Seeking out Majorana under the microscope Patrick A. Lee Science 346, 545 (2014); DOI: 10.1126/science.1260282
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here.
The following resources related to this article are available online at www.sciencemag.org (this information is current as of November 5, 2014 ): Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/346/6209/545.full.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/346/6209/545.full.html#related This article cites 9 articles, 1 of which can be accessed free: http://www.sciencemag.org/content/346/6209/545.full.html#ref-list-1 This article appears in the following subject collections: Physics http://www.sciencemag.org/cgi/collection/physics
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2014 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on November 5, 2014
Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.
To obtain clear spectra of the conformers of H2CO3, Reisenauer et al. developed a very clean source: the pyrolysis of esters of H2CO3 (8). This method had been demonstrated to produce carboxylic acids and also suggested, but not proven experimentally, to produce H2CO3 (see the scheme) (11). The products of the pyrolysis were trapped in solid argon at extremely low temperature (8 K), and characterization by IR spectroscopy revealed that the more stable cis,cis and the cis,trans conformers were formed. The spectra perfectly matched the spectra obtained by sublimation of β-H2CO3 and trapping the products in solid argon, but markedly differed from those of α-H2CO3 These results suggested that β-H2CO3 indeed is solid H2CO3, whereas α-H2CO3 has a different constitution. Reisenauer et al. also showed that “α-H2CO3” is actually its monomethyl ester, CO(OH)(OCH3). By using acidic CH3OH for the protonation, the classic conditions for esterification of carboxylic acids, the monomethyl ester instead of the free carbonic acid is formed. O O
H
R⬘
O
O H
O
R⬘ O O
H O H
+
R⬘
H O + H
R
R⬘ Ester pyrolysis to produce carbonic acid, where R is an alkyl group.
This finding also explains why aqueous HCl and HCO3⫺ in the absence of CH3OH produced solid H2CO3 as expected, which was formerly identified as the β-polymorph. Thanks to the clean synthesis of H2CO3 via ester pyrolysis, the fog has lifted, and we can now look forward to a clearer exploration of the chemical landscape of this fascinating compound. ■ REFERENCES
1. B. Jonsson et al., J. Am. Chem. Soc. 99, 4628 (1977). 2. J. K. Terlouw, C. B. Lebrilla, H. Schwarz, Angew. Chem. Int. Ed. Engl. 26, 354 (1987). 3. M. H. Moore, R. K. Khanna, Spectrochim. Acta A Mol. Biomol. Spectrosc. 47, 255 (1991). 4. W. Hage, A. Hallbrucker, E. Mayer, J. Am. Chem. Soc. 115, 8427 (1993). 5. W. Hage, A. Hallbrucker, E. Mayer, J. Mol. Struct. 408-409, 527 (1997). 6. W. Hage, K. R. Liedl, A. Hallbrucker, E. Mayer, Science 279, 1332 (1998). 7. J. Bernard, R. G. Huber, K. R. Liedl, H. Grothe, T. Loerting, J. Am. Chem. Soc. 135, 7732 (2013). 8. H. P. Reisenauer, J. P. Wagner, P. R. Schreiner, Angew. Chem. Int. Ed. 53, 11766 (2014). 9. T. Loerting et al., Angew. Chem. Int. Ed. 39, 891 (2000). 10. K. Adamczyk, M. Prémont-Schwarz, D. Pines, E. Pines, E. T. J. Nibbering, Science 326, 1690 (2009). 11. G. Bucher, Eur. J. Org. Chem. 2010, 1070 (2010).
10.1126/science.1260117
PHYSICS
Seeking out Majorana under the microscope A chain of iron atoms on lead may reveal a signature of the elusive Majorana particle By Patrick A. Lee
T
he Dirac equation was initially developed to give a quantum mechanical description of particles such as electrons, but ended up predicting the existence of positrons—the antiparticle of the electron. A year before he disappeared under mysterious circumstances in 1938, the young Italian physicist Ettore Majorana discovered a solution to the Dirac equation that implied the existence of particles, or states of matter, that are their own antiparticles. This finding was contrary to Dirac’s solution, in which particles (electrons) and their antiparticles (positrons) are distinct. It has long been suspected, but not proven, that neutrinos are Majorana particles (1). In the past several years, the Majorana state has attracted the attention of the condensed matter physics community, but a definitive sighting has remained elusive. On page 602 of this issue, NadjPerge et al. (2) report considerable progress toward creating the Majorana state in the laboratory. There are a number of reasons why this seemingly esoteric problem of creating and confirming Majorana particles is of interest to condensed matter physicists. First, the quasi-particles in a superconductor are natural candidates for Majorana physics, because they constitute a quantum mechanical admixture of particles and holes. If the admixture has equal amplitude, the antiparticles and particles become identical. Second, it was pointed out by Kitaev (3) that Majorana bound states (MBSs) form at the ends of a superconductor chain if the wave function of the electron pair formed in the superconductor is antisymmetric (called p-wave), as opposed to symmetric (s-wave) in conventional superconductors. The MBS wave functions are localized at each end of the chain and are exactly at zero energy. They hybridize to form a conventional quasi-particle (complex fermion) with energy Eo ≈ exp(–L/ξ), where L is the length of the chain and ξ is the coherence length of the superconductor. In the limit L >> ξ, Eo → 0 and the addition or removal of the quasi-particle costs no energy; that is,
SCIENCE sciencemag.org
the MBSs become part of the ground-state manifold. More generally, the presence of 2N MBSs far apart implies the existence of N low-energy quasi-particles, each of which can either be empty or occupied, leading to a ground-state degeneracy of 2N. It is as if each complex fermion has been split into two real MBSs that are spatially far apart. Because the association of MBS pairs into quasi-particles is a matter of choice, the following phenomenon occurs. Suppose we have four MBSs (see the figure, panel A) and the MBS pairs (1, 2) and (3, 4) form quasi-particles, which are both occupied. If we exchange the positions of MBSs 2 and 3, the state will change to a different state, in this case a linear superposition of states where the two quasi-particles are both occupied or both empty. For conventional particles, the exchange of their positions can only lead to a sign change (for fermions) or in certain cases a change in the phase of the wave function (for so-called anyons). But now we end up with a different state entirely. The MBSs obey what is referred to as non-Abelian statistics—an exotic possibility that has not been observed experimentally. Apart from being a fascinating example of quantum weirdness, the ability to spatially decompose quasi-particle states has been proposed to be the basis of a fault-tolerant quantum computer and memory (3). Superconductors with p-wave pairing are rare in nature; forming a one-dimensional chain out of them presents a formidable task. Fortunately, it may be possible to create structures with the desired properties using conventional s-wave superconductors. Starting with the proposal of Fu and Kane (4) to couple the newly discovered surface state of topological insulators to conventional superconductors by the proximity effect, many schemes have been proposed to build structures using a variety of more or less conventional materials, and the race to create MBS in the laboratory is on. Up to now, the most convincing sighting of MBS used a scheme that places a semiconducting nanowire on top of a superconductor (5). A zero-bias peak was observed in the tunneling conductance under conditions consistent with theoreti31 OCTOBER 2014 • VOL 346 ISSUE 6209
Published by AAAS
545
INSIGHTS | P E R S P E C T I V E S
1
2
3
4
Exotica under the microscope. (A) Non-Abelian statistics: The exchange of Majorana bound states 2 and 3 leads to a new state, not just a change in the phase of the wave function. (B) Representation of the experiment of Nadj-Perge et al. (2) to search for the Majorana bound state. A chain of Fe atoms is fabricated on top of a Pb substrate and probed by a STM tip.
STM probe
Fe atom
B
Pb substrate
cal predictions. However, the energy scale of the spin-orbit coupling—a key parameter responsible for the formation of the MBS— is very small (~0.05 meV) in the semiconductor nanowire and raises the question of whether disorder or other conventional effects may give alternative explanations of the zero-bias peak (6, 7). Additional experiments are clearly desirable. Nadj-Perge et al. report the observation of Majorana fermions in a chain of iron (Fe) atoms on the surface of superconductive lead (Pb) (see the figure, panel B). Remarkably, the Fe chains grow out of a central island along atomic rows on the crystalline Pb surface. If the chain of Fe is ferromagnetic, theory predicts that the strong spinorbit coupling in Pb will lead to an effective p-wave component in the induced superconductivity in the Fe chain and hence a realization of the Kitaev model. Indeed, using a scanning tunneling microscope (STM), the group discovered zero-bias peaks at the end of the chains, but not in the middle. The wave function of the state is localized near the chain end to a surprising degree, so that the zero-bias peaks disappear over a distance of 10 Å—a signature consistent with MBSs. The Fe chain system has an advantage over the semiconductor nanowire: Because we are dealing with atomic-scale devices, the energy scales are several orders of magPhysics Department, Massachusetts Institute of Technology, Cambridge, MA 02139, USA. E-mail: [email protected]
546
nitude higher. For example, the spin-orbit coupling is estimated to be 100 meV (versus 0.05 meV), and the spin splitting due to exchange is 900 meV (versus a few meV) from the magnetic field in the nanowire system. Thus, there are reasons to believe that the MBSs are more robust. The experiment is clearly a tour de force combining the fabrication of structures on an atomic scale and the ability to probe the electronic structure with high energy and spatial resolution. Unlike the earlier experiment, the spatial location of the zero-bias state is clearly identified. However, there are still a number of loose ends to contend with. The energy gap of the induced superconductivity on the Fe chain is estimated to be very small: 0.2 meV, 20% of the Pb bulk gap. This corresponds to about 2 K, not much higher than the temperature where the experiment is carried out, 1.4 K. As a result, no clear gap structure is seen and the tunneling conductance shows only a modest reduction of no more than 50% at low energies. The zero-bias peak is a small structure on top of this large background. All the interesting phenomena associated with MBSs require the state to be well isolated from the quasi-particle excitations—a condition that is far from being satisfied in this experiment. A second problem is that the chain length is short, about 300 Å. On the other hand, the superconductivity coherence length is expected to be much longer because of the small energy gap. Thus, we are in the opposite limit, where
L > ξ, may be beyond the reach of the atomic chain system. On the other hand, there are other interesting phenomena, such as the crossed Andreev effect (8) and noise correlation between leads attached to opposite ends or a chain (9), that require a voltage and temperature smaller than the energy splitting. It will be exciting to see whether the next generation of experiments capable of reaching much lower temperature will reveal these and other features of Majorana bound states. Meanwhile, other schemes that may be more scalable to long lengths are being pursued. Notable among them is a return to the original Fu-Kane scheme using a new type of two-dimensional topological insulators fabricated out of semiconductor heterostructures (10, 11). The story of the search for Majorana particles is far from over. ■ REF ERENCES AND NOTES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
F. Wilczek, Nat. Phys. 5, 614 (2009). S. Nadj-Perge et al., Science 346, 602 (2014). A. Kitaev, Ann. Phys. 303, 2 (2003). L. Fu, C. L. Kane, Phys. Rev. Lett. 100, 096407 (2008). V. Mourik et al., Science 336, 1003 (2012). J. Liu, A. C. Potter, K. T. Law, P. A. Lee, Phys. Rev. Lett. 109, 267002 (2012). H. O. H. Churchill et al., Phys. Rev. B 87, 241401 (2013). J. Nilsson, A. R. Akhmerov, C. W. Beenakker, Phys. Rev. Lett. 101, 120403 (2008). J. Liu, F. C. Zhang, K. T. Law, Phys. Rev. B 88, 064509 (2013). L. Du, I. Knez, G. Sullivan, R. R. Du, http://arxiv.org/ abs/1306.1925 (2013). V. S. Pribiag et al., http://arxiv.org/abs/1408.1701 (2014).
ACKNOWL EDGMENTS
I thank K. T. Law for discussions. Supported by the John Templeton Foundation and U.S. Department of Energy grant DE-FG-02-03-ER46076.
10.1126/science.1260282
sciencemag.org SCIENCE
31 OCTOBER 2014 • VOL 346 ISSUE 6209
Published by AAAS
CREDIT: V. ALTOUNIAN/SCIENCE
A
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here.
The following resources related to this article are available online at www.sciencemag.org (this information is current as of November 5, 2014 ): Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/346/6209/545.full.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/346/6209/545.full.html#related This article cites 9 articles, 1 of which can be accessed free: http://www.sciencemag.org/content/346/6209/545.full.html#ref-list-1 This article appears in the following subject collections: Physics http://www.sciencemag.org/cgi/collection/physics
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2014 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on November 5, 2014
Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.
To obtain clear spectra of the conformers of H2CO3, Reisenauer et al. developed a very clean source: the pyrolysis of esters of H2CO3 (8). This method had been demonstrated to produce carboxylic acids and also suggested, but not proven experimentally, to produce H2CO3 (see the scheme) (11). The products of the pyrolysis were trapped in solid argon at extremely low temperature (8 K), and characterization by IR spectroscopy revealed that the more stable cis,cis and the cis,trans conformers were formed. The spectra perfectly matched the spectra obtained by sublimation of β-H2CO3 and trapping the products in solid argon, but markedly differed from those of α-H2CO3 These results suggested that β-H2CO3 indeed is solid H2CO3, whereas α-H2CO3 has a different constitution. Reisenauer et al. also showed that “α-H2CO3” is actually its monomethyl ester, CO(OH)(OCH3). By using acidic CH3OH for the protonation, the classic conditions for esterification of carboxylic acids, the monomethyl ester instead of the free carbonic acid is formed. O O
H
R⬘
O
O H
O
R⬘ O O
H O H
+
R⬘
H O + H
R
R⬘ Ester pyrolysis to produce carbonic acid, where R is an alkyl group.
This finding also explains why aqueous HCl and HCO3⫺ in the absence of CH3OH produced solid H2CO3 as expected, which was formerly identified as the β-polymorph. Thanks to the clean synthesis of H2CO3 via ester pyrolysis, the fog has lifted, and we can now look forward to a clearer exploration of the chemical landscape of this fascinating compound. ■ REFERENCES
1. B. Jonsson et al., J. Am. Chem. Soc. 99, 4628 (1977). 2. J. K. Terlouw, C. B. Lebrilla, H. Schwarz, Angew. Chem. Int. Ed. Engl. 26, 354 (1987). 3. M. H. Moore, R. K. Khanna, Spectrochim. Acta A Mol. Biomol. Spectrosc. 47, 255 (1991). 4. W. Hage, A. Hallbrucker, E. Mayer, J. Am. Chem. Soc. 115, 8427 (1993). 5. W. Hage, A. Hallbrucker, E. Mayer, J. Mol. Struct. 408-409, 527 (1997). 6. W. Hage, K. R. Liedl, A. Hallbrucker, E. Mayer, Science 279, 1332 (1998). 7. J. Bernard, R. G. Huber, K. R. Liedl, H. Grothe, T. Loerting, J. Am. Chem. Soc. 135, 7732 (2013). 8. H. P. Reisenauer, J. P. Wagner, P. R. Schreiner, Angew. Chem. Int. Ed. 53, 11766 (2014). 9. T. Loerting et al., Angew. Chem. Int. Ed. 39, 891 (2000). 10. K. Adamczyk, M. Prémont-Schwarz, D. Pines, E. Pines, E. T. J. Nibbering, Science 326, 1690 (2009). 11. G. Bucher, Eur. J. Org. Chem. 2010, 1070 (2010).
10.1126/science.1260117
PHYSICS
Seeking out Majorana under the microscope A chain of iron atoms on lead may reveal a signature of the elusive Majorana particle By Patrick A. Lee
T
he Dirac equation was initially developed to give a quantum mechanical description of particles such as electrons, but ended up predicting the existence of positrons—the antiparticle of the electron. A year before he disappeared under mysterious circumstances in 1938, the young Italian physicist Ettore Majorana discovered a solution to the Dirac equation that implied the existence of particles, or states of matter, that are their own antiparticles. This finding was contrary to Dirac’s solution, in which particles (electrons) and their antiparticles (positrons) are distinct. It has long been suspected, but not proven, that neutrinos are Majorana particles (1). In the past several years, the Majorana state has attracted the attention of the condensed matter physics community, but a definitive sighting has remained elusive. On page 602 of this issue, NadjPerge et al. (2) report considerable progress toward creating the Majorana state in the laboratory. There are a number of reasons why this seemingly esoteric problem of creating and confirming Majorana particles is of interest to condensed matter physicists. First, the quasi-particles in a superconductor are natural candidates for Majorana physics, because they constitute a quantum mechanical admixture of particles and holes. If the admixture has equal amplitude, the antiparticles and particles become identical. Second, it was pointed out by Kitaev (3) that Majorana bound states (MBSs) form at the ends of a superconductor chain if the wave function of the electron pair formed in the superconductor is antisymmetric (called p-wave), as opposed to symmetric (s-wave) in conventional superconductors. The MBS wave functions are localized at each end of the chain and are exactly at zero energy. They hybridize to form a conventional quasi-particle (complex fermion) with energy Eo ≈ exp(–L/ξ), where L is the length of the chain and ξ is the coherence length of the superconductor. In the limit L >> ξ, Eo → 0 and the addition or removal of the quasi-particle costs no energy; that is,
SCIENCE sciencemag.org
the MBSs become part of the ground-state manifold. More generally, the presence of 2N MBSs far apart implies the existence of N low-energy quasi-particles, each of which can either be empty or occupied, leading to a ground-state degeneracy of 2N. It is as if each complex fermion has been split into two real MBSs that are spatially far apart. Because the association of MBS pairs into quasi-particles is a matter of choice, the following phenomenon occurs. Suppose we have four MBSs (see the figure, panel A) and the MBS pairs (1, 2) and (3, 4) form quasi-particles, which are both occupied. If we exchange the positions of MBSs 2 and 3, the state will change to a different state, in this case a linear superposition of states where the two quasi-particles are both occupied or both empty. For conventional particles, the exchange of their positions can only lead to a sign change (for fermions) or in certain cases a change in the phase of the wave function (for so-called anyons). But now we end up with a different state entirely. The MBSs obey what is referred to as non-Abelian statistics—an exotic possibility that has not been observed experimentally. Apart from being a fascinating example of quantum weirdness, the ability to spatially decompose quasi-particle states has been proposed to be the basis of a fault-tolerant quantum computer and memory (3). Superconductors with p-wave pairing are rare in nature; forming a one-dimensional chain out of them presents a formidable task. Fortunately, it may be possible to create structures with the desired properties using conventional s-wave superconductors. Starting with the proposal of Fu and Kane (4) to couple the newly discovered surface state of topological insulators to conventional superconductors by the proximity effect, many schemes have been proposed to build structures using a variety of more or less conventional materials, and the race to create MBS in the laboratory is on. Up to now, the most convincing sighting of MBS used a scheme that places a semiconducting nanowire on top of a superconductor (5). A zero-bias peak was observed in the tunneling conductance under conditions consistent with theoreti31 OCTOBER 2014 • VOL 346 ISSUE 6209
Published by AAAS
545
INSIGHTS | P E R S P E C T I V E S
1
2
3
4
Exotica under the microscope. (A) Non-Abelian statistics: The exchange of Majorana bound states 2 and 3 leads to a new state, not just a change in the phase of the wave function. (B) Representation of the experiment of Nadj-Perge et al. (2) to search for the Majorana bound state. A chain of Fe atoms is fabricated on top of a Pb substrate and probed by a STM tip.
STM probe
Fe atom
B
Pb substrate
cal predictions. However, the energy scale of the spin-orbit coupling—a key parameter responsible for the formation of the MBS— is very small (~0.05 meV) in the semiconductor nanowire and raises the question of whether disorder or other conventional effects may give alternative explanations of the zero-bias peak (6, 7). Additional experiments are clearly desirable. Nadj-Perge et al. report the observation of Majorana fermions in a chain of iron (Fe) atoms on the surface of superconductive lead (Pb) (see the figure, panel B). Remarkably, the Fe chains grow out of a central island along atomic rows on the crystalline Pb surface. If the chain of Fe is ferromagnetic, theory predicts that the strong spinorbit coupling in Pb will lead to an effective p-wave component in the induced superconductivity in the Fe chain and hence a realization of the Kitaev model. Indeed, using a scanning tunneling microscope (STM), the group discovered zero-bias peaks at the end of the chains, but not in the middle. The wave function of the state is localized near the chain end to a surprising degree, so that the zero-bias peaks disappear over a distance of 10 Å—a signature consistent with MBSs. The Fe chain system has an advantage over the semiconductor nanowire: Because we are dealing with atomic-scale devices, the energy scales are several orders of magPhysics Department, Massachusetts Institute of Technology, Cambridge, MA 02139, USA. E-mail: [email protected]
546
nitude higher. For example, the spin-orbit coupling is estimated to be 100 meV (versus 0.05 meV), and the spin splitting due to exchange is 900 meV (versus a few meV) from the magnetic field in the nanowire system. Thus, there are reasons to believe that the MBSs are more robust. The experiment is clearly a tour de force combining the fabrication of structures on an atomic scale and the ability to probe the electronic structure with high energy and spatial resolution. Unlike the earlier experiment, the spatial location of the zero-bias state is clearly identified. However, there are still a number of loose ends to contend with. The energy gap of the induced superconductivity on the Fe chain is estimated to be very small: 0.2 meV, 20% of the Pb bulk gap. This corresponds to about 2 K, not much higher than the temperature where the experiment is carried out, 1.4 K. As a result, no clear gap structure is seen and the tunneling conductance shows only a modest reduction of no more than 50% at low energies. The zero-bias peak is a small structure on top of this large background. All the interesting phenomena associated with MBSs require the state to be well isolated from the quasi-particle excitations—a condition that is far from being satisfied in this experiment. A second problem is that the chain length is short, about 300 Å. On the other hand, the superconductivity coherence length is expected to be much longer because of the small energy gap. Thus, we are in the opposite limit, where
L > ξ, may be beyond the reach of the atomic chain system. On the other hand, there are other interesting phenomena, such as the crossed Andreev effect (8) and noise correlation between leads attached to opposite ends or a chain (9), that require a voltage and temperature smaller than the energy splitting. It will be exciting to see whether the next generation of experiments capable of reaching much lower temperature will reveal these and other features of Majorana bound states. Meanwhile, other schemes that may be more scalable to long lengths are being pursued. Notable among them is a return to the original Fu-Kane scheme using a new type of two-dimensional topological insulators fabricated out of semiconductor heterostructures (10, 11). The story of the search for Majorana particles is far from over. ■ REF ERENCES AND NOTES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
F. Wilczek, Nat. Phys. 5, 614 (2009). S. Nadj-Perge et al., Science 346, 602 (2014). A. Kitaev, Ann. Phys. 303, 2 (2003). L. Fu, C. L. Kane, Phys. Rev. Lett. 100, 096407 (2008). V. Mourik et al., Science 336, 1003 (2012). J. Liu, A. C. Potter, K. T. Law, P. A. Lee, Phys. Rev. Lett. 109, 267002 (2012). H. O. H. Churchill et al., Phys. Rev. B 87, 241401 (2013). J. Nilsson, A. R. Akhmerov, C. W. Beenakker, Phys. Rev. Lett. 101, 120403 (2008). J. Liu, F. C. Zhang, K. T. Law, Phys. Rev. B 88, 064509 (2013). L. Du, I. Knez, G. Sullivan, R. R. Du, http://arxiv.org/ abs/1306.1925 (2013). V. S. Pribiag et al., http://arxiv.org/abs/1408.1701 (2014).
ACKNOWL EDGMENTS
I thank K. T. Law for discussions. Supported by the John Templeton Foundation and U.S. Department of Energy grant DE-FG-02-03-ER46076.
10.1126/science.1260282
sciencemag.org SCIENCE
31 OCTOBER 2014 • VOL 346 ISSUE 6209
Published by AAAS
CREDIT: V. ALTOUNIAN/SCIENCE
A
