Smoking supernovae dusty galaxies.
THE LATEST PORTION of the electromagnetic spectrum to be opened up for astronomy is the submillimeter waveband, where wavelengths of light range from a millimeter down to 100 microns--about the thickness of a human hair. By sheer good luck, I have a unique perspective on the opening up of this final electromagnetic frontier. That's because I was the first astronomer to observe with the world's first submillimeter telescope, the James Clerk Maxwell Telescope (JCMT), which opened for business in 1987.
At the time this didn't strike me as an important historical event. As a young postdoctoral researcher, I spent three years in Hawaii trying frenetically to use as many different telescopes as possible, and the JCMT was only one among a crowd of telescopes on Mauna Kea. Moreover, its pointing and tracking was so poor during this inaugural observing run that for years I had on my office floor a pile of spectra of the galaxy M82--spectra that were useless because I had no idea which part of the galaxy I had observed. (The pile--possibly an important historical document--eventually vanished during one of my many moves.)
This article, however, is not really about submillimeter astronomy. It is about the stuff that submillimeter astronomers observe--interstellar dust--and the role that this underappreciated constituent of the cosmos has played in solving one of astronomy's thorniest puzzles.
Discerning Cosmic Dust
Sky & Telescope readers have seen interstellar dust at work many times, because dust is responsible for some of the prettiest pictures in astronomy. Two examples are the Horsehead Nebula (below) and the nearby galaxy NGC 891, shown on the facing page. In both cases, the apparent absence of stars (toward the horse's head and along the galaxy's midplane, respectively) is caused by tiny grains of solid material. This dust--which actually is more like cigarette smoke than household dust--hides stars that lie beyond it.
Although we now take the existence of interstellar dust for granted, this ingredient of the universe was overlooked for centuries. After all, astronomers widely reasoned, the apparent absence of stars in dark nebulae might be real. Only in the last century did astronomers infer that invisible matter was diminishing the light of numerous stars.
Now, though, astronomers have a way to view cosmic dust directly. The starlight absorbed by a dust grain heats that grain, typically to a temperature about 20[degrees]C above absolute zero (that is, about 20[degrees] Kelvin). The wavelengths at which a body in thermal equilibrium radiates are determined solely by its temperature. As it turns out, grains at 20[degrees]K radiate predominantly in the submillimeter waveband.
By observing submillimeter radiation from dust, astronomers can find hidden energy sources throughout space. Starbirth, for example, is hidden from traditional optical (visible- and ultraviolet-light) telescopes because it occurs in dusty gas clouds such as the Orion Nebula (above). Only the glow of ever-so-slightly warmed-up dust allows astronomers to find and study newborn stars. Dust also hides energy generated within individual galaxies such as NGC 891. But how important is dust to our understanding of the universe as a whole? In the mid-1990s, it became clear that the answer is very.
A Dusty Universe
The primary evidence for this came courtesy of a NASA space mission, the Cosmic Background Explorer satellite, or COBE. COBE is best known for measuring the cosmic microwave background radiation, which heralds from a time only 400,000 years after the Big Bang. But one of its other scientific goals was to measure the summed radiation from all galaxies, near and far, in the submillimeter waveband (which includes wavelengths that some astronomers refer to as the far infrared).
When we inspect the spectrum of the night sky, which astronomers have laboriously measured throughout the electromagnetic spectrum, we see two strong peaks in the optical and submillimeter wavebands. (The cosmic background radiation is far weaker in every other waveband.) The exciting result from COBE's measurement of the latter peak was that about half the optical radiation emitted since the Big Bang by all the objects in the universe--stars, galaxies, quasars, and so forth--has been absorbed by dust and then reradiated as submillimeter emissions (S&T: May 1998, page 18). This discovery convinced astronomers that dust is not just a source of pretty pictures but has overarching implications for cosmology.
The Enigma of Ellipticals
Around the time that this COBE discovery was announced in 1996, my colleagues and I were becoming increasingly convinced that dust might also solve a problem we were having in writing the history of galaxies.
The thing that makes my chosen research field (galaxy evolution) possible is the fact that light travels fast--but not infinitely fast. Most readers know that we see what the stars tracing the constellations looked like tens or hundreds of years ago rather than what they look like today. My colleagues and I use this fact to look not hundreds but billions of years into the past, and by observing galaxies at different distances we can see how galaxies have changed throughout cosmic history.
The big practical problem, of course, is that galaxies billions of light-years away are very faint, and this kind of cosmic archaeology really became possible only in the 1990s with the commissioning of telescopes such as the Hubble Space Telescope and the gigantic Keck Telescopes. These instruments led to rapid progress in deciphering the history of galaxies (S&T: October 1998, page 46). But there remained one outstanding problem: elliptical galaxies.
Ellipticals include both the biggest and smallest galaxies in the universe, and they differ from spirals and irregular galaxies because almost all their stars are very old. The stars in nearby ellipticals typically are around 10 billion years old, and this means that they must have formed in a relatively short period right after the Big Bang (itself now thought to have taken place about 14 billion years ago). Objects in which large numbers of stars form very quickly should be very luminous, and for several decades astronomers have expected to find these newborn ellipticals if only they looked far enough out into space (and thus far enough back in time).
However, by the mid-1990s there was still no sign of these objects, even in the Hubble Deep Field, the deepest image that had ever been taken of the sky at visual wavelengths (July issue, page 42). One possible explanation of this failure was that these objects were hidden by dust. The obvious way to test this idea was to obtain submillimeter images of the sky and look for sources of this kind of radiation that had no counterparts in optical images like the Hubble Deep Field.
Fortunately, at just about this time the world's first submillimeter camera, SCUBA (Submillimetre Common User Bolometer Array), was installed on the JCMT, making it possible for the first time to take proper submillimeter pictures of the sky. In the last seven years SCUBA has been responsible for many discoveries, but I want to describe just two. One is about objects billions of years in the past; the other is about something that happened only 325 years ago. Surprisingly, they are closely connected, and together they may solve the puzzle posed by elliptical galaxies.
Two images dramatize the first of these discoveries. One is the Hubble Deep Field image made with the orbiting observatory in 1996. The other is a submillimeter image of the same bit of sky that David H. Hughes (then at the Royal Observatory Edinburgh) and his collaborators made with SCUBA. The two images are very different (see the next page). The Hubble image contains about 1,000 galaxies; the SCUBA image contains only five discrete sources of submillimeter radiation. This fundamental difference shows that while the optical and submillimeter portions of the cosmic background radiation are about equally strong, the latter is produced by much rarer objects.
What are these rare objects? Much of my time and that of many other astronomers has been spent trying to answer this question over the last few years. The submillimeter radiation is from dust, and unfortunately dust poses a problem for many observational techniques. On optical images, there is often only an exceedingly faint object present at the position of a submillimeter source. This means that the standard technique for measuring redshifts, spectroscopy, has proved very difficult. Only within the last year have Scott Chapman (Caltech) and his collaborators succeeded in using the Keck Observatory telescopes to measure redshifts for 10 SCUBA sources. Those redshifts lie between 1 and 4, which means that when we observe these objects we are looking 8 to 12 billion years back in time.
The dust also hides these objects' ultimate sources of energy. The high redshifts show that these objects are much more luminous than today's normal galaxies. Might they simply be quasars hidden by dust? This seems unlikely. The ultimate energy source in a quasar is gas swirling toward a supermassive black hole, a process that produces copious X-ray emissions. Telescopes such as the Chandra X-ray Observatory and XMM-Newton have failed to detect strong X-ray emissions from the SCUBA sources.
Rather, it now seems almost certain that the dust hides large numbers of newborn stars, with about 1,000 solar masses of gas being converted to stars yearly within a typical SCUBA galaxy. In a universe with hundreds of billions of galaxies, each containing hundreds of billions of stars, this may not seem like much. But a star-formation rate of 1,000 solar masses per year is enough to make all the stars in a galaxy like our own in just 100 million years--a cosmological eyeblink in today's 14-billion-year-old universe.
Dust-shrouded objects that prodigiously formed stars in the early universe--these are the very properties expected for the ancestors of elliptical galaxies. It now seems difficult to avoid the conclusion that an elliptical starts its life as an object full of gas, dust, and young stars; that the gas and dust are consumed or dispersed by the births of more stars; and that eventually, after 10 billion years or so, an object like one of today's elliptical galaxies remains.
An Explosive Origin for Distant Dust
Thus, SCUBA has provisionally solved one long-standing astronomical mystery. But in answering a question about ellipticals' origins, it raised another, equally confounding, one. And that is where the second discovery--the one about a 325-year-old astronomical object--comes in.
About three years ago, my Cardiff University research group began to shift its attention from that which is being hidden by the dust to the dust itself. Dust was being found at higher and higher redshifts, and thus closer and closer in time to the Big Bang. We began to realize that there might be a problem making so much dust in such a short time.
For dust grains to condense out of a gas, the conditions have to be just right: too high a temperature and the atoms or molecules will remain in the gas phase; too low a density and a gas atom will follow a lonely path through interstellar space, never encountering another gas atom to bind to.
The conditions are just right in the atmospheres of cool, aged stars--red giants and stars on the asymptotic giant branch of the Hertzsprung-Russell diagram, which charts the stages of stellar evolution (S&T: March 1999, page 43). Though rarefied by terrestrial standards, these stars' atmospheres are still quite dense, at least when compared with the interstellar medium; but the temperatures in their outer parts are below the melting points of many of the compounds making up interstellar dust.
Textbooks generally state that these stars are the main sources of cosmic dust. However, the process is not very quick--the Sun already is 5 billion years old and will not become a red giant for another 5 billion years. Stars more massive than the Sun will reach this stage faster, admittedly; but even so, we doubted that large amounts of dust could be made this way so soon after the Big Bang. We decided to look at another possible source: supernovae.
A supernova * is the explosion that occurs when a massive star comes to the end of its life. A supernova explosion is one of the most spectacular sights in the cosmos. If the massive star nearest our Sun, Orion's Betelgeuse, went supernova, it would briefly shine in our skies as brightly as the full Moon. For a few months a supernova can outshine all the other stars in its home galaxy combined.
We have a surprisingly intimate connection to supernovae.
As Joni Mitchell sang, "We are stardust." Many of the elements making up our bodies and the planet on which we live were formed in the centers of massive stars and then thrown out into space by supernova explosions.
Since supernovae are a key source of the chemical elements that make up cosmic dust, it seemed natural to suspect that supernovae might be a significant source of dust itself, with grains forming as the remains of the shattered star expand and cool. Because massive stars have short lives, we realized that supernovae could skirt the time-scale problem posed by red giants and asymptotic-giant-branch stars (stars passing through a post-red-giant phase in the evolution of relatively low-mass stars).
Of course, supernovae had been suspects for a long time, but they had generally been ruled out because only tiny amounts of dust had been found in them. We began to wonder whether dust had not been detected in supernova remnants because it was very cold. And with SCUBA we had the ideal instrument to search for cold dust. Along with three other Cardiff astronomers--Michael Edmunds, Loretta Dunne, and Haley Morgan--and with Robert Ivison (Royal Observatory Edinburgh), I set out to use SCUBA to search for dust in supernova remnants.
Ironically, the first problem we faced was posed by the very thing we were hoping to observe, namely dust. Although a supernova explosion is thought to occur in our Milky Way galaxy roughly once every 30 years, the last definite sighting was Johannes Kepler's, in 1604. That is because while supernova explosions are intrinsically very luminous, interstellar dust scattered throughout our galaxy's midplane hides most of them from view. Fortunately, 400 years is not long in astronomical terms, and we calculated that if we detected dust in the expanding gas cloud marking the position of Kepler's supernova, that dust would have to have been formed by the supernova rather than being preexisting dust swept up by the expanding cloud.
We also decided to observe Cassiopeia A, the remains of another supernova that English astronomer John Flamsteed may have seen in 1680 (S&T: May 2003, page 40). Whether Flamsteed indeed spied this "guest star" or not, we know that Cas A is about 325 years old because radio astronomers have measured the speed at which it is expanding (radio waves are not obscured by dust).
We went out to the JCMT to observe the remnant of Kepler's supernova and were fortunate to discover that Cas A had already been observed several years earlier, and that the data were in the JCMT archive. Soon we had the first-ever submillimeter pictures of both Kepler's remnant and Cas A (shown above) and discovered that both remnants are bright sources of submillimeter waves.
This alone was not enough, however, to clinch the case for copious dust, because another process can produce submillimeter radiation. Kepler's remnant and Cas A both are bright sources of radio waves, the result of energetic electrons spiraling around magnetic-field lines. These electrons also produce submillimeter radiation. Dunne and Morgan removed this contaminating signal from our JCMT images by using existing radio images to estimate how much submillimeter radiation likely came from the electrons. From the strength of the remaining submillimeter emission, we calculated that Kepler's remnant and Cas A each contain about 4 x [10.sup.30] kilograms of dust--twice the Sun's mass, and roughly 1,000 times more dust than previously had been detected in these supernova remnants.
A Bright Future for Dusty Supernovae
If our conclusion is correct, we have solved the problem of dust being found shortly after the Big Bang. However, supernova remnants are for technical reasons quite tricky to observe, and SCUBA, the world's first submillimeter camera, is quite a primitive instrument. Moreover, so far we have successfully observed only two nearby supernova remnants, and these may not be typical.
Fortunately, the discovery that so much of the universe is hidden from traditional telescopes has been responsible for a spectacular expansion in submillimeter astronomy: currently there are 10 submillimeter telescopes or telescope arrays on the drawing board or in construction (one, the eight-element Submillimeter Array, recently was completed atop Mauna Kea, near the JCMT).
Furthermore, SCUBA is soon to be replaced with SCUBA-2, a much larger array of radiation detectors that will greatly enhance the JCMT's performance. These new instruments will allow us to confirm our results (we hope!) and to investigate, through techniques such as spectroscopy and polarimetry, the properties of the newly formed dust, which may well be different from the dust that has been meandering around in the interstellar medium for millions of years.
One groundbreaking telescope that already is working is the Spitzer Space Telescope (April issue, page 30). A team led by Dean Hines (University of Arizona) recently observed Cas A using Spitzer's Multiband Imaging Photometer. Taken at a mid-infrared wavelength of 24 microns, the Spitzer image, shown at left, reveals dust. Confoundingly, Spitzer sees much smaller quantities of dust than SCUBA did. However, if the dust is very cold, we would not expect to detect it in the Spitzer image.
Other important upcoming telescopes include the European Space Agency's Herschel Space Observatory, which will be launched in 2007, and the multinational Atacama Large Millimeter Array, or ALMA. ALMA will be the ultimate submillimeter telescope in every way: in cost (about a half billion dollars); in location (at 5,400 meters, fully 1,200 meters higher than Mauna Kea); and in size and complexity (64 interlinked antennas, each spanning 12 meters). When completed in 2012, ALMA will allow us to study nearby dusty supernova remnants and distant dusty galaxies alike with unprecedented sensitivity and in exquisite detail.
* The supernovae described here are core-collapse supernovae. A fundamentally different kind of stellar explosion called a Type Ia supernova occurs in interacting binary-star systems (S&T: November 2002, page 28).
An astrophysicist at Cardiff University, STEVE EALES studies galaxy formation and evolution at submillimeter wavelengths.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Milky Way Galaxy|
|Publication:||Sky & Telescope|
|Date:||Aug 1, 2004|
|Previous Article:||An old friend in a new light.|
|Next Article:||Starlight detectives: The birth of celestial spectroscopy.|