IRAC will be dark calibrated by taking repeated images of the sky near the ecliptic pole. We must choose a region relatively free of stars, since these foreground objects complicate all of the forseen data reduction techniques which will be used to reduce the data. Unfortunately, the NEP is at galactic coordinates 96.4+29.8, which is actually relatively close to the galactic plane. The density of 2MASS stars (which are fairly bright by IRAC standards) is on the order, typically, of 8 per IRAC field at this latitutde. It is desirable to find locations where the number of stars are minimized.

The NEP is chosen because it is the location of (one of) the CVZ, and hence always accessible to SIRTF. The CVZ itself extends down from the pole nearly 10 degrees. In the following I am going to limit myself to a region 6 degrees on a side centered on the NEP. I think it is better to stay as close to the pole as possible, in particular in case on-orbit variances in the pointing constraints tighten the size of the CVZ.

Because the fields will be observed all year, the orientation of IRAC will sweep through a full 360 degrees. The most efficient way to use IRAC for this task is to take the data simulataneously through both FOVs (as opposed to picking a single region and covering it twice, once with each FOV). Unfortunately, this requires a fairly large star-free zone. Specifically, we need 5+2+5'=12' just for the 2 FOVs. If we assume that we are executing a small map (needed in order to dither out stars), then we need something more like a circular region 15'in diameter relatively free of stars.

Matt Ashby has sent me a diagram prepared at SAO, showing the central 2 degrees of the NEP (figure 1). Marked on this map are some proposed field locations (labeled 1-4). In blue I have shown the actual region that will need to be star free. In red is a region where the 2MASS data has not yet been released.

Figure 1 - central 2 degrees of NEP. Shown are the original field locations proposed by Matt.

2. Field Identification

As a starting point I have obtained from Tom Jarrett a specially cleaned pre-release version of the 2MASS point source catalog for a square region 6 degrees on a side, centered on the NEP. In order to score the fields, I generate a 360x360 grid of cells, with each cell one arcminute on a side. This is sufficiently small that given the size of the search area it should more than adequately sample the search region. I then step through the 127,000 stars in the region, and calculate each one's "badness" contribution to the cell in which it is located. In this case the badness penalty was just 10^(k/-a), where I set a=3, so that brighter objects pay an ever increasing penalty. Of course, for a=2.5 this is just an intensity map filtered on 1 arcminute size scales.

Then, for every cell I calculate the total penalty in a box 15' on a side. Ideally this would be a circle, but it is easier for me to program as a square. This is a conservative approach, since the actual area is larger than the circular region. The best regions are those where this function in minimized. It's basically like making an image with 1' pixels from the 2MASS point source catalog, and then boxcar filtering it.

Shown below is a closeup of the NEP, with the four fields proposed by Matt shown as red crosses. Brian has also proposed 4 fields, which are shown in green. The actual NEP is at grid location 180,180. The axis units are arcminutes.

Matt's Initial Field Selection

FIELD   RA J2000 DEC
---------------------------
1.      268.9   66.8
2.      268.6   66.6
3.      270.5   66.40
4.      270.15  66.83
        
Brian's Proposed Fields

FIELD   RA J2000 DEC
---------------------------
1.      17:59:13 66:12:33
2.      18:02:29 67:04:11
3.      17:57:19 66:45:03
4.      18:05:33 66:09:35

Figure 2a - Ashby fields in "image" space. The blue regions represent the area swept out over the year by the IRAC dark AOR.	Figure 2b - Scores for SAO fields. Ashby fields in red, Patten fields in green.

Of the red fields proposed by Matt, number 2 (westernmost) is probably the best. The other three (particularly number 3) will catch bright stars at certain times of the year. Of the green fields proposed by Brian, number one (second westernmost) is the best. Below in figure 3 I show the entire NEP region. It's not very obvious that there is a particularly compelling "best" location, although there are obviously some locations that are poor!

Figure 3 - 6x6 degree region centered on the NEP. NE (in RA/DEC) is at top left. Darker colors indicate a worse score.

The above table shows the total "badness" score for several fields. Lower numbers are better. The absolute best location is at 17h40m3.6s 69d00m15.5s. It's score is roughly 1/30 that of Ashby Field 3 at 270.15d+66.83d. In the immediate vicinity of the NEP the best place is 17h55m8.4s 66h30m15.5s. It's score is within a factor of two of the lowest point available. Unfortunately, it has the galaxy CGCG 322-021 within it's field. Next on the list is 18h00m10s 66d18m15.5s, which scores very similarly to 17h55m8.4s 66h30m15.5s. Shown below are some images taken from the DSS showing the candidate fields. They are 15 arcminutes on a side.

Ashby Field 1	Ashby Field 2	Ashby Field 3	Ashby Field 4

Patten Field 1	Patten Field 2	Patten Field 3	Patten Field 4

NEP	17h40m3.6s 69d00m15.5s Best Location	17h55m8.4s 66h30m15.5s	18h00m10s 66d18m15.5s Best near NEP