He Y.,Zhejiang Normal University |
Zhou W.,Center for Neutron Research |
Zhou W.,University of Maryland University College |
Qian G.,Zhejiang University |
Chen B.,University of Texas at San Antonio
Chemical Society Reviews | Year: 2014
Natural gas (NG), whose main component is methane, is an attractive fuel for vehicular applications. Realization of safe, cheap and convenient means and materials for high-capacity methane storage can significantly facilitate the implementation of natural gas fuelled vehicles. The physisorption based process involving porous materials offers an efficient storage methodology and the emerging porous metal-organic frameworks have been explored as potential candidates because of their extraordinarily high porosities, tunable pore/cage sizes and easily immobilized functional sites. In this view, we provide an overview of the current status of metal-organic frameworks for methane storage. This journal is © the Partner Organisations 2014.
News Article | March 16, 2016
Behaviors in complex systems are often more than the sum of their parts, whether in the brain where neurons enable expression of personality, or in fast data storage devices enabled by magnetic interactions. Physicists call these distinctive collective behaviors “emergent” properties. MIT assistant professor of physics Joseph G. Checkelsky is working at the intersection of materials synthesis and quantum physics to discover new materials that host these emergent phenomena, which in turn may foster new technologies. “In the context of solid state physics, emergent behavior often takes the form of collective behavior of electrons,” Checkelsky explains. “For example, looking at an individual electron, one would not imagine that it could form a coherent superconducting condensate by partnering with other electrons in a solid. Nor would one expect that different crystalline solids could give rise to the dynamic family of superconductors we know from experiment. Our goal is to look at materials where the quantum mechanical nature of the underlying electrons is brought out in new types of such collective properties.” Checkelsky recently won a National Science Foundation (NSF) CAREER award to pursue research into particular kinds of quantum materials that combine so-called frustrated atomic lattices and mobile electrons. An ordinary magnet gets its magnetism from the quantum mechanical property that electrons are indistinguishable and may lower their energy by aligning their magnetic moments. In some magnetic materials, known as antiferromagnets, magnetic moments of ions alternate in a repeating up/down pattern to lower their energy. The notion of a frustrated lattice, Checkelsky explains, is to create a repeating structure in the arrangement of atoms that confuses this ordering with competing interactions so that the system is unable to find a suitable way to orient itself. It is expected that this will lead to electrons with stronger quantum mechanical interactions that promote collective behavior. “By using these atomic arrangements one can force the electrons to have to deal with each other,” Checkelsky says. In Checkelsky’s lab, graduate student Linda Ye and postdoc Takehito Suzuki are studying an example of this with various arrangements of iron and tin atoms on the Kagome lattice structure, a series of overlapping triangles and hexagons. This work was inspired by theoretical work from 2011 by Cecil and Ida Green Professor of Physics Xiao-Gang Wen, Checkelsky notes. Many frustrated systems are electrically insulating. Whereas the study of those types of frustrated materials is established (including seminal contributions from MIT), in these new experiments the researchers are adding free electrons to these systems and studying how they behave. Ye explains that theorists proposed that ferromagnetic electrons on this Kagome lattice with strong spin-orbit coupling might give rise to topological bands with fractional excitations. “Generally this is a combination of strong spin-orbit interaction, magnetism, and the lattice degree of freedom,” she says. “It turns out to be a very interesting electronic system.” It differs from much of the low-temperature work in the lab because it displays interesting properties at room temperature. “It has strong ferromagnetism even at room temperature,” she says. The crystal lattice, in a ratio of three iron to two tin atoms, is made with the chemical vapor transport method. The obtained single crystals are hexagonally shaped plates with dimensions of about 1 millimeter (about four one-hundreths of an inch), Ye says. “There is one trick about growing this crystal. It’s not a stable phase at room temperature,” she says. Her approach is to cool down the crystals rapidly from high temperature in cold water so the lattice doesn’t change its phase. It can take up to a full month in the furnace to grow these single-crystal compounds. Ye is studying the mechanism underlying a phenomenon known as the anomalous Hall effect by measuring voltage made in the direction transverse (perpendicular) to a current flowing through a crystal of this material. Ye presented a report on the work today an American Physical Society meeting in Baltimore. Ye, 26, grew up in the city of Chengdu in Sichuan, China, and earned her bachelor’s degree at Tsinghua University in Beijing. She has a master’s in engineering (applied physics) from the University of Tokyo. Checkelsky was named a Moore Foundation Fellow in Materials Synthesis in August 2013. The award helped him to establish a lab that combines the growth of new materials with analyzing their properties and creating new devices that exploit their unique behavior. “The Moore Foundation has allowed our laboratory the capabilities to both synthesize and study materials. This ranges from thin film growth by molecular beam epitaxy to cryogenic measurements. Because of their support at the inception of our lab, we have been able to make fast progress towards realizing exotic quantum materials and studying their properties,” Checkelsky explains. The lab features a synthesis side with furnaces for growing materials and an analysis side with equipment such as cryogenic refrigerators and superconducting magnets for testing them. “It’s one room, but it has two distinct halves,” he says. “Every day the students are walking across the line between the two sides — we are trying to make this border disappear.” The group includes postdoc Takehito Suzuki, graduate students Linda Ye and Aravind Devarakonda and undergraduate Christina Wicker. A second postdoc is expected to join the group this summer, and Checkelsky plans to add more graduate and undergraduate students. “It’s very exciting to have seen all the big equipment come one by one, and put [it] together,” Ye says. “Many of the things in the lab we design and make by ourselves.” She is making a high vacuum probe for the thermal measurements inside the cryogenic refrigerator. Accurate measurements require using modern tools of nanoscience to probe the electronic properties of materials, Checkelsky says. “One type of life cycle for a project begins with an idea for physics of interest that we have in the office or learn through discussions with our theory colleagues. We then consider what materials systems are most likely to support such behavior and try to synthesize them. With the characterized material in hand we then design ways to incisively probe the physics of interest, which can involve high magnetic field, scattering, or making and measuring nanodevices from the material.” Checkelsky adds, “Accidental discoveries more interesting than our original targets also happen in the lab — which we are more than happy to embrace.” Besides equipment in his own lab, Checkelsky and his research team use facilities on the MIT campus at the Microsystems Technology Lab (MTL), as well as tools at the Harvard Center for Nanoscale Systems, the NIST Center for Neutron Research, and the National High Magnetic Field Laboratory in Tallahassee. The Florida lab features the world’s largest static magnet. “When we finally have our best materials, we bring them down to the magnet lab. They have been very supportive of our projects,” he adds. Checkelsky says he is looking forward to completion of MIT.nano, where he hopes to make use of the state-of-the-art facilities. “Then we will be able to go from thinking of projects to combining the powders to making the crystals to walking them next door to MIT.nano for the next step. It will be a really key facility for our projects,” he says. Checkelsky also is part of the Center for Integrated Quantum Materials. He works closely with MIT experimentalists Nuh Gedik, Ray Ashoori, and Pablo Jarillo-Herrero and theorists Patrick Lee, Liang Fu, Senthil Todadri, Leonid Levitov, and Xiao-Gang Wen. Graduate student Aravind Devarakonda is studying magnetic behavior of heavy metal compounds where static magnetism provides a source of correlation for the electrons. “The basic idea is that certain spinel compounds have a fixed level of magnetic order which can persist even with changing composition of other lattice constituents. By going from lighter to heavier elements surrounding these magnetic atoms, we can introduce topological electronic features into the material,” Devarakonda says. “If we want to combine electronic correlation with the physics of topological insulators and related materials we have to make progress on several fronts,” Checkelsky explains. “In addition to following theoretical predictions, an experimental voyage into material systems is a key component of this. And indeed thus far the work has been experimentally driven. In this sense the theme has been to look at systems where the presence of non-trivial electronic topology is not well-resolved from theory due to the complications of electronic correlation but which we can grow and examine experimentally.” Devarakonda, who earned his bachelor’s in applied science (physics) at Rutgers University, has become expert in the techniques of crystal growth, cryogenics, electrical measurements, nanofabrication, metal work, and first principles calculations — “the full tool belt,” Checkelsky says. Checkelsky received his BS in physics at Harvey Mudd College and PhD in physics at Princeton University in 2010. Before joining the faculty at MIT in January 2014, Checkelsky was a postdoc at Japan’s Institute for Physical and Chemical Research (RIKEN) and a lecturer at the University of Tokyo. His work and continued collaboration there includes studies of the surface properties of topological insulators and more recently unusual electronic properties of a compound material with layers of three elements — bismuth, tellurium, and iodine. “The atoms line up in the same stacking sequence through the whole material,” Checkelsky explains. “So, the system breaks a spatial symmetry and has a direction associated with it.” “This type of compound is called polar because of this directionality. That structural difference has profound implications for the electronic behavior,” he says. In particular this polarity changes the spin degree of freedom of the electron so that the way electrons move in the solid very strictly depends on the spin direction. Experiments published in a 2015 paper demonstrated a spin up channel and a spin down channel with different energy levels in measurement of their electrical resistance under a magnetic field. This work could have implications for future spintronic devices. Checkelsky is in his third term of teaching 8.02 (Electromagnetism) in the Department of Physics. With a typical class size of 100 students, 8.02 is taught in a Technology Enabled Active Learning (TEAL) classroom with students around the teacher. “It gives you not only face-to-face communication with the students, but a technological link as well. They have a clicker with which they can respond to questions, which can offer a view into what they’re thinking,” Checkelsky explains. “I had the good fortune of starting my teaching in 8.02 with Peter Dourmashkin, who is a constant source of insight into how to approach the class. I really enjoy working with the students and trying to convey not only my enthusiasm for physics but also practical, effective ways for them to tackle the course material.” Checkelsky’s lab has a partnership with the Wilson Creek Elementary School located in Johns Creek, Georgia, near his hometown north of Atlanta. With the help of their teachers, students there submit questions about science, and Checkelsky’s team comes up with the best answers they can. Analysis of student submissions showed that the most frequently occurring word in the students’ questions is “possible.” “They’re really at a very young age interested in what is possible for human beings and technology,” Checkelsky says. “We approach their questions as seriously as any kind of scientific question that comes up in our research. We have to read papers and discuss with colleagues; we really try to come up with a complete answer for students.” In addition to the support of the teachers and Principal Andrea Cushing, it is no doubt a great benefit that Checkelsky’s mom, Eileen Checkelsky, is the current principal’s secretary at Wilson Creek Elementary School. Checkelsky’s office windows are lined with towering plants inherited from Tom Greytak, retired professor of physics, who previously had the office. “At first I had to figure out what the names of the plants were. But taking care of them has become a serious hobby of mine.” This dovetails with his interest in cooking. “I think it’s curious that what I do at MIT is synthesizing and studying inorganic materials and my hobbies are basically synthesizing and processing organic ones. It all revolves around finding the right recipe,” he says.
News Article | September 21, 2016
Thin films of (LuFeO ) /(LuFe O ) were grown by reactive-oxide molecular-beam epitaxy in a Veeco GEN10 system on (111) (ZrO ) (Y O ) (or 9.5 mol% yttria-stabilized zirconia) substrates, denoted YSZ, held at a temperature of about 700 °C. This substrate temperature was determined by an optical pyrometer focused on an opaque, 200-nm-thick, platinum layer on the backside of the YSZ substrates used to absorb the heat from the SiC radiative substrate heater. Lutetium and iron were evaporated from elemental sources each at a flux of approximately 1 × 1013 atoms per square centimetre per second. The fluxes were first calibrated (approximately) with a quartz crystal microbalance. The lutetium flux was then more accurately refined using the reflection high-energy electron diffraction (RHEED) intensity oscillations that occurred when (111) Lu O was deposited on (111) YSZ. With the known lutetium flux, the iron flux was adjusted to produce a stoichiometric LuFeO film as judged by the out-of-plane lattice parameter measured by XRD33. The background oxygen partial pressure was meticulously varied during the superlattice growth to provide an environment sufficiently oxidizing for the Fe3+ in the LuFeO without over-oxidizing the Fe2.5+ in the LuFe O layers. The background partial pressure of a mixture of about 2% O and O was varied between 3 × 10−7 Torr and 9 × 10−7 Torr using a capacitance manometer to control the inlet pressure of a piezoelectric leak valve. Each growth sequence commenced with at least five monolayers of the LuFeO structure before the first LuFe O layer to avoid over-oxidation of the LuFe O by oxygen supplied by the YSZ substrate. (LuFeO ) /(LuFe O ) superlattices with thicknesses ranging from 28 nm to 35 nm were studied. The structure was characterized by XRD using a four-circle Rigaku SmartLab diffractometer equipped with a Ge(220) × 2 monochromator on the incident side and a Ge(220) × 2 analyser on the diffracted side, with Cu K radiation. XRD θ–2θ scans of all samples presented as well as a representative rocking curve are shown in Extended Data Fig. 1. Cross-sectional TEM specimens were prepared using an FEI Strata 400 Focused Ion Beam (FIB) with a final milling step of 2 keV to reduce surface damage. High-resolution HAADF-STEM images were acquired on a 100-keV Nion UltraSTEM, a fifth-order aberration-corrected microscope. The images in Fig. 1b were acquired from a single sample containing the representative layering patterns in this study. The lutetium distortions were quantified from HAADF-STEM images. Several images were averaged to reduce scan noise and the lutetium atomic positions were determined through an iterative two-dimensional Gaussian fitting procedure and segmentation using a watershed algorithm. The position of each atom was then compared to its neighbouring atom on each side in the same atomic plane. The magnitude and direction of the distortion was calculated by comparing the height difference of the three atoms as shown in Extended Data Fig. 6a. The schematics show different patterns of three-atom sets corresponding to the marked polarization. This distortion directly correlates with the ferroelectric polarization as shown in Extended Data Fig. 2. The resulting polarization at different locations in the structure is plotted in Extended Data Fig. 5 and in the aggregate in Extended Data Fig. 6b, c. In sum, the distortions from over 90,000 lutetium atoms were analysed to generate the results shown in Fig. 2e. The temperature dependence of the lutetium distortions was measured using a Protochips Aduro double-tilt heating holder in a 200-keV FEI Tecnai F20-ST microscope as shown in Extended Data Fig. 6d. A cross-sectional TEM specimen was prepared by FIB as above and then mounted onto a ceramic heating chip. The temperature was controlled on the chip by passing a current, which was calibrated to be within 50 K of the intended temperature. HAADF-STEM images were recorded at intervals as the sample was heated from 290 K to 675 K. The ferroelectric distortions were estimated from the strength of the superlattice reflections observed in the fast Fourier transform of the image corresponding to the tripling of the unit cell from the lutetium distortions. We label regions where ferroelectric distortions were observed, where weak distortions were observed and where the distortions were not observed (although they may be present, they are below our resolution limit). Magnetic measurements were performed with a Quantum Design MPMS3 SQUID magnetometer. M–T curves with the sample field-cooled in 1 kOe and zero-field-cooled were measured for each sample, along with a bare YSZ substrate that had undergone identical deposition conditions as a superlattice (same oxygen partial pressure and time at growth temperature, but not exposed to the lutetium or iron molecular beams). The M–T curves of the YSZ substrate were used to subtract the diamagnetic and paramagnetic backgrounds from the (LuFeO ) /(LuFe O ) M–T curves. The curves in Fig. 2a were normalized to facilitate comparison of the Curie temperatures (T ) at which the superlattices order ferromagnetically (or ferrimagnetically). The ferromagnetic (or ferrimagnetic) Curie temperature was taken to be the intersection between a line extending from the point of steepest descent of the field-cooled curve and a line extending from the high-temperature background. M–H loops (where H is the magnetic field) out to 70 kOe were measured for each sample at 50 K, 100 K, 200 K and 300 K. The linear diamagnetic background for each of these loops was subtracted with a linear fit to the high-field data. At higher temperature the loops saturate, but below 130 K the loops did not saturate, and an additional M–H loop was measured with the sample field-cooled in 70 KOe to measure the saturated magnetic moment at these temperatures. The known bulk magnetic moment of LuFeO (ref. 11) was subtracted from the total measured magnetic moment to obtain the LuFe O magnetic moments shown in Fig. 2d. Extended Data Fig. 3c displays ‘excess magnetization’ as a function of composition at 50 K, 100 K and 200 K. This value was found by subtracting the magnetization due to the end-member LuFeO and LuFe O signals from the total moment observed. Neutron diffraction was performed on the BT-4 triple-axis spectrometer at the NIST Center for Neutron Research using pyrolytic graphite monochromator and analyser crystals with neutrons of initial and final energies of 14.7 meV (λ = 2.359 Å). Relaxed horizontal collimation settings of the neutron optics in stream order were open–monochromator–40′–sample–40′–analyser–open–detector to maximize the intensity of the very weak scattered signal. The sample was mounted onto a silicon wafer using a small amount of aluminium foil and sealed in an aluminium canister with 4He gas to ensure thermal equilibration with the closed-cycle refrigerator. Neutron diffraction was used to determine the onset of long-range magnetic order of a (LuFeO ) /(LuFe O ) superlattice. Neutron scattering is not sensitive to small impurity phases that may influence other bulk characterization techniques. Magnetic reflections of the (LuFeO ) /(LuFe O ) superlattice were observed in neutron diffraction by scanning along the [10L] reciprocal lattice direction, where the Miller index L is along the c axis or growth direction, at several temperatures between 5 K and 325 K. We found a single 101 magnetic reflection (indexed to the magnetic unit cell of 10.4 nm, or two superlattice repeat distances) at 5 K that was not present at room temperature. This peak is observed to show considerable change in intensity between 5 K and room temperature as shown in Extended Data Fig. 4a. The centre of the peak is offset slightly from the 101 reflection owing to a slight misalignment of the sample. A Gaussian function is fit to the signal to determine the integrated intensity as a function of temperature. Disregarding finite-size effects of the films along the growth direction, the full width of the resolution function is shown as the horizontal line (black), demonstrating that the peak is not resolution-limited. The magnetic correlation length from this is found to be ξ ≈ 30 nm, or roughly six superlattice units along the c direction. The onset of magnetic order was determined by fitting the temperature-dependent integrated intensity of the peak with a mean-field order parameter. Shown in Extended Data Fig. 4b, we find T = 238 ± 28 K. This is consistent with the onset of ferromagnetism (or ferrimagnetism) determined from magnetometry of a (LuFeO ) /(LuFe O ) superlattice (246 K). To understand the origin of this 101 reflection, we calculated the magnetic structure factors for the magnetic moment arrangement obtained from DFT for the COII and Fe3+-doped configurations of the (LuFeO ) /(LuFe O ) superlattice. These calculations show that 10L-type reflections are particularly sensitive to the onset of ferrimagnetic order along the c axis in the LuFe O layers in the charge-ordered state. X-ray absorption spectroscopy on the Fe L edge was performed at Beamline 4.0.2 at the Advanced Light Source at Lawrence Berkeley National Laboratory. Spectra were acquired with 100% linearly polarized light oriented nearly parallel to and perpendicular to the c axis of the sample for an angle of X-ray incidence of 20° to the sample surface. The normalized difference between these spectra was collected for temperatures ranging from 300 K to 700 K for (LuFeO ) /(LuFe O ) , (LuFeO ) /(LuFe O ) , (LuFeO ) /(LuFe O ) and (LuFeO ) /(LuFe O ) . All temperatures measured were above the magnetic transition temperatures of the corresponding films and so the dichroic signal records an asymmetry in the electronic structure, as previously recorded for ferroelectric thin films31. As shown in Extended Data Fig. 7b, the (LuFeO ) /(LuFe O ) sample displays dichroism of about 20% at 300 K due to the structural anisotropy between the in-plane and out-of-plane electronic configurations in the hexagonal structure. A smaller anisotropy was also previously recorded for LuFe O (ref. 34). Previous calibration of the endstation has identified a systematic error of approximately ±5% in the measured dichroism due to the variation of the degree of X-ray polarization, the angle of X-ray incidence and photon energy with time. To further determine the uncertainty associated with these measurements, we measured the dichroic signal from the (LuFeO ) /(LuFe O ) sample multiple times before sample heating and again after the sample heating/irradiation. From comparison of the dichroic signal, we find a standard error of the mean of approximately 0.016 (4% in un-normalized units), within the uncertainty estimated from the previous calibration. The (LuFeO ) /(LuFe O ) , (LuFeO ) /(LuFe O ) and (LuFeO ) /(LuFe O ) samples, which were identified as displaying ferroelectric lutetium distortions by HAADF-STEM, demonstrated a dichroism of about 40% at 300 K. The temperature-dependent fits of the recorded difference spectra as a linear combination of the structural (LuFeO ) /(LuFe O ) component and (LuFeO ) /(LuFe O ) component at 300 K are shown in Fig. 2f. There is a drop in the recorded signal for the (LuFeO ) /(LuFe O ) and (LuFeO ) /(LuFe O ) samples indicative of a ferroelectric transition. The XLD signal does not decrease fully to the signal observed in (LuFeO ) /(LuFe O ) , suggesting that there may be an additional structural component in the (LuFeO ) /(LuFe O ) films that is not captured by the (LuFeO ) /(LuFe O ) component. The data shown in Fig. 2f were fitted to an order parameter plus a constant background using a Bayesian fitting algorithm. From this, we identify a transition T at about 550 K and 500 K in the (LuFeO ) /(LuFe O ) and (LuFeO ) /(LuFe O ) sample, respectively. Assuming the same constant background for the (LuFeO ) /(LuFe O ) sample, we estimate a transition near 900 K, beyond the measurement range of this experiment. The variable-temperature HAADF-STEM images (Extended Data Fig. 6d) recorded from these layering patterns are also consistent with transitions at comparable temperatures, although slightly different transitions between the m = 3 and m = 5 superlattices are found. The high-temperature transition in the (LuFeO ) /(LuFe O ) superlattice is consistent with the observation of superlattice reflections in RHEED corresponding to the trimer distortion24 during thin-film deposition at the growth temperature for this film. Vertical (or out-of-plane) PFM was used to measure the ferroelectric polarizability of the (LuFeO ) /(LuFe O ) superlattice shown in Figs 2g and 4a35–37. A bottom electrode was added to the sample after thin-film deposition. The substrate was first thinned using tripod mechanical polishing at a 1° angle; the measured region had a substrate thickness of approximately 150 nm. A copper adhesion layer and subsequent platinum layer were deposited onto the thinned substrate by direct-current magnetron sputtering in an argon background of 3 × 10−3 Torr at room temperature. This provided a conducting back electrode for the PFM experiments without disturbing the epitaxy of the (LuFeO ) /(LuFe O ) film on the insulating YSZ substrate. In Fig. 2g, a 20 μm × 20 μm region was poled with a d.c. bias of −15 V to switch the sample to the ‘upward’-oriented out-of-plane polarization. A second switch with a d.c. bias of +15 V was then performed to write the pattern shown with the corresponding ‘downward’-oriented polarization. The resulting structure was imaged directly after poling and again 100 h later using a 2-V a.c. voltage. The image acquired after 100 h was rotated after acquisition to follow the orientation in the original image. The persistence of the written structure demonstrates the ability to reversibly switch the ferroelectric polarization with an electric field. In Fig. 4a, a 15 μm × 15 μm region was poled using a single switch of a d.c. bias at +15 V and −15 V (applied to the cantilever tip), generating a ‘downward’- and ‘upward’-oriented out-of-plane polarization, plotted in red and turquoise, respectively. The magnetic structure presented in Fig. 4b–d was measured nearly 200 h after the initial poling shown in Fig. 4a, further corroborating the persistence of the switched state. Variable-temperature and magnetic field transport measurements were made in the van der Pauw geometry and performed using the combination of a Quantum Design PPMS and external Keithley electronics. Measurements were made of the in-plane resistivity, for example, parallel to the (0001) planes. The resistivity increases substantially as the temperature is lowered until it is beyond our measurement abilities below about 200 K. Resistivity values for the entire stack of the two samples we measured, (LuFeO ) /(LuFe O ) and (LuFeO ) /(LuFe O ) , are 4.3–86.5 Ω cm over the 330–198 K temperature range. Plotting the natural logarithm of the conductivity as various powers of inverse temperature suggests that conduction is primarily through thermally activated carriers. No anomaly in the resistivity at T or any changes in resistivity with the application of a magnetic field (0 T, 5 T and 9 T) were observed. We attempted, but were unable (the sample was too resistive), to measure a bare YSZ substrate that had undergone identical deposition conditions as a superlattice (the same oxygen partial pressure and time at growth temperature, but not exposed to the lutetium or iron molecular beams) to determine if the primary conduction path was mediated by oxygen vacancies generated in the substrate. The enhancement of T in region I could be caused by strain or oxygen stoichiometry effects. Lutetium atom distortions were not observed in these superlattices. A slight mismatch between the LuFeO in-plane lattice (a = 5.989 ± 0.005 Å in thick relaxed epitaxial films25 or 3.458 Å for the primitive cell) and the bulk LuFe O (a = 3.4406 Å; ref. 38) produces a tensile strain of up to 0.25% on the LuFe O layers; epitaxial strain has previously been shown to affect ferromagnetic transitions7, 39, 40. In region I, T reaches 250 K, approximately the transition observed in LuFe O single crystals17. Previous experiments have shown that the magnetic properties of LuFe O are extremely sensitive to deviations in oxygen stoichiometry41 and thus a small amount of oxygen diffusion between the LuFe O and LuFeO layers could have also caused the increase in T to 250 K. We use first-principles DFT calculations to elucidate the origin of the increase in the magnetic transition temperature T with the number of LuFeO layers (Fig. 2c) and, correspondingly, the magnitude of the electrical polarization (Fig. 2e). In particular, we perform DFT calculations on what we call ‘model LuFe O ’ systems: bulk-like LuFe O structures with the addition of the trimer distortion Q. This trimer distortion is characteristic of LuFeO and, as shown in Fig. 1b, is experimentally observed in the (LuFeO ) /(LuFe O ) superlattices, yet is not observed in bulk LuFe O . This allows us to derive a simple model that elucidates the origin of the effect in the superlattices, without performing DFT calculations directly on these larger structures. Previous experimental work17 has suggested that bulk LuFe O has a net polarization due to a charge disproportionation between the composing iron bilayers. In the first layer (A), an Fe2+ ion is at the centre of a hexagon composed of six Fe3+ ions, leading to a 1:2 ratio of the charges. The second layer (B) has a similar arrangement with a majority of Fe2+. It was proposed that the bilayers order to form an ABAB stacking along the c direction. We denote this structure as ‘COII’, which is displayed in Fig. 3b and reproduced in Extended Data Fig. 10b. This stacking sequence breaks inversion symmetry and so is compatible with a ferroelectric state (space group Cm). We have also considered multiple additional charge-order arrangements, including those compatible with a non-ferroelectric state, consistent with later experimental results42, 43. We introduce only the low-energy arrangements here. One such arrangement has the same bilayer structure as COII, but an opposite stacking of the bilayers, for example, ABBA ordering. In this case, a mirror plane is located at the Lu–O plane, preserving inversion symmetry (space group C2/m). We refer to this structure, displayed in Fig. 3a and reproduced in Extended Data Fig. 10a, as ‘COI’. In additional, the layers can be stacked in an AABB arrangement (proposed in ref. 42, but studied theoretically for the first time here), which we denote ‘COIII’ (space group C2/m) and display in Extended Data Fig. 10c. In this configuration, the AA and BB bilayers have excess charge and a mirror plane located at the centre between two A (B) layers. COI and COIII are non-polar (space group C2/m)42, 43 and therefore not ferroelectric. We find that COI is the lowest-energy configuration, followed by COII, which is higher than COI by only about 4 meV per formula unit. COIII is higher in energy than COI by more than about 50 meV per formula unit and therefore will not be considered further. The small difference in energy between COI and COII is expected because the two structures differ only in long-range inter-bilayer Coulomb interactions. Structurally, however, COI and COII have a quite distinct (and relatively easy to experimentally discern) difference: the COII configuration induces the lutetium trimer distortions, which are forbidden by symmetry in COI. One difficulty in definitively determining the magnetic ground state of LuFe O from first principles is the presence of multiple magnetic configurations of COI and COII with similar energy. The energetics of these magnetic configurations in COI and COII are shown in Extended Data Fig. 9b. The ground-state magnetization identified for COI is M = 0.5μ /Fe, whereas the lowest energy of COII has M = 1.2μ /Fe. Both COI and COII have low-energy states with M ranging from 0μ /Fe to 1.2μ /Fe. According to experimental measurements44, LuFe O undergoes a ferrimagnetic transition at T = 240–250 K, resulting in an Ising-like net magnetic moment along the c direction. The reported moment ranges from 0.8μ /Fe to 1.4μ /Fe and is particularly sensitive to oxygen stoichiometry41. We thus consider an Ising Hamiltonian to describe the magnetic properties of the LuFe O system: Here J is the symmetric super-exchange interaction and the spin S = ±2 (S = 5/2) for Fe2+ (Fe3+) in trigonal bipyramidal coordination with high spin. The six super-exchange interactions considered correspond to nearest-neighbour interactions for iron (2+/3+) located either in the same plane (J ), in neighbouring planes within a bilayer (J ) or in neighbouring planes between different bilayers (J ); see Extended Data Fig. 8a. These parameters are estimated from first principles. The magnetic ground states and Curie transition temperatures are then determined from Monte Carlo simulations. The simulated T for COI (T ≈ 500 K) is found to be about a factor of two higher than the experimentally reported value (T ≈ 240 K). The possible origins of this discrepancy are: (1) correlations among electrons not captured with DFT + U that strongly influence the magnetic exchange interactions or (2) relativistic interactions between spins that are not included in the present model, yet that often strongly compete with symmetric exchange interactions. We therefore introduce a normalization factor of approximately 0.48 to match the calculated T of COI with the corresponding experimentally reported value and use this same factor throughout our study. Although, in essence, the T of bulk (COI) LuFe O becomes a parameter, the changes in T (for example, from COI to COII) are calculated from a parameter-free, first-principles theory. The corresponding results are displayed in Fig. 3a and b for COI and COII (with Q ≈ 1), respectively, where it is seen that the ferrimagnetic transition temperature of COII is substantially larger (T ≈ 300 K) than that of COI. Noting that COII induces lutetium trimer distortions, we repeat our Monte Carlo simulations of the magnetic transition for different values of Q. As shown in Fig. 3b, we observe that T monotonically increases with the magnitude of this distortion. Calculated super-exchange values as a function of distortion are shown in Extended Data Fig. 8b. The largest super-exchange interactions correspond to in-plane interactions and follow a linear trend as a function of Q. Although the Fe3+–Fe3+ interactions increase and become more antiferromagnetic, the Fe2+–Fe2+ interactions follow the opposite trend. The Fe2+–Fe3+ interactions do not exhibit substantial variation as a function of Q. This result is consistent with the Goodenough–Kanamori–Anderson (GKA) rules considering the change in angle for the Fe–O–Fe super-exchange in-plane paths as illustrated in Extended Data Fig. 8a. As the distortions are increased, on average the Fe–O–Fe angle becomes larger than the undistorted value of 118° for Fe3+–Fe3+ interactions, smaller for Fe2+–Fe2+ and unchanged for Fe2+–Fe3+. Because the Fe2+ and Fe3+ ions both have partially filled d orbitals, the GKA rules predict antiferromagnetic and ferromagnetic super-exchange interactions for 180°and 90° angles, respectively, consistent with the trend observed in Extended Data Fig. 8b. The observed control of T by the trimer distortion is universal, parameter-free and is independent of any specific detail of the first-principles methods. DFT calculations were also performed on the (LuFeO ) /(LuFe O ) periodic supercells for superlattices with m = 1, 3 and 5. The size of these supercells varies depending on whether m is odd or even. For odd values of m, the number of (LuFeO ) /(LuFe O ) blocks has to be doubled along the out-of-plane direction to obtain a periodic supercell. For even values, the number of blocks has to be tripled. Additionally, taking into account that trimerization of the LuFeO structure and charge order of the LuFe O structure demand a supercell containing three formula units per plane, the total number of atoms for the different periodic (LuFeO ) /(LuFe O ) supercells is N = 6(5m + 7) and N = 9(5m + 7), for odd and even values of m, respectively. Owing to computational constraints, most of our calculations were performed on (LuFeO ) /(LuFe O ) (72 atoms) and (LuFeO ) /(LuFe O ) (132 atoms) supercells. Calculations were also performed on (LuFeO ) /(LuFe O ) supercells (192 atoms)—without considering full structural optimizations—to confirm the key result obtained for m = 3, that is, stabilization of hole-doped LuFe O blocks with head-to-head domain walls occurring in the middle of LuFeO blocks, as discussed below. In a single ferroelectric domain of a (LuFeO ) /(LuFe O ) superlattice, the most stable charge-order configuration is COII, as shown in Extended Data Fig. 10d for m = 3. The lowest-energy magnetic configuration corresponds to the same ferrimagnetic arrangement found in bulk LuFe O with the COII pattern, characterized by M = 1.2μ /Fe. Once LuFeO is added to LuFe O —that is, (LuFeO ) /(LuFe O ) superlattices with m ≥ 1—the COI pattern becomes unstable. The single-domain configuration with the COII pattern of the m = 3 (LuFeO ) /(LuFe O ) superlattice is metastable against multidomain configurations, even under E = 0 electrical boundary conditions. We refer to the multidomain configurations as (1) ‘undoped-type’ (Extended Data Fig. 10e) and (2) ‘doped-type’ (Extended Data Fig. 10f). For the m = 1 composition, the difference in energy between the domain types is negligible. By contrast, for the m = 3 superlattice, the doped-type domain configuration is about 8 meV per formula unit more stable than the undoped-type, which itself is about 8 meV per formula unit more stable than the single-domain configuration, demonstrating the added stability of this configuration with increasing m. We also verified that the doped-type domain configuration remains the most stable for the m = 5 superlattice. The undoped-type structure corresponds to the COII ordering pattern in which the AB bilayers are stacked in ABAB sequence along the c direction. Here the superlattice exhibits charged tail-to-tail and head-to-head ferroelectric domain walls. For the m = 3 superlattice, the tail-to-tail domain wall occurs at the interface between a bilayer and LuFeO , whereas the head-to-head domain wall always occurs in the LuFeO block next to the second bilayer, as labelled in Extended Data Fig. 10e. The magnetism of this structure is similar to the single ferroelectric domain case shown in Extended Data Fig. 10d. The starting point to understanding the doped-type domain structure is again the COII ordering pattern of the bulk LuFe O . In this case, however, nearly an electron of charge from a Fe2+ cation located in the A layer of LuFe O is transferred to an Fe3+ cation located in the LuFeO layer, as illustrated in Extended Data Fig. 10f. This results in Fe3+-rich iron bilayers and a corresponding stacking sequence of AAAA (see Extended Data Fig. 10f). Therefore, we can consider the head-to-head domains in this configuration as being stabilized as a result of LuFe O layers electron-doping the LuFeO layers. Indeed, in the doped-type domains, the tail-to-tail domain wall occurs at the interface between a bilayer and LuFeO , whereas the head-to-head domain wall occurs right at the doping layer in the LuFeO block, as shown in Extended Data Fig. 10f. Alternatively, we can consider the LuFeO layers as hole-doping the LuFe O . In this doped-type of ferroelectric domain, the only contribution to the total magnetization in the hole-doped LuFe O layers is from ferromagnetically aligned Fe2+ ions located at the centre of the Fe3+ hexagons. Because the in-plane Fe3+–Fe3+ super-exchange interactions are always found to be antiferromagnetic, there is no net contribution due to the Fe3+ ions. This leads to a magnetization of M = 1.33μ /Fe for each Fe3+-doped bilayer, as shown in Extended Data Fig. 9c. Although this charged domain-wall configuration might appear energetically unfavourable, mobile carriers redistribute to screen the excess of bound charges. The calculated density of states for each iron cation in the heterostructure is shown in Extended Data Fig. 10e. Our calculations demonstrate that electrons migrate from the LuFe O block to occupy iron states in contiguous LuFeO layers, which nevertheless remain insulating. Owing to the shift in the electrostatic potential caused by LuFeO dipoles, conducting Fe d states appear at the head-to-head domain walls (Extended Data Fig. 10f). Because the potential drop across the structure should increase with superlattice periodicity, we expect the electrical conductivity of the walls to grow monotonically with m. The same qualitative trends are observed in the m = 5 superlattices. Structural relaxations of bulk LuFe O were performed using the DFT + U method45 with the Perdew–Burke–Ernzerhof (PBE)46 form of exchange correlation functional as implemented in VASP code47. We used the projector-augmented plane-wave method48. We considered Lu 4f states in the core; for Fe 3d states we chose U = 4.5 eV (where U is the screened Coulomb interaction) and J = 0.95 eV (where J is the Hunds coupling on the iron site). The choice of U and J are based on our previous study on the LuFeO system16, but all results remain qualitatively similar for choices of U > 4.5 eV. We used a 6 × 6 × 2 k-point mesh and a kinetic-energy cut-off value of 500 eV. The Hellman–Feynman forces were converged to 0.001 eV Å−1. To account for the different charge orders (COI, COII and COIII), a supercell of the high-symmetry cell was considered. Considering optimized structures with VASP, spin–spin exchange interactions were estimated by fitting the Ising model with the energies of different magnetic configurations calculated with the linearized augmented plane-wave (LAPW) method as implemented in the Wien2k code49. We considered 25 magnetic configurations to construct the model Hamiltonian for the high-symmetry phase and each charge-ordered state. To be consistent, we used the same U and J values and k-point mesh as for VASP. We used U = 8.8 eV and J = 0.95 eV for Lu 4f states. The plane-wave cut-off, which is defined as the ‘muffin tin’ radius multiplied by k , is 7.0. For (LuFeO ) /(LuFe O ) superlattices, all internal parameters as well as out-of-plane lattice constants were fully relaxed (except for the in-plane lattice constant, which was fixed to the average value of the corresponding LuFeO and LuFe O lattice constants, a = 3.46 Å). All calculations involving superlattices were performed in VASP. Magnetic imaging was performed using cryogenic PEEM at the Advanced Light Source at Beamline 11.0.150, taking advantage of XMCD at the Fe L edge51. Magnetic images are obtained by dividing images recorded with left and right circular polarization. The resulting dark and bright contrast is a measure of the projection of the magnetization direction on the X-ray polarization vector, which has a 30° angle of incidence relative to the surface of the sample. To probe the coupling between the ferroic orders, XMCD-PEEM measurements were made on a region of the (LuFeO ) /(LuFe O ) sample that was poled with a PFM tip to form distinct regions of ‘up’ and ‘down’ c-oriented ferroelectric polarization. The PFM poling was performed eight days before PEEM imaging to ensure that the polarization configuration that was imaged was robust. The PFM poling was performed at 300 K. In the PEEM, the sample was first cooled to 200 K to increase the magnetic contrast from the buried LuFe O layer. As shown in Fig. 4b, the magnetic structure displays the same distinct pattern as the ferroelectric polarization. Finally, to confirm that the magnetic image shown in Fig. 4b could not be due to extrinsic effects, the sample was heated to 320 K. As shown in Fig. 4d, the resulting dichroism dropped by about 70%, consistent with the drop in the saturation magnetization between these temperatures identified in SQUID. Because the overall XLD contrast was constant between those two temperatures (not displayed in Fig. 2f, but acquired using the same experimental configuration) and any extrinsic chemical contrast would similarly be constant, this indicates that the strong dichroism observed at 200 K must arise from the magnetic order.
News Article | February 15, 2017
Tom Dombeck, an innovative and versatile physicist and project manager, passed away in Kāneʻohe, Hawaii, on 4 November 2016. His legacy includes many measurements in particle physics, the development of new techniques for the production of ultra-cold neutrons and substantial contributions to the management of several major scientific projects. Tom received a BA in physics from Columbia University in 1967 and a PhD in particle physics from Northwestern University in 1972, and his career saw him hold prominent roles at numerous institutes. He was a research associate at Imperial College London from 1972 to 1974 and visiting scientist at Dubna in the former USSR in 1975. Following six years at the University of Maryland, from 1981 to 1988 Tom held various roles at Los Alamos National Laboratory (LANL) after which he spent a year working in the US Department of Energy in the office of the Superconducting Supercollider (SSC). Afterwards he became a staff physicist and ultimately deputy project manager for operations at the SSC laboratory in Texas, where he led the successful “string test”. In 1994 he moved to a role as project manager for the Sloan Digital Sky Survey at the University of Chicago. Tom was deputy head for the technical division at Fermilab from 1997 to 1999, and project manager for the Next Linear Collider project at Fermilab between 2000 and 2002. From 2003 to 2006 he was project manager for the Pan-STARRS telescope at the University of Hawaii and an affiliated graduate faculty member there until 2016. Tom began his scientific research with bubble chambers and was a key participant in the experiment that observed the first neutrino interaction in a hydrogen filled bubble chamber in 1970 at the ZGS at Argonne National Laboratory. For many years he pursued measurements of the electric dipole moment (EDM) of the neutron and was also involved in the development of ultra-cold neutrons by Doppler shifting at pulsed sources. He proposed a new method for a neutron EDM measurement that involved Bragg scattering polarised neutrons from a silicon crystal and led an initial effort at the Missouri University Research Reactor, after which he initiated an experiment using the reactor at the NIST Center for Neutron Research. While at LANL, Tom led a neutrino-oscillation search that involved constructing a new beamline and neutrino source at LAMPF and provided improved limits on muon-neutrino to electron-neutrino oscillations. He carried these fundamental physics interests and abilities to his later work as a highly effective scientific programme manager. Tom was able to see the connections between disparate scientific areas and bring together new ideas and approaches that moved the field forwards. He could inspire people around him with his enthusiasm and kindness, and his wry sense of humour and wicked smile were trademarks that will long be remembered by his friends and colleagues. Tom was a devoted family man and is missed greatly by his wife Bonnie, his two children, Daniel and Heidi, and his four grandchildren. Sidney David Drell, professor emeritus of theoretical physics at SLAC National Accelerator Laboratory, senior fellow at Stanford’s Hoover Institution and a giant in the worlds of both academia and policy, died on 21 December 2016 at his home in Palo Alto, California. He was 90 years old. Drell made immense contributions to his field, including uncovering a process that bears his name and working on national and international security. His legacy as a humanitarian includes his friendship and support of Soviet physicist and dissident Andrei Sakharov, who won the Nobel Peace Prize in 1975 for his opposition of the abuse of power in the Soviet Union. Drell was also known for his welcoming nature and genuine, albeit perhaps unwarranted, humility. Drell’s commitment to arms control spanned more than 50 years. He served on numerous panels advising US Congress, the intelligence community and military. He was an original member of JASON, a group of academic scientists created to advise the government on national security and defence issues, and from 1992 to 2001 he was a member of the President’s Foreign Intelligence Advisory Board. He was also the co-founder of the Center for International Security and Cooperation at Stanford, and in 2006 he and former Secretary of State George Shultz began a programme at the Hoover Institution dedicated to developing practical steps towards ridding the world of nuclear weapons. In 1974, Drell met Sakharov at a conference hosted by the Soviet Academy of Sciences and they quickly became friends. When Sakharov was internally exiled to Gorky from 1980 to 1986 following his criticism of the Soviet invasion of Afghanistan, Drell exchanged letters with him and called on Soviet leader Mikhail Gorbachev for his release. He also organised a petition to allow another Soviet physicist and dissident, Nohim Meiman, to emigrate to Israel, and obtained the signatures of 118 members of the US National Academy of Sciences. Having graduated with a bachelor’s degree from Princeton University in 1946, Drell earned a master’s degree in 1947 and a PhD in physics in 1949 from the University of Illinois, Urbana-Champaign. He began at Stanford in 1950 as an instructor in physics, leaving to work as a researcher and assistant professor at the Massachusetts Institute of Technology and then returning to Stanford in 1956 as a professor of physics. He served as deputy director of SLAC from 1969 until his retirement from the lab in 1998. Drell’s research was in the fields of quantum electrodynamics and quantum chromodynamics. While at SLAC, he and research associate Tung-Mow Yan formulated the famous Drell–Yan Process, which has become an invaluable tool in particle physics. His theoretical work was critical in setting SLAC on the course that it took. As head of the SLAC theory group, Drell brought in a host of younger theoretical physicists who began creating the current picture of the structure of matter. He played an important role in developing the justification for experiments and turning the results into what became the foundation of the Standard Model of particle physics. For his research and lifetime of service to his country, Drell received many prestigious awards, including: the National Medal of Science; the Enrico Fermi Award; a fellowship from the MacArthur Foundation; the Heinz Award for contributions in public policy; the Rumford Medal from the American Academy of Arts and Sciences; and the National Intelligence Distinguished Service Medal. Drell was one of 10 scientists honoured as the founders of satellite reconnaissance as a space discipline by the US National Reconnaissance Office. He was elected to the National Academy of Sciences, the American Academy of Arts and Sciences and the American Philosophical Society, and was president of the American Physical Society in 1986. Drell was also an accomplished violinist who played chamber music throughout his life. He is survived by his wife, Harriet, and his children, Daniel, Virginia, Persis and Joanna. Persis Drell, a former director of SLAC who is also a physicist at Stanford and dean of the School of Engineering, will be the university’s next provost. • Based, with permission, on the obituary published on the Stanford University website on 22 December 2016. Mambillikalathil Govind Kumar Menon, a pioneer in particle physics and a distinguished statesman of science, passed away peacefully on 22 November at his home in New Delhi, India. He graduated with a bachelor of science from Jaswant College, Jodhpur, in 1946, and inspired by Chandrasekhara Venkata Raman, studied under the tutelage of spectroscopist Nanasaheb R Tawde before joining Cecil Powell’s group at the University of Bristol, UK, in 1949. Menon’s first important contribution was to establish the bosonic character of the pion through a study of fragments emerging from π-capture by light nuclei. He then focused his attention on the emerging field of K-meson physics. Along with his colleagues at Bristol, notably Peter Fowler, Cecil Powell and Cormac O’Ceallaigh, Menon discovered K+ → π+ π0 and K+ → π+ π– π+ events in nuclear emulsion indicating parity non-conservation, (the τ – θ puzzle). He also identified a sizeable collection of events showing the associated production of kaons and hyperons. In 1955 Menon joined the Tata Institute of Fundamental Research (TIFR), where he worked on cosmic-ray research programmes initiated by Homi Bhabha. Following Bhabha’s death in an air crash over Mont Blanc in 1966, the responsibility of the directorship of TIFR fell squarely on his shoulders, along with the wide-ranging initiatives for national development that Bhabha had started. Notwithstanding these additional demands on his time, his focus on particle physics never wavered. He continued with his research, establishing a collaboration with Arnold W Wolfendale at the University of Durham, UK, and Saburo Miyake of Osaka City University, Japan, for the study of particle physics with detectors deployed deep underground; he detected events induced by cosmic-ray neutrino interactions; and he also launched a dedicated effort to test the early predictions of violation of baryon-number conservation leading to proton decay. During his Bristol years, Menon established a close friendship with William O Lock, who had moved to CERN in 1959. This facilitated collaboration between TIFR and CERN, leading to the development of bubble-chamber techniques to study mesons produced in proton–antiproton collisions. These initial studies eventually led to highly successful collaborations between Indian researchers and the L3 experiment at LEP, and the CMS, ALICE and ATLAS experiments at the LHC. Menon won several awards including the Cecil F Powell and C V Raman medals, and was elected to the three scientific academies in India. He was elected as a fellow of the Royal Society in 1970, and subsequently to the Pontifical Academy of Sciences, American Academy of Arts and Sciences, the Russian Academy of Sciences and as an honorary fellow of the Institute of Physics and the Institution of Electrical & Electronics Engineers. He also served two terms as president of the International Council of Scientific Unions, and stimulated its participation in policy issues, including climate change. Menon held a firm conviction that science can bring about technological development and societal progress, which motivated him to work with Abdus Salam in founding the Third World Academy of Sciences. He held several high positions in the Indian government, and thus contributed to the growth of science and technology in India. Alongside his scientific achievements, M G K Menon was also very close to his wife Indumati and their two children Preeti and Anant Kumar. Our warmest thoughts go out to them and to innumerable others whose lives he touched in so many important ways. Helmut Oeschler, an active member of the ALICE collaboration, passed away from heart failure on 3 January while working at his desk. Born in Southern Germany, he received his PhD from the University of Heidelberg in 1972 and held postdoc positions at the Niels Bohr Institute in Copenhagen, and in Strasbourg, Saclay and Orsay in France. From 1981 he was at the Institute for Nuclear Physics of TU Darmstadt. He held a Doctorate Honoris Causa from Dubna University, Russia, and in 2006 he received the Gay-Lussac-Humboldt prize. Oeschler’s physics interests concerned the dynamics of nuclear reactions over a broad energy range, from the Coulomb barrier to ultra-relativistic collisions. He was a driving force for building the kaon spectrometer at the GSI in Darmstadt, which made it possible to measure strange particles in collisions of heavy nuclei. From the late 1990s he was actively involved in addressing new aspects of equilibration in relativistic nuclear reactions. Oeschler became a member of the ALICE collaboration at CERN in 2000 and made important contributions to the construction of the experiment. Together with his students, he was involved in developing track reconstruction software for measuring the production of charged particles in lead–lead collisions at the LHC. He also led the analysis efforts for the measurements of identified charged hadrons in the LHC’s first proton–proton collisions. From 2010 to 2014 he led the ALICE editorial board, overseeing the publication of key results relating to quark-gluon matter at the highest energy densities. His deep involvement in the data analysis and interpretation continued unabated and he made important contributions to several research topics. Advising and working in close collaboration with students was a much loved component of Helmut’s activity and was highly appreciated among the ALICE collaboration. Helmut Oeschler was a frequent visitor of South Africa and served there on numerous international advisory committees. He was instrumental in helping the South African community develop the physics of heavy-ion collisions and collaboration with CERN. With Helmut Oeschler we have lost an internationally renowned scientist and particular friend and colleague. His scientific contributions, especially on the production of strange particles in high-energy collisions, are important achievements.
News Article | December 20, 2016
The material, a formulation of iron, sodium, copper and arsenic created by Rice graduate student Yu Song in the laboratory of physicist Pengcheng Dai, is described this week in the journal Nature Communications. Dai said Song's recipe—which involves mixing ingredients in a pure argon atmosphere, sealing them in niobium canisters and baking them at nearly 1,000 degrees Celsius—produces a layered alloy in which iron and copper separate into alternating stripes. This striping is critical for the material's usefulness in explaining the origins of high-temperature superconductivity, said RCQM Director Qimiao Si. "By forming this regular pattern, Yu Song has physically removed disorder from the system, and that is crucially important for being able to say something meaningful about what's going on electronically," said Si, a theoretical physicist who has worked to explain the origins of high-temperature superconductivity and similar phenomena for nearly two decades. High-temperature superconductivity was discovered in 1986. It occurs when electrons pair up and flow freely in layered alloys like Song's new creation. Dozens of high-temperature superconducting alloys have been created. Most are complex crystals that contain a transition metal—typically iron or copper—and other elements. High-temperature superconductors are typically terrible conductors at room temperature and only become superconductors when they are cooled to a critical temperature. "The central problem of high-temperature superconductivity is to understand the precise relationship between these two fundamental states of matter and the phase transition between them," said Dai, professor of physics and astronomy at Rice. "The macroscopic change is evident, but the microscopic origins of the behavior are open to interpretation, largely because there are many variables in play, and the relationship between them is both synergistic and nonlinear." Dai said two schools of thought "developed from the very beginning of this field. One was the itinerant camp, which argues that both states ultimately arise from itinerant electrons. After all, these materials are metals, even if they may be poor metals." The other camp is the localized camp, which argues that fundamentally new physics arise—due to electron-electron interactions—at the critical point at which the materials transition from one phase to the other. Dai said measurements on Song's new material support the localized theory. In particular, the new material is the first member of a class of iron-based superconductors called pnictides (pronounced NIK-tides) that can be tuned between two competing phases: the superconducting phase in which electrons flow with no resistance, and a "Mott insulating" phase in which electrons become locked in place and do not flow at all. "The discovery that Yu Song made is that this material is more correlated, which is evident because of the Mott insulating phase," Dai said. "This is the first time anyone has reported an iron-based superconductor that can be continuously tuned from the superconducting phase to the Mott insulating phase." Samples were made and some tests were performed at RCQM. Additional tests were performed at Chalk River Laboratories' Canadian Neutron Beam Center in Ontario, the National Institute for Standards and Technology's Center for Neutron Research in Maryland, Brookhaven National Laboratory in New York, Oak Ridge National Laboratory's High Flux Isotope Reactor in Tennessee and the Paul Scherrer Institute's Advanced Resonant Spectroscopies beamline in Switzerland. "In the paper, we showed that if the interaction was weak, then even replacing 50 percent of the iron with copper would still not be sufficient to produce the insulating state," Si said. "The fact that our experimentalists have managed to turn the system to be Mott insulating therefore provides direct evidence for strong electron-electron interactions in iron pnictides. That is an important step forward because it suggests that superconductivity should be tied up with these strong electron correlations." Explore further: Material turns 'schizophrenic' on way to superconductivity More information: Yu Song et al. A Mott insulator continuously connected to iron pnictide superconductors, Nature Communications (2016). DOI: 10.1038/ncomms13879
News Article | September 13, 2016
In a paper published in Nature Communications, they demonstrate how they synthesised nanometre-sized cage molecules that can be used to transport charge in proton exchange membrane (PEM) applications. Proton-exchange membrane fuel cells (PEMFCs) are considered to be a promising technology for clean and efficient power generation in the twenty-first century. PEMFCs contain proton exchange membrane (PEM), which carries positively-charged protons from the positive electrode of the cell to the negative one. Most PEMs are hydrated and the charge is transferred through networks of water inside the membrane. To design better PEM materials, more needs to be known about how the structure of the membrane enables protons to move easily through it. However, many PEMs are made of amorphous polymers, so it is difficult to study how protons are conducted because the precise structure is not known. Scientists from the University's Department of Chemistry synthesised molecules that enclose an internal cavity, forming a porous organic cage into which other smaller molecules can be loaded, such as water or carbon dioxide. When the cages form solid materials, they can arrange to form channels in which the small 'guest' molecules can travel from one cage to another. The material forms crystals in which the arrangement of cages is very regular. This allowed the researchers to build an unambiguous description of the structure using crystallography, a technique that allows the positions of atoms to be located. The molecules are also soluble in common solvents, which means they could be combined with other materials and fabricated into membranes. They measured the protonic conductivity of these porous organic cages after loading the channels with water, to assess their viability as PEM materials. The cages exhibited proton conductivities of up to 10-3 S cm1, which is comparable to some of the best porous framework materials in the literature. In collaboration with researchers from the University of Edinburgh, Center for Neutron Research at National Institute of Standards and Technology (NIST), and (Defence Science and Technology Laboratory (DSTL), they used a combination of experimental measurements and computer simulations to build a rich picture of how protons are conducted by the cage molecules. Two distinctive features of the proton conduction in organic cage crystals were highlighted as design principles for future PEM materials. First, the cages are arranged so that the channels extend in three dimensions. This means that the movement of the protons is not limited to a particular direction, as in the case of many porous materials tested so far. Second, the cages direct the movement of the water molecules, which means that protons can be passed between them quickly. Also, the cages are flexible enough to allow the water to reorganize, which is also important when protons are transported from one water molecule to the next over longer distances. Dr Ming Liu who led the experimental work, said: "In addition to introducing a new class of proton conductors, this study highlights design principles that might be extended to future materials. "For example, the 'soft confinement' that we observe in these hydrated solids suggests new anhydrous proton conductors where a porous cage host positions and modulates the protonic conductivity of guest molecules other than water. This would facilitate the development of high temperature PEMFCs, as water loss would no longer be a consideration." Liverpool Chemist, Dr Sam Chong, added: "The work also gives fundamental insight into proton diffusion, which is widely important in biology." Dr Chong has recently been appointed as a lecturer in the University's Materials Innovation Factory (MIF). Due to open in 2017, the £68M MIF is set to revolutionise materials chemistry research and development through facilitating the discovery of new materials which have the potential to save energy and natural resources, improve health or transform a variety of manufacturing processes. The paper 'Three-dimensional Protonic Conductivity in Porous Organic Cage Solids' is published in Nature Communications. Explore further: New technique developed to separate complex molecular mixtures More information: Ming Liu et al, Three-dimensional protonic conductivity in porous organic cage solids, Nature Communications (2016). DOI: 10.1038/ncomms12750
News Article | April 6, 2016
By chemically modifying and pulverizing a promising group of compounds, scientists at the US National Institute of Standards and Technology (NIST) have potentially brought safer, solid-state rechargeable batteries two steps closer to reality. Unlike the traditional liquid electrolytes used in rechargeable batteries, these compounds are stable, solid materials that do not pose a risk of leaking or catching fire. They are based on commonly-available substances known as lithium and sodium closo-borate salts, which are made primarily from hydrogen, boron and either lithium or sodium. Since first discovering the properties of these compounds in 2014, a team led by NIST scientists has sought to enhance their performance in two key ways: increasing their current-carrying capacity and ensuring that they can operate in a sufficiently wide temperature range to be useful in real-world environments. Considerable advances have now been made on both fronts, according to Terrence Udovic of the NIST Center for Neutron Research, whose team has published a pair of scientific papers that detail each improvement, in Advanced Energy Materials and Energy Storage Materials respectively. The first advance came when the team found that the original compounds were even better at carrying current with a slight change to their chemical makeup. Replacing one of the boron atoms with a carbon atom improved their ability to conduct ions, which are what carry charge inside a battery. As the team reported in the Advanced Energy Materials paper, this switch made the compounds about 10 times better at conducting. But perhaps more important was clearing the temperature hurdle. The compounds now conducted ions well enough to operate in a battery – as long as they were in an environment typically hotter than boiling water. Unfortunately, there's not much of a market for such high-temperature batteries. By the time the compounds cooled to room temperature, their favorable chemical structure had often changed to a less conductive form, decreasing their performance substantially. One solution turned out to be crushing the compound's particles into a fine powder. The team had been investigating particles that are measured in micrometers, but as nanotechnology research has demonstrated time and again, the properties of a material can change dramatically at the nanoscale. The team found that pulverizing the compounds into nanometer-scale particles resulted in materials that could still perform well at room temperature and far below. "This approach can remove worries about whether batteries incorporating these types of materials will perform as expected even on the coldest winter day," says Udovic, whose collaborators on the Energy Storage Materials paper include scientists from Japan's Tohoku University, the University of Maryland and Sandia National Laboratories. "We are currently exploring their use in next-generation batteries, and in the process we hope to convince people of their great potential." This story is adapted from material from NIST, with editorial changes made by Materials Today. The views expressed in this article do not necessarily represent those of Elsevier. Link to original source.
News Article | November 1, 2016
Fashion is changing in the avant-garde world of next-generation computer component materials. Traditional semiconductors like silicon are releasing their last new lines. Exotic materials called topological insulators (TIs) are on their way in. And when it comes to cool, nitrogen is the new helium. This was clearly on display in a novel experiment at the National Institute of Standards and Technology (NIST) that was performed by a multi-institutional collaboration including UCLA, NIST and the Beijing Institute of Technology in China. Topological insulators are a new class of materials that were discovered less than a decade ago after earlier theoretical work, recognized in the 2016 Nobel Prize in physics, predicted they could exist. The materials are electrical insulators on the inside and they conduct electricity on the outer surface. They are exciting to computer designers because electric current travels along them without shedding heat, meaning components made from them could reduce the high heat production that plagues modern computers. They also might be harnessed one day in quantum computers, which would exploit less familiar properties of electrons, such as their spin, to make calculations in entirely new ways. When TIs conduct electricity, all of the electrons flowing in one direction have the same spin, a useful property that quantum computer designers could harness. The special properties that make TIs so exciting for technologists are usually observed only at very low temperature, typically requiring liquid helium to cool the materials. Not only does this demand for extreme cold make TIs unlikely to find use in electronics until this problem is overcome, but it also makes it difficult to study them in the first place. Furthermore, making TIs magnetic is key to developing exciting new computing devices with them. But even getting them to the point where they can be magnetized is a laborious process. Two ways to do this have been to infuse, or "dope," the TI with a small amount of magnetic metal and/or to stack thin layers of TI between alternating layers of a magnetic material known as a ferromagnet. However, increasing the doping to push the temperature higher disrupts the TI properties, while the alternate layers' more powerful magnetism can overwhelm the TIs, making them hard to study. To get around these problems, UCLA scientists tried a different substance for the alternating layers: an antiferromagnet. Unlike the permanent magnets on your fridge, whose atoms all have north poles that point in the same direction, the multilayered antiferromagnetic (AFM) materials had north poles pointing one way in one layer, and the opposite way in the next layer. Because these layers' magnetism cancels each other out, the overall AFM doesn't have net magnetism -- but a single layer of its molecules does. It was the outermost layer of the AFM that the UCLA team hoped to exploit. Fortunately, they found that the outermost layer's influence magnetizes the TI, but without the overwhelming force that the previously used magnetic materials would bring. And they found that the new approach allowed the TIs to become magnetic and demonstrate all of the TI's appealing hallmarks at temperatures far above 77 Kelvin -- still too cold for use as consumer electronics components, but warm enough that scientists can use nitrogen to cool them instead. "It makes them far easier to study," says Alex Grutter of the NIST Center for Neutron Research, which partnered with the UCLA scientists to clarify the interactions between the overall material's layers as well as its spin structure. "Not only can we explore TIs' properties more easily, but we're excited because to a physicist, finding one way to increase the operational temperature this dramatically suggests there might be other accessible ways to increase it again. Suddenly, room temperature TIs don't look as far out of reach."
News Article | August 22, 2016
In the pursuit of material platforms for the next generation of electronics, scientists are studying new compounds such as topological insulators (TIs), which support protected electron states on the surfaces of crystals that silicon-based technologies cannot. Dramatic new physical phenomena are being realized by combining this field of TIs with the subfield of spin-based electronics known as spintronics. The success within spintronics of realizing important magnetic technologies such as the spin valve have increased the expectations that new results in TIs might have near-term applications. However, combining these two research threads has relied on “shoehorning” magnetism by forcing magnetic atoms to partially occupy elemental positions in TIs or by applying a conventional magnetic field. Realizing an integrated material that is both intrinsically magnetic and has a topological character has proven more challenging. Recently a team of researchers based in the group of Joseph G. Checkelsky, assistant professor of physics at MIT, and collaborators at the NIST Center for Neutron Research (NCNR), Carnegie Mellon University, and the Beijing Institute of Technology have experimentally demonstrated a “hybrid material” solution to this problem. They studied a compound of three elements, gadolinium, platinum and bismuth, known together as a ternary compound. In their compound, gadolinium supplies the magnetic order while the platinum-bismuth components support a topological electronic structure. These two components acting in concert make a correlated material that is more than the sum of its parts, showing quantum mechanical corrections to electrical properties at an unprecedented scale. Their results were reported July 18 in Nature Physics. Proving this delicate interplay between the constituent elements of this compound required studying it from different perspectives. With experimental efforts led by research scientist Takehito Suzuki, the MIT group (including physics graduate student Aravind Devarakonda and physics undergraduate Yu-Ting Liu) synthesized single crystals and studied their electronic and magnetic properties. The team found these crystals at the same time were exotic magnets and exhibited signatures of electronic topology. The latter was observed through the so-called Berry phase corrections to electronic behavior, where they saw the largest such response reported to date in this type of magnet, which is known as an antiferromagnet. The Berry phase reflects the quantum mechanical nature of the charge-carrying electrons in metals and is influenced by magnetic order. They identified an antiferromagnetic transition temperature of 9.2 kelvins (-443 degrees Fahrenheit). At or below this temperature, magnetic moments of the gadolinium atoms align in an alternating pattern of spin up and spin down. Interestingly, for temperatures significantly higher than this they could observe remnants of this magnetic order in both magnetic and electronic properties, a possible hallmark of the underlying frustrated geometry of the crystal lattice. While these experiments were enticing, the team wanted to be sure that what they were observing originated from the topological properties that would connect this material to potentially ground-breaking types of future electronic devices. Experimentally, this involved work at national facilities including the NCNR, where Suzuki worked with Robin Chisnell PhD ’14 and NIST Fellow Jeffrey W. Lynn. Using a triple axis spectrometer (BT-7), they studied the scattering of neutrons from carefully aligned single crystals in the low temperature antiferromagnetic phase including in different magnetic field conditions. These experiments provided the ability to map the behavior of the magnetic gadolinium spins to precisely know their orientation and response to temperature and magnetic field. In order to do these experiments, naturally occurring gadolinium could not be used due to its overwhelming neutron scattering cross-section, and the researchers used instead a costly isotope known as gadolinium -160. In general, growing single crystals of these compounds is challenging because of their high melting temperature; the growth process known as a “flux method” requires high-temperature centrifuging to remove the crystals from a bath of liquid bismuth. “It’s a bit like growing rock candy sugar crystals, except that we use liquid bismuth instead of water,” says Checkelsky. The process is well-known to solid state chemists, but can have a high failure rate. The team was able to obtain enough of the isotope for just two growth runs, both of which turned out to be successful. The subsequent experiments at NCNR were able to provide critical information about the gadolinium moments that shaped the team’s understanding of their results. The team also made use of the National High Magnetic Field Laboratory (NHMFL) based in Tallahassee, Florida. There, the MIT team brought the crystals to measure their response to extreme magnetic conditions involving magnetic fields in excess of 30 T (among the largest DC fields available in the world). The extreme conditions available at the NHMFL also allowed the group to broadly map the electronic and magnetic properties of the crystals to complete the picture of the magnetic order. In particular, they were able to observe a previously unreported phase transition for the gadolinium spins near 25 T that appeared to finally “break” the antiferromagnetic state. The final aspect of the collaborative effort was with professors Di Xiao of Carnegie Mellon University and Wanxiang Feng of Beijing Institute of Technology, who provided first principles electronic structure calculations based on the experimental data taken at MIT, NCNR, and the NHMFL to determine the underlying electronic character of this new materials system. “The authors combine high-quality crystal growth, transport measurements, neutron spectroscopy, and theoretical calculations to establish the magnetic ordering and its profound effect on electrical properties in a topological material,” says Liang Fu, assistant professor of physics at MIT, who was not involved in this research. “This seminal work reveals surprising quantum phenomena arising from the interplay between electron topology and correlation. This type of correlated topological phenomena is long sought after, but has been difficult to find in real materials. By identifying the right material, Joe Checkelsky's group and collaborators have found a new, promising platform for fundamental research and potential spintronics applications.”