To attain a true idea of the nature and composition of this science [astronomy], it is indispensable...to mark the boundaries of the positive knowledge that we are able to gain of the stars....We can never by any means investigate their chemical composition.
Auguste Comte, Cours de philosophie positive (1842)
Philosophy students may call Auguste Comte (1798-1857) a great thinker, yet astronomers frequently remember him for the prediction he got wrong. Had Comte lived two years longer, he would have witnessed the discovery which revealed the ingredients of which stars are made.
For most of human history, all the information we had about stars we learned from their light. To a modern physicist, "light" covers a broad category: there is of course the "visible spectrum", which covers the colors we are used to seeing, but this is only one part of a much wider range. All this radiation, including visible light, consists of oscillations in a slightly mysterious medium called the electromagnetic field. For our purposes, the field's exact properties aren't too important; the critical point is that it can vibrate in various ways, much like the way ripples travel over a smooth lake. The "wiggles" in this field all travel at the same speed, roughly 300,000 kilometres (186,000 miles) per second. Although the speed of all "electromagnetic waves" is the same, the waves can have different wavelengths: the distance from peak to peak can vary. The shortest wavelengths, belonging to the light we call gamma rays, can be less than one ten-billionth of a metre. Radio waves, on the other hand, are light with wavelengths that can be tens, hundreds or thousands of metres long.
If we restrict our attention to the kind of light that human eyes can easily see, we find that red has the longest wavelength (about 700 nanometres, where a nanometre is one billionth of a metre). Violet has the shortest wavelength, around 400 nanometres. In between are all the familiar colors, the "ROY G BIV" we all learnt in elementary school. Light whose wavelength is just slightly longer than red is termed infrared, while light whose wavelength is shorter than violet is called ultraviolet. The cells in our eyes which receive light are weakly sensitive both to ultraviolet and infrared. Wearing goggles made from material that blocks all light except infrared lets the wearer observe the infrared radiation that one normally doesn't notice. The cornea, a transparent surface at the front of the eye, screens out ultraviolet (which would otherwise be harmful to the eye's interior). Patients who have had corneas removed sometimes report odd glows, coming from objects like newly washed white socks. (Laundry detergents often include ingredients that glow in visible light when exposed to ultraviolet, a phenomenon called fluorescence. This makes clothes appear "whiter than white", but has unintended consequences when people can actually see ultraviolet.)
"White light" is actually a mixture of all the ROY G BIV colors. If one color is more predominant--if the mix has a stronger dose of light in one particular range of wavelengths--then the white light is tinged a certain hue. For example, the Sun shines yellowish. This concept of light is an old one: in the thirteenth century, Theodoric of Freiburg proposed a theory of the rainbow, according to which rain left tiny droplets of water in the air that bent light. Water bent different colors of light to different degrees, producing the separation we see. The great English scientist Isaac Newton was the first to make the idea explicit, however, by shining sunlight through a glass prism and observing the resulting separation.
We can learn a surprising amount about starlight by watching what happens with everyday fire. This is a trifle odd, because the nuclear reactions which drive stars are far more powerful than the chemical reactions which produce ordinary fire. (The Sun would have gone dark after a few thousand years if it relied upon even the most efficient chemical fuel, yet all evidence indicates that it has shone steadily for a timespan a millionfold longer.) However, nature is often remarkable in how it presents the solution to distant puzzles in phenomena we find close at hand. Therefore, on we go:
To start with, what affects the color of fire? In other words, what is responsible for making more light of one wavelength than of another? First of all, we note that flames from different sources have radically different appearances. The flame from a gas burner on a stove is bluish and not terribly illuminating, while wood burning in a fireplace or campfire is a dancing mix of reds, yellows and oranges. Observing more closely, we might catch a blue glow near the base of the campfire flame, near the burning wood. Furthermore, if we let the fire die out, we might see bark on the fuel logs turn from orange-yellow to a dull cherry red, before becoming cold and black. Perhaps a red glow indicates a lower temperature than orange or yellow, and a blue flame indicates hotter burning still.
This turns out to be the case, although explaining why stumped scientists for an awfully long time. In 1860, the German physicist Gustav Robert Kirchhoff (1824-1887) suggested studying solid black objects, which absorb all colors of light equally well. Kirchoff had noticed that objects tended to absorb certain colors while cool and emit those same colors when heated, so a black body--which absorbed all colors equally well--should radiate light of all wavelengths. Deciding how much of each wavelength light the black body would emit--in other words, which color would predominate in its "emission spectrum"--was a more difficult problem.
The Austrian physicist Josef Stefan (1835-1893) studied a wide variety of substances heated to many temperatures, and in 1879 he announced that the total amount of radiation a body emitted depended upon the fourth power of its "absolute temperature". By this point, scientists had already deduced that heat was the jittery motion of molecules, and that the minimum possible temperature--absolute zero--must be when all molecular motion ceased. The British physicist William Thomson (1824-1907), who inherited the title Lord Kelvin, proposed a temperature scale based on this fact. Each degree would be the same size as a Celsius degree (1.8 Fahrenheit degrees), but the zero point would be at absolute zero, the coldest anything in the Universe could ever be. (Later, the German physical chemist William Nernst (1864-1941) reasoned that indicated absolute zero could never be reached, a conclusion called the "Third Law of Thermodynamics".) Lord Kelvin reckoned that 0 degrees Celsius, the freezing point of water, was about 273 degrees above absolute zero. Later scientists made the figure more precise and named the scale in Kelvin's honor, so that one can say the freezing point of water is 273.15 kelvins. (Officially, the phrase "degrees Kelvin" is no longer used, and we simply say "kelvins" instead.)
So, according to Stefan, if an object radiates a certain amount of light at 1,000 kelvins, if we double the temperature to 2,000 kelvins, the object will radiate 24 = 2 x 2 x 2 x 2 = 16 times as much. If we tripled the temperature to 3,000 kelvins, the total radiation output would increase by 34 = 81 times.
This observation is nowadays called the Stefan or Stefan-Boltzmann Law. Using this rule and careful observations of the Sun's energy output, Stefan deduced that its surface temperature was around 5,700 K. At that temperature, Earthly objects also radiate yellowish-white light. (Lightning bolts give off flashes of purer white, which might suggest that they are hotter than the Sun's surface. This turns out to be the case: lightning bolts have been recorded five times hotter than the solar surface--but the temperature inside the Sun's core is much hotter still.)
The next task was to figure out why some wavelengths are emitted more predominantly than others, and why this changes with temperature. Throughout the latter decades of the nineteenth century, physicists struggled to devise an equation that would predict how a hot ember glows orange, then turning red as it cools. In 1900, the German physicist Max Planck (1858-1947) found an equation which reproduced the experimental observations that had been made of blackbody radiation. He then set about trying to find a way to deduce his equation from first principles. Like others before him, he found that this was impossible, according to the way light was then understood. He found that he was only able to reproduce the observed result if he assumed that the radiating blackbody gave off light in little pieces, instead of continuously, and that the energy of each discrete piece was inversely proportional to the wavelength. For example, violet light has roughly half the wavelength of red light, so a "lump" of violet light would carry twice as much energy as a "lump" of red light. Planck called these energy units quanta, from a Latin word meaning "how much?" (the singular is quantum, which became the adjective to describe Planck's theory).
Planck suspected that quanta might merely be a mathematical trick for solving this one particular problem, but five years later, Albert Einstein (1879-1955) used the idea of quanta to explain how some metals can eject electrons when exposed to light. This phenomenon, the photoelectric effect, had likewise eluded earlier attempts at explanation, until Einstein resolved it. Because the photoelectric effect had nothing to do with blackbody radiation, it suddenly appeared that light really was made up of quanta, and that they were not a mere trick good for only solving one problem. (In fact, according to Einstein's Special Theory of Relativity, which he published in the same year, if you assume that light comes in quanta, the energy of each quantum has to be inversely proportional to the wavelength, just as Planck had originally imagined.)
Thanks to Planck, if we observe an object to radiate a certain color, we can deduce its temperature, if it is a reasonably close match to the idealized blackbody. However, not all light sources fall into that category. Let us return to the campfire and observe another effect: if we throw a candy wrapper or a wadded-up magazine page onto the fire, it may flare a bright green before crumpling up into a heap of ashes. Surely, the magazine paper is burning at about the same temperature as newsprint, so why the difference?
Suspicion falls on the chemical distinctions. Magazine paper is made with fillers to make it glossy, and it is often covered with various inks. Perhaps the ink reacts to heat in a way that gives off light wavelengths different from Planck's blackbody spectrum?
Clearly, the thing to do is to test chemical substances for the flame colors they produce. In the mid-nineteenth century, the German chemist Robert Bunsen invented a new type of burner specially adapted to this purpose. Gas made by cooking coal flowed up a pipe and out a nozzle, where it was ignited. Normally, the gas burned blue, just like a gas stove's flames, but Bunsen found that mixing air with the gas before burning it produced a colorless, non-luminous flame. Therefore, anything placed within the flame and ignited would only display its own colors, not those affected by the flame itself.
Gustav Kirchhoff, who proposed the study of blackbody radiation, made use of Bunsen's burner. In 1859, he announed that all the chemical elements he had studied each produced a distinct spectrum, which he could identify by passing the flame's light through a prism. The prism spread out the light, but instead of seeing a continuous rainbow, Kirchhoff observed a pattern of distinct lines. The pattern of lines was different for each element, and what's more, Kirchhoff saw that when light passed through a vapor of some element, the element absorbed the same colors as it emitted when it was heated. (This led to his interest in the blackbody spectrum.)
Kirchhoff realized that if a new rock sample were heated and its spectrum found to contain lines not characteristic of any known element, the sample must contain a new element, hitherto unknown. Using this technique, he discovered cesium a year later, in 1860. He named it after a Latin word meaning "sky blue", because a spectral line of that color gave away the new element's existence. A year after that, he discovered the element rubidium, which he named for the Latin word meaning "red".
Familiar with spectral analysis, Kirchhoff also pointed out that dark lines visible in the Sun's spectrum must be from elements in the solar atmosphere. He reasoned that gases in the Sun's atmosphere must be cooler than those on the surface or deep inside, so that their atoms would absorb their characteristic wavelengths and make dark lines across an otherwise continuous spectrum. Using this technique, Kirchhoff detected the presence of hydrogen, sodium and several other elements.
Now, the spectrum lines are light absorbed. This and the above image illustrate the difference between "emission" spectra and "abosrption" spectra; spectroscopy is quite the versatile thing.
In 1868, the French astronomer Pierre-Jules-Cesar Janssen (1824-1907) took advantage of a total eclipse to study the spectra of solar prominences, eruptions in the Sun's surface which are normally only visible when the main part of the solar disc is blocked from view. (Otherwise, they are lost in the Sun's overall brightness.) Janssen detected a spectral line which did not match any known element. The British astronomer Joseph Norman Lockyear (1836-1920), who also worked on solar spectra, decided that this line must represent a new element. The new element, named helium from the Greek word for "sun", was the first one discovered off the Earth before it was observed on our own planet. (In later years, other unidentifiable lines have been observed in various celestial sources, but they have turned out to emanate from known elements in unusual conditions. Helium remains unique in this respect.)
Helium remained unknown on Earth until 1895, when the British physicist John William Strutt, Lord Ramsay (1842-1919), discovered trace amounts of it inside a sample of clevite, a uranium ore. Several more years passed before the helium's presence was explained; the source turned out to be radioactive decay.
Spectral lines are useful for more than identifying elements. When a source of waves moves relative to your position, the frequency of waves you receive changes. This phenomenon was first explained by the Austrian physicist and mathematician Christian Doppler (1803-1853), and it is therefore known as the Doppler Effect. The classic example is a train whistle: as the train approaches, you can hear its whistle rise in pitch. The sound's pitch falls again as the train recedes into the distance. The pitch of a sound is the sound wave's frequency, just like a light wave's frequency gives its color. The Doppler Effect applies to light as well, although because the speed of light is almost nine hundred thousand times greater than that of sound, the frequency shift is harder to notice. Stars and galaxies often move with sufficient speed that the Doppler shift is noticeable, and the Doppler Effect is therefore a valuable measurement tool.
You might notice a problem with using the Doppler Effect. If a star shines white light, which contains all colors and thus all frequencies, how can we detect a shift? Won't the ultraviolet frequencies just beyond the visible range be shifted into the violet, and the red frequencies into the infrared, leaving nothing to be observed? Not quite. First, recall that a "black body" radiates more light at some frequencies (i.e., colors) than at others. The "frequency distribution" will peak at a particular place, which (as Kirchhoff observed) depends upon the temperature. The Doppler Effect shifts this peak towards the red or the blue end of the spectrum, "sliding" the frequency graph one way or the other. It turns out that a Doppler-shifted blackbody spectrum from a body at one temperature doesn't look exactly like a blackbody spectrum produced at any other temperature, so the shift can still be detected. A second way to notice the Doppler shift is the following: spectral lines, remember, cross the frequency band at definite places, depending upon the star's chemical composition. If we find the familiar lines for hydrogen, say, in an unexpected part of the spectrum, we can reasonably say that the star was moving relative to us and therefore had its light Doppler-shifted to new frequencies. By measuring the displacement between the lines of stationary hydrogen and those of the moving star, we can work out how the star is moving through space.
The astronomer Maarten Schmidt (1929- ) working at Caltech in Pasadena, California, applied this method to strange objects astronomers had been unable to explain. Radio astronomers, using large antennae to receive signals, had detected unexpectedly strong radio sources in deep space. Because these objects looked like small, starlike points of light, they were termed "quasi-stellar radio sources". This mouthful was soon contracted to the short word quasar. Quasars showed spectral lines in the visible light region which no one could identify. Schmidt realized that these peculiar lines were in fact typical ones for heated hydrogen normally observed in the ultraviolet range. In order for these UV lines to appear where they were, deep in the visible, the quasars' light would have to have been red-shifted to a fantastic degree. This in turn implied that the quasars were moving away from Earth at an incredible speed, a considerable fraction of the speed of light. We now know that quasars are billions of light-years distant from our home galaxy. Because it takes light billions of years to travel that distance, we see quasars now as they were in the distant past; they are among the earliest objects we can find in the Universe. For their light not to have faded entirely as it traveled one hundred billion trillion kilometres to reach us, quasars must have been extraordinarily bright. The source of their energy still puzzles astronomers and physicists. We would not have had this mystery to solve, however, without a careful study of their spectral lines.
For a clear account of the way scientists interpret the "wiggles" of the electromagnetic field, see Carl Sagan's Pale Blue Dot (Random House, 1994). The accomplishments of Theodoric of Freiburg are told in James Burke's The Day the Universe Changed, in both the TV series and the companion book (Little, Brown; 1985). Burke, in turn, draws upon William A. Wallace's The Scientific Methodology of Theodoric of Freiburg (Freiburg: The University Press, 1959).
Researched, written and maintained by Blake Stacey.