Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.atnf.csiro.au/pasa/15_1/ferrara/paper.pdf Дата изменения: Thu May 14 08:27:12 1998 Дата индексирования: Sun Apr 10 04:27:02 2016 Кодировка: Поисковые слова: dust disk

Publ. Astron. Soc. Aust., 1998, 15, 19­23
.

Probing ISM Mo dels with H Observations
Andrea Ferrara
Osservatorio Astrofisico di Arcetri, Largo. E. Fermi 5, Firenze, Italy, 50125 ferrara@arcetri.astro.it Received 1997 September 8, accepted 1997 November 4

Abstract: I review the capabilities of H observations to constrain some aspects of the current models of the interstellar medium. In particular, it is shown that turbulence is a necessary ingredient of any viable model, since most of the energy produced by supernova explosions and ionising radiation is stored in kinetic form in the ISM. Various forms of turbulent energy dissipation, including cloud collisions, are analysed. Two additional aspects, concerning the existence of galactic fountains and their relation with high-velocity Clouds, and the extended ionised layer of spiral galaxies are discussed; some crucial experiments are suggested. Keywords: interstellar medium -- turbulence

1 Relevance of H Observations for the ISM In spite of the impressive observational and theoretical advances that have occurred in the last decade, there are still a number of important issues concerning the structure and evolution of the Galactic interstellar medium (ISM) which require more study. Among these, three particularly relevant issues are discussed below. First, one would like to know to what extent the energy budget of the ISM is regulated by turbulence. Energy in this form is certainly supplied to the gas by the concurrent action of supernova explosions, winds from massive stars and HII regions. However, it is a difficult problem to ascertain the relative powers of these sources and the total amount of energy stored in different forms (i.e. thermal, radiative, magnetic and cosmic ray energy). Next, we understand that the disk of the Milky Way is not a closed system but is connected to the Galactic halo by different forms of energy transfer involving mass entrainment (particularly hot gas but also cold gas and dust) and photons leaking from the production regions. It is therefore important, both for energetic and evolutionary (chemical and dynamical) considerations, to assess the relevance of these disk/halo interactions, which are often referred to in brief as the Galactic Fountain. Finally, after more than 10 years since its firm discovery, the mystery of the existence of a vertically extended, ionised gas distribution, the so-called `Reynolds layer' has not yet been dispelled. The aim of this work is to show how H observations can help us in making progress in each of these areas. 2 Turbulent Models of the ISM Massive stars are probably the most important energy sources for the ISM. They inject power in both radiative (with ionising photons creating
Astronomical Society of Australia 1998

HII regions) and mechanical (supernova explosions) forms. The rate of kinetic energy density deposited via photoionisation is
( Wuk) = (2)

Ї n2 E2 e
25

1 · 9 в 10- erg cm
-3

n2 e 0 · 1 cm

-6

s

-1

;

(1)

the analogous quantity for a supernova explosion is W
(k) s

2 · 2 в 10- в VG 78 kpc3

25

0 · 04 yr- erg cm

1

-1

-3

s

-1

,

(2)

where is the total (both Type I and II) supernova rate and VG is the volume of the Galaxy (McKee 1990; Ricotti, Ferrara & Miniati 1997). Assuming efficiencies of kinetic energy conversion into the ISM of u 1% and s 3% (Spitzer 1978) for radiative and mechanical input, respectively, we find the total rate per unit volume at which energy available for motions in the ISM is produced: (k) (k) (k) Wi = u Wu + s Ws 8 в 10-27 erg cm-3 s-1 . Kinetic energy is mostly dissipated by cloud collisions at a rate W
(k) c

2 · 2 в 10-27 n в ti 1 · 1 в 107 yr
(k)

vr 14 km s
-1

-1 -3 -1

erg cm

s

0 · 25Wi

,

(3)
1323-3580/98/010019$05.00


20

A. Ferrara

where n, vr and ti are the typical cloud density, relative velocity and the mean time interval between collisions, respectively. Thus there is clearly enough energy production to support the observed motions in the ISM.

out by Norman & Ferrara (1996). The authors calculated the detailed grand source function (shown in Figure 2) for the conventional sources of turbulence from supernovae, superbubbles, stellar winds and HII regions. As seen from Figure 2, superbubbles are the main contributors to interstellar turbulence. In addition, from the study of the general properties of the turbulent spectrum, using an approach based on a spectral transfer equation derived from the hydrodynamic Kovasznay approximation, they conclude that the turbulent pressure calculated from the grand source function is Pturb 10­100Pg . Also, given the scale-dependent energy dissipation from a turbulent cascade, the multi-phase medium concept has to be generalised to a more natural continuum description where density and temperature are functions of scale.

Figure 1--Thermal pressure contributions Pg,i to the total pressure Ptot (determined from hydrostatic equilibrium) from the known gaseous phases of the ISM, where i = wn, wi, h indicate warm neutral, warm ionised, hot components, respectively. The difference Ptot - Pg shows the importance of non-thermal (turbulent, magnetic, cosmic ray) forms of energy density.

These bulk motions have a strong impact on the structure of the Galactic ISM. For example, they are shown to largely regulate and reproduce the vertical distribution of the HI in the gravitational potential of the Galaxy (Lockman & Gehman 1991), once the effect of radiation pressure on dust grains embedded in clouds (the so-called `photolevitation', Ferrara 1993) is properly taken into account. In addition, turbulence may be the most important form of energy storage in the ISM, as can be appreciated from Figure 1: the thermal pressure contributed by all known gaseous ISM phases, Pg , appears to be at most 13% of the total pressure, Ptot , as derived by imposing gas hydrostatic equilibrium in the Galactic gravitational field. Such a large ratio of the turbulent to thermal pressure implies a large porosity factor, Q, of the hot gas. McKee (1990) estimated that 4 Q 1 · 1(Pturb /Ptot ) 3 , implying Q = 0 · 92 from the above estimates if the nonthermal energy is predominantly in turbulent form; this corresponds to a hot gas filling factor f 0 · 6. Given these arguments it seems necessary to revise the current ISM models to include turbulence. A first attempt in this direction has been carried

Figure 2--Normalised turbulent source functions Si (k) for supernovae, superbubbles, winds and HII regions. Solid lines show the sum of primary and secondary shock contributions for each source; dashed lines show secondary shocks only (see Norman & Ferrara 1996). The thick line is the total grand source function.

As recalled above, cloud collisions represent the most efficient dissipation mechanism of large-scale (k) given turbulence. The simple estimate for Wc above assumes that al l the kinetic energy of the clouds is radiated away by the post-shock gas, i.e. an inelastic collision. This hypothesis is correct only in a restricted region of the collision parameters (velocity and mass ratio of the colliding clouds, magnetic field strength, gas metallicity). Ricotti et al. 1997 have studied the dependence of the elasticity (defined as the ratio of the final to the initial kinetic energy of the clouds) on such parameters (recently extended to


Probing ISM Models

21

include pre-interaction with the intercloud medium by Miniati et al. 1997). They find that (i) the collision elasticity is a maximum for a cloud relative velocity vr 30 km s-1 ; and (ii) the elasticity is ZL2 , where Z is the metallicity and Lc is the cloud c size: the larger ZL2 , the more dissipative (inelastic) c the collision will be. During the collision the warm post-shock gas will radiate a substantial fraction of its internal energy in the H line depending on vr and Lc . Figure 3 shows the H luminosity for a collision occurring 1 kpc away from us for different values of vr and Lc . Thus H observations can be used in principle as a powerful indicator of large-scale turbulent motion dissipation.

Figure 4--IR surface brightness from dust in HVCs with HI column density NHI = 5 в 1019 cm-2 . Numbers refer to fixed grain temperatures (solid) or radiation flux intensity (dotted), where F is the Galactic ISRF. Also shown are the Wakker & Boulanger (1986) limits and ISOPHOT sensitivity (see text).

Figure 3--H luminosity evolution from cloud collisions at a distance of 1 kpc as a function of time in units of collision time (tc = Lc /vr ). The upper (lower) group of lines refers to a cloud size Lc = 1 pc (Lc = 0 · 1 pc) for different values of the relative velocity of the collision: vr = 6 km s-1 (solid), vr = 16 km s-1 (long-dashed), vr = 31 km s-1 (short-dashed) and vr = 63 km s-1 (dotted).

3 Does a Galactic Fountain exist? One of the ma jor predictions and elements in favour of the Galactic Fountain (GF) model (Shapiro & Field 1976; subsequently detailed by Bregman 1980) is the existence of the high velocity clouds (HVCs, recently reviewed by Wolfire et al. 1995) in the halo, as a by-product of the cooling of the hot fountain gas. Any constraint on the origin of HVCs would be highly valuable in terms of understanding the global disk/halo circulation. Ferrara & Field (1994) have investigated the ionisation and thermal structure of HVCs due to the extragalactic background radiation field and calculated the H emission from the partially ionised edge of a given cloud. Comparing the model results to the available H observations, the authors found that the observed H intensity is larger than the predicted one for all the different cases

considered. One possible cause of this discrepancy is that some of the observed emission is H light coming from the disk of the Galaxy and subsequently backscattered by dust in the clouds. Assessing the presence of dust in HVCs is a young research field and very little is known about this issue. However, there are theoretical bases on which to expect dust to be present. The efficiency of dust destruction in a shock (presumably the heating source for the fountain gas) is likely to be less than 10% (McKee 1989); the subsequent thermal sputtering in the hot gas might destroy some of the smallest grains: in a 106 K gas, the sputtering time is shorter than the gas cooling time only for grain sizes smaller than 0 · 03 µm. However, previous low-sensitivity searches using IRAS data have produced only upper limits (Wakker & Boulanger 1986) on HVCs, dust contents. In addition to the unlikely complete dust depletion, there are at least two other possibilities which might explain the non-detection: (i) HVCs are far above the Galactic plane (10 kpc); and (ii) the dust is too cold to emit substantially in the IRAS bands. Since, at least for some of the clouds, the distance is bracketed around a much lower value for the distance, (i) is an unlikely explanation. It is quite possible, instead, that the dust is cold, due to the very diluted halo radiation field. The expected IR emission from a NHI = 5 в 1019 cm-2 HVC, assuming either a given temperature of the grains or a galactic ISRF diluted by a given factor (1­100 times) is shown in Figure 4 together with ISOPHOT surface brightness limits (128 s of integration time, S/N = 10) in the various bands. Clearly, dust hotter than 10 K and/or heated by an ISRF diluted by at


22

A. Ferrara

Figure 5--The z -axis cuts through the centre of NCG891 for (left column) the H luminosity profiles corresponding to contributions of HII regions and DIG component, (middle) the ratio of the two components, and (right) the profiles of the linear polarisation degree. Plots refer to the cases of x = 0 · 7, 0 · 6, 0 · 5 (from top to bottom), where x is the ratio between H luminosity from H II regions to the total (DIG+H II regions) luminosity.

most 10 times could be detectable. Such a detection would bring strong support to the idea that HVCs are not of intergalactic origin and that they are very likely part of a global disk/halo circulation. 4 H from the Galactic DIG The thick (z 1 kpc) Reynolds layer of diffuse ionised gas (DIG) discovered in the Galaxy and in external ones poses some of the most challenging problems for our understanding of the large-scale structure of the Galactic ISM (Reynolds 1995). One of the best ways to study this component is represented by spectroscopic observations of emission ° lines like H, [NI I ](6583 ° and [SI I ]( 6716 A). A) The different excitation conditions found in a given galaxy as a function of height above the plane pose a relevant question concerning the amount of light originating in the disk (where the most obvious

ionisation sources are located) and light scattered back into the line of sight by dust. This aspect has been investigated recently by Ferrara et al. (1996), using Monte Carlo simulations to calculate the radiation transfer of H line emission, produced both by HII regions in the disk and in the DIG, through the dust layer of the galaxy NGC891. The amount of light originating in the HII regions of the disk and scattered by EGD can be then compared with the emission produced by recombinations in the DIG. The cuts of photometric and polarimetric maps along the z -axis show that scattered light from HII regions is still 10% of that of the DIG at z 600 pc (Figure 5), whereas the degree of linear polarisation is small ( < 1%). This could explain the observed behaviour of emission line ratios as a function of height (Ferrara et al. 1996).


Probing ISM Models

23

Acknowledgments I deeply thank all my collaborators who have been part of this pro ject: S. Bianchi, R. Dettmar, G. Field, C. Norman and M. Ricotti.
Bregman, J. N. 1980, ApJ, 236, 577 Ferrara, A. 1993, ApJ, 407, 157 Ferrara, A., & Field, G. B. 1994, ApJ, 423, 665 Ferrara, A., Bianchi, S., Dettmar, R.-J., & Giovanardi, C. 1996, ApJL, 467, 69 Lockman, F. J., & Gehman, C. S. 1991, ApJ, 382, 182 McKee, C. F. 1989, in Interstellar Dust, IAU Symp. 135, ed. L. J. Allamandola & A. G. G. M. Tielens (Dordrecht: Kluwer), p. 431

McKee, C. F. 1990, in The Evolution of the Interstellar Medium, ed. L. Blitz, (San Francisco: PASP), p. 3 Miniati, F., Jones, T. W., Ferrara, A., & Ryu, D. 1997, ApJ, 491, 216 Norman, C. A., & Ferrara, A. 1996, ApJ, 467, 280 Reynolds, R. J. 1995, in The Physics of the Interstellar Medium and Intergalactic Medium, ASP Conf. Ser. 80, ed. A. Ferrara et al. (San Francisco: ASP), p. 388 Ricotti, M., Ferrara, A., & Miniati, F. 1997, ApJ, 485, 254 Shapiro, P. R., & Field, G. B. 1976, ApJ, 205, 762 Spitzer, L. 1978, Physical Processes in the Interstellar Medium (New York: Wiley) Wakker, B. P., & Boulanger, F. 1986, A&A, 170, 84 Wolfire, M. G., Hollenbach, D., McKee, C. F., Tielens, A.G.G.M., & Bakes, E. L. O. 1995, ApJ, 443, 152