Visualizing the Center of the Milky Way

By Angela DePace

There is an enormous black hole at the center of our galaxy, the Milky Way. Our black hole is 4 million times the Sun's mass, but modest compared to other galaxies whose black holes can be up to a thousand times more massive. All large galaxies are thought to have giant black holes at the center, and a fundamental question is how this high-gravity environment affects the space around it. We spoke with Dr. Susan Stolovy about her survey of the Milky Way's galactic center, and Dr. Robert Hurt who generated images of her data about the challenges of looking deep into space.

How was the data gathered?

Susan Stolovy: These data were gathered using the Spitzer Space Telescope, which was launched in 2003. You can't see the galactic center at all with visible light telescopes - too much dust obscures the center and only nearby stars can be seen. But infrared light can penetrate the dust, so collecting data at these wavelengths (about 10 times longer than light visible to us), gives us an clear view of the center. Spitzer is equipped with an Infrared Array Camera (IRAC), that captures four different colors of infrared light with wavelengths ranging from 3.6 to 8.0 microns. Each color reflects different physical processes - for example stars are visible at the shorter wavelengths, while the organic molecules in interstellar dust are visible at longer wavelengths.

Generally, objects that lie across the middle of the image are located at the Galactic Center (about 25,000 light years away from us) and objects towards the top and bottom of the image are located in the foreground. Because there are so many stars at the galactic center and they are so bright in the infrared, we had to use quite short exposure times to avoid saturating the detectors (about 1.2 seconds per pixel). A very small portion of the survey was so bright that we observed it with only 0.02 sec exposure times; one of these regions is the dynamical center of the galaxy near the black hole, which is surrounded by a high concentration of stars and dust. We collected all of the data for these mosaic images (comprised of about 12,000 individual exposures) very efficiently--the entire survey used less than 16 hours of telescope time.

What's unique about this dataset?

Susan Stolovy: This is the highest resolution view of the galactic center at these wavelengths. You can see from the image comparing data from the 1996 MSX satellite to our new IRAC data, our spatial resolution is about 10 times better! We're able to see a fantastic number of new features in unprecedented detail, and learn a great deal about star formation under unusual conditions. For example, you can see bright filamentary clouds, similar to the "Pillars of Creation" documented by Hubble, only in this case we are four times further away (8 micron image, bottom row, second inset from the left). One can can also see dark molecular clouds scattered throughout the image that are so opaque that even these infrared wavelengths are absorbed. Overall, we were able to detect about a million individual stars in this dataset, and there were many more that were too blended together to detect.

What are the biggest challenges in representing data from the Spitzer telescope?

Robert Hurt: Unlike Hubble, which operates in the visible part of the spectrum, Spitzer sees infrared light far beyond human vision. Naturally we must display this data as visible light just to see it ourselves, so there is always some level of visual reinterpretation.  

We may in some cases have more than three infrared images we want to present in a single composite, while in others there may be only one. Thus we must find different ways to best utilize the three colors our eyes can see (red, green, and blue) to portray the infrared view.

Dynamic range is also a concern in many images. Telescope detectors can record a far wider range of brightness than a digital camera or even film. When photographing a sunlit landscape through a window in a dark room you have to choose between exposing for either the shadows or the highlights. To see the full range of brightness, a telescope's data can be "compressed" using various mathematical functions. These Spitzer images show a range of brightness and features the human eye could not take in all at once, even if it could see in infrared.

How are these images created?

Robert Hurt: Observations from the Spitzer Space Telescope are distributed in a FITS file format that is common to the field of astronomy. To make the jump from data to image, many observatories now use the FITS Liberator plugin for Photoshop. This freely-distributed import filter acts like a digital darkroom for the image. You can preview the image, set the black and white points, and even select between several functions to compress the dynamic range of the image.

Going from FITS file to image presents many of the same challenges a photographer faces in selecting exposures for an image. One needs to identify all of the key features in the image and choose level settings and dynamic range compression that brings them out in the most visually-appealing manner possible.

How do you choose the colors for a particular image?

Robert Hurt: When we're using just one channel of an observation, we usually apply a "pseudocolor" gradient to the image, letting color show brightness. In effect this is "colorizing" what would otherwise be a simple black and white image, but the result is more appealing and can help to better see features.

When we assign the wavelengths of data to color channels we often maintain the relative ordering of the colors, known as chromatic ordering. Since blue light has a shorter wavelength than red light, we organize the infrared colors in the same way with the shortest wavelengths going into the blue channel, and the longest into the red. This gives us an image that represents the colors we might see if we could shift our color eyesight from visible to infrared and preserves a lot of our color sensibilities.

For instance, a sunset looks red because the gas and dust in the atmosphere scatters out the blue light fastest, leaving the red to reach our eyes. This process continues on down the spectrum into the infrared regime. By using chromatic ordering with infrared images, we will perceive the same reddening effect as stars fall behind dust clouds. This is something we get intuitively, and produces images that are visually easier for us to understand.

What are the differences between these two representations?

Robert Hurt: In one image we have taken all four datasets from Spitzer's IRAC image and assigned them to red, green, and blue channels. One of the four is blended in as an additional orange channel, contributing a bit to both red and green. The resulting full color image represents the real color variations seen in the infrared. Starlight appears blue, and the dust red.

The second image shows only the longest-wavelength infrared color at 8.0 microns. But instead of presenting it in greyscale, colors have been added as a function of brightness. This helps us better distinguish a far greater range of detail in the dust surrounding the Galactic Center than is possible in a multi-channel composite.

How is each image successful?

Robert Hurt: I think that each of these images contributes a unique view into the Spitzer observations of the Galactic Center. Using multiple channels yields a color image where different colors highlight different physical processes, but at the expense of the detail in each individual channel.  In this case we felt that we got a more complete story by presenting both.

The full composite of all four IRAC bands gives us a dramatic window onto the center of our Galaxy. The torrent of blue-white stars conveys the fantastic density of stars here, and leads the eye to the brilliant cluster at the very heart of the Milky Way. The dark foreground dust clouds blocking our view stand out almost three-dimensionally from the more distant glowing clouds behind. The colors give it a real photographic sense, and show us something forever hidden in visible light.

The single-band image trades photographic reality for a richness of detail in the many layers of interstellar dust. The numerous faint filamentary structures, finger-like pillars, and dark clouds are incredibly vivid. Using a gradient of reds, oranges, and yellows to help trace differences in brightness dramatically increases contrast over what is possible in just a single channel of an RGB image. Plus the specific colors chosen help to convey the idea of infrared wavelengths; the exact same image colored with blues or greens would visually come across very differently.

Overall the two images make a compelling set, complementing each other. But there are many other ways these datasets could be rendered and colored, each conveying its own narrative. It's like there is a whole gallery of images waiting to be discovered in this one observation.