Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 347, 455-472 (1999)

Previous Section Next Section Title Page Table of Contents

3. Correction for differential reddening

In order to correct the CMDs for differential reddening, we used a refined version of the method described by Grebel et al. (1995). The entire frame is divided into subframes. These are determined by covering the whole frame with a regular subgrid and dividing the grid cells further until the number of stars in one cell becomes too small to define the CMD structure. The CMDs will be shifted according to the reddening vector (described below) with respect to CMDs from neighbouring cells. The shift in colour supplies the differential reddening. If two neighbouring subframes have the same reddening, these subframes are merged. There are two problems with this method: First, one has to be careful to use the HB as a means for comparing two CMDs, as the HB may be intrinsically elongated. Useful results can only be achieved by comparing the RGBs and TOPs, as far as they are accessible. Second, the size of the subfields must be large enough to render meaningful CMDs. Fig. 22 shows the resulting extinction maps for the seven clusters. The smallest subfields have a size of about [FORMULA]. But as some of them still showed differential reddening, the scale of the structures responsible for the differential reddening is be expected to be even smaller. The smallest scales we got from a comparison of coordinates of stars with different reddening amounted to values of [FORMULA].

For the correction of the CMDs we need the extinction


However, assuming a uniform reddening law led to CMDs which in some cases showed the corrected HBs having larger or smaller slopes than the uncorrected HBs. Moreover, as there is some uncertainty in the literature regarding [FORMULA], with values varying between [FORMULA] (Savage & Mathis 1979) and [FORMULA] (Grebel & Roberts 1995), we determined the slope of the reddening vector via the tilted HBs of our CMDs. This leads to reasonable results only if the HBs are intrinsically clumpy. This assumption is corroborated by the fact that the well dereddened CMDs (Fig. 23, 27, 29) have clumpy HBs indeed. Table 1 shows the slopes [FORMULA] for each cluster. In Fig. 21, the slopes are plotted against the galactic longitude. These variations, although at the margin of the errors, confirm earlier observations by Meyer & Savage (1981) and Turner (1994). Meyer & Savage determined via two-color-diagrams the deviation of single stars in the extinction behaviour from the galactic mean extinction law. Turner demonstrated the inapplicability of a mean galactic reddening law for objects lying close to the galactic plane. To correct the diagrams for differential reddening, we referred all sub-CMDs of one cluster to the one with detected minimal reddening and we shifted all other sub-CMDs onto that. As we thus use the minimal absolute reddening as a point of reference, the absolute reddening determined later on will be smaller than value given in the literature. The differentially dereddened CMDs are shown in Figs. 7 through 19. As the correction led to a clearly improved appearance for all clusters, the corrected versions of the CMDs will be used for further investigation.

[FIGURE] Fig. 6. Unselected CMD for NGC 5927. RGB and HB are clearly visible as well as some stars of the field population.

[FIGURE] Fig. 7. Differentially dereddened CMD for NGC 5927. The RGB-bump is well discernible now.

[FIGURE] Fig. 8. Unselected CMD for NGC 6316. The AGB/RGB of the field population is well discernible.

[FIGURE] Fig. 9. Differentially dereddened CMD for NGC 6316. Due to the only slight differential reddening, the effect of the correction is not as distinct as in NGC 5927, for example. Moreover, the dereddened CMD contains stars of the inner [FORMULA] around the cluster's center only (See extinction maps in paragraph 4).

[FIGURE] Fig. 10. Unselected CMD for NGC 6342. TOP and upper MS are reached.

[FIGURE] Fig. 11. Differentially dereddened CMD for NGC 6342. The effect of the correction is best seen at the TOP-region.

[FIGURE] Fig. 12. Unselected CMD for NGC 6441. Strong contamination by the field population.

[FIGURE] Fig. 13. Differentially dereddened CMD for NGC 6441. The effect of the correction shows best in the narrower lower RGB. The HB still partly overlaps with the RGB, which means, that the correction was not completely successful. This is mostly due to the strong field population. 

[FIGURE] Fig. 14. Unselected CMD for NGC 6760. The effect of differential reddening is clearly visible.

[FIGURE] Fig. 15. Differentially dereddened CMD for NGC 6760. The HB lies well to the blue of the RGB, the RGB-bump clearly on the RGB below the HB.

[FIGURE] Fig. 16. Unselected CMD for NGC 6528. Below the AGB/RGB there are traces of the background population.

[FIGURE] Fig. 17. Differentially dereddened CMD for NGC 6528. The HB still overlaps strongly with the RGB. Thus, the correction was not completely successful due to the strong contamination by field stars. However, the RGB narrowed perceptibly, and the RGB-bump is now visible below the HB.

[FIGURE] Fig. 18. Unselected CMD for NGC 6553. Strong contamination by the field population is visible.

[FIGURE] Fig. 19. Differentially dereddened CMD for NGC 6553. RGB and RGB-bump are clearly defined now, and the HB lies well to the blue of the RGB.

[FIGURE] Fig. 20. NGC 5927: Unselected HST-CMDs (PC, WF2, WF3 and WF4). The TOP is well resolved.

[FIGURE] Fig. 21. Slopes of the HBs, i.e. the reddening vectors against galactic longitude (here in cartesian coordinates).

[FIGURE] Fig. 22. Extinction maps for NGC 5927 to NGC 6553, derived from plotting CMDs for each of the areas. The numbers in the subfields give the absolute reddening derived via isochrone fitting. The differential reddening was determined using the minimal absolute reddening as a point of reference. As the derived scale of reddening variation strongly depends on the number of stars, the resulting scales can only be estimates for upper boundaries. The maps cover [FORMULA], i.e. they cover the whole area of the original frames. An exception is made with NGC 6136, where we constrained the map to the inner [FORMULA], because of the clusters small size.


Table 1. Maximum differential reddening and slope of the reddening vector.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: June 30, 1999