The Lives of Stars

Wednesday, 5 February 2014 - 1:00pm
Museum of London



Stars are a ubiquitous feature of our local Galactic environment. They do not last forever – but form, develop and evolve over timescales of millions or billions of years, eventually to expire in dramatic style. We look at the lifecycles of different kinds of stars, and in particular, what they reveal about our own origins.


Transcript of the lecture

5 February 2014
The Lives of Stars
Professor Carolin Crawford
This talk concerns stars, the most easily observable objects in the night sky. The nearest to us − and the most easily studied example − is the Sun, which will be a topic for one of next year’s lectures. They are all massive spheres of gas that exist in the form of stars for only as long as they can fight the inexorable inward pull of gravity by producing energy deep within their core. Stars differ from one another in fundamental properties that may not be so obvious to the first glance, but which dictate that they live out very different kinds of lives and deaths. Stars are not eternal and everlasting, though they may as well seem so compared to human lifespans. Our Galaxy is a dynamic environment - stars are being born, and stars are dying even as you read this; and during their lives they generate all the elements that prove so crucial to the formation of planets, and to our own existence.
To the unaided eye, it is immediately apparent that not all stars shine with the same brightness, and that they are not distributed evenly across the night sky. If you are fortunate to have a clear dark night, you may even see that they are not all simply white in colour: for example there is a marked contrast between the very orangey-red Betelgeuse at one of Orion’s shoulders and the very blue-white Rigel marking his opposite knee. The difference in star colours is best recorded in digital images, as there is not enough light from stars to trigger colour vision in the human eye. 
From fairly simple observations we can infer much about the intrinsic physical properties of stars, and it soon becomes apparent that the Sun is only a fairly average and (thankfully!) pretty dull star. Other stars display an enormous range of: brightness (stars shine with luminosities that are tens of thousands of times fainter and over a million times brighter than our Sun); of surface temperature (from below 3,000K to greater than 30,000K); of size (with diameters of about ten km for neutron stars, about 1.4 million km for our Sun, and up the largest known star at just under two billion km); of mass (from about a tenth the mass of our Sun to around 100 times greater); of age (some still being formed, while others are perhaps eleven billion years old); of chemical composition (some are of primordial composition, some have more of the heavier elements present in their atmosphere); and of environment (not all are as isolated as the Sun - many live in binary systems, where two stars continually orbit one another; others are gathered into loose open clusters containing between tens to hundreds of young stars, or into the tightly packed round globular clusters alongside hundreds of thousands of companions). 
A fundamental key to understanding how different types of stars develop through their lives is the observation that these properties are related to one another. But to make a meaningful comparison of stellar properties, let alone how they then determine the process that is known (a little misleadingly) as stellar evolution, we need to have knowledge of the intrinsic or true, physical properties. This is not a trivial process, and depends on being able to first establish the distance to a star. 
We can only observe how bright a star appears to be. A star shining with the same luminosity will appear fainter if viewed at a larger distances. How much of the light from a star is actually captured by a telescope at the Earth is diluted down by the inverse square of its separation from us. The fact that many of these stars appear so bright in the night sky despite their colossal distances immediately demonstrates the incredibly luminosities under consideration. [A discussion of how we measure distances to distant stars is interesting in its own right, and is postponed to another lecture in April on How the Earth moves.]
If you ever dip a toe into the world of amateur astronomy, you rapidly run into the archaic measure we use for the brightness of stars, known as their magnitude. This measurement scale was originally devised by Hipparchus, a 2nd Century BC Greek astronomer who introduced it as a system to compare the brightness of different stars: the very brightest stars that could be seen with the unaided eye were labelled ‘first’ magnitude, the very faintest marked ‘sixth’ magnitude, and all other stars were distributed between these two extremes according to their brightness. As this is a scale that stems from observations with the unaided eye, it is an example of a logarithmic scale, because visual perception does not follow a linear relationship. Second magnitude stars are about as half as bright as first-magnitude stars, whereas a sixth magnitude star is about 100 times fainter. The scale is also confusing as it works ‘backwards’, in that dimmer stars have greater magnitudes – and it doesn’t help that since the scale was introduced, some stars were subsequently measured to be brighter than first magnitude, so negative values of magnitude had to be introduced (for example, Sirius, the brightest star in the night sky, has a magnitude of -1.5). The scale was refined when astronomers were able to better record the light from stars in the nineteenth century, and a difference of magnitude of one is now defined to correspond to a drop in brightness by a factor of 2.512. We still use the magnitude scale, even though we no longer have to rely on the human eye for recording our observations; the largest telescopes in use now can regularly objects observe down to magnitude of 27. 
The problem remains that magnitudes are still an observed rather than an intrinsic measure of the luminosity of the star. To be able to compare the true brightness of stars we scale them to an ‘absolute’ magnitude, which is the magnitude a star would have if it were placed at the (not entirely arbitrary) distance of 32.6 light-years away from us. The range of absolute magnitudes of stars varies between -6 to 16, with the Sun featuring at only a very modest ‘five’. 
A further complication is that stars have different colours: some will produce far more blue light than red, and vice versa. The colour a star emits most strongly depends on its surface temperature; hotter objects give off more of their light at shorter (ie bluer) wavelengths of light, and cooler objects glow with redder colours. If we plot the amount of light radiated at different wavelengths by a star (ie its spectrum), it follows a characteristic curve known as a black body. This spectrum is generated by thermal radiation, and where curve peaks can give a good estimate of the temperature of the body giving off the light. So at first pass, the dominant colour of a star reveals its temperature. Most of the light that we receive from the surface of Sun comes from a region known as the photosphere, and its distinct yellowy-white colour (better viewed from space than filtered through our atmosphere) reveals it to be at a temperature of around 5800K. Blue-white stars are much hotter, with surface temperatures around 12,000K, while redder stars radiate from surfaces at about 3000K. The colour of a star thus complicates the measurement of a star’s brightness, as different colour stars will show different magnitudes at different wavelengths; and so a measured magnitude always has to be quoted for the specific colour band it was recorded in. Conversely, this can also be a good way to estimate a star’s temperature; different black-body curves will produce different values for the magnitudes compared in two or more colours. 
Nearly all stars appear as points of light, even when viewed through the most powerful telescopes available. Only a few hundred of the larger and closer stars can be resolved into a disc, and in those cases you can infer the real diameter of the star from a measurement of its angular width (as long as you have a reliable estimate of its distance away from us). This measurement is much easier to make for larger and closer stars – the star Betelgeuse is a good example: at a distance of ‘only’ 650 light-years away from us, it is also intrinsically a large star (of a type known as a red supergiant), and we can even resolve features such as starspots on its surface. But stars like Betelgeuse remain very much the exception. Instead we have to rely on other ways to infer the diameter of stars, and we can use their luminosity and temperature (as inferred from observations of their brightness and colour). In particular, for an object emitting black-body radiation, the luminosity emitted from the whole surface area of a star radius R is given by the equation L=4πR2σT4  (where T is its temperature and σ is a constant). Thus if we know a star’s luminosity L, and its temperature T, we can infer its size R. This relation implies that for a given luminosity, a hotter star must be smaller; or that for a given temperature, a brighter star must be larger. 
Thus most easily measured quantities of a star’s brightness at different colours enable us to characterise it in terms of its surface temperature (T) and luminosity (L), which are related to one another through the radius of the star. If we want to understand more about the differences between stars, then the best place to start is by comparing the distribution of these properties in a graph of L against T. Whether plotted with the empirical measurements of colour and magnitude, or translated into the intrinsic properties of temperature and luminosity on the two axes, it is known as the Hertzsprung-Russell (HR) diagram after the work of Ejnar Hertzsprung (1873-1967) and Henry Norris Russell (1877 – 1957). 
The whole area of the graph is not filled uniformly, and stars populate certain combinations of luminosity and temperature more than others. Most noticeably, there is a long thin track that stretches diagonally from upper left to lower right across the diagram, moving from hot and bright stars down to ones which are cool and faint; there are only a few stars occupying the areas to either side. About 90% of all stars lie on this main sequence, including our Sun, which is located midway along the track. As each pair of (L, T) measurements corresponds uniquely to a stellar radius, the different concentrations of stars in this plot offer different size classes as well. There are luminous cool stars that occupy a region above the lower part of the main sequence, and the relationship between L and T determines that they must be huge − consequently they are known as giants. Typically they are ten – 100 times wider than the main sequence stars of comparable temperatures, and can be 100 – 1000 times more luminous. The coolest are known as red giants, and examples of this kind of star that we can see in the night sky include Aldebaran and Arcturus. Above and to the left of the red giants are the much rarer supergiant stars – named because they are more luminous, and thus must be larger than the red giants of comparable temperature. Their radii extend up to about 1500 times that of the Sun, and examples in the night sky include Betelgeuse (a red, and thus cool, supergiant) and Rigel a hot (blue-white) supergiant.  The diameter of Betelgeuse is greater than the size of Jupiter’s orbit around the Sun; although the star’s surface temperature is much cooler than that of the Sun, at only about 3,500K, the fact that it is radiating through such an enormous surface area means that it is over 9000 times more luminous. Together, giants and supergiants comprise only about 1% of all stars. Most of the remaining 9% are hot but very dim; thus they must be small – probably not much bigger than the Earth – and are termed white dwarfs. 
Except for certain types of star known as variable stars, or the last mad hurrah of a massive star as it explodes in a supernova, most stars do not appear to change their luminosity or brightness on human timescales. We can expect that stars will change any of their luminosity, size or temperature, however, as they evolve through their lives. If they do so, then they will change their position in the HR diagram. The tendency of stars to concentrate in certain parts of the colour-magnitude diagram then suggests that many stars occupy those regions for a comparatively large fraction of their lives − the longer that phases lasts, then the greater the number of stars will be observed with those properties. Conversely, sparser areas that contain relatively few stars represent only a very short-lived period of a star’s life. Thus the HR diagram shows us that a star spends most of its life sitting on the main sequence, and spends only a short period in any of the giant, supergiant and white dwarf phases. Whilst we cannot track the evolution of any single star, we can match the observed distribution of stars in this plot to predictions that emerge theoretical models of their development and evolution. By now we can confidently map out well-determined tracks through the HR diagram followed out by different kinds of stars as they live out their life; consequently we can now invert this information to use a star’s position in the HR diagram to tell us where it is in its stellar life cycle. The only caveat is that if at any stage of its life a star is obscured from view – if it is surrounded by dust clouds, for example – then it will not feature on this diagram. 
There is another physical property of a star, which (unlike the temperature, size or luminosity) we can assume remains broadly constant throughout its life, and that is its mass. The mass of a star cannot be observed directly, and neither can it be simply inferred from its size, as stars can have different densities. It is most easily calculated where we see two stars orbiting each other in a binary system. Each is accelerated by the gravitational attraction of the other, and so their relative masses can be calculated from observations of their speed of revolution around each other and their separation. Using this method, stellar masses have been measured to range from about a tenth to a hundred times the mass of the Sun. Very low mass stars are thought to be much more common than those of high mass. The lower limit of about 8% of a Solar mass comes about because objects of lesser mass never achieve a high enough core temperature for nuclear fusion to begin. Low mass objects are thought to commonly form in the same way as, and alongside, stars and they are known as brown dwarfs. At the other extreme, the upper limit to a star’s mass is somewhere between 100-150 Solar masses. Stars any more massive than this will emit so much light that the outwards pressure exerted by this radiation (which is negligible in stars like our Sun) would be so great as to blow the star apart. 
Once we incorporate the knowledge of stellar masses into the HR diagram, it becomes apparent that there is a simple relationship of the mass of a star to its place on the main sequence. This follows the form L ∝ M3 ie more massive stars on the main sequence are more luminous – and thus hotter. This makes sense in terms of a star needing to sustain a stable balance between its self-gravity and the outward pressure due to the energy it produces. The more massive a star is, the higher the inward pressure due to gravity it will experience, and so the more energy the star will have to produce at its core to counteract it; the increased rate of nuclear fusion both raises its central temperature and releases more energy to make the star shine more brightly. Supergiants are more massive than the red giants, which are in turn more massive than white dwarfs; but unlike for stars on the main sequence, there does not seem to be any correlation of mass with the luminosity or temperature within each of these classes. 
So what does the mass variation along the main sequence tell us? We assume that once formed, a star does not undergo a significant change from its initial mass. It is only right at the very end of its life that it will undergo a dramatic loss of matter. (We shall ignore any mass loss through stellar winds that stream from the surface of a star all through its life, as it accumulates to a total which is only a tiny fraction of its original mass; likewise we shall not consider here exchanges of matter between stars such as happen to companions in a binary system, as such interactions can accelerate and distort the natural stellar development.) If a star does not change its mass, then its location on the HR diagram cannot move up or down the main sequence during its life; thus during this phase a star remains constant in its luminosity and surface temperature. 
Stars are born from the material of the interstellar medium. Interstellar space is not truly empty, but filled with a tenuous gas at a range of temperatures. It is comprised of neutral or ionised atoms, mixed with molecules and tiny particles of solid matter known as dust. The composition of the gas is mainly hydrogen (90%) and helium (8%), with the remainder (2%) in the form of heavier elements. Some of this gas is primordial, never having been incorporated into stars; as we shall see, some of the matter has already been part of stars, before being recycled back into the interstellar medium. 
Much of the space between the stars is filled with what is known as the hot, and warm intercloud medium, which is completely transparent at visible wavelengths. Although very widespread, it is at such a low density that it does not contain the bulk of the matter, and it is mostly fully ionised hydrogen. Embedded within the intercloud medium are distinct diffuse clouds, each smaller, denser and cooler than their surroundings. These cooler clouds take up a lot less space (only about 3% of the volume), but their higher density means that they contain much more of the mass of the interstellar medium. The diffuse clouds are again mostly transparent, made up of mostly neutral atoms that can only be detected in the radio, IR and microwave wavebands, along with from the emission from molecules and dust. Within these diffuse clouds are the densest and coolest pockets of gas, so condensed that the dust carried within them renders them opaque in the visible and UV. The deepest internal reaches of these clouds drop to temperatures where the atoms have joined up to build many different (and mostly carbon-based) molecules, so they are known as molecular clouds. Even within these clouds are small clumps or cores, each containing a mass of material anywhere between a half to a thousand or a million Solar masses. Even though these clumps are the very densest patches of the interstellar space, they still contain well over a trillion times fewer gas molecules and atoms compared to the high density of Earth’s atmosphere. However, they provide the most favourable conditions for the formation of stars. 
Even though the mass of each particle in an interstellar dust cloud is tiny, they combine to make a large total mass distributed throughout the cloud. Every atom (or molecule) in a gas cloud experiences the gravitational attraction of the combined mass of all the others. If there is the slightest over-density of matter in a diffuse cloud, it will make the gravity slightly stronger in one location than in another, and will preferentially pull some of the surrounding matter toward it. This will increase the overdensity, enhance the gravitational pull, and attract more matter… leading gradually to a runaway process of gravitational collapse. Energy has to be conserved within this process, and so as matter falls under gravity, the potential energy released is converted into heat. The contraction will continue until gravity has squeezed the cloud so tight that in the central region it reaches a density and temperature sufficient for nuclear fusion to be triggered – this requires temperatures of about eighteen million degrees K, and this threshold marks the formation of a proto-star. Of course this is very much a simplification; the heat released during the contraction phase will warm the surrounding material, slowing its collapse. The star formation will pause until any dust particles in the cloud have been evaporated – then the material stops absorbing the heat, the radiation can escape away and the contraction under gravity can resume. There are also other factors – such as the strength of any magnetic field threading through the gas, the amount of rotation present, or the fact that the cloud will not be uniform in density or temperature or shape – which can all also inhibit a cloud’s contraction under gravity. 
We see star formation is an ongoing process throughout the disc of our galaxy.  Many of these diffuse gas clouds have existed for billions of years, so how have they resisted gravitational collapse up to now? The reason is that even if they are only a few degrees above absolute zero, all the atoms in an interstellar cloud are in continual motion (after all, the temperature of a cloud is just a measure of the average kinetic energy of the random and microscopic motions of its constituent particles). This constant jiggling produces a net outwards pressure, which, if large enough, enables the particles in a cloud to resist the local pull of self-gravity. Thus whether or not a cloud of a certain size will collapse to form stars is dictated by a delicate balance between its density and temperature: the warmer a cloud, the higher the density that is required for self-gravity to overcome the thermal pressure. So it is the densest and coolest regions of the interstellar medium that provide the most favourable conditions for stars to form. All the dark clouds that have not just collapsed already are just slightly too sparse, or too hot to collapse spontaneously. A cloud in the disc of our galaxy can thus exist in an undisturbed ‘equilibrium state’ for many millions or billions of years until only the slightest triggering mechanism tips it over into gravitational collapse. A cloud will only form stars if either the density or the temperature changes significantly.  A decrease in temperature is hard to effect quickly, whereas an increase in density can be achieved rapidly; the gas in a cloud can be squeezed suddenly, for example by the gravitational tidal effect, or a collision with another cloud, by the push of stellar winds, or compression from a supernova blast wave. 
Many stars are born in small groups or larger clusters, where all are formed from the collapse of the same diffuse cloud. The original cloud will not have been uniform, and any enhancements in density will become exaggerated as the material collapses. It will fragment into smaller denser pockets which each start a runaway gravitational collapse on their own. Each of these clumps will form an individual star, along with any accompanying planetary system. Binary or multiple star systems might form as different fragments collapse nearby each other. One parent gas cloud can form an ‘open’ cluster of young stars, where tens to hundreds of stars are formed at around the same time, just with a range of masses. The young hot stars will not always stay together. Gravitational ties are tenuous, and within a few million years the random motions of individual stars will cause them to disperse and become independent of each other. 
Star formation is inefficient, and clusters of newly-formed stars remain embedded in the remnants of their parent cloud. Their winds and UV radiation erode the surrounding matter, and excite the gas atoms contained in the clouds to glow with their own light. Such star formation nurseries can be clearly seen as the emission ‘nebulae’, and in a galaxy like our own, they occur in the flat disc, following a distinctive spiral pattern. This spiral is not simply created by the different orbital speeds of material at difference distances from the centre, as the spiral arms would wind up in a relatively short time. Instead a ‘density wave’ is thought to sweep around the disc, travelling more slowly than the stars and gas rotate around the centre of the galaxy. As the wave moves through portions of the disc, it compresses all the material that passes through it, triggering the gravitational collapse of gas clouds, and leaving in its wake an arc of newly formed bright blue star clusters. 
The exact process of star birth is impossible to observe directly, as it is shielded from observation behind opaque clouds of gas and dust. It also happens quickly, compared to the eventual lifetime of the star itself. Only the very core of a dense pocket of gas reaches high enough temperatures for nuclear fusion to commence – and the surrounding layers of dust and gas are later dispersed by the winds and radiation. More massive stars form more quickly; while a Solar-mass star is estimated to take about ten million years to form from the gas cloud, a fifteen Solar-mass star may take only 100,000 years to complete the same process. As the dense clump of gas originally condensed from the cloud, it will have spun up and flattened out to form a thick circumstellar disc around the protostar (which will itself perhaps end up as an accompanying planetary system). Even when this torus of gas and dust veils the early star from view, the presence of the young stellar objects is revealed by twin jets of gas streaming out in opposite directions along an axis perpendicular to the disc. The gas in these jets moves at high speeds, at up to 50-100 km/s, and while they may last only for about 10,000 years, the jets carry significant amounts of mass away from the system. They stretch out for up to a light year above and below the protostar, ending only where they ram into surrounding interstellar gas, forming glowing knots of compressed gas. 
A protostar arrives suddenly onto the HR diagram at the bottom right, as it only has low temperature and luminosity to begin with. It takes a while to balance out how much energy it needs to generate at its core to counteract gravity, and as it stabilises it follows a path on the HR diagram that lands it onto the place on the main sequence appropriate for its mass. The length of time a star will spend on the main sequence – ie the speed with which it will get through its life – depends entirely on its mass. Stars of different masses don’t evolve at the same rate: remember, the more massive the star, the more energy it has to produce to counteract the greater self-gravity. Even though it may have started with more fuel originally, it will burn so furiously that it exhausts that supply far more rapidly: more massive stars live furiously and die young. 
A high-mass star burns with a bluer colour because it is hotter, and has a main sequence lifetime of only a few tens of millions of years. Thus when we see groups of blue stars within a galaxy, we know these are young stars, and so that they trace the most recent star formation events. The HR diagram also gives us a way to age the stars in a cluster; these are assumed to have formed at about the same time, from the same cloud of hot gas and dust. They are thus all the same age, and of similar chemical composition; the only difference between them is the mass of each individual star. If you plot their positions in the HR diagram, the main sequence is still apparent, but it is truncated at a place that depends on the cluster’s age. More massive stars have shorter lifetimes, and many of them in the cluster have already exhausted their fuel to move away from the main sequence; the age of the most massive stars still remaining on the main sequence can thus translate into an age for the cluster. 
Stars like our Sun (with masses between about 0.8 – 10 Solar masses) behave in a similar way to each other. For most of their life, and while on the main sequence, they convert hydrogen into helium. This happens via a process of nuclear fusion that converts four hydrogen nuclei (ie protons) into a single helium nucleus; each reaction also produces two positrons, two neutrinos and energy in the form of gamma rays. As a helium nucleus has less mass than the original four protons, the difference in mass (about 0.7%) is converted into energy (mass and energy are equivalent to each other in the principle encapsulated by the equation by E=mc2). The Sun loses mass to energy at the astonishing rate of four million kg of mass every second – while this may seem enormous, the loss of mass remains insignificant compared to the Sun’s total mass of around 2 x 1030 kg. As protons are each positively charged, they must be smashed together at high speed to overcome the electric repulsion of like charges. The hotter a gas, the faster all the particles in the gas are moving, so the fusion process can only operate at temperatures over ten million degrees. The conversion of hydrogen to helium thus cannot occur throughout a star, but only where the temperature is the greatest, at the very centre and in a surrounding region known as the core. Nuclear reactions place at the centre of the Sun only where the temperature is about fifteen million K and the density is around 160,000 kg/m3.
During its 5-billion lifetime, the Sun has already converted about half the hydrogen in its core into helium, and it has sufficient fuel to continue doing so for about another six billion years or so. Once it runs out of sufficiently hot hydrogen, however, it can no longer produce enough energy to withstand the squeeze of gravity. There is no mixing of material between the core and the rest of the star, so the supply of fuel at the centre is not replenished with fresh hydrogen. The star can no longer support the weight of its outer layers, which now begin to compress the core. As it shrinks, however, gravitational potential energy is converted to thermal energy, and the core increases in temperature to a point where fusion of hydrogen can be restarted – but in the depleted core. Hydrogen fusion will now take place in the shell of unprocessed material immediately surrounding the original core, as this region has also increased in temperature. There is now a cycle of fuel depletion, shrinking, heating and re-ignition that repeats as subsequent layers are exhausted, sustaining hydrogen fusion in a concentric shell around the core that moves gradually outward through the star. This process adds more layers of helium ‘ash’ to the central core, which by now is entirely composed of helium. Very low mass stars will reach a point where the cycle of compression no longer heats the layer of hydrogen around the core sufficiently, and nuclear reactions will then cease. 
Stars with mass above a half-Solar mass will continue until the process of the gravitational compression has heated the core to a temperature of 100 million degrees, enabling a new range of nuclear processes. Helium fusion reactions are initiated which create yet heavier nuclei such as carbon and oxygen – and again the associated release of energy temporarily halts the contraction of the core, stabilising the star. Less energy is released in each reaction, however, and so the helium in the core is exhausted much more rapidly than was the case for the hydrogen; it is consumed within about 10% of the star’s lifetime on the main sequence, around only about 100 million years. Once the helium fusion stops at the centre, it might also then continue in a shell around the core, now surrounded by a slightly outer shell of hydrogen fusion.  
By now the changes in the structure of the star, and the development of its temperature and luminosity, mean that it has moved to a new location on the HR diagram; it no longer sits on the main sequence. It has become a red giant, as much more energy is now directed into the star’s outer layers than before, and its atmosphere puffs up to many times its original size. The outermost layers are now much further away from the source of energy, and the surface cools down to appear dim and red. When it reaches this stage, the Sun’s atmosphere will grow so large as to engulf Mercury and Venus. A red giant has expanded so much that its outer layers experience a much-reduced surface gravity, and become vulnerable to escape away from the star and into space. 
Heavier elements can be created within the cool distended envelopes of red giant stars. Neutrons released from the nuclear reactions going on in the shells can combine with any existing heavy nuclei that may be already present in the gas layers (where these nuclei were not created by the nuclear reactions in that star, but are merely present as they were in the original cloud that the star formed from, if it had been enriched by earlier generations of stars). Adding a single neutron to a nucleus increases its mass by one, but does not change its chemical properties, so it forms a different isotope of the same element. If that isotope is unstable, however, given time it will eject an electron, converting one of its neutrons to a proton to become a different element. The whole process can repeat to build up a variety of heavier elements, but this a very slow process, as there is a low density of neutrons being released to be available for capture. The long time between neutron captures allows time for some radioactive decay, and only certain elements (for example, strontium zirconium barium lanthanum and lead) can be made in this way. The cool envelopes of red giants are also an ideal environment for the tiny solid particles of carbonate and silicate that make up the interstellar dust. 
A star of a mass similar to the Sun also reaches a limit as to how much the core can be heated from successive cycles of gravitational compression, and eventually all nuclear reactions in both core and shells can no longer be re-ignited. The star puffs away its outer envelope  – often in several stages thousands of years apart – and in this way loses a substantial amount of its mass; this expands away to form a shell, leaving the central core of the star exposed. 
The core of the star exposed by the dispersing outer atmosphere can contain about half the mass of the Sun, all packed into a volume about the size of the Earth. This is a stable body known as a white dwarf, and it is able to resist further gravitational collapse through electron degeneracy pressure – the fact that electrons resist being squeezed too tightly together. Although it is hot, at a temperature between 5,000 − 30,000K, there is insufficient hydrogen and helium remaining to continue fusion, and any nuclear reactions involving heavier elements require much higher temperatures. So in the absence of any further fusion, the white dwarf will cool and fade; its small size means that it will radiate its energy away only slowly, and it will take a timescale of billions of years for it to become a cold, dark and dense ball of compressed matter, rich in carbon and oxygen. 
The outer shell of gas expanding away from the core forms a temporary structure known as a planetary nebula (which, despite the name, has nothing to do with planets!). This nebula will eventually disperse into space within a few tens of thousands of years; but while they are close to the hot, bright white dwarf, the gas atoms glow and shine, illuminating a very pleasing variety of shapes and patterns within the nebulae. Some (such as the famous Ring nebula) take the simple form of a bubble surrounding the white dwarf, while others show much more intricate structures indicating shells or winds and even far outer halos created from material shrugged off during earlier episodes in the star's evolution. Some planetary nebulae have a bipolar or ‘hourglass’ shape, where the outflow has been modified and constricted by a ring of opaque matter that obscures the white dwarf. 
More massive stars – with masses some 10-50 times that of our Sun – will begin life on the main sequence in the same way, but the greater gravitational compression they experience will raise the temperature in the core so much higher that they can convert hydrogen to helium using a more efficient process. Here the nuclear reactions involve carbon, nitrogen and oxygen as catalysts to speed things up, causing the stars to burn 10,000 to 100,000 times brighter, but also get through their main sequence life very much faster. The much greater mass present in these stars means that they can continue to fusion processes that involve and create much heavier elements – each time the fusion of one fuel is exhausted at the core, the gravitational compression increases the central temperature enough that its products can then be used in further reactions, thus making elements of ever increasing atomic weight. The internal structure of the centre of such a star near the end of its life is complicated, resembling the structure of an onion. The outermost shells would contain hydrogen fusing into helium, surrounding inner and deeper layers where helium is fused to carbon and oxygen, where carbon is then fused into neon and magnesium, and so on. If temperatures reach about eight billion degrees at the very centre, a core of iron is created from the fusion of silicon. Each step in this chain forming ever-heavier elements at higher temperatures lasts for a progressively tinier fraction of the star’s life; the stage of iron production may only last for about a week. But the formation of iron-type elements marks the point where nuclear fusion is no longer a viable energy source; iron is the most tightly bound atomic nucleus, and any heavier elements only release energy when split into lighter elements. The star now has no further means to create energy from fusion reactions. 
Once the options for a central energy supply are completely depleted, the inward pull of gravity finally wins. The core of the star becomes so compressed by the overlying layers that it will reach densities as high as a trillion kg per cubic metre, and temperatures in excess of ten billion degrees K. At that point the matter undergoes a process known as photo-disintegration, whereby photons split the iron nuclei into protons and neutrons. Any free electrons combine with these protons to form even more neutrons, releasing a huge flux of neutrinos in the process. The iron core of about 12,000 km diameter then rapidly collapses to form a neutron star only about 10- 20 km across; at this point the matter can resist any further gravitational pull through neutron degeneracy pressure. The matter is far denser than that in a white dwarf, and a single teaspoonful would weigh a billion tonnes on Earth. 
The catastrophic collapse of the core into a neutron star happens in less than a second, and it leaves the outermost layers of the star temporarily unsupported. They crash down, only to bounce on the incompressible surface of the neutron star, and the potential energy released from this collapse is then released instantly to form a massive explosion known as a supernova. So much energy is liberated that the event can outshine the light of its entire host galaxy for a few days/weeks/months – this is even though only 0.01% of the energy is released as light! Most of it is carried away in the form of neutrinos.  A supernova creates an intense flood of neutrons which can bombard the ions and atoms of the stellar debris. This allows a much more rapid neutron capture process, where two or three neutrons can be acquired by the atomic nuclei in the debris before it undergoes radioactive decay, and a different set of heavy elements can be created than from the slower process operating in the envelopes of red giant stars. 
The hot debris from the outer layers of the star explodes out through space at speeds of about 5000 km/s, rapidly sweeping out a huge bubble of hot gas, surrounded by layers of material swept up in the expanding blast wave. As the material ploughs into its surroundings it creates shocks which heat and compress the gas – both in the debris and the interstellar medium – to millions of degrees. The expansion of the material slows down only after about 10,000 to 20,000 years to form a spherical shell known as a supernova remnant. Such a remnant can remain visible for tens of thousands of years as a colourful network of filaments extending many tens of light-years across. Long after the initial explosion has faded, these remain the only indication of the dramatic event that has taken place, and eventually this material becomes completely mixed with the dust and gas clouds of the interstellar medium. 
Stars with mass between 50 – 100 Solar masses have the shortest lives of all. They will progress through the same basic steps of generating energy by nuclear fusion of ever heavier elements, but a key difference is that they will lose a lot of their mass through a stellar wind. With mass-loss rates of up to a millionth of a solar mass a year, sometimes these rarest but most massive stars can lose so much matter that the core is left exposed. This then drastically alters the final stages of their development as they are not able to undergo shell burning. They thus never turn into red giants, and are known as Wolf-Rayet stars, where expanding shells of gas surround the exposed core of the star.  Although they still end their lives in a giant supernova explosion, the light from the event will indicate a distinct lack of hydrogen in the ejecta. This is known as a ‘type 1b’ supernova, and indicates that most of the outermost (and hydrogen-rich) material had been lost long before the explosion in the strong stellar wind. 
With so much mass, the final stages gravitational compression are so great that they can overcome neutron degeneracy pressure, and there is no mechanism beyond that able to halt the contraction. The collapse continues until all the mass of the star is squashed into an infinitely small space to form a singularity – a region of the most extreme gravity anywhere in the Universe known as a black hole.
The elements made in stars are returned back to interstellar space through a variety of processes. First of all, through the stellar winds that flow from all stars during their lives, although massive hot stars and red giants are the most effective at returning material to the interstellar medium. Winds are a fairly gentle form of dispersal, and can account for a significant return of material to the environment. The creation of a planetary nebulae is a slightly more dramatic shedding of material, which is both hotter and denser – but both winds and planetary nebulae distribute material that is only modestly enriched in heavy elements. Although supernovae provide the most violent ejection of stellar material, their comparative rarity means that they return less material to the interstellar medium than one might expect. However, supernovae are not only the major source of heavy elements, but the energy of the expanding gas shell is so great that the ejected material is mixed much more thoroughly with its surroundings, and over a much larger volume of space. 
While the very first stars are expected to have formed from primordial gas containing only hydrogen and helium, once they run through their rapid lives to become supernovae, the process of contamination of the interstellar medium with heavy elements is begun. Each generation of short-lived and massive stars steadily increases the proportion of heavy elements in the clouds with time. But there is less enrichment than one might suppose. High mass stars are far less common than the longer-lived low mass objects, and even though the nuclear reactions in stars will have been running for nearly all of the Universe’s 13.8 billion years, only about 5% of all stars ever born have evolved past the first stage of converting hydrogen into helium. In addition, not all material is recycled – many heavy elements remain locked into neutron stars, black holes, brown and white dwarfs, and in any planets that are not absorbed by its star at the end of its life. But slowly, the interstellar medium is enriched beyond its original primordial chemical composition with a smattering of heavier elements, to evolve in its chemical composition. This affects the chemical composition of the future stars that will condense from that reservoir of material. Thus the observed ‘metallicity’ of a star can provide an indication of the level of enrichment in heavy elements that had occurred in its host cloud prior to the star’s formation. A very old and long-lived star would have formed when the metallicity of the interstellar medium was lower than when the Sun was formed. Indeed, we see a gradient in behaviour supporting this interpretation, where stars in the Milky Way show different levels of enrichment in heavy elements that represent different epochs of star formation: younger stars have high metallicities, and older stars tend to have lower metallicities. Thus the chemical composition of individual stars in a galaxy can start to teach us about the formation of that galaxy.  
© Professor Carolin Crawford, 2014