It’s remarkable that astronomers can tell precisely what stars are made of, even though they cannot extract physical samples from them. The way they learned to do this is one of the great stories of astronomy, ranking in significance alongside the invention of the telescope. It starts with the unexpected villain of the piece, a 19th-century Parisian philosopher by the name of Auguste Comte. In many respects, he is a hero of science, since his philosophy of reason and the importance of rigorously testing ideas is at the heart of the scientific method. In 1835, however, he let himself down in regard to our understanding of the stars, when he confidently asserted that we would ‘never be able by any means to study their chemical composition’, and that such attributes as their density and temperature would be ‘forever denied to us’.
Well, never say never. Particularly since, in that same year, scientists were already taking steps to understand the means by which we might investigate those matters. In August 1835, the English scientist Charles Wheatstone carried out a telling demonstration at the fifth meeting of the British Association for the Advancement of Science, held in Dublin. He used a prism, whose ability to deconstruct sunlight into a band of rainbow colours from deep violet to deep red had been established by Isaac Newton 170 years earlier, leading Newton to invent the term ‘spectrum’.
But instead of using the prism to split sunlight into its component colours, Wheatstone pointed it towards an electric spark formed between two metal electrodes. Rather than a spectrum composed of a continuous band of colour, the prism revealed a set of discrete narrow lines of light, each an image of the spark itself, but composed of a single colour. It was as if the other colours between the lines had been rubbed out. We call these features ‘emission lines’, and now know that each one corresponds to light of a differing microscopic wavelength, with the violet lines having wavelengths about half those of the red. So, while plain white light, from, say, an incandescent lamp, is composed of gazillions of adjoining wavelengths producing what is predictably known as a continuous spectrum, the discrete bright lines of the emission spectrum are produced by the atoms of the metal excited by the spark.
Different metals emit different patterns of bright lines, as Wheatstone gleefully pointed out in Dublin. And that is the key to being able to determine remotely what stars – and many other classes of celestial object – are made of. In fact, earlier work by other English scientists had already shown that differing salts burned in a flame also produced differing sets of emission lines, but it was Wheatstone’s demonstration that created interest in the topic.
Then, barely two years after Comte’s death in 1857, a not-quite-household-name physicist at the University of Heidelberg by the name of Gustav Kirchhoff carried out a detailed analysis of the subject. He worked closely with a chemist who really is a household name – Robert Bunsen, of Bunsen burner fame. Together, they devised an improved device for viewing the spectra of light sources – a spectroscope – and used it to make the crucial discovery that every element has its own unique emission-line spectrum, not just a few metals. It is as if nature itself has hidden an identifying barcode in the light of every chemical imaginable. Once that barcode has been revealed by the spectroscope, the identity of the material is known.
There is a well-known photograph of these two great scientists standing together, probably taken early in their collaboration during the 1850s. The statuesque Bunsen towers over his younger, more slightly built colleague, giving them the vaguely comical appearance of a Laurel and Hardy of spectroscopy. Be that as it may, their collaboration yielded the fundamental rules of light analysis, embodied in what are known as Kirchhoff’s Laws.
Briefly, they are (1) that incandescent bodies such as a white-hot lump of metal or a glowing electrical filament emit a continuous spectrum, (2) that materials excited in a spark or flame emit their own characteristic emission-line spectrum, as we have seen, and (3) that if you view the continuous spectrum of a hot object through a cooler gas, you’ll get what is known as an absorption spectrum. What’s that? Almost miraculously, the colours (that is, wavelengths) that would be emitted by the gas if it was excited are subtracted from the continuous spectrum of the background source, producing a ribbon of colour crossed not by bright lines, but dark ones. Not surprisingly, they are called absorption lines, since the light of the background source has been absorbed at those wavelengths – absorbed by the intervening gas. And, once again, the pattern of dark lines unambiguously identifies the gas through which the light has travelled.
In 1861, Kirchhoff and Bunsen were able to show that dark lines in the continuous spectrum of the Sun – recognised since 1802, but never understood – were absorption lines produced by known elements in the Sun’s atmosphere. The atmosphere is at a lower temperature than the underlying luminous gas, which, by the way, is called the photosphere – the visible ‘surface’ of the Sun. At a stroke, the two scientists had definitively established what the Sun is made of, despite the intervening 150 million kilometres. It is mostly hydrogen, but the spectral signatures of many other elements are there, too. The confident proclamations of Auguste Comte were, by now, seriously under threat.
THE FINAL BLOW CAME LATER IN THE 1860s. ANOTHER Englishman, a fellow subsequently aided by a wife who was at least as able as he was, had sold his family business in order to pursue his interest in astronomy. Equipped with a telescope capable of serious research, he greeted the news of Kirchhoff and Bunsen’s work on the solar spectrum with enthusiasm, and resolved to investigate whether stars showed the same kinds of spectroscopic signatures as the Sun. His name was William Huggins, and he enlisted the help of a friend – a professor of chemistry from King’s College, London, by the name of William Miller. Together, Huggins and Miller built a spectroscope for the telescope, and then embarked on a tour of the heavens, checking out everything bright enough to reveal a spectrum to their eager eyes.
What they found amazed them. While the Moon and planets exhibited essentially the spectrum of the Sun (as expected, given that they shine by reflected sunlight), the spectra of stars varied significantly. We now recognise that this is due principally to differing sizes and temperatures, a subtlety unknown to Huggins, but he had no difficulty grasping the main message. The barcode signatures of familiar earthly elements were there before his eyes in the absorption lines of the stars. As he later wrote, ‘a common chemistry…exists throughout the universe’. What a breakthrough. Huggins and Miller published their catalogue of the spectra of 50 stars in 1864, and the new science of astrophysics was born.
It used to be thought that Huggins’ wife, Margaret, whom he married in 1875, only assisted him, but recent studies have shown clearly that they were an equal partnership, with several jointly authored papers to their credit. Moreover, Margaret’s technical interests, which predated her marriage to William, enabled her to make innovations that significantly furthered their research. She was the one, for example, who promoted the idea of photography in studying the spectra of the stars, attaching a camera to a spectroscope to make what is still known as a spectrograph. Today’s instruments are equipped with state-of-the-art electronic sensors rather than photographic plates, and are as sensitive as the laws of physics allow. But they work on the same principle as Margaret Huggins’ spectrograph.
OF THE HUGGINSES’ DISCOVERIES, WE SHALL HEAR MORE in this chapter, but there was one observation that eluded them. It had been known since the 1840s that starlight should exhibit something known as the Doppler effect. Most people are familiar with it, even if they might not be able to put a name to it. When applied to sound waves, it’s the change in pitch that occurs when a sound source moves, most commonly heard when a fire truck or ambulance speeds by with its siren blaring. The sound is higher pitched when the emergency vehicle is approaching, and lower as it recedes, and the effect is caused by the wave-motion of sound.
The fact that exactly the same thing happens with light waves means astronomers can measure the speeds of objects along the line of sight, whether these are planets, stars, galaxies or whatever. They look for a shift in the spectrum lines, and by measuring it, can deduce the object’s velocity in the radial direction (that is, towards or away from us – ‘towards’ producing a blue-shift, and ‘away’ a red-shift). The spectroscope or spectrograph in effect becomes a celestial speedometer.
These are delicate measurements to make, however, and while the Hugginses attempted them several times from 1868, it was not until 1889 that Hermann Carl Vogel, Director of the Astrophysical Observatory in Potsdam, obtained the first reliable measurements of stellar radial velocities photographically. In fact, Vogel’s initial work concerned the bright star Algol in the northern-hemisphere constellation of Perseus. This star varies in brightness, and was already known to be what is called a binary system – that is, two stars orbiting a common centre of mass. Vogel detected a periodic shift in the spectrum lines, which he correctly interpreted as being due to the brighter of the stars exhibiting radial velocity changes as it orbited its companion. Such objects are known as spectroscopic binaries, because it is usually the case that the regular velocity changes are the only symptom of their duality. In their visual appearance, they are indistinguishable from single stars.
LET ME MENTION A FEW MORE APPLICATIONS OF THE spectroscopic technique. As Vogel’s work on Algol suggests, the Doppler effect can be used to deduce whether – and how fast – things are rotating. We saw a few chapters ago that it was used in the 1890s to show that Saturn’s rings rotate not like a solid object, but as a swarm of particles. The technique extends across the whole gamut of astronomy, from rotating planets, stars and gas clouds to whole galaxies of billions of stars.
Astronomers today are also using the effect to find things that are completely invisible. The planets of stars in the Sun’s neighbourhood are, for the most part, too faint to see directly, even with the largest telescopes, but they can reveal themselves by the way they tug on their parent stars as they orbit around. The resulting backwards and forwards component of the star’s motion is very small, ranging from several metres per second in the case of Jupiter-sized planets to just a few centimetres per second for Earth-like objects. Despite that, the velocities can be detected with advanced equipment, and this so-called ‘Doppler wobble technique’ is routinely used at several of the world’s major observatories, including the Anglo-Australian Telescope at Siding Spring. Actually, the biggest problem is calibration, since you have to compare super-precise observations taken days or sometimes weeks or months apart as the planets move around their parent stars, and you need to be sure that all the spectrum lines are measured against the same zero-point. Some novel and exotic optical devices are used for this – iodine cells and photonic combs, for example.
You might also be surprised to learn that magnetism can be detected by its effect on light. It was a Dutch physicist, Pieter Zeeman, who noticed in 1896 that spectrum lines (both emission and absorption lines) split into several components when the light is emitted in a magnetic field. This so-called Zeeman effect allows astronomers to probe the magnetism of the Sun and stars. And, by combining the Zeeman effect with the Doppler shift, it is possible to make maps of magnetised spots on stars (like the sunspots visible on the Sun), even though the stars are too far away for their discs to be visible. This complex but highly effective technique is called Zeeman Doppler imaging, and it’s also carried out at the Anglo-Australian Telescope, principally by colleagues from the University of Southern Queensland.
And, finally, there’s the expansion of the Universe. In 1929, American astronomer Edwin Hubble used a spectrograph to discover that galaxies are flying away from us with speeds that are proportional to their distances. Rather than being due to the Doppler effect (which is caused by the motion of an object through space), these so-called ‘recession velocities’ are interpreted as being due to the Universe itself expanding. In other words, space is getting bigger, and it carries the galaxies along with it. In homage to its discoverer, we call that overall expansion the Hubble flow.
Because the light from these galaxies has been travelling for hundreds of millions if not billions of years, the Universe has expanded significantly since it was emitted. The light waves themselves have participated in the expansion, so they arrive at our telescopes stretched to a longer wavelength than when they set out. That means the light spectrum – including the barcode of emission or absorption lines – is shifted to the red. This effect is called the ‘cosmological redshift’ to distinguish it from the simple Doppler effect, and it is one of the most remarkable tools available to astronomers.
Its effect is to date-stamp the light with the time when it was emitted. Because we know what the barcode of spectrum lines looked like when it left its source, we can measure directly how much redshift it has experienced. Thus, astronomers can deduce how much smaller the Universe was when the light set off, relative to today’s Universe. And, knowing how the size of the Universe changes with time, they can calculate when the light left the galaxy that emitted it. Once again, this work is a major part of investigations at the Anglo-Australian Telescope, where the technique is used to make detailed three-dimensional maps of the Universe. They reveal the structure imprinted by the Big Bang – the event in which the Universe is believed to have been created some 13.8 billion years ago. And they are also being used to investigate some of the most pressing questions in modern astronomy, concerning the nature of dark matter and dark energy.
THE CONSTRUCTION OF SUCH DETAILED MAPS INVOLVES one further trick of the trade, and it’s something I’ve been deeply involved with during my career in astronomy. When Huggins, Hubble and countless other astronomers throughout history made their spectroscopic observations of stars and galaxies, they had no alternative but to make them one at a time. And each observation took approximately forever. I have always admired the work of an early 20th-century American astronomer called Vesto Slipher, who carried out some of the observations of galaxy spectra used by Edwin Hubble in formulating his work on the expanding Universe. Slipher’s catalogue of the spectra of 25 galaxies, published in 1917, required between 20 and 40 hours of observing for each galaxy. That meant observing the same object night after night to build up enough information on one of the crude photographic plates then in use, before developing it to reveal the faint spectrum.
Today’s catalogues of galaxies are measured in hundreds of thousands, and will soon be in multi-millions. And the same is true of catalogues of stars in our own Galaxy. How are such totals achieved? The amount of exposure time per observation has fallen from tens of hours to tens of minutes by virtue of bigger telescopes, more efficient spectrographs and super-sensitive electronic image-sensors. However, even with such advances, astronomers would still be limited to observing their targets one at a time if it were not for the trick of the trade I spoke of. And that is to use clever technology to permit astronomers to observe hundreds of objects at a time – which, very soon, will increase to many thousands.
Most large telescopes have a reasonably wide field of view – that is, they see a significant chunk of sky with each observation. In a truly wide-field telescope like the United Kingdom Schmidt Telescope at Siding Spring Observatory, you can see an area of sky six degrees across – a dozen times the diameter of the full Moon, and big enough to encompass the whole Hyades Cluster. Most telescopes have a smaller field size than this, but you get the idea.
So – the telescope is really effective at presenting you with images of a large number of target stars or galaxies in its field of view, but how do you transfer individual samples of their light into the spectrograph? The answer is with optical fibres, thin strands of glass-like material that are as flexible as guitar strings, but transport the light from one end to the other with virtually no loss of intensity. If you have a bundle of hundreds of fibres, and can position one end of each accurately on a selected object, the flexibility of the fibres allows you to take the light to a convenient and stable location, often metres from the telescope, and then arrange the whole lot neatly in a straight line at the other end. Why? Because that is just what is needed to analyse them all simultaneously in a spectrograph. And the trick works like a dream.
The one difficulty with the multi-fibre spectroscopy technique is that each fibre has to be aligned exactly with its selected target in the telescope. That has to be accurate to a tiny fraction of a millimetre, and demands sophisticated robotic technology which has taken several decades to perfect through successive phases – most of which I have been directly involved with. Some of my colleagues have kindly referred to me as one of the pioneers of multi-fibre spectroscopy, and I guess it’s true that I was the first to do various things, like using the technique to observe stars rather than galaxies (in 1982), and building some ground-breaking instruments for various large telescopes during the 1980s and 90s. I can’t lay claim to writing the world’s first PhD thesis on the topic, though. That honour goes to an eminent US astronomer by the name of John Hill. Mine was the second.
I don’t think it’s too immodest to say that the discoveries that have been made using this technique have revolutionised astronomy, with multi-fibre instruments now being used on most of the world’s biggest telescopes. I’ve already mentioned large-scale galaxy surveys that help us understand both the way galaxies evolve and the way the Universe as a whole has evolved. But the spectroscopic observation of large numbers of stars is giving us similar insights into the structure and evolution of our own Milky Way Galaxy. From 2003 to 2013, the UK Schmidt Telescope I mentioned earlier was occupied by a survey called RAVE – the RAdial Velocity Experiment. Remember radial velocities? I was the project manager for RAVE, and I’m delighted that the final catalogue of half a million star velocities and other characteristics is just about to be published. Meanwhile the larger Anglo-Australian Telescope is undertaking a survey of a million stars known as GALAH – which might sound like a much-maligned Australian parrot, but is actually GALactic Archaeology with HERMES. Of course, HERMES is itself an acronym for the super-sensitive home-grown spectrograph being used, while galactic archaeology is the investigation of the history of our Galaxy by measuring the exact chemistry of as many stars as possible. And there are new surveys in the offing, using new technology that will extend the capabilities of multi-fibre spectroscopy well into the 2030s. I feel privileged to have been so closely involved with this revolution.
LET ME RETURN FINALLY TO ONE OTHER DISCOVERY MADE by William Huggins. While it wasn’t really his fault, it became an example of unjustified scientific hubris rivalling that of Auguste Comte. In the early 1860s, when Huggins was wondering whether there was anything his new science of astronomical spectroscopy couldn’t achieve, he turned his attention to one of the great scientific problems of the time. That concerned the nature of nebulae – ill-defined misty patches in the sky that were neither stars nor planets. The Big Question was whether they were made of myriads of stars too faint to be seen individually, or something else, such as a cloud of glowing gas. Or (as we now know to be the case), an assortment of both.
Huggins directed his spectroscope towards one of these nebulae in the August of 1864, and was amazed by what he saw. Emission lines – the clear signal of an excited gas – rather than the absorption-line spectrum of a cloud of stars. As he recalled three decades later, ‘The riddle of the nebulae was solved. The answer, which had come to us in the light itself, read: Not an aggregation of stars, but a luminous gas.’ At the age of 40, Huggins had revolutionised the astronomy of his time, and his place in history was assured.
But there was a snag. Very soon, astronomers realised that some of the emission lines they could see in various nebulae didn’t belong to any known element on Earth. Yes, hydrogen was there, but what was this bright green line that didn’t correspond to anything they’d seen already? And others? It was, indeed, a puzzle – but there was a precedent in a strange yellow line that had been observed in the spectrum of the Sun during an eclipse in 1868. Two English scientists, Norman Lockyer and Edward Frankland, had deduced that this was the signature of an unknown element that was present in the Sun, but not on Earth. Dubbed ‘helium’, it was expected to reveal itself some day in the inventory of terrestrial chemical elements. And so it did – in 1895, in the hands of a Scottish chemist by the name of William Ramsay, who isolated it from a mineral known as cleveite. It was the first chemical element to be discovered in space rather than on Earth – a triumph for astronomical spectroscopy.
It’s no wonder, then, that astronomers should take it for granted that the unidentified emission lines in the spectrum of nebulae, including the mysterious green line, were the spectral signature of another unknown element. With supreme confidence, they dubbed it ‘nebulium’, a name Margaret Huggins first recorded in 1898, but probably did not invent. Using the measured wavelengths of the lines, and improvements in the understanding of atoms, scientists worked hard to discover the properties of nebulium. In an impressive research paper published in 1914, for example, a trio of eminent French astronomers even deduced that it must be two different elements, but got no nearer to identifying what they were.
At last, with improvements during the early 20th century in our understanding of why emission lines occur at all, the mist started to clear. The fact that excited atoms emit light with very specific wavelengths comes about because of specific energy levels occupied by the electrons clouding around their nuclei. Particular wavelengths are emitted when the electrons jump from one energy level to another, emitting a certain ‘quantum’ of light. Sound familiar? Yes, it’s the foundation of quantum theory. But one of the theory’s quirks is that it incorporates so-called selection rules. Some of those energy transitions are permitted, while others are forbidden. They don’t happen. Actually, they do, but only if the excited atoms belong to a gas at a pressure much, much lower than anything possible in a laboratory on Earth. A rarefied gas in the depths of space, for example.
It was a 28-year-old genius at the California Institute of Technology by the name of Ira Sprague Bowen who, in 1927, was busy calculating the theoretical wavelengths of light that would be emitted by the electron transitions of various elements. Of course, he followed the selection rules – until he realised that the forbidden lines weren’t really forbidden, but just extremely unlikely at the gas pressures encountered on Earth. In a moment of brilliance, he thought of nebulium. Going back to his calculations, he worked out what forbidden emission lines might be emitted if the selection rules didn’t prohibit them. And sure enough, when he looked at oxygen, the forbidden wavelengths matched those of nebulium perfectly – including that bright green line. Eureka! Inspired, Bowen feverishly calculated the forbidden lines that other elements would emit, and obtained similar outcomes. His results eventually appeared in a seminal paper in 1928. And nebulium was consigned to the history books. Ira Bowen went on to have an outstanding career in the astronomy of the United States, masterminding some huge advances in both the science and technology of astronomy, until his death in 1973.