Annu. Rev. Astron. Astrophys. 1994. 32: 531-590 Copyright © 1994 by . All rights reserved

9.5. Infrared Searches for Brown Dwarfs

Even though brown dwarfs do not burn hydrogen, they still generate some luminosity in the infrared. They radiate first by gravitational contraction (for about 10⁷ y) and then by degenerate cooling. If the disk or halo dark matter is in the form of brown dwarfs, it is therefore important to consider whether they can be detected via this infrared emission. Current constraints on BDs are rather weak (Low 1986, van der Kruit 1987, Beichmann et al 1990, Nelson et al 1993) but the prospects of detection will be much better with impending space satellites such as ISO and SIRTF.

The problem has been addressed in various contexts by several authors. Karimabadi & Blitz (1984) have calculated the expected intensity from BDs with a discrete IMF comprising an = 1 cosmological background. Adams & Walker (1990) have discussed the possibility of detecting the collective emission of the brown dwarfs in our own Galactic halo for both a discrete and power-law IMF. Daly & McLaughlin (1992) have considered the prospects of detecting the emission of individual halo brown dwarfs of a given mass and age in the Solar vicinity, as well as the collective emission of brown dwarfs in other galaxy halos. Kerins & Carr (1994) have considered the possibility that the BDs are assembled into dark clusters and also discuss how infrared observations at different wavelengths could be used to probe the mass spectrum of the brown dwarfs.

As an illustration of the feasibility of detecting radiation from BDs, let us consider the prospects of detecting the nearest one in our halo. If the BDs all have the same mass m, then the local halo density (₀ = 0.01 M pc^-3) implies that the expected distance to the nearest one is 0.55(m / 0.01 M)^1/3pc. The expected spectra are shown in Figure 9 and compared to the sensitivities of IRAS and ISO. This assumes the temperature and luminosity of Stevenson (1986) where the BD age and opacity are taken to be 10¹⁰y and 0.01 cm²g^-1 (corresponding to electron-scattering). Although IRAS gives no useful constraints (it is too weak by a factor of 2 even for the optimal mass of 0.07 M), the ISOCAM instrument on ISO could detect 0.08 M BDs in a few hours, 0.04 M BDs in a few days, and 0.02 M BDs in a few months. Note that disk BDs, would be younger, locally more numerous, and more opaque than halo BDs, increasing the peak flux by 6 and decreasing the peak wavelength by 0.6. IRAS results already imply that BDs with a discrete IMF could provide the disk dark matter only if their mass is below 0.01 M.

Figure 9. Expected flux from the nearest halo BD for various values of BD mass. The IRAS point source sensitivity at 12 µ is shown; this is a factor of two above the predicted flux even in the optimal case. The expected 3 ISO 6.75 µ sensitivity is also shown, assuming an observation time of 10 days and a 100 s integration time.

One might expect the BDs to be easier to detect if they are in clusters. This is because, although the distance to the nearest source is increased by a factor (M_c / m)^1/3, the luminosity is increased by (M_c / m), giving an increase in flux of (M_c / m)^1/3. Rix & Lake (1993) have already used this to exclude the cluster scenario. However, they assume that the clusters are point sources and, as illustrated in Figure 4, the dynamical constraints discussed in Section 6.4 imply that the clusters will always be extended sources. In fact, the IRAS extended source sensitivity (EES) at 12 µ is too low to permit the detection of clusters. The ISO extended source sensitivity at 6.75 µ will suffice, but the time required to find these clusters is very sensitive to their mass and radius. Note that the halo clusters will cover the sky if they are large enough, corresponding to the line K_h > 1 in Figure 4, in which case detecting the clusters is equivalent to detecting the halo background. ISO would take several months to detect the Galactic background, even in the optimal case with m = 0.08 M (Kerins & Carr 1994). The background spectrum in this case is also indicated in Figure 6.