Adventures with Maffei 1 & 2

Our fearless leader M. Skrutskie was kind enough to authorize early observation of the notable galaxy pair discovered by Maffei over 30 years ago. These galaxies, known as Maffei 1 and 2, are unique in the sense that they are nearby (thus big and bright) but are vailed by the disk of our Galaxy. On the optical POSS images they appear as faint smudges -- reminiscent of low surface brightness galaxies. This is what Maffei saw way back in the days of the Beatles.

Now with our infrared eyes, we see something quite different. Roc has composed a nice 3-band JHK image of the pair. Check out Maffei 1 and Maffei 2 . Since 2MASS will survey the entire Galactic plane, we now have the means to find additional (if any) galaxies like the Maffei twins. So for now, the Maffei galaxies represent a sort of benchmark -- Can GALWORKS detect them? The answer is Yes. But the story is rather interesting. Stay tuned.

Maffei 1 had a very different detection path from that of Maffei 2, primarily because Maffei 2 is in my input catalog of known galaxies but Maffei 1 is not. The reason is that the catalog has a size limit of about 1 arcmin in diameter. Maffei 1 does not satisfy this limit in the optical (in the IR it is booming bright, as we shall see). In retrospect the catalog should probably be expanded to include smaller galaxies. I will pursue this angle at another time. In any case, Maffei 2, being in the catalog, was automatically detected and carried along in GALWORKS as a known entity. It was processed and parameterized in the usual fashion, with the information and stamp images written to disk. The stamp images are shown later in this memo. In contrast, Maffei 1 had a more unusual path.

At the onset, Maffei 1 was already different from any other galaxy encountered by GALWORKS thus far (note: relatively speaking, this is a miniscule sample): It is so bright in the core (K < 8) that it passes the flux thresholds for bright star blanking. That is to say, GALWORKS initially thought it was a bright star. The object, accordingly, was blanked from the coadd images during the bright star masking procedure. This was not a good turn of events. It was also not anticipated. Why? Because any galaxy that is bright enough to look like a bright star (to the point source processor) should in principle be big enough to be included in the catalog of known galaxies (like Maffei 2; see above). But Maffei 1 exposes the weakness in this assumption. It is big and bright, but the catalog does not know this (it is primarily an optical sample) -- thanks to Galactic extinction. There are undoubtedly other "gotchas" that I have not anticipated.

Maffei 1 is not a bright star, it is an extended object. It exhibits a "confusion radius" signature quite different from that of bright stars. GALWORKS applies a set of algorithms to find the optimum image masking of bright stars; see Automated Bright Star Masking . The first component of bright stars that is computed is the confusion radius -- corresponding to the PSF of the bright star plus residual (halo) emission symmetrically distributed about the star. The method is to compute the median value in annuli extended from the star, with the confusion radius set at about the 0.5 to 1-sigma level of the sky background. The confusion radius, however, cannot grow beyond 2.5 times the initial guess of the confusion radius (i.e., we apply a boundary limit). This limit should leave plenty of room for the confusion radius to converge. The exception occurs when either (1) the initial flux of the bright star is so far off that the initial guess at the confusion radius is severely underestimated (this can happen with very bright stars that saturate leaving a very poorly determined R1 mag), or (2) the object is not really a bright star (i.e., the initial flux of the object -- a pt. source estimated flux -- bears no resemblance to the true flux of the object). The latter case applies to Maffei 1. The confusion radius did not converge for this object (there was plenty of galaxy beyond 2.5 * estimated confusion radius). So, the obvious modification to the Automated Bright Star Masking routine was to set three confusion radii flags (one for each band), 1 for convergence and 0 for non-convergence. We then do not mask any object that does not converge. Maffei 1 satisfies this last conditional -- it is not blanked as a bright star.

We have happily solved the problem of Maffei 1 being too bright (and landing amongst the bright star blanking set). This is not to say that all galaxies that have initial K < 9 (or J < 10) will fail to have a confusion radius that converges. We can imagine core bright galaxies (like seyfert I types) that will satisfy the bright star masking algorithm thresholds. The solution for these galaxies is unclear at this time (but we note that they must be rare because they have yet to be encountered).

With Maffei safely protected from brutal masking, the object was successfully detected and parameterized by GALWORKS. Let the sunshine in. The images below show the postage stamps of the galaxy pair. Note the large number of stars subtracted from the images (red circles) -- the stellar number density is very high in this field. Also notice that the integrated flux for Maffei 1 is about 5th mag!! Maffei 2 is about 6th mag. Still, these isophotal fluxes severely underestimate the true total flux of the galaxies because they measure only the photons coming from the inner 80 pixels (in radius) -- the GALWORKS working limit. The galaxy pair is considerably bigger than 2 arc min in radius. The third image in the gif below is a star that is located in the outer halo of Maffei 2 (the bigger of the pair).

JHK stamp images of Maffei 1 & 2, and a "star" in the halo of Maffei 2