Next Contents Previous

5. Photoionization Models for a System of Clouds

Having calculated the emergent spectrum of isolated clouds, we now consider the gas distribution and the different ways to combine the emission from such clouds into a composite spectrum.

5.1. Spherical BLR Models

Consider a spherically symmetric system of clouds, extending from rin to rout. The mass of the individual clouds is conserved, but it is not necessarily the same for all clouds. The following physical parameters are all functions of r: the cloud number density nc(r), the cloud particle density N, the cloud emissivity jc(r), and the cloud velocity v(r).

Most cases of interest involve optically thick clouds, where energy conservation requires that the total emergent line and diffuse continuum luminosity equal the energy absorbed by the gas. In this case, the geometrical cross section of the cloud as seen from the center, Ac(r), is another useful parameter.

Designate epsilonl(r) as the flux emitted by the cloud, in a certain emission line l, per unit projected surface area (erg s-1 cm-2), we have the following relation for the cloud emission:

Equation 49 (49)

We now make the following simplified assumptions about the radial dependence of the different parameters:

Equation 50 (50)
Equation 51 (51)
Equation 52 (52)
Equation 53 (53)


Equation 54 (54)

Using these definitions, the radial dependence of the ionization parameter is:

Equation 55 (55)

5.1.1. Covering factor. For optically thick clouds, it is useful to introduce the covering factor, C(r), which is the fraction of sky covered by photoionized gas clouds, as seen from the center. For a single cloud

Equation 56 (56)

and for a thin shell of thickness dr

Equation 57 (57)

The integrated covering factor can be estimated from the Lyman continuum observations (chapter 2) or from comparing the line and continuum luminosities (chapter 6). For the more luminous AGNs it is about 0.1 for the BLR and less than that for the NLR. It is thus justified to neglect obscuration and proceed on the assumption that the flux reaching the clouds depends only on r.

5.1.2. Integrated line fluxes. Given the calculated epsilonl(r) for individual clouds (chapter 4), we can now integrate over r to obtain cumulative line fluxes. In the most general case

Equation 58 (58)

There are interesting cases where the number of radius dependent parameters is considerably reduced. An important example is a system of spherical clouds, of radius Rc(r), in pressure equilibrium with a confining medium of pressure P. The kinetic temperature of a photoionized gas is only a weak function of U, and we can safely assume that the radial dependences of P and N are identical (propto r-s). Since R3c N = const., the cloud cross-section is

Equation 59 (59)

(i.e. s = -3/2q) and the column density is

Equation 60 (60)

Assume further that the clouds are moving in or out with their virial velocity. Mass conservation (ncvr2 = const.) requires that p = 2-t and substituting t = 1/2 we get the following radial dependence of the covering factor:

Equation 61 (61)

General considerations (chapter 9) suggest that 1 leq s leq 5/2 in many cases of interest. In particular, s = 9/4 corresponds to a case where the covering factor of a thin shell is proportional to the shell thickness.

The total flux, in all forms of emission, from a shell of thickness dr is proportional to dC(r) (4), and the cumulative flux is proportional to

Equation 62 (62)

This is, in fact, a good approximation for many individual lines (i.e. m = 2 is a good approximation for many lines).

We conclude that in those cases with s > 3/4, the cumulative covering factor, and therefore the integrated line emission, increases rapidly outward, and the innermost parts of the atmosphere do not contribute much to the line emission. For s leq 3/4 much of the line emission takes place close to the center and there is a natural boundary to the cloud distribution. Note, however, some potential difficulties. For example, the column density of ionized material depends on U and r, and clouds that are optically thick near rout may become transparent closer in, or vice versa.

5.1.3. Specific models. Fig. 11 shows cumulative BLR line fluxes, as a function of rout for a constant ionization parameter model, s = 2. They were calculated as explained in chapter 4, and added according to the prescription explained above. In this particular case, L(ionizing continuum) = 1046 erg s-1 U = 0.3 for all clouds and the normalization is such that Ncol = 2 × 1023 cm-2 where N = 1010 cm-3. The integration starts at rin = 1016.5 cm, where the density is 1012 cm-3, and carried out to large radii. Vertical cuts in the diagram correspond to a certain rout and give the integrated line fluxes up to that radius.

Figure 11

Figure 11. Cumulative broad line fluxes, as a function of rout, for a model with U = 0.3 and s = 2. The model is calculated for the continuum shown in Fig. 7, assuming L(ionizing luminosity) = 1046 erg s-1. the value of rout for other luminosities is obtained by noting that rout propto L1/2.

The narrow line fluxes can be integrated in a similar way, and Fig. 12 shows a model with s = 1, normalized such that U = 3.6 × 10-2 where N = 105 cm-3 and Ncol = 1022 cm-2. Here again L(ionizing luminosity) = 1046 erg s-1, the inner radius is at 1019.5 cm and the density there is 106 cm-3. Note that here, and in the BLR model of Fig. 11, the line intensity shown is normalized in such a way that the covering factor is unity at the outermost radius shown.

Figure 12

Figure 12. Cumulative narrow line fluxes, as a function of rout, for a model with s = 1 and parameters as explained in the text.

4 The radial dependence for individual lines is different, unless m = 2. Back.

Next Contents Previous