Deep Washington Photometry of the HST M31-halo field Edward Olszewski, Abi Saha, and Andy Dolphin Many thanks to: MMT staff, SAO Instrument builders, Steward.

Deep Washington Photometry of the HST M31-halo field

Edward Olszewski, Abi Saha, and Andy Dolphin

Many thanks to:MMT staff, SAO Instrument builders, Steward technical staff, Especially…

Brian McLeodDennis SmithMo ConroyNelson CaldwellDan FabricantTim PickeringMMT telescope operatorsDavid Dean (Steward)Steward Machine Shop

Background

M31 and the Milky Way are [approximately] twins, yet

M31 has a large bulgeM31’s halo is [much] more metal-rich than ours

M31 has evidence for massive streams of destroyed galaxies.We have evidence for streams, but not as extreme.

Our halo is old and metal poor.

Streams in our Milky Way and the future of the Milky Way

The halo is “polluted” with remnants of “visitors from outer space”

The most impressive example is Sagittarius (Sgr).

Sgr has been almost completely destroyed and now has a “comet tail”of stars tidally ripped from it that wraps more than once around itsorbit.

It is injecting both metal-poor and metal-rich stars into our halo,and is injecting globulars.

In a few billion years, the Milky Way will swallow and destroythe Magellanic Clouds, and the Clouds will inject many relativelymetal-rich stars into our halo, in fact, more stars than are there now.

And not too long after that, we and M31 will swallow each other,possibly creating a single elliptical galaxy.

M31, its bulge, its halo

The relatively huge bulge of M31 and the more impressivestreams in M31’s halo point to more impressive galaxycannibalism.

In 2003, Tom Brown et al published their HST H-R diagramof a tiny piece of the M31 hal. 39.1 hours in a broad yellowfilter (F606W) and 45.4 hours in a broad red filter (F814W)

These are among the deepest images ever taken, and reach50% completeness at 30th mag.

They, unfortunately, have not released the data productsnor the photometry catalogs.

But, here is their H-R diagram…

You’ll see the image in a movie in a while…

They conclude that the typical M31 halo star is relatively“intermediate age”, 6-10 Gyr old, and relatively metal rich,in contrast to a typical halo star in our Milky Way.

Their analysis is fine as far as it goes. But ground-baseddata can offer value added.

They only know brightnesses and one color. Stellarevolution of stars of different ages (masses) and abundancescan make stars of a specific color.

So they made a not-unreasonable assumption that youngerstars are more metal rich, and let a stellar populationscomputer code tell us the mixture of ages and abundances.

But wouldn’t it be nice to KNOW the mixture of abundances,and force the code to mimic that empirical mixture?We W

We can measure abundances of the red giants.

Properties of red giants and of the Washington System

Red giants in M31 range in brightness from V=22 or so, to V=25 or so.

Unfortunately, there are lots of faint galaxies that can mimic starsIn ordinary seeing.

Washington system designed to measure abundances of red giantsusing broad band photometry. Broad-band = more light = abilityto work on fainter stars.

Specifically, the Washington C filter, in the near UV, with centralwavelength 3860A, when coupled with filters that give youthe stellar temperature (say R-I), measures abundances.

Given good enough data, we get abundances of all of the red giantsIn the HST field at once, without using spectroscopy. Getting abundancesOf 24th mag giants with spectroscopy is hard/impossible.

The data

We used minicam in 2003, and megacam in 2004 and 2005, andWIYN’s minimo in 2005/2006.

Despite bad weather and telescope problems, we collected30000 sec at MMT in Washington C, and 3600 sec in Rand 1800 sec in I.

We are in the early stages of the final reductions, and have somegorgeous less than ideally calibrated data (thin clouds).

Today we show megacam 2005 data, roughly 15000 secand roughly 0.8 arcsec seeing of the stack. We also showWIYN R data at 0.5 arcsec seeing.

Here is a movie (I think we need to step out of powerpoint),of HST data at 0.05 arcsec seeing, WIYN R at 0.5, and WashingtonC at 0.8…

The movie

0.8 arcsec seeing in “U” is pretty darn good, especially for extended periods.

But imagine 0.5, et’s do what it takes to get there…

Histogram of C magnitudes

As we said, the calibration is less than ideal right now because of clouds.

But we’re close

26th mag at C is pretty darn impressive…

Using the WIYN data, here are color-magnitude diagrams…

First, R versus R-I, then C versus C-R

Finally, here is the diagnostic diagram

R-I gives temperature, and C-R gives abundance.BUT, C-R has temperature sensitivity, so you needexquisite R-I. We don’t have that.

We therefore must re-reduce a subset of the HST datato get good-enough temperature.

We must also co-add the 2004 and 2005 data, andsomehow average in the minicam C and the WIYN C,and we need to get improved calibration.

We will re-reduce the data using HST positions forcedupon the ground-based data, which will improvethe photometry.

Final philosophical comments

We used specialized filters, but never got enough clear weather.It’s easy to parcel out data in numbers of nights that pretty muchguarantee that the data will not be calibrated or competitive.

It’s a nontrivial problem to solve this telescope-time dilemma.