INSIGHTS | P E R S P E C T I V E S
B
C
On the surface. From minerals that react with molecules in the environment (A), to advanced electronic devices (B), to nanoscale heterogeneous catalysts (C), metal oxide surfaces with their sometimes unexpected structures (D) play a key role.
D
SURFACE SCIENCE
Stability at the surface The structure beneath a surface is a key factor in its electronic and chemical function By Scott A. Chambers
M
etal oxides are ubiquitous as minerals in the terrestrial environment, as well as in a variety of technologically important structures such as electronic devices and heterogeneous catalysts. Within these various contexts, interfaces between oxides and gases, liquids, and solids drive many critically important phenomena ranging from the uptake of contaminants in groundwater by redox-active minerals to the switching of the millions of transistors found in every cell phone and computer. Function is tied to structure. Therefore, fundamental understanding of the structure of oxide surfaces and interfaces is crucial to the comprehension of a plethora of phenomena involving this broad class of materials (see the figure). On page 1215 of this issue, Bliem et al. (1) report a remarkable structural insight into the surface of the (001)-oriented crystal face of magnetite (Fe3O4), the first magnetic material discovered in antiquity. Physical Sciences Division, Pacifc Northwest National Laboratory, Richland, WA 99352, USA. E-mail: sa.chambers@ pnnl.gov
1186
It is most convenient to think of solid surfaces as simple truncations of the bulk crystal structure. However, by combining experimental and theoretical methods, Bliem et al. make a convincing case for the absence of Fe atoms at certain lattice sites in the third layer below the surface where they are expected, and the presence of Fe at certain interstitial sites in the second layer below the surface where they are not expected. The chemical composition within the first three atomic planes reported here (Fe11O16) is intermediate between that of bulk Fe2O3 and Fe3O4, and is driven by the oxidative environment in which the surface is prepared. Bliem et al. used highly surface-sensitive electron diffraction and scanning probe microscopy, along with first-principles modeling, to establish the detailed atomic structure within the top ~0.4 nm. What they found was not at all expected based on the simple models of surface structure found throughout the literature. In conventional thinking, the surface is modeled as a termination of the bulk on some atomic plane, with lateral and/or transverse displacements within the top few layers. Such displacements generally have minimal impact on the electronic, mag-
netic, and chemical properties of the surface. As a result, limiting the descriptions of surfaces (and the interfaces these surfaces form when in contact with other materials) to such simple models has led to predictions of everything from chemical reactivity to electronic properties to magnetization that are often far from reality. The study by Bliem et al. reminds us that nature does not always generate the simple, easily conceptualized structures that we like to envision. Rather, nature does what must be done to achieve equilibrium, which is that minimum-energy configurational space in which materials typically exist and function. In the case of oxides, the implications are far-reaching. For example, the optimized deposition of Cr metal onto semiconducting SrTiO3(001) should lead to an atomically abrupt interface in which the Cr atoms crystallize into the body-centered cubic structure found in bulk Cr, but with a 45° rotation relative to the perovskite lattice of SrTiO3 to minimize the interfacial energy. This electrical junction should be rectifying. However, even though the two lattices line up as expected, the junction is perfectly ohmic, and the contact resistance is as low as any measured for any metal on SrTiO3 (2). The reason for this completely unexpected electrical behavior is an unanticipated interface structure in which a fraction of the first atomic layer of Cr atoms diffuses into the SrTiO3, occupies interstitial sites in the lattice, transfers charge to the adjacent Ti ions, and metallizes the interface (3). To what extent does this process occur because of Sr metal vacancies present in the subsurface region of the SrTiO3? We do not yet know. Or consider the observation of conductivity at the interface of two oxide insulators, SrTiO3 and LaAlO3 (4), which has been the subject of more than 300 scientific journal articles and conference proceedings over the past decade. The original explanation, which still has many adherents, involves the assumption of an abrupt or nearly abrupt interface of two perfectly stoichiometric oxides. The nonzero charges on alternating layers of (LaO)+ and (AlO2)– are thought to generate a diverging electric sciencemag.org SCIENCE
5 DECEMBER 2014 • VOL 346 ISSUE 6214
Published by AAAS
Downloaded from www.sciencemag.org on December 4, 2014
A
potential within the LaAlO3, which triggers what is essentially dielectric breakdown and charge transfer to the SrTiO3 at some critical thickness, thereby producing free carriers and conductivity. Yet, the expected potential drop across the LaAlO3 film has not been detected by the most sensitive and direct techniques (5–8). Moreover, the interface has been shown to be chemically diffuse rather than atomically sharp (9, 10), strikingly so at times (6), and certain cation rearrangements can eliminate the diverging potential that is thought to drive charge transfer and conductivity (6). To what extent does cation mixing depend on defects in the SrTiO3? Again, we do not yet know. Additionally, recent experiments reveal that defects related to nonstoichiometry in the LaAlO3 themselves play a crucial if not enabling role in activating conductivity, even though a comprehensive mechanistic understanding has not yet been developed (11, 12). The problem is that accurate detection and characterization of defects are hard work. Point defects, such as vacancies, are often detected indirectly by the structural perturbations they cause to surrounding atoms in the crystal lattice. If the defect structures do not exhibit long-range crystallographic order, as is often the case, their detection, and the theoretical treatment of their effect on structure and functional properties, is more difficult. However, with a combination of definitive experimental characterization and accurate theoretical modeling, like that presented by Bliem et al., defects can be detected, understood, and eventually harnessed for scientific and technological advantage. In the case of metal oxides, defects are an important component of the associated equilibrium structures that are typically present. The sooner we acknowledge that surfaces and interfaces are not necessarily simple terminations and junctions of the bulk material, the sooner we can come to a realistic understanding of these important material structures and engineer metal oxides to the fullest possible extent. ■
IILLUSTRATION: V. ALTOUNIAN/SCIENCE
REFERENCES
1. R. Bliem et al., Science 346, 1215 (2014). 2. C. Capan, G. Y. Sun, M. E. Bowden, S. A. Chambers, Appl. Phys. Lett. 100, 052106 (2012). 3. S. A. Chambers et al., Adv. Mater. 25, 4001 (2013). 4. A. Ohtomo, H. Y. Hwang, Nature 427, 423 (2004). 5. Y. Segal et al., Phys. Rev. B 80, 241107(R) (2009). 6. S. A. Chambers et al., Surf. Sci. Rep. 65, 317 (2010). 7. M. Takizawa, S. Tsuda, T. Susaki, H. Y. Hwang, A. Fujimori, Phys. Rev. B 84, 245124 (2011). 8. E. Slooten et al., Phys. Rev. B 87, 085128 (2013). 9. P. R. Willmott et al., Phys. Rev. Lett. 99, 155502 (2007). 10. A. S. Kalabukhov et al., Phys. Rev. Lett. 103, 146101 (2009). 11. M. P. Warusawithana et al., Nat. Commun. 4, 2351 (2013). 12. E. Breckenfeld et al., Phys. Rev. Lett. 110, 196804 (2013). 10.1126/science.aaa1543
HUMAN GENOMICS
Sleeping dogs of the genome Retrotransposable elements may be agents of somatic diversity, disease, and aging tive retrotransposon called long interspersed nuclear element–1 (LINE-1) Homo sapiens (L1Hs). The L1Hs-encoded proteins, includetrotransposons have a way of staying a reverse transcriptase, are essential for ing under the radar. These mobile the retrotransposition process and form a DNA elements are well-known agents ribonucleoprotein particle with the mRNA. of genome instability and evolution A second abundant human retrotranspobut are mostly thought of as oddities son, the Alu element, depends on L1Hs for occasionally associated with an interits movement as it does not encode any esting phenotype or worse, because of their proteins, and its transcripts are assembled high copy number, confound the analysis with L1Hs-encoded proteins. Both L1Hs of genome sequences. However, as much as and Alu ribonucleoprotein particles then two-thirds or more of the human genome is enter the nucleus and by a process called derived from repetitive sequences (1). Most target primed reverse transcription, insert of these are retrotransposon sequences that new DNA elements at quasi-random sites throughout the genome (see the second figure). Stable genome As the “selfish gene” theory L1 repressed predicts (2), for a transposable element to succeed in the evolutionary race, the critical battleground is the germ line. Defense mechanisms: There are numerous examples of insertions that affect human Stress, DNA SIRT6, small RNAs damage, Heterochromatin health (3). Because of this conAging telomere DNA methylation stant threat of genetic mayhem, shortening Antiviral defenses cells have an impressive armaAutophagy mentarium to combat mobile elements. Where does this battle Weakened defenses stand today? On one hand, the human retrotransposon load has L1 activated been reduced to perhaps as few as 100 active elements; yet, twothirds of the human genome is Genomic instability scarred by the evidence of milDeregulated transcription lions of years of warfare against Cancer, Aging mobile DNA elements, and new insertions occur at a frequency Aging guards invite a jailbreak. With aging, increased stress, DNA of 1 per 95 to 270 live births for damage, and telomere shortening weaken the multiple systems that keep L1Hs, and 1 in 20 for Alu (3). retrotransposons in check. Aged cells lose repressive heterochromatin, What happens outside the SIRT6 relocalizes away from L1 promoters, and autophagy becomes less germ line, in somatic tissues? efficient. Other defense pathways (see box) may also lose their effectiveHistorically, little attention has ness. The consequent unleashing of L1 elements could lead to profound been given to this question, besomatic damage, driving age-associated cell and tissue dysfunction. cause somatic retrotransposition is evolutionarily a dead-end prowere active in the distant evolutionary past cess. However, rampant genomic instability and are now present as fossils that litter our and mutagenesis are deleterious to all cells, genomes. But this image belies the damage and somatic cells mount the same spectrum that can be wreaked by evolutionarily recent 1 transposable elements. Department of Biology, University of Rochester, Rochester, NY 14627, USA. 2Institute for Systems Genetics, New York Retrotransposons can be copied into University Langone Medical Center, New York, NY 10016, mRNA and then back to DNA that can inteUSA. 3Department of Molecular Biology, Cell Biology and grate into the genome. The human genome Biochemistry, Brown University, Providence, RI 02912, USA. E-mail: [email protected] harbors ~100 copies of an autonomously acBy Vera Gorbunova,1 Jef D. Boeke,2 Stephen L. Helfand,3 John M. Sedivy3
R
SCIENCE sciencemag.org
5 DECEMBER 2014 • VOL 346 ISSUE 6214
Published by AAAS
1187
Stability at the surface Scott A. Chambers Science 346, 1186 (2014); DOI: 10.1126/science.aaa1543
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here. The following resources related to this article are available online at www.sciencemag.org (this information is current as of December 4, 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/346/6214/1186.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/346/6214/1186.full.html#related This article cites 12 articles, 1 of which can be accessed free: http://www.sciencemag.org/content/346/6214/1186.full.html#ref-list-1
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 December 4, 2014
This copy is for your personal, non-commercial use only.
B
C
On the surface. From minerals that react with molecules in the environment (A), to advanced electronic devices (B), to nanoscale heterogeneous catalysts (C), metal oxide surfaces with their sometimes unexpected structures (D) play a key role.
D
SURFACE SCIENCE
Stability at the surface The structure beneath a surface is a key factor in its electronic and chemical function By Scott A. Chambers
M
etal oxides are ubiquitous as minerals in the terrestrial environment, as well as in a variety of technologically important structures such as electronic devices and heterogeneous catalysts. Within these various contexts, interfaces between oxides and gases, liquids, and solids drive many critically important phenomena ranging from the uptake of contaminants in groundwater by redox-active minerals to the switching of the millions of transistors found in every cell phone and computer. Function is tied to structure. Therefore, fundamental understanding of the structure of oxide surfaces and interfaces is crucial to the comprehension of a plethora of phenomena involving this broad class of materials (see the figure). On page 1215 of this issue, Bliem et al. (1) report a remarkable structural insight into the surface of the (001)-oriented crystal face of magnetite (Fe3O4), the first magnetic material discovered in antiquity. Physical Sciences Division, Pacifc Northwest National Laboratory, Richland, WA 99352, USA. E-mail: sa.chambers@ pnnl.gov
1186
It is most convenient to think of solid surfaces as simple truncations of the bulk crystal structure. However, by combining experimental and theoretical methods, Bliem et al. make a convincing case for the absence of Fe atoms at certain lattice sites in the third layer below the surface where they are expected, and the presence of Fe at certain interstitial sites in the second layer below the surface where they are not expected. The chemical composition within the first three atomic planes reported here (Fe11O16) is intermediate between that of bulk Fe2O3 and Fe3O4, and is driven by the oxidative environment in which the surface is prepared. Bliem et al. used highly surface-sensitive electron diffraction and scanning probe microscopy, along with first-principles modeling, to establish the detailed atomic structure within the top ~0.4 nm. What they found was not at all expected based on the simple models of surface structure found throughout the literature. In conventional thinking, the surface is modeled as a termination of the bulk on some atomic plane, with lateral and/or transverse displacements within the top few layers. Such displacements generally have minimal impact on the electronic, mag-
netic, and chemical properties of the surface. As a result, limiting the descriptions of surfaces (and the interfaces these surfaces form when in contact with other materials) to such simple models has led to predictions of everything from chemical reactivity to electronic properties to magnetization that are often far from reality. The study by Bliem et al. reminds us that nature does not always generate the simple, easily conceptualized structures that we like to envision. Rather, nature does what must be done to achieve equilibrium, which is that minimum-energy configurational space in which materials typically exist and function. In the case of oxides, the implications are far-reaching. For example, the optimized deposition of Cr metal onto semiconducting SrTiO3(001) should lead to an atomically abrupt interface in which the Cr atoms crystallize into the body-centered cubic structure found in bulk Cr, but with a 45° rotation relative to the perovskite lattice of SrTiO3 to minimize the interfacial energy. This electrical junction should be rectifying. However, even though the two lattices line up as expected, the junction is perfectly ohmic, and the contact resistance is as low as any measured for any metal on SrTiO3 (2). The reason for this completely unexpected electrical behavior is an unanticipated interface structure in which a fraction of the first atomic layer of Cr atoms diffuses into the SrTiO3, occupies interstitial sites in the lattice, transfers charge to the adjacent Ti ions, and metallizes the interface (3). To what extent does this process occur because of Sr metal vacancies present in the subsurface region of the SrTiO3? We do not yet know. Or consider the observation of conductivity at the interface of two oxide insulators, SrTiO3 and LaAlO3 (4), which has been the subject of more than 300 scientific journal articles and conference proceedings over the past decade. The original explanation, which still has many adherents, involves the assumption of an abrupt or nearly abrupt interface of two perfectly stoichiometric oxides. The nonzero charges on alternating layers of (LaO)+ and (AlO2)– are thought to generate a diverging electric sciencemag.org SCIENCE
5 DECEMBER 2014 • VOL 346 ISSUE 6214
Published by AAAS
Downloaded from www.sciencemag.org on December 4, 2014
A
potential within the LaAlO3, which triggers what is essentially dielectric breakdown and charge transfer to the SrTiO3 at some critical thickness, thereby producing free carriers and conductivity. Yet, the expected potential drop across the LaAlO3 film has not been detected by the most sensitive and direct techniques (5–8). Moreover, the interface has been shown to be chemically diffuse rather than atomically sharp (9, 10), strikingly so at times (6), and certain cation rearrangements can eliminate the diverging potential that is thought to drive charge transfer and conductivity (6). To what extent does cation mixing depend on defects in the SrTiO3? Again, we do not yet know. Additionally, recent experiments reveal that defects related to nonstoichiometry in the LaAlO3 themselves play a crucial if not enabling role in activating conductivity, even though a comprehensive mechanistic understanding has not yet been developed (11, 12). The problem is that accurate detection and characterization of defects are hard work. Point defects, such as vacancies, are often detected indirectly by the structural perturbations they cause to surrounding atoms in the crystal lattice. If the defect structures do not exhibit long-range crystallographic order, as is often the case, their detection, and the theoretical treatment of their effect on structure and functional properties, is more difficult. However, with a combination of definitive experimental characterization and accurate theoretical modeling, like that presented by Bliem et al., defects can be detected, understood, and eventually harnessed for scientific and technological advantage. In the case of metal oxides, defects are an important component of the associated equilibrium structures that are typically present. The sooner we acknowledge that surfaces and interfaces are not necessarily simple terminations and junctions of the bulk material, the sooner we can come to a realistic understanding of these important material structures and engineer metal oxides to the fullest possible extent. ■
IILLUSTRATION: V. ALTOUNIAN/SCIENCE
REFERENCES
1. R. Bliem et al., Science 346, 1215 (2014). 2. C. Capan, G. Y. Sun, M. E. Bowden, S. A. Chambers, Appl. Phys. Lett. 100, 052106 (2012). 3. S. A. Chambers et al., Adv. Mater. 25, 4001 (2013). 4. A. Ohtomo, H. Y. Hwang, Nature 427, 423 (2004). 5. Y. Segal et al., Phys. Rev. B 80, 241107(R) (2009). 6. S. A. Chambers et al., Surf. Sci. Rep. 65, 317 (2010). 7. M. Takizawa, S. Tsuda, T. Susaki, H. Y. Hwang, A. Fujimori, Phys. Rev. B 84, 245124 (2011). 8. E. Slooten et al., Phys. Rev. B 87, 085128 (2013). 9. P. R. Willmott et al., Phys. Rev. Lett. 99, 155502 (2007). 10. A. S. Kalabukhov et al., Phys. Rev. Lett. 103, 146101 (2009). 11. M. P. Warusawithana et al., Nat. Commun. 4, 2351 (2013). 12. E. Breckenfeld et al., Phys. Rev. Lett. 110, 196804 (2013). 10.1126/science.aaa1543
HUMAN GENOMICS
Sleeping dogs of the genome Retrotransposable elements may be agents of somatic diversity, disease, and aging tive retrotransposon called long interspersed nuclear element–1 (LINE-1) Homo sapiens (L1Hs). The L1Hs-encoded proteins, includetrotransposons have a way of staying a reverse transcriptase, are essential for ing under the radar. These mobile the retrotransposition process and form a DNA elements are well-known agents ribonucleoprotein particle with the mRNA. of genome instability and evolution A second abundant human retrotranspobut are mostly thought of as oddities son, the Alu element, depends on L1Hs for occasionally associated with an interits movement as it does not encode any esting phenotype or worse, because of their proteins, and its transcripts are assembled high copy number, confound the analysis with L1Hs-encoded proteins. Both L1Hs of genome sequences. However, as much as and Alu ribonucleoprotein particles then two-thirds or more of the human genome is enter the nucleus and by a process called derived from repetitive sequences (1). Most target primed reverse transcription, insert of these are retrotransposon sequences that new DNA elements at quasi-random sites throughout the genome (see the second figure). Stable genome As the “selfish gene” theory L1 repressed predicts (2), for a transposable element to succeed in the evolutionary race, the critical battleground is the germ line. Defense mechanisms: There are numerous examples of insertions that affect human Stress, DNA SIRT6, small RNAs damage, Heterochromatin health (3). Because of this conAging telomere DNA methylation stant threat of genetic mayhem, shortening Antiviral defenses cells have an impressive armaAutophagy mentarium to combat mobile elements. Where does this battle Weakened defenses stand today? On one hand, the human retrotransposon load has L1 activated been reduced to perhaps as few as 100 active elements; yet, twothirds of the human genome is Genomic instability scarred by the evidence of milDeregulated transcription lions of years of warfare against Cancer, Aging mobile DNA elements, and new insertions occur at a frequency Aging guards invite a jailbreak. With aging, increased stress, DNA of 1 per 95 to 270 live births for damage, and telomere shortening weaken the multiple systems that keep L1Hs, and 1 in 20 for Alu (3). retrotransposons in check. Aged cells lose repressive heterochromatin, What happens outside the SIRT6 relocalizes away from L1 promoters, and autophagy becomes less germ line, in somatic tissues? efficient. Other defense pathways (see box) may also lose their effectiveHistorically, little attention has ness. The consequent unleashing of L1 elements could lead to profound been given to this question, besomatic damage, driving age-associated cell and tissue dysfunction. cause somatic retrotransposition is evolutionarily a dead-end prowere active in the distant evolutionary past cess. However, rampant genomic instability and are now present as fossils that litter our and mutagenesis are deleterious to all cells, genomes. But this image belies the damage and somatic cells mount the same spectrum that can be wreaked by evolutionarily recent 1 transposable elements. Department of Biology, University of Rochester, Rochester, NY 14627, USA. 2Institute for Systems Genetics, New York Retrotransposons can be copied into University Langone Medical Center, New York, NY 10016, mRNA and then back to DNA that can inteUSA. 3Department of Molecular Biology, Cell Biology and grate into the genome. The human genome Biochemistry, Brown University, Providence, RI 02912, USA. E-mail: [email protected] harbors ~100 copies of an autonomously acBy Vera Gorbunova,1 Jef D. Boeke,2 Stephen L. Helfand,3 John M. Sedivy3
R
SCIENCE sciencemag.org
5 DECEMBER 2014 • VOL 346 ISSUE 6214
Published by AAAS
1187
Stability at the surface Scott A. Chambers Science 346, 1186 (2014); DOI: 10.1126/science.aaa1543
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here. The following resources related to this article are available online at www.sciencemag.org (this information is current as of December 4, 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/346/6214/1186.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/346/6214/1186.full.html#related This article cites 12 articles, 1 of which can be accessed free: http://www.sciencemag.org/content/346/6214/1186.full.html#ref-list-1
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 December 4, 2014
This copy is for your personal, non-commercial use only.
