Astron. Astrophys. 329, 522-537 (1998)

8. Constraints on the halo

In this section, we describe the method to obtain the exclusion diagram and the results of the EROS project, then we compare with the limit obtained by the MACHO project (in the low mass range only).

8.1. The exclusion diagram

From the expected and observed number of events, we obtain exclusion diagrams, i.e. the fraction of halo compatible with our results. Assuming a Poisson statistic, the expected number of events compatible at 95% CL with zero observed event is 3.0.

8.2. EROS limits

Fig. 20 presents the limits obtained from our CCD data with the eleven models of Galaxy.

Fig. 20. Limit on the Halo fraction comprised of planetary mass objects, using our 10 Galaxy models plus the model of (Ansari et al, 1996) and the CCD detection efficiency including the finite source size and the blending effects.

The limit on the halo fraction in the range of mass [ ] M is between 10 and 25 %. This limit is stronger for spherical halos than for flattened halos, for light disks than heavy disks. Nevertheless, even for a Galaxy composed of a flattened halo and a heavy disk, we give a strong constraint on the maximal contribution of planetary mass objects to the Halo dark matter.

Fig. 21 presents the limits with the model of Ansari et al, 1996. We use efficiencies which take into account, or not, blending and finite source size effects, and either our CCD resultsalone or the combination of the Schmidt plates and CCD results.

Fig. 21. Limit on the Halo fraction made with planetary mass objects, using the Galaxy model of (Ansari et al, 1996) and the CCD detection efficiency including the finite size and the blending effects. We also show the result with the simple efficiency and combining the CCD results with the Schmidt plates ones.

The blending and the finite source size effects do not affect the limits for objects with a mass larger than 2 10^-6 M . Still, the source size effect (and the blending effect to a lesser extent) supresses all sensitivity to objects with a mass lower than 10^-7 M . Without this effect we would be able to put constraints down to 10^-8 M . But the evaporation limit of brown dwarfs is certainly nearer to 10^-6 M than to 10^-8 M (de Rújula et al, 1992); of course, the very existence of brown dwarfs with a mass around 10^-7 M is improbable.

Combining the CCD limits with the Schmidt plates results, we observe that the fraction of halo in form of objects of mass between 5 10^-7 and 2 10^-3 M is lower than 10 % to 20 %. This limit is more or less flat: it thus remains valid whatever the mass function inside this mass interval.

8.3. Comparing EROS and MACHO results

We compare the results of the two teams searching for microlensing events towards the Magellanic Clouds. The limit on lower mass objects by the MACHO collaboration was obtained from a "spike" analysis because of a sampling of at most 2 or 3 points per night (Alcock et al, 1996b). Here we present all limits for deflectors with mass ranging from 10^-7 to 10^-2 M : no microlensing events were detected in this range that corresponds to short timescales. We use only the spherical model with limits at 95 %CL. Fig. 22 shows the limits from both teams. A complete study of the (small) overlaps in time and in line of sight between the MACHO and EROS observations will be done in a further common paper to give stronger constraints on planetary mass objects.

Fig. 22. Comparaison between the EROS limit (solid line, CCD+Schmidt plates) and the MACHO one (dotted line) on the part of the Halo which can be composed of planetary mass objects.

© European Southern Observatory (ESO) 1998

Online publication: December 8, 1997
helpdesk@link.springer.de