The rapid neutron-capture process, or so-called r-process, is a series of successive nuclear reactions in nuclear astrophysics that is responsible for the creation (nucleosynthesis) of approximately half the abundances of the atomic nuclei heavier than iron. The r-process entails a succession of rapid neutron captures (hence the name) by one or more seed nuclei, typically beginning with nuclei near 56Fe that are the heaviest produced in stars. Captures must be rapid enough such that another neutron may be captured before beta decay has sufficient time to occur. The sequence of captures is halted only when the increasingly neutron-rich nuclei cannot physically retain another neutron as they approach the neutron drip line, or at peaks of abundance at magic numbers where binding energy and stability increase.[1] The r-process therefore can occur only in locations where there exists a high density of free neutrons, on the order of 1024 free neutrons per cm3.[2][3] This amounts to almost a gram of free neutrons in every cubic centimeter, an astonishing number requiring extreme locations with temperatures of about one billion degrees. The r-process is known to occur in core-collapse supernovae (as part of supernova nucleosynthesis)[4] or decompression of neutron-star matter thrown off by a binary neutron star merger,[5] though the relative contributions of these sources to the astrophysical abundance of r-process elements is a matter of ongoing research.[6]

The r-process contrasts with the s-process, the other predominant mechanism for the production of heavy elements, in which neutron captures occur at a much slower rate. Together, the r- and s-processes account for almost the entire abundance of chemical elements heavier than iron. The s-process primarily occurs within ordinary stars, particularly AGB stars, where the neutron flux is sufficient to cause neutron captures to recur in accumulated heavy seed nuclei in intervals longer than beta decay half-lives. The r-process, however, requires many neutron captures to occur faster than beta-decay,[3] and no sufficient quantities of heavy seed nuclei are thus created in s-process nucleosynthesis to contribute to the production of significant amounts of heavy elements. Furthermore, as the r-process generates its own seed nuclei and can occur much farther from the valley of stability, it is entirely responsible for the production of the most neutron-rich isotopes of each element as well as all elements heavier than bismuth.[3] Hence, in heavy element formation, the s-process is secondary to the r-process.

A limited r-process-like series of neutron captures occurs to a minor extent in thermonuclear weapon explosions. These led to the discovery of the elements einsteinium (element 99) and fermium (element 100) in nuclear weapon fallout. The historical challenge has been to locate physical settings appropriate for their time scales; thus, it remains open whether the r-process can artificially be induced and if it is capable of producing elements heavier than fermium.

History

edit

Following pioneering research into the Big Bang and the formation of helium in stars, an unknown process responsible for producing heavier elements found on Earth from hydrogen and helium was suspected to exist. One early attempt at explanation came from Chandrasekhar and Henrich who postulated that elements were produced at temperatures between 6×109 and 8×109 K. Their theory accounted for elements up to chlorine, though there was no explanation for elements of atomic weight heavier than 40 amu at non-negligible abundances.[7] This became the foundation of a study by Fred Hoyle, who hypothesized that conditions in the core of collapsing stars would enable nucleosynthesis of the remainder of the elements via rapid capture of densely packed free neutrons. However, there remained unanswered questions about equilibrium in stars that was required to balance beta-decays and precisely account for abundances of elements that would be formed in such conditions.[7]

The need for a physical setting providing rapid neutron capture, which was known to almost certainly have a role in element formation, was also seen in a table of abundances of isotopes of heavy elements by Hans Suess and Harold Urey in 1956.[8] Their abundance table revealed larger than average abundances of natural isotopes containing magic numbersCite error: There are <ref> tags on this page without content in them (see the help page). of neutrons as well as abundance peaks about 10 amu lighter than stable nuclei containing magic numbers of neutrons which were also in abundance, suggesting that radioactive neutron-rich nuclei having the magic neutron numbers but roughly ten fewer protons were formed. These observations also implied that rapid neutron capture occurred faster than beta decay, and the resulting abundance peaks were caused by so-called waiting points at magic numbers.[2]Cite error: There are <ref> tags on this page without content in them (see the help page). This process, rapid neutron capture by neutron-rich isotopes, became known as the r-process, whereas the s-process was named for its characteristic slow neutron capture. A table apportioning the heavy isotopes phenomenologically between s-process and r-process isotopes was published in 1957 in the B2FH review paper,[2]  which named the r-process and outlined the physics that guides it. Alastair G. W. Cameron also published a smaller study about the r-process in the same year.[9]

Mechanism

edit

Initiation

edit

There are three candidate sites for r-process nucleosynthesis where the required conditions are thought to exist: low-mass supernovae, Type II supernovae, and neutron star mergers.[10]

Immediately after the severe compression of electrons in a Type II supernova, beta-minus decay is blocked. This is because the high electron density fills all available free electron states up to a Fermi energy which is greater than the energy of nuclear beta decay. But nuclear capture of those free electrons still occurs, and causes increasing neutronization of matter. There results an extremely high density of free neutrons which cannot decay, and as a result a large neutron density (on the order of 1024 neutrons per cm3)[2] and high temperatures. As this re-expands and cools, neutron capture by still-existing heavy nuclei occurs much faster than beta-minus decay.

Waiting points and beta-decay

edit

Neutron captures in r-process nucleosynthesis leads to the formation of neutron-rich, weakly bound nuclei with neutron separation energies as low as 2 MeV.[1][2] At this stage, closed neutron shells at N = 50, 82, and 126 are reached and neutron capture is temporarily paused. These waiting points are characterized by increased binding energy relative to heavier isotopes, leading to low neutron capture cross sections and a buildup of semi-magic nuclei that are more stable toward beta decay.[11] These nuclei are then allowed to beta decay toward stability before further neutron capture can occur,[2] resulting in a slowdown or freeze-out of the reaction.[11]

Abundance of elements

edit
 
Predicted decay modes of heavy and superheavy nuclei. Green denotes nuclei decaying by spontaneous fission, at A ~ 275 and A ≥ 290, that block r-process nucleosynthesis and decay towards the island of stability (nuclei with white borders)

Limits

edit

Beyond the heaviest primordial elements thorium and uranium, it is unknown if heavier elements may be produced naturally via the r-process. The concept of an island of stability, located in the superheavy mass region near the hypothesized magic neutron number N = 184 and isotopes 291Cn and 293Cn with relatively long half-lives, has fascinated many researchers.[12] It is hypothesized that spontaneous fission will limit the r-process before an island of stability is reached. The Finite Range Droplet Model (FRDM) predicts a limit at Z ~ 94 and N ~ 186, after which neutron-induced fission becomes the dominant reaction over neutron capture. This is caused in part by the N = 184 shell closure and a calculated shift of the position of the drip line closer to beta-stability, which would render further produced nuclei too weakly bound.[13] It is also possible, according to the Extended Thomas-Fermi Model (ETFSI), that neutron captures may continue until Z ~ 105 and N ~ 220 before succumbing to beta-delayed fission.[13] In both models, the decreasing height of fission barriers plays an important role in the stability of superheavy nuclei as fission half-lives quickly decrease relative to beta half-lives. This then would terminate the r-process entirely as neutron density would become too low to sustain the reaction and lower-mass fission products would instead be populated. Hence, it is likely that no nuclei beyond A ~ 280 are produced in the r-process or are allowed to decay toward stability.[13] Even if these fission-dominated regions may be bypassed, especially the gaps of instability known at 258-260Fm and predicted at A ~ 275, the natural abundance of superheavy elements would be on the order of 10-12 relative to lead.[12]

Astrophysical occurrences

edit
 
Periodic table showing the cosmogenic origin of each element. The elements heavier than iron with origins in supernovae are typically those produced by the r-process.

Supernovae

edit

Neutron-star mergers

edit

Artificial processes similar to the r-process

edit

Original article

edit

History

edit

The need for a physical setting providing rapid capture of neutrons was seen from the relative abundances of isotopes of heavy chemical elements given in a table of abundances by Hans Suess and Harold Urey in 1956.[14] Not only did their abundance table reveal larger than average abundances of natural isotopes containing magic numbersCite error: There are <ref> tags on this page without content in them (see the help page). numbers of neutrons. These came to be called magic numbers.</ref> of neutrons but also abundance peaks about 10 amu lighter than those containing magic numbers of neutrons in their structure. They also realized that captures of free neutrons should be part of any explanation because of the lack of electric repulsion between nuclei and chargeless neutrons. This phenomenology suggested that these lighter subsidiary abundance peaks could result from radioactive nuclei having the magic neutron numbers but roughly ten fewer protons. To achieve this would require radioactive neutron-rich isotopes to capture another neutron faster than they can undergo beta decay in order to create abundance peaks that will decay subsequently to germanium, xenon, and platinum, elements prominent in the r-process abundance peaks. According to the nuclear shell model, radioactive nuclei that would decay into isotopes of these elements must have closed neutron shells near the neutron drip line, where more neutrons cannot be added. The neutron-capture flow must therefore wait for beta decay at those so-called waiting points, which therefore grow larger in abundance similar to water in a dammed up river. For those hitherto unexplained abundance peaks, which are approximately 10 u lighter than the s-process abundance peaks, to be created by rapid neutron capture implied that other neutron-rich nuclei would also be synthesized by the same process. That process, rapid neutron capture by neutron-rich isotopes, is called the r-process. A table apportioning the heavy isotopes phenomenologically between s-process and r-process isotopes was published in 1957 in the famous B2FH review paper[2] which named the r-process and outlined the physics that guides it. Alastair G. W. Cameron also published a smaller study about the r-process in the same year.[15]

The stationary r-process as described by the B2FH paper was first demonstrated in a time-dependent calculation at Caltech by Phillip A. Seeger, William A. Fowler and Donald D. Clayton,[16] who found that no single temporal snapshot matched the solar r-process abundances, but, that when superposed, did achieve a successful characterization of the r-process abundance distribution . Shorter-time distributions emphasize abundances at atomic weights less than A = 140, whereas longer-time distributions emphasized those at atomic weights greater than A = 140.[17] Subsequent treatments of the r-process reinforced those temporal features. Seeger et al. were also able to construct more quantitative apportionment between s-process and r-process of the abundance table of heavy isotopes, thereby establishing a more reliable abundance curve for the r-process isotopes than B2FH had been able to define. Today, the r-process abundances are determined using their technique of subtracting the more reliable s-process isotopic abundances from the total isotopic abundances and attributing the remainder to r-process nucleosynthesis.[18] That r-process abundance curve (vs. atomic weight) has provided for many decades the target for theoretical computations of abundances synthesized by the physical r-process.

The creation of free neutrons by electron capture during the rapid collapse to high density of a supernova core along with quick assembly of some neutron-rich seed nuclei makes the r process a primary nucleosynthesis process, meaning a process that can occur even in a star initially of pure H and He, in contrast to the B2FH designation as a secondary process building on preexisting iron. Primary stellar nucleosynthesis begins earlier in the galaxy than does secondary nucleosynthesis. Alternatively the high density of neutrons within neutron stars would be available for rapid assembly into r-process nuclei if a collision were to eject portions of a neutron star, which then rapidly expands freed from confinement. That sequence could also begin earlier in galactic time than would s-process nucleosynthesis; so each scenario fits the earlier growth of r-process abundances in the galaxy. Each of these scenarios is the subject of active theoretical research. Observational evidence of the early r-process enrichment of interstellar gas and of subsequent newly formed of stars, as applied to the abundance evolution of the galaxy of stars, was first laid out by James W. Truran in 1981.[19] He and subsequent astronomers showed that the pattern of heavy-element abundances in the earliest metal-poor stars matched that of the shape of the solar r-process curve, as if the s-process component were missing. This was consistent with the hypothesis that the s-process had not yet begun to enrich interstellar gas when these young stars missing the s-process abundances were born from that gas, for it requires about 100 million years of galactic history for the s-process to get started whereas the r process can begin after two million years. These s-process-poor, r-process-rich stellar compositions must have been born earlier than any s-process, showing that the r-process emerges from quickly-evolving massive stars that become supernovae and leave neutron-star remnants that can merge with another neutron star. The primary nature of the early r-process thereby derives from observed abundance spectra in old stars[6] that had been born early, when the galactic metallicity was still small, but that nonetheless contain their complement of r-process nuclei.

 
Periodic table showing the cosmogenic origin of each element. The elements heavier than iron with origins in supernovae are typically those produced by the r-process, which is powered by supernovae neutron bursts

Either interpretation, though generally supported by supernova experts, has yet to achieve a totally satisfactory calculation of r-process abundances because the overall problem is numerically formidable; but existing results are supportive. And in 2017 new data about the r-process was discovered when the LIGO and Virgo gravitational-wave observatories discovered a merger of two neutron stars ejecting r-process matter.[20] See Astrophysical sites below.

Noteworthy is that the r-process is responsible for our natural cohort of radioactive elements, such as uranium and thorium, as well as the most neutron-rich isotopes of each heavy element.

Nuclear physics

edit

Immediately after the severe compression of electrons in a core-collapse supernova, beta-minus decay is blocked. This is because the high electron density fills all available free electron states up to a Fermi energy which is greater than the energy of nuclear beta decay. But nuclear capture of those free electrons still occurs, and causes increasing neutronization of matter. There results an extremely high density of free neutrons which cannot decay, and as a result a large neutron density (on the order of 1024 neutrons per cm3[2]) and high temperatures. As this re-expands and cools, neutron capture by still-existing heavy nuclei occurs much faster than beta-minus decay. As a consequence, the r-process runs up along the neutron drip line and highly-unstable neutron-rich nuclei are created.

Three processes which affect the process of climbing the neutron drip line are; a notable decrease in the neutron-capture cross section at nuclei with closed neutron shells, the inhibiting process of photodisintegration, and the degree of nuclear stability in the heavy-isotope region. This last phenomenon terminates the r-process when its heaviest nuclei become unstable to spontaneous fission, which is currently believed to be in the neutron-rich region of the table of nuclides when the total number of nucleons approaches 270; even before this, the fission barrier may be low enough that neutron capture might induce fission instead of continuing up the neutron drip line.[21] After the neutron flux decreases, these highly unstable radioactive nuclei undergo a rapid succession of beta decays until they reach more stable, neutron-rich nuclei.[22] So, while the s-process creates an abundance of stable nuclei having closed neutron shells, the r-process in neutron-rich predecessor nuclei creates an abundance of radioactive nuclei about 10 amu below the s-process peaks after their decay back to stability.[23]

The r-process also occurs in thermonuclear weapons, and was responsible for the initial discovery of neutron-rich almost stable isotopes of actinides like plutonium-244 and the new elements einsteinium and fermium (atomic numbers 99 and 100) in the 1950s. It has been suggested that multiple nuclear explosions would make it possible to reach the island of stability, as the affected nuclides (starting with uranium-238 as seed nuclei) would not have time to beta decay all the way to the quickly spontaneously fissioning nuclides at the line of beta stability before they absorbed more neutrons in the next explosion, thus providing a chance to reach neutron-rich superheavy nuclides like copernicium-291 and -293 which should have half-lives of centuries or millennia.[12]

Astrophysical sites

edit

The most probable candidate site for the r-process has long been suggested to be core-collapse supernovae (spectral types Ib, Ic and II), which may provide the necessary physical conditions for the r-process. However, the very low abundance of r-process nuclei in the interstellar gas limits the amount each can have ejected. It requires either that only a small fraction of supernovae eject r-process nuclei to the interstellar medium, or that each supernova ejects only a very small amount of r-process material. The ejected material must be relatively neutron-rich, a condition which has been difficult to achieve in models,[4] so that astrophysicists remain uneasy about their adequacy for successful r-process yields

Entirely new astronomical data about the r process was discovered in 2017 when the LIGO and Virgo gravitational-wave observatories discovered a merger of two neutron stars.[20] The localization on the sky of the source of the gravitational waves radiated by the collapse of the two neutron stars into a black hole, but with significant spun off mass of highly neutralized matter, enabled several teams[24][25][26] to discover and study the optical counterpart of the merger, finding spectroscopic evidence of r-process material thrown off by the merging neutron stars. The bulk of this material seems to consist of two types: hot blue masses of highly radioactive r-process matter of lower-mass-range heavy nuclei (A < 140) and cooler red masses of higher mass-number r-process nuclei (A > 140) rich in actinides (such as uranium, thorium, californium etc). When released from the huge internal pressure of the neutron star, these ejecta expand and form seed heavy nuclei that rapidly capture free neutrons, and radiate detected optical light for about a week. Such duration of luminosity would not be possible without heating by internal radioactive decay, which is provided by r-process nuclei near their waiting points. Two distinct mass regions (A < 140 and A > 140) for the r-process yields have been known since the first time dependent calculations of the r-process.[16] Because of these spectroscopic features it has been argued that such nucleosynthesis in the Milky Way has been primarily ejecta from neutron-star mergers rather than from supernovae.[5]

These results offer a new possibility for clarifying six decades of uncertainty over the site of origin of r-process nuclei. Confirming relevance to the r-process is that it is radiogenic power from radioactive decay of r-process nuclei that maintains the visibility of these spun off r-process fragments. Otherwise they would dim quickly. Such alternative sites were first seriously proposed in 1974[27] as decompressing neutron star matter. It was proposed such matter is ejected from neutron stars merging with black holes in compact binaries. In 1989[28] (and 1999[29]) this scenario was extended to binary neutron star mergers (a binary star system of two neutron stars that collide). After preliminary identification of these sites,[30] the scenario was confirmed in GW170817.

Notes and references

edit
Notes
References
  1. ^ a b Cite error: The named reference thoennes was invoked but never defined (see the help page).
  2. ^ a b c d e f g h Burbidge, E. M.; Burbidge, G. R.; Fowler, W. A.; Hoyle, F. (1957). "Synthesis of the Elements in Stars" (PDF). Reviews of Modern Physics. 29 (4): 547–650. Bibcode:1957RvMP...29..547B. doi:10.1103/RevModPhys.29.547.
  3. ^ a b c Riquelme, M.A. (2009). "R-process nucleosynthesis and its site" (PDF).
  4. ^ a b Thielemann, F.-K.; et al. (2011). "What are the astrophysical sites for the r-process and the production of heavy elements?". Progress in Particle and Nuclear Physics. 66 (2): 346–353. Bibcode:2011PrPNP..66..346T. doi:10.1016/j.ppnp.2011.01.032.
  5. ^ a b Kasen, D.; Metzger, B.; Barnes, J.; Quataert, E.; Ramirez-Ruiz, E. (2017). "Origin of the heavy elements in binary neutron-star mergers from a gravitational-wave event" (PDF). Nature. 551 (7678): 80–84. arXiv:1710.05463. Bibcode:2017Natur.551...80K. doi:10.1038/nature24453. PMID 29094687.
  6. ^ a b Frebel, A.; Beers, T. C. (2018). "The formation of the heaviest elements". Physics Today. 71 (1): 30–37. arXiv:1801.01190. Bibcode:2018PhT....71a..30F. doi:10.1063/pt.3.3815. Nuclear physicists are still working to model the r-process, and astrophysicists need to estimate the frequency of neutron-star mergers to assess whether r-process heavy-element production solely or at least significantly takes place in the merger environment.
  7. ^ a b Hoyle, F. (1946). "The Synthesis of the Elements from Hydrogen". Monthly Notices of the Royal Astronomical Society. 106 (5): 343–383. doi:10.1093/mnras/106.5.343.{{cite journal}}: CS1 maint: unflagged free DOI (link)
  8. ^ 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.
  9. ^ Cameron, A. G. W. (1957). "Nuclear reactions in stars and nucleogenesis". Publications of the Astronomical Society of the Pacific. 69 (408): 201. Bibcode:1957PASP...69..201C. doi:10.1086/127051.
  10. ^ Bartlett, A.; Görres, J.; Mathews, G.J.; Otsuki, K.; Wiescher, W. (2006). "Two-neutron capture reactions and the r process" (PDF). Physical Review C. 74: 015082. doi:10.1103/PhysRevC.74.015802.
  11. ^ a b Eichler, M.A. (2016). Nucleosynthesis in explosive environments: neutron star mergers and core-collapse supernovae (PDF) (Doctoral thesis). University of Basel.
  12. ^ a b c Zagrebaev, Valeriy; Karpov, Alexander; Greiner, Walter (2013). "Future of superheavy element research: Which nuclei could be synthesized within the next few years?" (PDF). Journal of Physics. 420. IOP Publishing Ltd: 012001. arXiv:1207.5700. doi:10.1088/1742-6596/420/1/012001. Cite error: The named reference "Zagrebaev" was defined multiple times with different content (see the help page).
  13. ^ a b c Petermann, I; Langanke, K.; Martínez-Pinedo, G.; Panov, I.V; Reinhard, P.G.; Thielemann, F.K. (2012). "Have superheavy elements been produced in nature?". European Physical Journal A. 48 (122). arXiv:1207.3432. doi:10.1140/epja/i2012-12122-6.
  14. ^ 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.
  15. ^ Cameron, A. G. W. (1957). "Nuclear reactions in stars and nucleogenesis". Publications of the Astronomical Society of the Pacific. 69 (408): 201. Bibcode:1957PASP...69..201C. doi:10.1086/127051.
  16. ^ a b Seeger, P. A.; Fowler, W. A.; Clayton, D. D. (1965). "Nucleosynthesis of heavy elements by neutron capture". Astrophysical Journal Supplement. 11: 121–66. Bibcode:1965ApJS...11..121S. doi:10.1086/190111.
  17. ^ See Seeger, Fowler & Clayton 1965. Figure 16 shows the short-flux calculation and its comparison with natural r-process abundances whereas Figure 18 shows the calculated abundances for long neutron fluxes.
  18. ^ See Table 4 in Seeger, Fowler & Clayton 1965.
  19. ^ Truran, J. W. (1981). "A new interpretation of the heavy-element abundances in metal-deficient stars". Astronomy and Astrophysics. 97 (2): 391–93. Bibcode:1981A&A....97..391T.
  20. ^ a b Abbott, B. P.; et al. (LIGO Scientific Collaboration and Virgo Collaboration) (2017). "GW170817: Observation of Gravitational Waves from a Binary Neutron Star Inspiral". Physical Review Letters. 119 (16): 161101. arXiv:1710.05832. Bibcode:2017PhRvL.119p1101A. doi:10.1103/PhysRevLett.119.161101. PMID 29099225.
  21. ^ Boleu, R.; Nilsson, S. G.; Sheline, R. K. (1972). "On the termination of the r-process and the synthesis of superheavy elements". Physics Letters B. 40 (5): 517–521. Bibcode:1972PhLB...40..517B. doi:10.1016/0370-2693(72)90470-4.
  22. ^ Clayton, D. D. (1968), Principles of Stellar Evolution and Nucleosynthesis, Mc-Graw-Hill, pp. 577–91, ISBN 978-0226109534, provides a clear technical introduction to these features. More technical can be found in Seeger, Fowler & Clayton 1965.
  23. ^ Figure 10 of Seeger, Fowler & Clayton 1965 shows this path of captures reaching magic neutron numbers 82 and 126 at smaller values of nuclear charge Z than it does along the stability path
  24. ^ Arcavi, I.; et al. (2017). "Optical emission from a kilonova following a gravitational-wave-detected neutron-star merger" (PDF). Nature. 551 (7678): 64–66. arXiv:1710.05843. Bibcode:2017Natur.551...64A. doi:10.1038/nature24291.
  25. ^ Pian, E.; et al. (2017). "Spectroscopic identification of r-process nucleosynthesis in a double neutron-star merger" (PDF). Nature. 551 (7678): 67–70. arXiv:1710.05858. Bibcode:2017Natur.551...67P. doi:10.1038/nature24298. PMID 29094694.
  26. ^ Smartt, S. J.; et al. (2017). "A kilonova as the electromagnetic counterpart to a gravitational-wave source" (PDF). Nature. 551 (7678): 75–79. arXiv:1710.05841. Bibcode:2017Natur.551...75S. doi:10.1038/nature24303. PMID 29094693.
  27. ^ Lattimer, J. M.; Schramm, D. N. (1974). "Black Hole–Neutron Star Collisions". Astrophysical Journal Letters. 192 (2): L145–147. Bibcode:1974ApJ...192L.145L. doi:10.1086/181612.
  28. ^ Eichler, D.; Livio, M.; Piran, T.; Schramm, D. N. (1989). "Nucleosynthesis, neutrino bursts and gamma-rays from coalescing neutron stars" (PDF). Nature. 340 (6229): 126–128. Bibcode:1989Natur.340..126E. doi:10.1038/340126a0.
  29. ^ Freiburghaus, C.; Rosswog, S.; Thielemann, F.-K (1999). "r-process in Neutron Star Mergers". Astrophysical Journal Letters. 525 (2): L121–L124. Bibcode:1999ApJ...525L.121F. doi:10.1086/312343. PMID 10525469.
  30. ^ Tanvir, N.; et al. (2013). "A 'kilonova' associated with the short-duration gamma-ray burst GRB 130603B" (PDF). Nature. 500 (7464): 547–9. arXiv:1306.4971. Bibcode:2013Natur.500..547T. doi:10.1038/nature12505. PMID 23912055.