S-process

Last updated

The slow neutron-capture process, or s-process, is a series of reactions in nuclear astrophysics that occur in stars, particularly asymptotic giant branch stars. The s-process is responsible for the creation (nucleosynthesis) of approximately half the atomic nuclei heavier than iron.

Contents

In the s-process, a seed nucleus undergoes neutron capture to form an isotope with one higher atomic mass. If the new isotope is stable, a series of increases in mass can occur, but if it is unstable, then beta decay will occur, producing an element of the next higher atomic number. The process is slow (hence the name) in the sense that there is sufficient time for this radioactive decay to occur before another neutron is captured. A series of these reactions produces stable isotopes by moving along the valley of beta-decay stable isobars in the table of nuclides.

A range of elements and isotopes can be produced by the s-process, because of the intervention of alpha decay steps along the reaction chain. The relative abundances of elements and isotopes produced depends on the source of the neutrons and how their flux changes over time. Each branch of the s-process reaction chain eventually terminates at a cycle involving lead, bismuth, and polonium.

The s-process contrasts with the r-process, in which successive neutron captures are rapid: they happen more quickly than the beta decay can occur. The r-process dominates in environments with higher fluxes of free neutrons; it produces heavier elements and more neutron-rich isotopes than the s-process. Together the two processes account for most of the relative abundance of chemical elements heavier than iron.

History

The s-process was seen to be needed from the relative abundances of isotopes of heavy elements and from a newly published table of abundances by Hans Suess and Harold Urey in 1956. [1] Among other things, these data showed abundance peaks for strontium, barium, and lead, which, according to quantum mechanics and the nuclear shell model, are particularly stable nuclei, much like the noble gases are chemically inert. This implied that some abundant nuclei must be created by slow neutron capture, and it was only a matter of determining how other nuclei could be accounted for by such a process. A table apportioning the heavy isotopes between s-process and r-process was published in the famous B2FH review paper in 1957. [2] There it was also argued that the s-process occurs in red giant stars. In a particularly illustrative case, the element technetium, whose longest half-life is 4.2 million years, had been discovered in s-, M-, and N-type stars in 1952 [3] [4] by Paul W. Merrill. [5] [6] Since these stars were thought to be billions of years old, the presence of technetium in their outer atmospheres was taken as evidence of its recent creation there, probably unconnected with the nuclear fusion in the deep interior of the star that provides its power.

Periodic table showing the cosmogenic origin of each element. The elements heavier than iron with origins in dying low-mass stars are typically those produced by the s-process, which is characterized by slow neutron diffusion and capture over long periods in such stars. Nucleosynthesis periodic table.svg
Periodic table showing the cosmogenic origin of each element. The elements heavier than iron with origins in dying low-mass stars are typically those produced by the s-process, which is characterized by slow neutron diffusion and capture over long periods in such stars.

A calculable model for creating the heavy isotopes from iron seed nuclei in a time-dependent manner was not provided until 1961. [7] That work showed that the large overabundances of barium observed by astronomers in certain red-giant stars could be created from iron seed nuclei if the total neutron flux (number of neutrons per unit area) was appropriate. It also showed that no one single value for neutron flux could account for the observed s-process abundances, but that a wide range is required. The numbers of iron seed nuclei that were exposed to a given flux must decrease as the flux becomes stronger. This work also showed that the curve of the product of neutron-capture cross section times abundance is not a smoothly falling curve, as B2FH had sketched, but rather has a ledge-precipice structure. A series of papers [8] [9] [10] [11] [12] [13] in the 1970s by Donald D. Clayton utilizing an exponentially declining neutron flux as a function of the number of iron seed exposed became the standard model of the s-process and remained so until the details of AGB-star nucleosynthesis became sufficiently advanced that they became a standard model for s-process element formation based on stellar structure models. Important series of measurements of neutron-capture cross sections were reported from Oak Ridge National Lab in 1965 [14] and by Karlsruhe Nuclear Physics Center in 1982 [15] and subsequently, these placed the s-process on the firm quantitative basis that it enjoys today.[ citation needed ]

The s-process in stars

The s-process is believed to occur mostly in asymptotic giant branch stars, seeded by iron nuclei left by a supernova during a previous generation of stars. In contrast to the r-process which is believed to occur over time scales of seconds in explosive environments, the s-process is believed to occur over time scales of thousands of years, passing decades between neutron captures. The extent to which the s-process moves up the elements in the chart of isotopes to higher mass numbers is essentially determined by the degree to which the star in question is able to produce neutrons. The quantitative yield is also proportional to the amount of iron in the star's initial abundance distribution. Iron is the "starting material" (or seed) for this neutron capture-beta minus decay sequence of synthesizing new elements.[ citation needed ]

The main neutron source reactions are:

13
6
C
 
+  4
2
He
 
  16
8
O
 
+ 
n
22
10
Ne
 
+  4
2
He
 
  25
12
Mg
 
+ 
n
The s-process acting in the range from Ag to Sb. S-process-elem-Ag-to-Sb.svg
The s-process acting in the range from Ag to Sb.

One distinguishes the main and the weak s-process component. The main component produces heavy elements beyond Sr and Y, and up to Pb in the lowest metallicity stars. The production sites of the main component are low-mass asymptotic giant branch stars. [16] The main component relies on the 13C neutron source above. [17] The weak component of the s-process, on the other hand, synthesizes s-process isotopes of elements from iron group seed nuclei to 58Fe on up to Sr and Y, and takes place at the end of helium- and carbon-burning in massive stars. It employs primarily the 22Ne neutron source. These stars will become supernovae at their demise and spew those s-process isotopes into interstellar gas.

The s-process is sometimes approximated over a small mass region using the so-called "local approximation", by which the ratio of abundances is inversely proportional to the ratio of neutron-capture cross-sections for nearby isotopes on the s-process path. This approximation is – as the name indicates – only valid locally, meaning for isotopes of nearby mass numbers, but it is invalid at magic numbers where the ledge-precipice structure dominates.

Chart representing the final part of the s-process. Red horizontal lines with a circle in their right ends represent neutron captures; blue arrows pointing up-left represent beta decays; green arrows pointing down-left represent alpha decays; cyan/light-green arrows pointing down-right represent electron captures. S-R-processes-atomic-mass-201-to-210.svg
Chart representing the final part of the s-process. Red horizontal lines with a circle in their right ends represent neutron captures; blue arrows pointing up-left represent beta decays; green arrows pointing down-left represent alpha decays; cyan/light-green arrows pointing down-right represent electron captures.

Because of the relatively low neutron fluxes expected to occur during the s-process (on the order of 105 to 1011 neutrons per cm2 per second), this process does not have the ability to produce any of the heavy radioactive isotopes such as thorium or uranium. The cycle that terminates the s-process is:

209
Bi
captures a neutron, producing 210
Bi
, which decays to 210
Po
by β decay. 210
Po
in turn decays to 206
Pb
by α decay:

209
83
Bi
 
+ 
n
 
  210
83
Bi
 
+ 
γ
210
83
Bi
 
    210
84
Po
 
+ 
e
 
+ 
ν
e
210
84
Po
 
    206
82
Pb
 
+  4
2
He

206
Pb
then captures three neutrons, producing 209
Pb
, which decays to 209
Bi
by β decay, restarting the cycle:

206
82
Pb
 
+ 3 
n
 
  209
82
Pb
209
82
Pb
 
    209
83
Bi
 
+  
e
 
+  
ν
e

The net result of this cycle therefore is that 4 neutrons are converted into one alpha particle, two electrons, two anti-electron neutrinos and gamma radiation:

  4 
n
 
  4
2
He
 
+ 2 
e
 
+ 2 
ν
e
 
+ 
γ

The process thus terminates in bismuth, the heaviest "stable" element, and polonium, the first non-primordial element after bismuth. Bismuth is actually slightly radioactive, but with a half-life so long—a billion times the present age of the universe—that it is effectively stable over the lifetime of any existing star. Polonium-210, however, decays with a half-life of 138  d to stable lead-206.

The s-process measured in stardust

Stardust is one component of cosmic dust. Stardust is individual solid grains that condensed during mass loss from various long-dead stars. Stardust existed throughout interstellar gas before the birth of the Solar System and was trapped in meteorites when they assembled from interstellar matter contained in the planetary accretion disk in early Solar System. Today they are found in meteorites, where they have been preserved. Meteoriticists habitually refer to them as presolar grains. The s-process enriched grains are mostly silicon carbide (SiC). The origin of these grains is demonstrated by laboratory measurements of extremely unusual isotopic abundance ratios within the grain. First experimental detection of s-process xenon isotopes was made in 1978, [18] confirming earlier predictions that s-process isotopes would be enriched, nearly pure, in stardust from red giant stars. [19] These discoveries launched new insight into astrophysics and into the origin of meteorites in the Solar System. [20] Silicon carbide (SiC) grains condense in the atmospheres of AGB stars and thus trap isotopic abundance ratios as they existed in that star. Because the AGB stars are the main site of the s-process in the galaxy, the heavy elements in the SiC grains contain almost pure s-process isotopes in elements heavier than iron. This fact has been demonstrated repeatedly by sputtering-ion mass spectrometer studies of these stardust presolar grains. [20] Several surprising results have shown that within them the ratio of s-process and r-process abundances is somewhat different from that which was previously assumed. It has also been shown with trapped isotopes of krypton and xenon that the s-process abundances in the AGB-star atmospheres changed with time or from star to star, presumably with the strength of neutron flux in that star or perhaps the temperature. This is a frontier of s-process studies in the 2000s.

Related Research Articles

In chemistry, an element is a pure substance consisting only of atoms that all have the same numbers of protons in their nuclei. Unlike chemical compounds, chemical elements cannot be broken down into simpler substances by any chemical reaction. The number of protons in the nucleus is the defining property of an element, and is referred to as its atomic number – all atoms with the same atomic number are atoms of the same element. All of the baryonic matter of the universe is composed of chemical elements. When different elements undergo chemical reactions, atoms are rearranged into new compounds held together by chemical bonds. Only a minority of elements, such as silver and gold, are found uncombined as relatively pure native element minerals. Nearly all other naturally occurring elements occur in the Earth as compounds or mixtures. Air is primarily a mixture of the elements nitrogen, oxygen, and argon, though it does contain compounds including carbon dioxide and water.

Big Bang nucleosynthesis Process during the early phases of the Universe

In physical cosmology, Big Bang nucleosynthesis is the production of nuclei other than those of the lightest isotope of hydrogen during the early phases of the Universe. Primordial nucleosynthesis is believed by most cosmologists to have taken place in the interval from roughly 10 seconds to 20 minutes after the Big Bang, and is calculated to be responsible for the formation of most of the universe's helium as the isotope helium-4 (4He), along with small amounts of the hydrogen isotope deuterium, the helium isotope helium-3 (3He), and a very small amount of the lithium isotope lithium-7 (7Li). In addition to these stable nuclei, two unstable or radioactive isotopes were also produced: the heavy hydrogen isotope tritium ; and the beryllium isotope beryllium-7 (7Be); but these unstable isotopes later decayed into 3He and 7Li, respectively, as above.

Nucleosynthesis is the process that creates new atomic nuclei from pre-existing nucleons and nuclei. According to current theories, the first nuclei were formed a few minutes after the Big Bang, through nuclear reactions in a process called Big Bang nucleosynthesis. After about 20 minutes, the universe had expanded and cooled to a point at which these high-energy collisions among nucleons ended, so only the fastest and simplest reactions occurred, leaving our universe containing about 75% hydrogen and 24% helium by mass. The rest is traces of other elements such as lithium and the hydrogen isotope deuterium. Nucleosynthesis in stars and their explosions later produced the variety of elements and isotopes that we have today, in a process called cosmic chemical evolution. The amounts of total mass in elements heavier than hydrogen and helium remains small, so that the universe still has approximately the same composition.

Stellar nucleosynthesis Process by which the natural abundances of the chemical elements within stars change due to nuclear fusion reactions

Stellar nucleosynthesis is the creation (nucleosynthesis) of chemical elements by nuclear fusion reactions within stars. Stellar nucleosynthesis has occurred since the original creation of hydrogen, helium and lithium during the Big Bang. As a predictive theory, it yields accurate estimates of the observed abundances of the elements. It explains why the observed abundances of elements change over time and why some elements and their isotopes are much more abundant than others. The theory was initially proposed by Fred Hoyle in 1946, who later refined it in 1954. Further advances were made, especially to nucleosynthesis by neutron capture of the elements heavier than iron, by Margaret and Geoffrey Burbidge, William Alfred Fowler and Hoyle in their famous 1957 B2FH paper, which became one of the most heavily cited papers in astrophysics history.

In astrophysics, silicon burning is a very brief sequence of nuclear fusion reactions that occur in massive stars with a minimum of about 8–11 solar masses. Silicon burning is the final stage of fusion for massive stars that have run out of the fuels that power them for their long lives in the main sequence on the Hertzsprung–Russell diagram. It follows the previous stages of hydrogen, helium, carbon, neon and oxygen burning processes.

Presolar grains Very old dust in space

Presolar grains are interstellar solid matter in the form of tiny solid grains that originated at a time before the Sun was formed. Presolar stardust grains formed within outflowing and cooling gases from earlier presolar stars.

<i>r</i>-process nucleosynthesis pathway

In nuclear astrophysics, the rapid neutron-capture process, also known as the r-process, is a set of nuclear reactions that is responsible for the creation of approximately half of the atomic nuclei heavier than iron; the "heavy elements", with the other half produced by the p-process and s-process. The r-process usually synthesizes the most neutron-rich stable isotopes of each heavy element. The r-process can typically synthesize the heaviest four isotopes of every heavy element, and the two heaviest isotopes, which are referred to as r-only nuclei, can be created via the r-process only. Abundance peaks for the r-process occur near mass numbers A = 82, A = 130 and A = 196.

The term p-process is used in two ways in the scientific literature concerning the astrophysical origin of the elements (nucleosynthesis). Originally it referred to a proton capture process which is the source of certain, naturally occurring, neutron-deficient isotopes of the elements from selenium to mercury. These nuclides are called p-nuclei and their origin is still not completely understood. Although it was shown that the originally suggested process cannot produce the p-nuclei, later on the term p-process was sometimes used to generally refer to any nucleosynthesis process supposed to be responsible for the p-nuclei.

Neutron capture Atomic nuclear process

Neutron capture is a nuclear reaction in which an atomic nucleus and one or more neutrons collide and merge to form a heavier nucleus. Since neutrons have no electric charge, they can enter a nucleus more easily than positively charged protons, which are repelled electrostatically.

Cosmic dust Dust floating in space

Cosmic dust, also called extraterrestrial dust or space dust, is dust which exists in outer space, or has fallen on Earth. Most cosmic dust particles measure between a few molecules and 0.1 mm. Larger particles are called meteoroids. Cosmic dust can be further distinguished by its astronomical location: intergalactic dust, interstellar dust, interplanetary dust and circumplanetary dust.

Supernova nucleosynthesis is the nucleosynthesis of chemical elements in supernova explosions.

Although there are nine known isotopes of helium (2He), only helium-3 and helium-4 are stable. All radioisotopes are short-lived, the longest-lived being 6
He
with a half-life of 806.92(24) milliseconds. The least stable is 10
He
, with a half-life of 260(40) yoctoseconds, although it is possible that 2
He
may have an even shorter half-life.

rp-process Process of nucleosynthesis

The rp-process consists of consecutive proton captures onto seed nuclei to produce heavier elements. It is a nucleosynthesis process and, along with the s-process and the r-process, may be responsible for the generation of many of the heavy elements present in the universe. However, it is notably different from the other processes mentioned in that it occurs on the proton-rich side of stability as opposed to on the neutron-rich side of stability. The end point of the rp-process is not yet well established, but recent research has indicated that in neutron stars it cannot progress beyond tellurium. The rp-process is inhibited by alpha decay, which puts an upper limit on the end point at 104Te, the lightest observed alpha-decaying nuclide, and the proton drip line in light antimony isotopes. At this point, further proton captures result in prompt proton emission or alpha emission, and thus the proton flux is consumed without yielding heavier elements; this end process is known as the tin–antimony–tellurium cycle.

Nuclear astrophysics

Nuclear astrophysics is an interdisciplinary part of both nuclear physics and astrophysics, involving close collaboration among researchers in various subfields of each of these fields. This includes, notably, nuclear reactions and their rates as they occur in cosmic environments, and modeling of astrophysical objects where these nuclear reactions may occur, but also considerations of cosmic evolution of isotopic and elemental composition (often called chemical evolution). Constraints from observations involve multiple messengers, all across the electromagnetic spectrum (nuclear gamma-rays, X-rays, optical, and radio/sub-mm astronomy), as well as isotopic measurements of solar-system materials such as meteorites and their stardust inclusions, cosmic rays, material deposits on Earth and Moon). Nuclear physics experiments address stability (i.e., lifetimes and masses) for atomic nuclei well beyond the regime of stable nuclides into the realm of radioactive/unstable nuclei, almost to the limits of bound nuclei (the drip lines), and under high density (up to neutron star matter) and high temperature (plasma temperatures up to 109 K). Theories and simulations are essential parts herein, as cosmic nuclear reaction environments cannot be realized, but at best partially approximated by experiments. In general terms, nuclear astrophysics aims to understand the origin of the chemical elements and isotopes, and the role of nuclear energy generation, in cosmic sources such as stars, supernovae, novae, and violent binary-star interactions.

Photodisintegration disintegration of atomic nuclei from high-energy EM radiation

Photodisintegration is a nuclear process in which an atomic nucleus absorbs a high-energy gamma ray, enters an excited state, and immediately decays by emitting a subatomic particle. The incoming gamma ray effectively knocks one or more neutrons, protons, or an alpha particle out of the nucleus. The reactions are called (γ,n), (γ,p), and (γ,α).

The B2FH paper was a landmark scientific paper on the origin of the chemical elements. The paper's title is Synthesis of the Elements in Stars, but it became known as B2FH from the initials of its authors: Margaret Burbidge, Geoffrey Burbidge, William A. Fowler, and Fred Hoyle. It was written from 1955 to 1956 at the University of Cambridge and Caltech, then published in Reviews of Modern Physics in 1957.

Neutron capture nucleosynthesis describes two nucleosynthesis pathways: the r-process and the s-process, for rapid and slow neutron captures, respectively. R-process describes neutron capture in a region of high neutron flux, such as during supernova nucleosynthesis after core-collapse, and yields neutron-rich nuclides. S-process describes neutron capture that is slow relative to the rate of beta decay, as for stellar nucleosynthesis in some stars, and yields nuclei with stable nuclear shells. Each process is responsible for roughly half of the observed abundances of elements heavier than iron. The importance of neutron capture to the observed abundance of the chemical elements was first described in 1957 in the B2FH paper.

Nuclear drip line Atomic nuclei decay delimiter

The nuclear drip line is the boundary delimiting the zone beyond which atomic nuclei decay by the emission of a proton or neutron.

p-nuclei are certain proton-rich, naturally occurring isotopes of some elements between selenium and mercury inclusive which cannot be produced in either the s- or the r-process.

Donald D. Clayton American astrophysicist

Donald Delbert Clayton is an American astrophysicist whose most visible achievement was the prediction from nucleosynthesis theory that supernovae are intensely radioactive. That earned Clayton the NASA Exceptional Scientific Achievement Medal (1992) for “theoretical astrophysics related to the formation of (chemical) elements in the explosions of stars and to the observable products of these explosions”. Supernovae thereafter became the most important stellar events in astronomy owing to their profoundly radioactive nature. Not only did Clayton discover radioactive nucleosynthesis during explosive silicon burning in stars but he also predicted a new type of astronomy based on it, namely the associated gamma-ray line radiation emitted by matter ejected from supernovae. That paper was selected as one of the fifty most influential papers in astronomy during the twentieth century for the Centennial Volume of the American Astronomical Society. He gathered support from influential astronomers and physicists for a new NASA budget item for a gamma-ray-observatory satellite, achieving successful funding for Compton Gamma Ray Observatory. With his focus on radioactive supernova gas Clayton discovered a new chemical pathway causing carbon dust to condense there by a process that is activated by the radioactivity.

References

  1. Suess, H. E.; Urey, H. C. (1956). "Abundances of the Elements". Reviews of Modern Physics . 28 (1): 53–74. Bibcode:1956RvMP...28...53S. doi:10.1103/RevModPhys.28.53.
  2. Burbidge, E. M.; Burbidge, G. R.; Fowler, W. A.; Hoyle, F. (1957). "Synthesis of the Elements in Stars". Reviews of Modern Physics . 29 (4): 547–650. Bibcode:1957RvMP...29..547B. doi: 10.1103/RevModPhys.29.547 .
  3. Hammond, C. R. (2004). "The Elements". Handbook of Chemistry and Physics (81st ed.). CRC Press. ISBN   978-0-8493-0485-9.
  4. Moore, C. E. (1951). "Technetium in the Sun". Science . 114 (2951): 59–61. Bibcode:1951Sci...114...59M. doi:10.1126/science.114.2951.59. PMID   17782983.
  5. Merrill, P. W. (1952). "Technetium in the stars". Science. 115 (2992): 484.
  6. George Sivulka (8 March 2017). "An Introduction to the Evidence for Stellar Nucleosynthesis". Stanford University. Retrieved 3 May 2018.
  7. Clayton, D. D.; Fowler, W. A.; Hull, T. E.; Zimmerman, B. A. (1961). "Neutron capture chains in heavy element synthesis". Annals of Physics . 12 (3): 331–408. Bibcode:1961AnPhy..12..331C. doi:10.1016/0003-4916(61)90067-7.
  8. Clayton, D. D.; Rassbach, M. E. (1967). "Termination of the s-process". The Astrophysical Journal . 148: 69. Bibcode: 1967ApJ...148...69C . doi: 10.1086/149128 .
  9. Clayton, D. D. (1968). "Distribution of neutron-source strengths for the s-process". In Arnett, W. D.; Hansen, C. J.; Truran, J. W.; Cameron, A. G. W. (eds.). Nucleosynthesis. Gordon and Breach. pp. 225–240.
  10. Peters, J. G.; Fowler, W. A.; Clayton, D. D. (1972). "Weak s-process Irradiations". The Astrophysical Journal . 173: 637. Bibcode: 1972ApJ...173..637P . doi: 10.1086/151450 .
  11. Clayton, D. D.; Newman, M. J. (1974). "s-process Studies: Exact Solution to a Chain Having Two Distinct Cross-Section Values". The Astrophysical Journal . 192: 501. Bibcode: 1974ApJ...192..501C . doi: 10.1086/153082 .
  12. Clayton, D. D.; Ward, R. A. (1974). "s-process Studies: Exact Evaluation of an Exponential Distribution of Exposures". The Astrophysical Journal . 193: 397. Bibcode: 1974ApJ...193..397C . doi: 10.1086/153175 .
  13. Ward, R. A.; Newman, M. J.; Clayton, D. D. (1976). "s-process Studies: Branching and the Time Scale". The Astrophysical Journal Supplement Series . 31: 33. Bibcode: 1976ApJS...31...33W . doi: 10.1086/190373 .
  14. Macklin, R. L.; Gibbons, J. H. (1965). "Neutron Capture Data at Stellar Temperatures". Reviews of Modern Physics . 37 (1): 166–176. Bibcode:1965RvMP...37..166M. doi:10.1103/RevModPhys.37.166.
  15. Kaeppeler, F.; Beer, H.; Wisshak, K.; Clayton, D. D.; Macklin, R. L.; Ward, R. A. (1982). "s-process studies in the light of new experimental cross sections". The Astrophysical Journal . 257: 821–846. Bibcode: 1982ApJ...257..821K . doi: 10.1086/160033 .
  16. Boothroyd, A. I. (2006). "Heavy elements in stars". Science . 314 (5806): 1690–1691. doi:10.1126/science.1136842. PMID   17170281. S2CID   116938510.
  17. Busso, M.; Gallino, R.; Wasserburg, G. J. (1999). "Nucleosynthesis in Asymptotic Giant Branch Stars: Relevance for Galactic Enrichment and Solar System Formation" (PDF). Annual Review of Astronomy and Astrophysics . 37 (1): 239–309. Bibcode:1999ARA&A..37..239B. doi:10.1146/annurev.astro.37.1.239.
  18. Srinivasan, B.; Anders, E. (1978). "Noble Gases in the Murchison Meteorite: Possible Relics of s-process Nucleosynthesis". Science . 201 (4350): 51–56. Bibcode:1978Sci...201...51S. doi:10.1126/science.201.4350.51. PMID   17777755. S2CID   21175338.
  19. Clayton, D. D.; Ward, R. A. (1978). "s-process studies: Xenon and krypton isotopic abundances". The Astrophysical Journal . 224: 1000. Bibcode: 1978ApJ...224.1000C . doi: 10.1086/156449 .
  20. 1 2 Clayton, D. D.; Nittler, L. R. (2004). "Astrophysics with Presolar Stardust" (PDF). Annual Review of Astronomy and Astrophysics . 42 (1): 39–78. Bibcode:2004ARA&A..42...39C. doi:10.1146/annurev.astro.42.053102.134022. S2CID   96456868. Archived from the original (PDF) on 2020-02-19.