THE QUASAR ABSORPTION LINE
FINE STRUCTURE EXPERIMENT

Fine Structure Home

Web Page by
Chris Churchill

describing work in collaboration with
J.K. Webb, M.T. Murphy, V.V. Flambaum, V.A. Dzuba,
J.D. Barrow, J.X. Prochaska, and A.M. Wolfe


This work funded in part by:
NASA, National Science Foundation, John Templeton Foundation
Listen to an audio presentation with accompanying Power Point slides.

In Brief: The Nuts and Bolts

The QSO absorption line fine structure experiment is, in prinicple, very simple. First, we point the world's most powerful (largest) telescopes at quasars (QSOs), which are the extremely powerful black holes that voraciously consume the central regions of their host galaxies. Due to this process, QSOs shine with the brilliace of several 100 billion suns and thus can easily be observed even out to the farthest reaches of the universe. We then spread the QSO light out into a spectrum, which provides a sensitive and precise measure of the QSO brightness as a function of the light's energy (which we measure as wavelength, or color). This is not much different than looking at a rainbow of the sun except with great sensitivity). Since a QSO is not equally luminous at all wavelengths, the spectrum is not uniform brightness, but shows "strong peaks" at certain wavelengths. These peaks allow us to measure the QSO redshift, or cosmic distance (more on that below).

Additional features are present in the spectrum as well. These features are called absorption lines and appear as narrow dips in the spectrum's brightness. These absorption lines are created in gas louds that lie between us and the QSO; they are not associated with the QSO itself. Each time the QSO light passes through a gas cloud, the atoms in the cloud remove some of the light. A beautiful gift of nature is that each atom, say silicon, magnesium, or iron, absorbs the light in a unique "finger print" pattern that is clearly recorded in the spectrum. For example, iron creates five absorption lines with wavelengths 2344, 2374, 2383, 2587, and 2600 angstoms (100 millionth of a centimeter). Magnesium on the otherhand creates a clear "doublet" at 2796 and 2803 angstroms. So, not only are the "colors" at which the lines appear unique, but their relative placement in the spectrum with respect to one another is also a powerful clue to the identiy of the absorbing atoms. Using these absorption lines, one can very accurately count iron and magnesium atoms in these intergalactic gas clouds without having ever seen the clouds themselves!

For this expermient we use transitions from magnesium, aluminum, silicon, chromium, irom, nickel, and zinc. Since the universe is expanding, the more distant an object is the more rapidly it is moving away from Earth. The speed of recession is measured as the redshift, or z. This recession of the absorbing gas clouds results in an ineresting phenomenon. At the gas cloud itself, an magnesium atom absorbs the 2796 and 2803 wavelengths of the QSO light. But, the expansion of the universe stretches the spectrum so that the wavelengths are larger by the factor 1 + z. The diagram below shows the observed wavelength as a function of cloud redshift. At z=0, the transitions are observed at with their laboratory wavelengths. As redshift increases, the wavelengths get longer.
For example, the iron 2344, 2374, 2383, 2587, and 2600 and the magnesium 2796 and 2803 transitions lines are shown in two spectra below. The top (red) spectrum is for a cloud at z=2 and the bottom (green) spectrum is for a cloud at z=1. There are two points to make here. First, the wavelengths of the line (labeled in angstroms across the bottom) have been shifted to the red (redshifted). Note that they match the values shown in the diagram above where the dotted line crosses the figure at z=1. Second, note that the lines for the z=2 cloud are shifted even further and they are spread out further, matching the z=2 cut across the above diagram. (The scale size of both spectra are equal to show that more spectrum is required for a z=2 cloud in order to observe the same transitions from a z=1 cloud.) This means that the separation between the lines is also increased, and this stretching is also by the factor 1 + z. Thus, by examining both the observed wavelengths and their relative separations, we can accurately identify the element and pinpint the redshift of the cloud.


It turns out that the redshift of an object is directly related to the cosmic time in which that object exists. Since it takes light time to travel across the universe (at the rate of 300,000 km/s or 186,000 mi/s), process that ocured 10 billion years ago in that cloud.

We have now established that absorption lines are readily identifiable, atomic configuration giving rise to them are readily decoded, and the time at which the absorption took place is read directly from the redshift. So what it left?

If there is no evolution in the fine structure constant, alpha, then everything described above holds true. But if alpha were slightly different at say 10 billions years ago, the the precise wavelengths of iron, silicon, and magnesium would be ever-so-slighly shifted. since the shifting is sl slight, the general finger print patterns are virtually unchanges. However, very subtle difference would be detectable. For example, in a given systemw ith both magnesium and iron, compared to the magnesium lines, the iron lines will appear very sligtly shifted. A measurement of this slight shift is how we measure a small change in alpha.


What are Quasar Absorption Lines?

The Physical Picture
Consider the black-background cartoon picture (below) showing a redshift z=3 QSO to the right and a blue Earth at z=0 to the left. The redshift increases from left to right and is labeled below the cartoon on a red scale. The yellow "beam" from the QSO to the Earth represents the path of the QSO light through the universe. This particular light beam is passing through seven hydrogen clouds (red slab-like objects) on its journey toward Earth. These so called Lyman-alpha clouds are at the approximate redshifts z=1.3, 1.7, 2.1, 2.3, 2.4, 2.6, and 2.8. Finally, at z=1, the light path passes very near to a galaxy (green object). However, most galaxies have extended "halos" filled with gas (shown as red dashed circle centered on the galaxy), and the QSO light does pass through this gaseous halo. So, by the time the QSO light has reached Earth (in this example), it has interacted with seven hydrogen clouds and one galaxy halo, each being at a unique redshift.
The QSO Spectrum and Its Redshift
The remainder of the above schematic is the spectrum of the QSO. The wavelength scale (in angstroms) is provided across the bottom in blue and the spectrum itself is red. Note the narrow peak at 4862 angstroms that is labeled "Ly alpha (z=3)". This is due to excited neutral hydrogen surounding the QSO itself. Hydrogen absorbs and emits light with wavelength 1215.67 angstroms. This "emission" feature is observed at 1215.67*(1+z) = 4862.68 angstroms, where z=3 is QSO redshift. There is also a little absorption feature at 4770 angstroms due to cooler hydrogen also very near the QSO (shown as red material in the QSO schematic). Note also that there is another clear emission feature labeled "C IV (z=3)" observed at 6195 angstroms. This is due to ionized carbon around the QSO. Not that it also shows some absorption at 6100 angstroms. These Ly-alpha and C IV features are intrinsic to the QSO and help to both identify the object as a QSO and to provide its redshift.

Mapping Cosmic Time with Absorption Lines

What are we Measuring?: Line Shifts and Fine Structure Evolution

Examples of Line Shifts

Caption for Figure 1. (to the right)

Consider absorption from the Mg II 2796, 2806 doublet and the Fe II 2344, 2382, and 2600 multiplets in a gas cloud at at redshift of z=1. The laboratory wavelengths of their absorption lines will be redshifted by the factor 1+z, and therefore they will be observed at much larger wavelengths (in this case twice as large as what would be observed in the laboratory, i.e. 2796.35*(1+z) = 5592.70 as seen in top panel).

Shown to the right are (from top to bottom) the Mg II 2796, Mg II 2803, FeII 2344, FeII 2382, and Fe II 2600 transitions for a z=1 gas cloud. The black spectra show the precise locations in wavelength (measured across the bottom of each panel) of the absorption lines if there we no evolution in alpha. A slightly different alpha would would yield slightly different laboratory wavelength of the transitions, which would then be redshift by the factor 1+z to slightly different locations in the observed spectrum.

The magnitude and direction (to smaller or larger wavelengths) of the shift is different for each transition. The red spectra show the predicted positions of the example transitions for a fractional change in alpha of 1 part in 10,000 (10 times larger than what we measured).

Note thet the MgII transitions (top two) shift very little, whereas the Fe II transitions all shift toward smaller wavelengths. The insignificant shifts of the Mg II lines allows us to use them as "anchors" by which to compare the larger Fe II shifts. That is, we are able to sensitively measure the relative shifts between all observed transition (the many-multiplet method); this provides a significant improvement in sensitivty over previous methods.

Caption for Figure 2. (to the right)

Shown is absorption from the same cloud at z=1 as presented in Figure 1, but for Si II 1526, Cr II 2056, Cr II 2066, Zn II 2026, and Zn II 2062. The black spectra show the observed wavelength locations for these transitions in the case of no evolution in alpha. The red spectra show their predicted locations for an fractional change of alpha of 1 part in 10,000 (a factor of 10 larger than what we found).

Note, that similar to the Mg II transitions shown in Figure 1, the Si II 1526 transition virtually has no dependence upon alpha for such a small change. Thus, Si II serves as an "anchor" by which the relative shifts of Zn II and Cr II can be measured very precisely.

An important point with regard to possible systematic errors in our anaylsis is that the Cr II lines shift to larger wavelengths, whereas the Zn II lines shift to short wavelengths. This all but eliminates the possibility that a systematic problem with the data calibration or analysis could be yielding a false result.

A further point is that the Fe II lines are all blueward of their Mg II anchor, whereas the Zn II and Cr II lines are all redward of their Si II anchor. This also makes it very difficult to understand any systematic problems in the data calibration or analysis.
Web page by Chris Churchill: last updated Aug 29, 2K+1
cwc@astro.psu.edu
525 Davey Laboratory
Department of Astonomy & Astrophysics
Penn State University, University Park, PA, 16802
ph. 814-863-2918; fax. 814-863-3399