronomy and Astrophysics 440 (2005) 775. Y. Fukui, et al., Publications of the Astronomical Society of Japan 51 (1999) 6. Y. Fukui, et al., Publications of the Astronomical Society of Japan 53 (2001) 6. S. Casanova, et al., Publications of the Astronomical Society of Japan 62 (2010) 769.