, convection zones with arrowed cycles and radiative zones with red flashes. To the left a low-mass
red dwarf, in the center a mid-sized
yellow dwarf and at the right a massive blue-white main-sequence star. Eventually the star's core exhausts its supply of hydrogen and the star begins to evolve off the
main sequence. Without the outward
radiation pressure generated by the fusion of hydrogen to counteract the force of
gravity, the core contracts until either
electron degeneracy pressure becomes sufficient to oppose gravity or the core becomes hot enough (around 100 MK) for
helium fusion to begin. Which of these happens first depends upon the star's mass.
Low-mass stars What happens after a low-mass star ceases to produce energy through fusion has not been directly observed; the
universe is around 13.8 billion years old, which is less time (by several orders of magnitude, in some cases) than it takes for fusion to cease in such stars. Recent astrophysical models suggest that
red dwarfs of may stay on the main sequence for some six to twelve trillion years, gradually increasing in both
temperature and
luminosity, and take several hundred billion years more to collapse, slowly, into a
white dwarf. Such stars will not become red giants as the whole star is a
convection zone and it will not develop a degenerate helium core with a shell burning hydrogen. Instead, hydrogen fusion will proceed until almost the whole star is helium. Slightly more
massive stars do expand into
red giants, but their helium cores are not massive enough to reach the temperatures required for helium fusion so they never reach the tip of the red-giant branch. When hydrogen shell burning finishes, these stars move directly off the red-giant branch like a post-
asymptotic-giant-branch (AGB) star, but at lower luminosity, to become a white dwarf.
Mid-sized stars Stars of roughly become
red giants, which are large non-
main-sequence stars of
stellar classification K or M. Red giants lie along the right edge of the Hertzsprung–Russell diagram due to their red color and large luminosity. Examples include
Aldebaran in the constellation
Taurus and
Arcturus in the constellation of
Boötes. Mid-sized stars are red giants during two different phases of their post-main-sequence evolution: red-giant-branch stars, with inert cores made of helium and hydrogen-burning shells, and asymptotic-giant-branch stars, with inert cores made of carbon and helium-burning shells inside the hydrogen-burning shells. Between these two phases, stars spend a period on the
horizontal branch with a helium-fusing core. Many of these helium-fusing stars cluster towards the cool end of the horizontal branch as K-type giants and are referred to as
red clump giants.
Subgiant phase When a star exhausts the hydrogen in its core, it leaves the main sequence and begins to fuse hydrogen in a shell outside the core. The core increases in mass as the shell produces more helium. Depending on the mass of the helium core, this continues for several million to one or two billion years, with the star expanding and cooling at a similar or slightly lower luminosity to its main sequence state. Eventually either the core becomes degenerate, in stars around the mass of the sun, or the outer layers cool sufficiently to become opaque, in more massive stars. Either of these changes cause the hydrogen shell to increase in temperature and the
luminosity of the star to increase, at which point the star expands onto the red-giant branch.
Red-giant-branch phase and
giant phases, until its outer envelope is expelled to form a
planetary nebula at upper right The expanding outer layers of the star are
convective, with the material being mixed by turbulence from near the fusing regions up to the surface of the star. For all but the lowest-mass stars, the fused material has remained deep in the stellar interior prior to this point, so the convecting envelope makes fusion products visible at the star's surface for the first time. At this stage of evolution, the results are subtle, with the largest effects, alterations to the
isotopes of hydrogen and helium, being unobservable. The effects of the
CNO cycle appear at the surface during the first
dredge-up, with lower 12C/13C ratios and altered proportions of carbon and nitrogen. These are detectable with
spectroscopy and have been measured for many evolved stars. The helium core continues to grow on the red-giant branch. It is no longer in thermal equilibrium, either degenerate or above the
Schönberg–Chandrasekhar limit, so it increases in temperature which causes the rate of fusion in the hydrogen shell to increase. The star increases in luminosity towards the
tip of the red-giant branch. Red-giant-branch stars with a degenerate helium core all reach the tip with very similar core masses and very similar luminosities, although the more massive of the red giants become hot enough to ignite helium fusion before that point.
Horizontal branch In the helium cores of stars in the 0.6 to 2.0 solar mass range, which are largely supported by
electron degeneracy pressure, helium fusion will ignite on a timescale of days in a
helium flash. In the nondegenerate cores of more massive stars, the ignition of helium fusion occurs relatively slowly with no flash. The nuclear power released during the helium flash is very large, on the order of 108 times the
luminosity of the Sun for a few days However, the energy is consumed by the thermal expansion of the initially degenerate core and thus cannot be seen from outside the star. Due to the expansion of the core, the hydrogen fusion in the overlying layers slows and total energy generation decreases. The star contracts, although not all the way to the main sequence, and it migrates to the
horizontal branch on the Hertzsprung–Russell diagram, gradually shrinking in radius and increasing its surface temperature. Core helium flash stars evolve to the red end of the horizontal branch but do not migrate to higher temperatures before they gain a degenerate carbon-oxygen core and start helium shell burning. These stars are often observed as a
red clump of stars in the colour-magnitude diagram of a cluster, hotter and less luminous than the red giants. Higher-mass stars with larger helium cores move along the horizontal branch to higher temperatures, some becoming unstable pulsating stars in the yellow
instability strip (
RR Lyrae variables), whereas some become even hotter and can form a blue tail or blue hook to the horizontal branch. The morphology of the horizontal branch depends on parameters such as metallicity, age, and helium content, but the exact details are still being modelled.
Asymptotic-giant-branch phase After a star has consumed the helium at the core, hydrogen and helium fusion continues in shells around a hot core of
carbon and
oxygen. The star follows the
asymptotic giant branch on the Hertzsprung–Russell diagram, paralleling the original red-giant evolution, but with even faster energy generation (which lasts for a shorter time). Although helium is being burnt in a shell, the majority of the energy is produced by hydrogen burning in a shell further from the core of the star. Helium from these hydrogen burning shells drops towards the center of the star and periodically the energy output from the helium shell increases dramatically. This is known as a
thermal pulse and they occur towards the end of the asymptotic-giant-branch phase, sometimes even into the post-asymptotic-giant-branch phase. Depending on mass and composition, there may be several to hundreds of thermal pulses. There is a phase on the ascent of the asymptotic-giant-branch where a deep convective zone forms and can bring carbon from the core to the surface. This is known as the second dredge up, and in some stars there may even be a third dredge up. In this way a
carbon star is formed, very cool and strongly reddened stars showing strong carbon lines in their spectra. A process known as hot bottom burning may convert carbon into oxygen and nitrogen before it can be dredged to the surface, and the interaction between these processes determines the observed luminosities and spectra of carbon stars in particular clusters. Another well known class of asymptotic-giant-branch stars is the
Mira variables, which pulsate with well-defined periods of tens to hundreds of days and large amplitudes up to about 10 magnitudes (in the visual, total luminosity changes by a much smaller amount). In more-massive stars the stars become more luminous and the pulsation period is longer, leading to enhanced mass loss, and the stars become heavily obscured at visual wavelengths. These stars can be observed as
OH/IR stars, pulsating in the infrared and showing OH
maser activity. These stars are clearly oxygen rich, in contrast to the carbon stars, but both must be produced by dredge ups.
Post-AGB , a
planetary nebula formed by the death of a star with about the same mass as the Sun These mid-range stars ultimately reach the tip of the asymptotic-giant-branch and run out of fuel for shell burning. They are not sufficiently massive to start full-scale carbon fusion, so they contract again, going through a period of post-asymptotic-giant-branch superwind to produce a planetary nebula with an extremely hot central star. The central star then cools to a white dwarf. The expelled gas is relatively rich in heavy elements created within the star and may be particularly
oxygen or
carbon enriched, depending on the type of the star. The gas builds up in an expanding shell called a
circumstellar envelope and cools as it moves away from the star, allowing
dust particles and molecules to form. With the high infrared energy input from the central star, ideal conditions are formed in these circumstellar envelopes for
maser excitation. It is possible for thermal pulses to be produced once post-asymptotic-giant-branch evolution has begun, producing a variety of unusual and poorly understood stars known as born-again asymptotic-giant-branch stars. These may result in extreme
horizontal-branch stars (
subdwarf B stars), hydrogen deficient post-asymptotic-giant-branch stars, variable planetary nebula central stars, and
R Coronae Borealis variables.
Massive stars , a red supergiant In massive stars, the core is already large enough at the onset of the hydrogen burning shell that helium ignition will occur before electron degeneracy pressure has a chance to become prevalent. Thus, when these stars expand and cool, they do not brighten as dramatically as lower-mass stars; however, they were more luminous on the main sequence and they evolve to highly luminous supergiants. Their cores become massive enough that they cannot support themselves by
electron degeneracy and will eventually collapse to produce a
neutron star or
black hole.
Supergiant evolution Extremely massive stars (more than approximately ), which are very luminous and thus have very rapid stellar winds, lose mass so rapidly due to radiation pressure that they tend to strip off their own envelopes before they can expand to become
red supergiants, and thus retain extremely high surface temperatures (and blue-white color) from their main-sequence time onwards. The largest stars of the current generation are about because the outer layers would be expelled by the extreme radiation. Although lower-mass stars normally do not burn off their outer layers so rapidly, they can likewise avoid becoming red giants or red supergiants if they are in binary systems close enough so that the companion star strips off the envelope as it expands, or if they rotate rapidly enough so that convection extends all the way from the core to the surface, resulting in the absence of a separate core and envelope due to thorough mixing. The core of a massive star, defined as the region depleted of hydrogen, grows hotter and denser as it accretes material from the fusion of hydrogen outside the core. In sufficiently massive stars, the core reaches temperatures and densities high enough to fuse carbon and heavier elements via the
alpha process. At the end of helium fusion, the core of a star consists primarily of carbon and oxygen. In stars heavier than about , the carbon ignites and
fuses to form neon, sodium, and magnesium. Stars somewhat less massive may partially ignite carbon, but they are unable to fully fuse the carbon before
electron degeneracy sets in, and these stars will eventually leave an oxygen-neon-magnesium
white dwarf. The exact mass limit for full carbon burning depends on several factors such as metallicity and the detailed mass lost on the
asymptotic giant branch, but is approximately . Although non-exploding red giants can produce significant quantities of elements heavier than iron using neutrons released in side reactions of earlier
nuclear reactions, the abundance of elements heavier than
iron (and in particular, of certain isotopes of elements that have multiple stable or long-lived isotopes) produced in such reactions is quite different from that produced in a supernova. Neither abundance alone matches that found in the
Solar System, so both supernovae,
neutron star mergers and ejection of elements from red giants are required to explain the observed abundance of heavy elements and
isotopes thereof. The energy transferred from collapse of the core to rebounding material not only generates heavy elements, but provides for their acceleration well beyond
escape velocity, thus causing a Type Ib, Type Ic, or Type II supernova. Current understanding of this energy transfer is still not satisfactory; although current computer models of Type Ib, Type Ic, and Type II supernovae account for part of the energy transfer, they are not able to account for enough energy transfer to produce the observed ejection of material. However, neutrino oscillations may play an important role in the energy transfer problem as they not only affect the energy available in a particular flavour of neutrinos but also through other general-relativistic effects on neutrinos. Some evidence gained from analysis of the mass and orbital parameters of binary neutron stars (which require two such supernovae) hints that the collapse of an oxygen-neon-magnesium core may produce a supernova that differs observably (in ways other than size) from a supernova produced by the collapse of an iron core. The most massive stars that exist today may be completely destroyed by a supernova with an energy greatly exceeding its
gravitational binding energy. This rare event, caused by
pair-instability, leaves behind no black hole remnant. In the past history of the universe, some stars were even larger than the largest that exists today, and they would immediately collapse into a black hole at the end of their lives, due to
photodisintegration. == Stellar remnants ==