An Efficient Energy Transfer Mechanism on the Nanoscale Using Plasmonic Nanoparticles

From Left to Right: (top row) Larousse Khosravi Khorashad, Lucas V. Besteiro, Eva-Maria Roller, (bottom row) Claudia Pupp, Tim Liedl, Alexander O. Govorov
 
Authors: Larousse Khosravi Khorashad¹, Lucas V. Besteiro¹, Eva-Maria Roller², Claudia Pupp², Tim Liedl², Alexander O. Govorov¹

Affiliation:
¹Department of Physics and Astronomy, Ohio University, Athens, Ohio 45701, United States 
²Department of Physics and Center for NanoScience, Ludwig-Maximilians-Universität München, 80539 Munich, Germany

Research in electronic devices is pushing the size of transistors below the 10 nm mark, but at this scale information processing faces difficulties in terms of heat dissipation, as well as in controlling quantum effects such as electron tunneling. An alternative to electron-based computing lies in the creation of all-optical computational devices. Although we are still far from producing functional processors, today we show how plasmonic structures are viable tools not only for the nanoscale localization of electromagnetic energy but also its transfer, contrary to what their typically high dissipation rates would have us believe.

Plasmonics, in general terms, is the field of Physics that studies the interaction of light with materials, mostly at the nanoscale, that support collective electronic oscillations [1]. These plasmonic materials are typically metallic nanoparticles (NPs) or heavily doped semiconductors that exhibit free-electron behavior. When light, as an electromagnetic wave, is incident on the NPs, the free electron gas of the NPs is excited and oscillates as a result of the polarization of the electromagnetic wave. These excited plasmonic oscillations typically have resonances in the visible spectral range and the resonance frequencies depend on the type, shape and size of the plasmon-supporting material.

Plasmonic materials afford us the capability of manipulating light and funneling its energy at the nanoscale, and has enabled the creation of novel devices, from optical metamaterials [2] to efficient nanoscale heat generators [3]. For all-optical computational devices, it is challenging to transfer and manipulate photonic energy at the nanoscale. Here, photon transfer is not feasible using traditional waveguides such as glass fibers, which require sizes comparable to the wavelength of the confined light. Plasmonic nanostructures are remarkably effective in localizing the electromagnetic energy of light, but these structures suffer from high energy losses due to the high rate of electron scattering with the background charges in the material. In our recent work [4] we show how this essential limitation can be successfully circumvented by carefully considering the interaction of different material systems with distinct plasmonic resonances.

Figure 1. TEM image of a trimer fabricated with the DNA origami technique, with a 30 nm AgNP placed between two 40 nm AuNPs. Below there is an energy diagram of the trimer. The AgNP resonance is not fully excited, acting as an efficient energy transmitter, coupling the outer AuNPs. Figure adapted from Ref. [4]

By carefully arranging two gold NPs (AuNPs) and a silver NP (AgNP) in a linear chain , our model system takes advantage of the different optical spectra of gold and silver [4]. Figure 1 shows a transmission electron microscopy (TEM) image of one of the experimentally fabricated trimers, illustrating the arrangement of our proposed hybrid nanostructure. The two 40 nm AuNPs are spatially separated by a 38 nm gap, partially occupied by a 30 nm AgNP that leaves two 4 nm interparticle gaps. To achieve this accurate NP arrangement we have used the DNA origami technique [5,6] to build a rigid bundle of 14 interconnected, parallel DNA double helices. At three points along this DNA structure, attachment sites for each particle are built, consisting of single-stranded DNA (ssDNA) extensions with orthogonal sequences for each material. Figure 1 also includes a schematic energy diagram of the trimer, which illustrates the main effect explored in the paper. The two AuNP plasmon states are only weakly coupled due to the distance separating them, but the connecting AgNP acts as a bridge for their interaction. More importantly, it mediates the transfer with a plasmonic resonance of different energy than the resonance of the AuNPs. In this way, the AgNP plasmon state is not fully occupied and it functions as a virtual state for the plasmon to travel from one AuNP to the other.

The presence of the AgNP is crucial for the coherent transfer of energy from one AuNP (transmitter) to another (receiver) with negligible energy dissipation. The AgNP bridges the large gap between the AuNPs with a virtual energy state that allows the transfer to take place. As we confirmed experimentally, the distance between the AuNPs is enough to impede their coupling in the absence of the mediating particle. In fact, the AuNPs would respond mostly as independent dipoles in the absence of the AgNP. With the three particles in place, the longitudinal plasmonic resonance of the trimer is far from the resonant frequency of the AgNP. Therefore, this mediating particle facilitates the coupling of the two AuNPs and the excitation of the global trimer plasmon mode, while avoiding the dissipation of heat that would appear in the AgNP resonant excitation. Importantly, our simulations show the hot spots that typically occur between the surfaces of metal particles in close proximity, allowing here the effective coupling of their near fields. The coupling is experimentally verified through the observation of the scattering spectra of the trimer, evidencing the hybridization of the plasmonic modes of the NPs. Figure 2 showcases some of the scattering experimental data, alongside results from the simulated system.

Figure 2: (click on the image to view with higher resolution) Experimental (a) and simulated (b) spectra of interacting trimers, compared with a weakly interacting gold NP dimer. The scattering signal is red shifted and strongly enhanced when introducing the mediating silver NP, creating a new resonance for the whole interacting system and evidencing the hybridization of their originally non-interacting states. Figure adapted from Ref. [4]

In a second set of classical electrodynamic simulations, we observed the time resolved transfer inside the trimer after being excited by a single point dipole source, located just outside the chain. We find an extremely fast plasmon passage between the two outer particles, orders of magnitudes faster than the energy transfer observed in typical Förster-type dipole-dipole interactions [7]. By our calculations, we estimate a transfer time of about 5 fs and the oscillation of energy between the two gold particles for more than 40 fs (Figure 3). We have also described the system with a theoretical quantum model in which the three NPs are considered as three quantum oscillators that are coupled to each other via Coulomb interactions. This quantum model predicts the energy shift of the system’s bright plasmon modes that are observed experimentally, as well as the dark modes found in a full electrodynamical simulation.

Figure 3: (click on the image to view with higher resolution) Evolution in time of the excitation in the two AuNPs as obtained from simulations, showing the total electric dipole in each NP after a pulse excitation by a dipole source close to the left AuNP. The diagram on top is a representation of the trimer and the source. The middle AgNP acts as a transmitter. The data in (b), shown in a logarithmic scale, is the total amplitude of the dipolar moment in each gold NP, in which one can appreciate the back-and-forth transfer of the excitation between the AuNPs. Figure adapted from Ref. [4]

In summary, in this research we presented a novel approach toward fast and almost loss-less plasmon transfer between metal nanoparticles, which can have a great impact in applications for energy and information transfer on the nanoscale. We used a simple combination of gold and silver nanoparticles to coherently transfer energy and overcome the large heat dissipation characteristic of plasmonic systems. Our proposed quantum model provides us with a clear understanding of the physics, and our full electrodynamic simulation is in excellent agreement with the experimental measurements performed on this exciting system. This work opens new doors for designing future optical-based computing devices and for the teleportation of data over longer distances at the nanoscale.

References
[1] Stefan A. Maier, "Plasmonics: fundamentals and applications" (Springer, 2007).
[2] Kan Yao and Yongmin Liu, “Plasmonic metamaterials”. Nanotechnology Rev.iews  3, 177 (2014).  Abstract.
[3] Larousse Khosravi Khorashad, Lucas V. Besteiro, Zhiming Wang, Jason Valentine, and Alexander O. Govorov, “Localization of Excess Temperature Using Plasmonic Hot Spots in Metal Nanostructures: Combining Nano-Optical Antennas with the Fano Effect”. Journal of Physical Chemistry C, 120, 13215 (2016). Abstract.
[4] Eva-Maria Roller, Lucas V. Besteiro, Claudia Pupp, Larousse Khosravi Khorashad, Alexander O. Govorov and Tim Liedl, “Hotspot-mediated non-dissipative and ultrafast plasmon passage”. Nature Physics (published online May 15, 2017). Abstract.
[5] Anton Kuzyk, Robert Schreiber, Zhiyuan Fan, Günther Pardatscher, Eva-Maria Roller, Alexander Högele, Friedrich C. Simmel, Alexander O. Govorov and Tim Liedl, “DNA-based self-assembly of chiral plasmonic nanostructures with tailored optical response". Nature, 483,  311 (2012). Abstract.
[6] Paul W. K. Rothemund, “Folding DNA to create nanoscale shapes and patterns”. Nature, 440,  297 (2006). Abstract.
[7] Volkhard May and Oliver Kühn, “Charge and Energy Transfer Dynamics in Molecular Systems” (3rd Ed. Wiley, 2011). 

Ultratransparent Media: Towards the Ultimate Transparency

From left to right: (top row) Jie Luo, Yuting Yang, Zhongqi Yao, Weixin Lu; (bottom row) Bo Hou, Zhi Hong Hang, C. T. Chan, and Yun Lai.

Authors: Jie Luo¹, Yuting Yang¹, Zhongqi Yao¹, Weixin Lu¹, Bo Hou¹, Zhi Hong Hang¹, Che Ting Chan², Yun Lai¹

Affiliation:
¹College of Physics, Optoelectronics and Energy & Collaborative Innovation Center of Suzhou Nano Science and Technology, Soochow University, Suzhou 215006, China
²Department of Physics and Institute for Advanced Study, Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong.

Transparent media are the foundation of almost all optical instruments, such as optical lens, etc. However, perfect transparency has never been realized in natural transparent solid materials such as glass because of the impedance mismatch with free space or air. As a consequence, there generally exist unwanted reflected waves at the surface of a glass slab, as illustrated in Fig. 1(a). It is well known that non-reflection only occurs at a particular incident angle for a specific polarization, which is known as the Brewster angle effect [1]. Our question is: is it possible to extend the Brewster angle from a particular angle to a wide range of or all angles, so that there is no reflection for any incident angle.

In addition, the virtual image formed by a glass slab placed in air is usually blurred to a certain extent [Fig. 1(a)]. Such a blur indicates the aberration of virtual images, and is caused by the mismatch of equal frequency contours (EFCs) between air (grey lines) and the glass (blue lines).

Figure 1: (a) Virtual image formation through a glass slab, with general reflection and aberration. (b) Aberration-free virtual image formation through an ultratransparent photonic crystal without any reflection due to omnidirectional impedance matching. The black arrows and blue dashed lines in represent the light rays from a point source, and the back tracing lines, respectively. The yellow dashed curves in (b) indicate equal phase surfaces of transmitted rays. The inset graphs show the corresponding EFCs. The figure is adapted from Reference [2].

The purpose of our work [2] is to explore the possibility of realizing the ultimate transparency by artificial optical structures such as photonic crystals (PhCs) [3] and metamaterials [4]. In other words, we pursue the realization of transparent media with the extreme property of omnidirectional impedance matching and the ability of forming aberration-free virtual images, which are hereby denoted as ultratransparent media.

In this work, we propose that omnidirectional impedance matching can be realized by utilizing effective medium with nonlocal parameters, i.e. permittivity and permeability that are dependent on the incident angle. Interestingly, such an effective medium can be realized by using pure dielectric PhCs. Moreover, the EFC of the ultratransparent PhC can be tuned to be a shifted ellipse (red lines) with the same height of the EFC of air (grey lines). By using ray optics, we prove that such an EFC endows the valuable ability of forming aberration-free virtual images, as presented in Fig. 1(b).

Because of the shift of EFC, the PhC is beyond the local medium framework, and effective parameters are nonlocal (i.e. spatially dispersive). Interestingly, such nonlocality leads to additional phase modulation p d, where p is the shift magnitude and d is the slab thickness [Fig. 1(b)].

Figure 2: (a) Illustration of the unit cell of the ultratransparent PhC. (b) The EFC of the PhC. (c) Transmittance through the PhC slab with N (=4, 5, 6, 15) layers of unit cells as functions of the incident angle. The figure is adapted from Reference [2].

An extreme example with almost complete transparency (T>99%) for nearly all incident angles (-89^o, +89^o) is shown in Fig. 2. The PhC is two-dimensional and its unit cell is shown in Fig. 2(a). For transverse electric polarization, at the working frequency, the EFC is a shifted ellipse (red dashed curve) with the same height as that in free space (grey dashed curve), as shown in Fig. 2(c). The transmittance through such a PhC slab is always near unity (>99%) for nearly all incident angles (<89^o), and is almost irrespective of the layer number, N.

Figure 3: (a) Photo of the simplified PhC composed of alumina bars (white) placed inside the microwave field mapper. (b) The EFC of the PhC. (c) Transmittance through the PhC slab in simulations (solid lines) and experiments (triangular dots) and an alumina slab having the same thickness (dashed lines) as the function of incident angles. The figure is adapted from Reference [2].

To prove the theory, we performed proof-of-principle microwave experiments by utilizing a simplified PhC consisting of rectangular alumina bars in a square lattice, as shown in Fig. 3(a). Such a PhC exhibits a shifted elliptical EFC [Fig. 3(b)] and a wide-angle impedance matching effect. The measured transmission data (triangular dots) and simulation results (solid lines) both show high transmittance, which is great enhancement compared to the transmittance through an alumina slab with the same thickness (dashed lines).

Finally, we also note that such ultratransparent media can extend transformation optics (TO) [5, 6] to the general realm of nonlocal media. The traditional TO was founded in the local medium framework and require local media. Here, we demonstrate that ultratransparent media with controllable refractive indexes are also good candidates for TO applications such as invisibility cloaks. Interestingly, due to the nonlocality, the ultratransparent media also enable additional freedom in phase modulation, which is absent in the traditional TO. At optical frequencies, ultratransparent PhCs exhibit the significant advantages of omnidirectional impedance matching, low loss and micro fabrication requirement.

The concept and theory of ultratransparency gives a guideline for realizing the ultimate transparency which is broadband, omnidirectional and polarization-insensitive. Recently, we designed broadband, wide-angle and polarization-insensitive transparent media by using one-dimensional dielectric PhCs [7]. In the future, ultratransparent solid materials may be optimized to exhibit an unprecedented level of transparency and produce no reflection at all in certain ranges of frequencies.

References:
[1] John D. Jackson, "Classical Electrodynamics" (3rd edition, Wiley, New York, 1975).
[2] J. Luo, Y. Yang, Z. Yao, W. Lu, B. Hou, Z. H. Hang, C. T. Chan, and Y. Lai, "Ultratransparent media and transformation optics with shifted spatial dispersions", Physical Review Letters, 117, 223901 (2016). Abstract.
[3] John D. Joannopoulos, Steven G. Johnson, Joshua N. Winn, Robert D. Meade, "Photonic Crystals: Molding the Flow of Light" (2nd edition, Princeton University Press, Princeton, USA, 2008).
[4] Yongmin Liu, Xiang Zhang, "Metamaterials: a new frontier of science and technology", Chemical Society Reviews, 40, 2494-2507 (2011). Abstract.
[5] J. B. Pendry, D. Schurig, and D. R. Smith, "Controlling electromagnetic fields", Science, 312, 1780-1782 (2006). Abstract.
[6] Ulf Leonhardt, "Optical conformal mapping", Science, 312, 1777-1780 (2006). Abstract.
[7] Zhongqi Yao, Jie Luo, Yun Lai, "Photonic crystals with broadband, wide-angle, and polarization-insensitive transparency", Optics Letters, 41, 5106 (2016). Abstract.

Carbon Nanotubes as Exceptional Electrically Driven On-Chip Light Sources

From left to right: (top row) Felix Pyatkov, Svetlana Khasminskaya, Valentin Fütterling; (bottom row) Manfred M. Kappes, Wolfram H. P. Pernice, Ralph Krupke

Authors: Felix Pyatkov^1,2, Svetlana Khasminskaya¹, Valentin Fütterling¹, Randy Fechner¹, Karolina Słowik^3,4, Simone Ferrari^1,5, Oliver Kahl^1,5, Vadim Kovalyuk^1,6, Patrik Rath^1,5, Andreas Vetter¹, Benjamin S. Flavel¹, Frank Hennrich¹, Manfred M. Kappes^1,7, Gregory N. Gol’tsman⁶, Alexander Korneev⁶, Carsten Rockstuhl^1,3, Ralph Krupke^1,2, Wolfram H.P. Pernice⁵

Affiliation:
¹Institute of Nanotechnology, Karlsruhe Institute of Technology, Germany
²Department of Materials and Earth Sciences, Technical University of Darmstadt, Germany
³Institute of Theoretical Solid State Physics, Karlsruhe Institute of Technology, Germany
⁴Institute of Physics, Nicolaus Copernicus University, Poland
⁵Department of Physics, University of Münster, Germany
⁶Department of Physics, Moscow State Pedagogical University, Russia
⁷Institute of Physical Chemistry, Karlsruhe Institute of Technology, Germany

Carbon nanotubes (CNTs) belong to the most exciting objects of the nanoworld. Typically, around 1 nm in diameter and several microns long, these cylindrically shaped carbon-based structures exhibit a number of exceptional mechanical, electrical and optical characteristics [1]. In particular, they are promising ultra-small light sources for the next generation of optoelectronic devices, where electrical components are interconnected with photonic circuits.

Few years ago, we demonstrated that electically driven CNTs can serve as waveguide-integrated light sources [2]. Progress in the field of nanotube sorting, dielectrophoretical site-selective deposition and efficient light coupling into underlying substrate has made CNTs suitable for wafer-scale fabrication of active hybrid nanophotonic devices [2,3].

Recently we presented a nanotube-based waveguide integrated light emitters with tailored, exceptionally narrow emission-linewidths and short response times [4]. This allows conversion of electrical signals into well-defined optical signals directly within an optical waveguide, as required for future on-chip optical communication. Schematics and realization of this device is shown in Figure 1. The devices were manufactured by etching a photonic crystal waveguide into a dielectric layer following electron beam lithography. Photonic crystals are nanostructures that are also used by butterflies to give the impression of color on their wings. The same principle has been used in this study to select the color of light emitted by the CNT. The precise dimensions of the structure were numerically simulated to tailor the properties of the final device. Metallic contacts in the vicinity to the waveguide were fabricated to provide electrical access to CNT emitters. Finally, CNTs, sorted by structural and electronic properties, were deposited from a solution across the waveguide using dielectrophoresis, which is an electric-field-assisted deposition technique.

Figure 1: (a) Schematic view of the multilayer device structure consisting of two electrodes (yellow) and a photonic waveguide (purple) that is etched into the Si3N4 layer. Its central part is underetched into the SiO2 layer to a depth of 1.5 µm and photonic crystal holes are formed. The carbon nanotube bridges the electrodes on top of the waveguide. (b,c) False colored scanning electron microscope images of the device. The figure is adapted from Reference [4].

The functionality of the device was verified with optical microscopy and spectroscopy, which allowed detection of light emitted by the CNT and also of the light coupled into the waveguide. An electrically biased CNT generates photons, which efficiently couple into the photonic crystal waveguide, as shown in Figure 2a. The emitted light propagates along the waveguide and is then coupled out again using on-chip grating couplers. Because of the photonic crystal the emission spectrum of the CNT is extremely sharp (Figure 2b) and the emission wavelength can be tailored by our manufacturing process. In addition, the nanotube responds very quickly to electrical signals and hence acts as a transducer for generating optical pulses in the GHz range (Figure 2c). The modulation rates of these CNT-based transducers can in principle be pushed to much higher frequencies up to 100 GHz using more advanced nanostructures.

Figure 2: (click on the image to view with higher resolution) CCD camera image of the electrically biased device. Light emission is observed from the nanotube and from the on-chip grating couplers, both connected with the emitter via the waveguide. (b) Emission spectra simultaneously measured at the grating coupler. (c) A sequence of the driving electrical pulses as well as the recorded waveguided emission pulses (red) in GHz frequency range. The figure is adapted from Reference [4].

Nanophotonic circuits are promising candidates for next-generation computing devices where electronic components are interconnected optically with nanophotonic waveguides. The move to optical information exchange, which is already routinely done in our everyday life using optical fibers, also holds enormous benefit when going to microscopic dimensions – as found on a chip. Essential elements for such opto-electronic devices are nanoscale light emitters which are able to convert fast electrical signals into short optical pulses. Using such ultrafast transducers will allow for reducing power requirements and eventually speed up current data rates. For achieving ultimately compact devices the emitter should be as small as possible and interface efficiently with sub-wavelength optical devices. It would also have to operate at a chosen design wavelength and at high speed.

CNTs integrated into a photonic crystal nanobeam waveguides fulfill these requirements and constitute a promising new class of transducers for on-chip photonic circuits. These novel emitters are particularly interesting because of their simplicity. In contrast to conventional laser sources, CNTs are made entirely from carbon, which is available in abundance and does not require expensive fabrication routines as needed for III-V technologies. Moreover, CNTs can also be readily combined with existing CMOS technology, which makes them attractive for a wide range of applications.

So far we spoke about traditional computers based on binary logic. Going beyond classical computation, quantum computers that exploit the enormous potential of quantum mechanics for complex calculations and cryptography hold promise to revolutionize current information processing approaches. Optical quantum systems that employ single photons to realize quantum bits (qubits) belong to the prominent candidates for such future quantum information processing systems. To build a photonic quantum computer, sources of single photons (e.g. single molecules, quantum dots and semiconducting CNTs [5]), optical quantum gates and single photon detectors are needed. These devices are capable of very fast and reliable emission and detection of distinct photons.

An experimental approach, which allows for showing that a light source emits one photon at a time, consists of measuring intensity correlations in the emitted light. We performed this experiment on a solid silicon-based chip with an electrically driven CNT -- actings as a non-classical light source, waveguides, and two detectors for single photons [6]. The nanophotonic circuit shown in Figure 3 includes these three components: a CNT, a dielectric waveguide for the low-loss light propagation and a pair of superconducting nanowires for the efficient detection of light. A single chip carries dozens of such photonic circuits. The device fabrication process was similar to the realization of photonic crystal waveguides, except that now travelling-wave nanowire detectors were also formed on top of the waveguide.

Figure 3: (click on the image to view with higher resolution) (a,b) Schematics and optical image of device with an electrically driven light emitting nanotube in the middle (E) and single-photon superconducting NbN-detectors at the ends (D) of waveguide. The figure is adapted from Reference [6].

The functionality of the device was verified at cryogenic conditions with a setup which allowed the ultra-fast detection of light that was emitted by the CNT and then coupled into the waveguide. An electrically biased semiconducting CNT generates single photons, which can propagate bidirectionally within the waveguide towards the highly sensitive detectors. The intensity of the emitted light was measured as a function of the electrical bias current through the nanotube (Figure 4a). If only one photon at a time is emitted, the simultaneous detection of two photons with both detectors is highly unlikely. This can be derived from the dip in the second-order correlation function shown in Figure 4b. The low probability of simultaneous many-photon detection underlines the non-classical nature of the light source, which is the first step towards a true single-photon emitter. In essence, we thus realized a fully-integrated quantum photonic circuit with a single photon source and detectors, both of which are electrically driven and scalable.

Figure 4: (a) Measurement of the CNT emission intensity in dependence of bias current. Within the marked region semiconducting CNTs reveal non-classical emitting properties. (b) A measured second order correlation function. The minimum value significantly below unity represents the low possibility for simultaneous emission of two photons. The figure is adapted from Reference [6].

References:
[1] Phaedon Avouris, Marcus Freitag, Vasili Perebeinos, "Carbon-Nanotube Photonics and Optoelectronics", Nature Photonics, 2, 341-350 (2008). Abstract.
[2] Svetlana Khasminskaya, Felix Pyatkov, Benjamin S. Flavel, Wolfram H. P. Pernice, Ralph Krupke ,"Waveguide-Integrated Light-Emitting Carbon Nanotubes", Advanced Materials, 26, 3465-3472 (2014). Abstract.
[3] Randy G. Fechner, Felix Pyatkov, Svetlana Khasminskaya, Benjamin S. Flavel, Ralph Krupke, Wolfram H. P. Pernice, "Directional Couplers with Integrated Carbon Nanotube Incandescent Light Emitters", Optics Express, 24, 966-974 (2016). Abstract.
[4] Felix Pyatkov, Valentin Fütterling, Svetlana Khasminskaya, Benjamin S. Flavel, Frank Hennrich, Manfred M. Kappes, Ralph Krupke, Wolfram H. P. Pernice, "Cavity-Enhanced Light Emission from Electrically Driven Carbon Nanotubes", Nature Photonics, 10, 420-427 (2016). Abstract.
[5] Alexander Högele, Christophe Galland, Martin Winger, Atac Imamoğlu, "Photon Antibunching in the Photoluminescence Spectra of a Single Carbon Nanotube", Physical Review Letters, 100, 217401 (2008). Abstract.
[6] Svetlana Khasminskaya, Felix Pyatkov, Karolina Słowik, Simone Ferrari, Oliver Kahl, Vadim Kovalyuk, Patrik Rath, Andreas Vetter, Frank Hennrich, Manfred M. Kappes, Gregory N. Gol’tsman, Alexander Korneev, Carsten Rockstuhl, Ralph Krupke, Wolfram H.P. Pernice "Fully Integrated Quantum Photonic Circuit with an Electrically Driven Light Source", Nature Photonics (2016). Abstract.

Efficient Long-Distance Heat Transport by Microwave Photons

Research team behind the original discovery (Reference [11] ) from left to right: Tuomo Tanttu, Joonas Goveenius, Mikko Möttönen, Matti Partanen, and Miika Mäkelä (Missing from the figure: Kuan Yen Tan and Russell Lake). Photo Credit: Vilja Pursiainen/Kaskas Media.

Authors: Matti Partanen and Mikko Möttönen

Affiliation: QCD Labs, Department of Applied Physics, Aalto University, Finland.

Link to the Quantum Computing and Devices (QCD) Group >>

Quantum computers are predicted to vastly speed up the computation for certain problems of great practical interest [1]. One of the most promising architectures for quantum computing is based on superconducting quantum bits [2], or qubits, which are the key ingredients in circuit quantum electrodynamics [3]. In such systems, the control of heat at the quantum level is extremely important, and remote cooling may turn out to be a viable option.

In one dimension, heat transport may be described by individual heat conduction channels -- each corresponding to a certain quantized profile of the heat carriers in the transverse direction. Importantly, the maximum heat power flowing in a single channel between bodies at given temperatures is fundamentally limited by quantum mechanics [4,5]. This quantum limit has previously been observed for phonons [6], sub-wavelength photons [7,8], and electrons [9]. Among these, the longest distance of roughly 50 μm [7,8] was recorded in the photonic channel [10]. Such short distance may be undesirable in cooling quantum devices which are sensitive to spurious dissipation.

In our recent work [11], we observe quantum-limited heat conduction by microwave photons flying in a superconducting transmission line of length 20 cm and 1 m. Thus we were able to extend the maximum distance 10,000 fold compared with the previous experiments.

Figure 1: (click on the figure to view with higher resolution) Sample structure and measurement scheme. The electron temperature of the right resistor is controlled with an external voltage while the temperatures of both resistors are measured. Microwave photons transport heat through the spiraling transmission line.

Our sample is shown in Figure 1. The heat is transferred between two normal-metal resistors functioning as black-body radiators to the transmission line [10,12]. To be able to fabricate the whole sample on a single relatively small chip, the transmission line has a double spiral structure. We have measured such spiraling transmission lines without resistors and confirmed that photons travel along the line; they do not jump through vacuum from one end to the other. Thus for heat transport, the distance should be measured along the line.

We measure the electron temperatures of both normal-metal resistors while we change the temperature of one of them [13]. The obtained temperature data agrees well with our thermal model, according to which the heat conduction is very close to the quantum limit.

In contrast to subwavelength distances employed in References [7,8], we need to match the resistance of the normal-metal parts to the characteristic impedance of the transmission line to reach the quantum limit. Furthermore, the transmission line itself has to be so weakly dissipative that almost no photons are absorbed even over distances of about a meter. However, we managed to develop nanofabrication techniques which enabled us to satisfy these conditions well. In fact, the losses in the transmission line are so weak they allow a further increment of the distance by several orders of magnitude.

We consider that long-distance heat transport through transmission lines may be a useful tool for certain future applications in the quickly developing field of quantum technology. If the coupling of a quantum device to a low-temperature transmission line can be well controlled in situ, the device may be accurately initialized without disturbing its coherence properties when the coupling is turned off [14]. Furthermore, the implementation of such in-situ-tunable environments opens an interesting avenue for the study of the detailed dynamics of open quantum systems and quantum fluctuation relations [15].

Acknowledgements: We thank M. Meschke, J. P. Pekola, D. S. Golubev, J. Kokkala, M. Kaivola and J. C. Cuevas for useful discussions, and L. Grönberg, E. Mykkänen, and A. Kemppinen for technical assistance. We acknowledge the provision of facilities and technical support by Aalto University at Micronova Nanofabrication Centre. We also acknowledge funding by the European Research Council under Starting Independent Researcher Grant No. 278117 (SINGLEOUT), the Academy of Finland through its Centres of Excellence Program (project nos 251748 and 284621) and grants (nos 138903, 135794, 265675, 272806 and 276528), the Emil Aaltonen Foundation, the Jenny and Antti Wihuri Foundation, and the Finnish Cultural Foundation.

References:
[1] T.D. Ladd, F. Jelezko, R. Laflamme, Y. Nakamura, C. Monroe, J.L. O'Brien, “Quantum computers”, Nature, 464, 45 (2010). Abstract.
[2] J. Kelly, R. Barends, A.G. Fowler, A. Megrant, E. Jeffrey, T.C. White, D. Sank, J.Y. Mutus, B. Campbell, Yu Chen, Z. Chen, B. Chiaro, A. Dunsworth, I.-C. Hoi, C. Neill, P.J.J. O’Malley, C. Quintana, P. Roushan, A. Vainsencher, J. Wenner, A.N. Cleland, John M. Martinis, “State preservation by repetitive error detection in a superconducting quantum circuit”, Nature, 519, 66 (2015). Abstract.
[3] A. Wallraff, D.I. Schuster, A. Blais, L. Frunzio, R.-S. Huang, J. Majer, S. Kumar, S.M. Girvin, R.J. Schoelkopf, “Strong coupling of a single photon to a superconducting qubit using circuit quantum Electrodynamics”, Nature, 431, 162 (2004). Abstract.
[4] J.B. Pendry, “Quantum limits to the flow of information and entropy”, Journal of Physics A: Mathematical and General, 16, 2161 (1983). Abstract.
[5] Luis G. C. Rego, George Kirczenow, “Fractional exclusion statistics and the universal quantum of thermal conductance: A unifying approach”, Physical Review B, 59, 13080 (1999). Abstract.
[6] K. Schwab, E.A. Henriksen, J.M.Worlock, M.L. Roukes, “Measurement of the quantum of thermal conductance”, Nature, 404, 974 (2000). Abstract.
[7] Matthias Meschke, Wiebke Guichard, Jukka P. Pekola, “Single-mode heat conduction by photons”, Nature, 444, 187 (2006). Abstract.
[8] Andrey V. Timofeev, Meri Helle, Matthias Meschke, Mikko Möttönen, Jukka P. Pekola, “Electronic refrigeration at the quantum limit”, Physical Review Letters, 102, 200801 (2009). Abstract.
[9] S. Jezouin, F.D. Parmentier, A. Anthore, U. Gennser, A. Cavanna, Y. Jin, and F. Pierre, “Quantum limit of heat flow across a single electronic channel”, Science, 342, 601 (2013). Abstract.
[10] D.R. Schmidt, R.J. Schoelkopf, A.N. Cleland, “Photon-mediated thermal relaxation of electrons in nanostructures”, Physical Review Letters, 93, 045901 (2004). Abstract.
[11] Matti Partanen, Kuan Yen Tan, Joonas Govenius, Russell E. Lake, Miika K. Mäkelä, Tuomo Tanttu, Mikko Möttönen, “Quantum-limited heat conduction over macroscopic distances”, Nature Physics, Advance online publication, DOI:10.1038/nphys3642 (2016). Abstract.
[12] L.M.A. Pascal, H. Courtois, F.W.J. Hekking, “Circuit approach to photonic heat transport”, Physical Review B, 83, 125113 (2011). Abstract.
[13] Francesco Giazotto, Tero T. Heikkilä, Arttu Luukanen, Alexander M. Savin, Jukka P. Pekola, “Opportunities for mesoscopics in thermometry and refrigeration: Physics and applications”, Reviews of Modern Physics, 78, 217 (2006). Abstract.
[14] P. J. Jones, J.A.M. Huhtamäki, J. Salmilehto, K.Y. Tan, M. Möttönen, “Tunable electromagnetic environment for superconducting quantum bits”, Scientific Reports, 3, 1987 (2013). Abstract.
[15] Jukka P. Pekola, “Towards quantum thermodynamics in electronic circuits”, Nature Physics, 11, 118 (2015). Abstract.

Single-Photon Sources Combine High Purity, Indistinguishability and Efficiency All Together

From left to right: Chao-Yang Lu, Jian-Wei Pan, Sven Höfling and Christian Schneider.

Authors: Chao-Yang Lu¹, Christian Schneider², Sven Höfling^1,2,3,  Jian-Wei Pan¹

Affiliation:
¹CAS Centre for Excellence and Synergetic Innovation Centre in Quantum Information and Quantum Physics, University of Science and Technology of China, Hefei, China.
²Technische Physik, Physikalisches Institat and Wilhelm Conrad Rontgen-Center for Complex Material Systems, Universitat Wurzburg, Germany.
³SUPA, School of Physics and Astronomy, University of St. Andrews, UK.

One-sentence summary: A single-photon source has been demonstrated which, for the first time, combines the features of high efficiency and near-perfect levels of purity and indistinguishabilty, opening the way to scalable multi-photon experiments on a semiconductor chip.

Spontaneous parametric down conversion has served as an excellent workhorse for fundamental test of quantum mechanics, quantum teleportation and optical quantum computing [1]. In this nonlinear optics process, the emission of photon pairs is probabilistic (with a probability of p) and inevitably accompanied by higher-order emission events (on the order of p²), which strongly limit the scalability for optical quantum information processing. So far, up to eight-photon entanglement—created from four independent photon pairs—have been demonstrated [2].

Past 2Physics article by Chao-Yang Lu and/or Jian-Wei Pan :
March 22, 2015: "Quantum Teleportation of Multiple Properties of A Single Quantum Particle" by Chao-Yang Lu and Jian-Wei Pan
January 04, 2015: "Achieving 200 km of Measurement-device-independent Quantum Key Distribution with High Secure Key Rate" by Yan-Lin Tang, Hua-Lei Yin, Si-Jing Chen, Yang Liu, Wei-Jun Zhang, Xiao Jiang, Lu Zhang, Jian Wang, Li-Xing You, Jian-Yu Guan, Dong-Xu Yang, Zhen Wang, Hao Liang, Zhen Zhang, Nan Zhou, Xiongfeng Ma, Teng-Yun Chen, Qiang Zhang, Jian-Wei Pan
June 30, 2013: "Quantum Computer Runs The Most Practically Useful Quantum Algorithm" by Chao-Yang Lu and Jian-Wei Pan.

In an attempt to overcome this obstacle, increasing attention has turned to single quantum emitters, such as self-assembled semiconductor quantum dots (QD), trapped atoms or ions, single defects in diamond or monolayer, and single molecules. In the past two decades, although many previous proof-of-principle experiments have established photon antibunching — an unambiguous evidence for single-photon emission, a scalable extension to multiple photonic quantum bits remain elusive.

To be useful for multi-photon applications such as Boson sampling, a perfect single quantum emitters should fulfill the following wish list: (1) High quantum efficiency: The decay of excited states should predominantly result in an emitted photon. (2) Deterministic generation: Upon a pulsed excitation, the source should deterministically emit one photon in a push-button fashion. (3) High purity: The emission should have a vanishing multi-photon probability. (4) High indistinguishability: Individual photons emitted at different trials should be quantum mechanically identical to each other. (5) High collection efficiency: The radiated photons should be extracted with a high efficiency to a single spatial mode.

Past 2Physics article by Sven Höfling :
May 17, 2015: "A Current Out Of Fluctuations" by Pierre Pfeffer, Fabian Hartmann, Sven Höfling, Martin Kamp, Lukas Worschech.

Among the discovered single quantum emitters so far, QDs have the highest quantum efficiency in solid state and narrowest linewidth at cryogenic temperature, and thus are promising as deterministic single-photon emitters. However, despite the extensive efforts, simultaneously fulfilling all the five criteria in the wish list proved challenging. Most previous experiments either relied on non-resonant excitation of a QD-microcavity that degraded the photon purity and indistinguishability [3,4], or used resonant excitation of a QD in a planar cavity that limited the extraction efficiency [5].

Figure 1: (a) Scanning electron microscopy image of a typical QD micropillar. (b) Numerical simulation of the photon emission from the QD-micropillar. (c) The photons collected into the first lens per pulse versus single-photon purity versus pump power.

Recently, the USTC-Wurzburg joint team exploited s-shell pulsed resonant excitation of a Purcell-enhanced QD-micropillar to deterministically generate resonance fluorescence single photons [6] which for the first time combines all the features in the wish list. The experiments were performed on an InAs/GaAs QD embedded inside a 2.5 micron diameter micropillar cavity (see Fig.1a) with a quality factor of 6124 and a Purcell factor of 6.3. Great efforts are made to find a single perfect QD at a sweet point where at 7.8 K the QD is to spatially coupled and spectrally resonant to the micropillar. At pi pulse, we detect 3.7 million single photon counts per second. The overall system efficiency is 4.6%. After correcting for detection efficiency and optical loss, we estimate that 66% of the generated single photons are collected into the first objective lens. Figure 1c summarizes the combined performance of the efficiency and single-photon purity as a function of pump power. It should be noted that the high generation and extraction efficiency are obtained with little compromise of the single-photon purity (g²(0) ≤ 0.009).

The overall system efficiency 4.6% — the highest reported in QDs — can be improved using techniques such as orthogonal excitation and detection of RF, near-unity-efficiency superconducting nanowire single-photon detection, and antireflection coatings of the optical elements. At this stage already, the performance of the single-photon source is already about ten time brighter than the triggered single-photon source used in eight-photon entanglement, but requires a pump power that is 7 orders of magnitude lower.

Figure 2: Quantum interference between two single photons separated by ~13 ns where the photon polarization set at cross (a) and parallel (b). A zoom-in near the zero time delay is shown in (c).

Another crucial demand is that the photons should possess a high degree of indistinguishability. We note that the pulsed resonant excitation is more critically needed for QDs with large Purcell factors where the reduced radiative lifetime approaches the time jitter. The single photons' indistinguishability is tested using two-photon Hong-Ou-Mandel interference. Figure 2a and 2b show histograms of normalized two-photon counts for orthogonal and parallel polarization at an emission time separation of ~13 ns, respectively. An almost vanishing zero-delay peak is observed for two photons with identical polarization (see Fig. 2c for a zoom-in). We obtain a degrees of indistinguishability to be 0.978.

Such a single-photon source can be readily used to perform multi-photon experiments on a solid-state platform. Immediate applications include implementation of Boson sampling [7] — an intermediate quantum computation where it is estimated that with 20-30 single photons one can demonstrate complex tasks that is difficult for classical computers. In addition to the photonic applications, the high-efficiency fluorescence extraction would also allow a fast high-fidelity single-shot readout of single electron spins, and efficiently entangling distant QD spins.

References:
[1] Jian-Wei Pan, Zeng-Bing Chen, Chao-Yang Lu, Harald Weinfurter, Anton Zeilinger, Marek Żukowski, "Multi-photon entanglement and interferometry", Review of Modern Physics, 84, 777–838 (2012). Abstract.
[2] Xing-Can Yao, Tian-Xiong Wang, Ping Xu, He Lu, Ge-Sheng Pan, Xiao-Hui Bao, Cheng-Zhi Peng, Chao-Yang Lu, Yu-Ao Chen, Jian-Wei Pan, "Observation of eight-photon entanglement", Nature Photonics, 6, 225–228 (2012). Abstract.
[3] Charles Santori, David Fattal, Jelena Vučković, Glenn S. Solomon, Yoshihisa Yamamoto, "Indistinguishable photons from a single-photon device", Nature, 419, 594–597 (2002). Abstract.
[4] Stefan Strauf, Nick G. Stoltz, Matthew T. Rakher, Larry A. Coldren, Pierre M. Petroff, Dirk Bouwmeester, "High-frequency single-photon source with polarization control", Nature Photonics, 1, 704 (2007). Abstract.
[5] Yu-Ming He, Yu He, Yu-Jia Wei, Dian Wu, Mete Atatüre, Christian Schneider, Sven Höfling, Martin Kamp, Chao-Yang Lu, Jian-Wei Pan, "On-demand semiconductor single-photon source with near-unity indistinguishability", Nature Nanotechnology, 8, 213–217 (2013). Abstract.
[6] Xing Ding, Yu He, Z.-C. Duan, Niels Gregersen, M.-C. Chen, S. Unsleber, S. Maier, Christian Schneider, Martin Kamp, Sven Höfling, Chao-Yang Lu, Jian-Wei Pan, "On-Demand Single Photons with High Extraction Efficiency and Near-Unity Indistinguishability from a Resonantly Driven Quantum Dot in a Micropillar", Physical Review Letters, 116, 020401 (2016). Abstract.
[7] Scott Aaronson, Alex Arkhipov, The computational complexity of linear optics, Proceedings of the 43rd annual ACM symposium on Theory of computing, 2011, San Jose (ACM, New York, 2011), p. 333. Full Article.

A Continuously Pumped Reservoir of Ultracold Atoms

From Left to Right: (top row) Jan Mahnke, Ilka Kruse, Andreas Hüper, (bottom row) Wolfgang Ertmer, Jan Arlt, Carsten Klempt.

Authors: Jan Mahnke¹, Ilka Kruse¹, Andreas Hüper¹, Stefan Jöllenbeck¹, Wolfgang Ertmer¹, Jan Arlt², Carsten Klempt¹

Affiliation:
¹Institut für Quantenoptik, Gottfried Wilhelm Leibniz Universität Hannover, Germany
²Institut for Fysik og Astronomi, Aarhus Universitet, Aarhus C, Denmark

During the last decades, the quantum regime could be accessed through very different systems, including trapped ions, micromechanical oscillators, superconducting circuits and dilute ultracold gases. Mostly, these systems show the desired quantum-mechanical features at ultra-low temperatures only. The lowest temperatures [1] to date are reached in dilute atomic gases by a combination of laser cooling and evaporative cooling. This approach has two disadvantages: It relies on the internal structure of the atoms due to the laser cooling and it can only cool discrete samples instead of continuous beams due to the evaporative cooling.

However, many applications would greatly benefit from a continuous source of cold atoms, for example sympathetic cooling [2] of molecules. Here, the molecules are brought into contact with a bath of cold atoms to redistribute the thermal energy through collisions. Ideally, such a cold bath is realized in absence of disturbing laser light as the rich internal structure of the molecules results in a broad absorption spectrum and any photon can potentially harm the cooling process. Another application of a continuous beam of cold atoms is continuous matter interferometry. Atom interferometry is already in use for the precise measurement of many observables, including time [3], gravity [4] and rotation [5]. These measurements could greatly benefit from a continuous observation instead of the sequential interrogation of discrete samples. Even though continuous sources are highly desired, no continuous sources without the application of laser light have been demonstrated in the microkelvin regime yet.

One possible realization of an ultracold continuous sample was proposed theoretically [6], where a conservative and static trap is loaded by an atomic beam. In this scheme, pre-cooled atoms are guided towards the entrance barrier of an elongated trap with a finite trap depth (see figure 1). If the atoms pass the entrance barrier, they follow the elongated potential until they are reflected by the end of the trap. The strong confinement in the radial direction ensures that most atoms collide with another atom before they reach the entrance barrier again. These collisions allow for a redistribution of the kinetic energy. Consequently, some atoms acquire a kinetic energy larger than the trap depth and escape the trap. Other atoms lose energy and stay trapped. If the trap parameters are chosen well, an equilibrium condition with a surprisingly large phase-space density may be reached.

Figure 1: 3D plot of the static trapping potential in the x–z-plane through the point of the trap minimum.

In our recent publication [7], we demonstrate the first experimental implementation of such a continuous loading of a conservative trap. Our realization is based on a mesoscopic atom chip (see figure 2 and Ref. [8]), a planar structure of millimeter-sized wires. The mesoscopic chip generates the magnetic fields for a three-dimensional magneto-optical trap, a magnetic waveguide and the static trapping potential described above. The three-dimensional magneto-optical trap is periodically loaded with an ensemble of atoms. These ensembles are launched into the magnetic waveguide, where they overlap and produce a continuous atom beam with varying intensity. This beam traverses an aperture which optically isolates the loading region from the static trapping region. In this trapping region, the atom beam is directed onto the elongated magnetic trap, where the atoms accumulate.

Figure 2: Photograph of the mesoscopic atom chip with millimeter-scale wires. The magneto-optical trap is in the lower left area and the static trap is generated in the top right area. The bend wires create a guide connecting the two regions.

With this loading scheme, we create and maintain an atomic reservoir with a total number of 3.8 × 10⁷ trapped atoms at a temperature of 102 µK, corresponding to a peak phase-space density of 9 × 10^-8 h^-3. This is the first continuously loaded cloud in the microkelvin regime without the application of laser light. Such a continuously replenished ensemble of ultracold atoms presents a new tool for metrological tasks and for the sympathetic cooling of other atomic species, molecules or nanoscopic solid state systems. The scheme is also very versatile in creating cold samples of atoms and molecules directly, as it does not rely on any internal level structure.

Figure 3 Photograph of the experimental setup. The atom chip is visible outside of the vacuum chamber at the top. In the front is the glass cell and the optics for the two-dimensional magneto-optical trap, which is used to load the three dimensional magneto-optical trap on the chip.

References:
[1] A. E. Leanhardt, T. A. Pasquini, M. Saba, A. Schirotzek, Y. Shin, D. Kielpinski, D. E. Pritchard,  W. Ketterle. “Cooling Bose-Einstein condensates below 500 picokelvin”, Science, 301, 1513 (2003). Abstract.
[2] Wade G. Rellergert, Scott T. Sullivan, Svetlana Kotochigova, Alexander Petrov, Kuang Chen, Steven J. Schowalter, Eric R. Hudson, “Measurement of a large chemical reaction rate between ultracold closed-shell ⁴⁰Ca atoms and open-shell ¹⁷⁴Yb⁺ ions held in a hybrid atom-ion trap”, Physical Review Letters, 107, 243201 (2011). Abstract.
[3] R. Wynands and S. Weyers. “Atomic fountain clocks”, Metrologia, 42 (3), S64 (2005). Abstract.
[4] A. Louchet-Chauvet, S. Merlet, Q. Bodart, A. Landragin, F. Pereira Dos Santos, H. Baumann, G. D'Agostino, C. Origlia, “Comparison of 3 absolute gravimeters based on different methods for the e-MASS project”, Instrumentation and Measurement, IEEE Transactions on, 60(7), 2527-2532 (2011). Abstract.
[5] J. K. Stockton, K. Takase, and M. A. Kasevich. “Absolute geodetic rotation measurement using atom interferometry”, Physical Review Letters, 107, 133001 (2011). Abstract.
[6] C. F. Roos, P. Cren, D. Guéry-Odelin, and J. Dalibard. “Continuous loading of a non-dissipative atom trap”, Europhysics Letters, 61, 187 (2003). Abstract.
[7] J. Mahnke, I. Kruse, A. Hüper, S. Jöllenbeck, W. Ertmer, J. Arlt, C. Klempt. “A continuously pumped reservoir of ultracold atoms”, Journal of Physics B: Atomic Molecular and Optical Physics, 48, 165301 (2015). Abstract.
[8] S. Jöllenbeck, J. Mahnke, R. Randoll, W. Ertmer, J. Arlt, C. Klempt. “Hexapole-compensated magneto-optical trap on a mesoscopic atom chip”, Physical Review A, 83, 043406 (2011). Abstract.

Metamaterial Shrinks Integrated-Photonics Devices

Rajesh Menon

Author: Rajesh Menon

Affiliation: Department of Electrical & Computer Engineering, University of Utah, USA.

Integrated electronics is the driving force behind the information revolution of the last 6 decades. A similar revolution is happening in photonics, where devices that manipulate the flow of light (or photons) are being miniaturized and integrated. The main challenge for integrated photonics is that the wavelength of light is far larger than the equivalent wavelength of electrons. This is the main reason that devices fundamental to integrated electronics are significantly smaller than those used in integrated photonics. Furthermore, no one had come up with a way to design devices close to this limit for integrated photonics.

We recently solved this problem by first coming up with a new design algorithm and then experimentally verifying that our devices work as intended [1-3]. One crucial advantage of our method is that our fabrication process is completely compatible with the very mature processes already developed for silicon electronics. This means that we can exploit the vast existing manufacturing infrastructure to enable integrated photonics.

In our recent publication, we demonstrated the smallest polarization beam-splitter to date [1]. This device (shown in the figure below) has 1 input and 2 outputs. The 2 outputs correspond to the 2 linear polarization states of light. The device is designed to take either polarization of light (or both) as the input and separate the 2 polarizations into the 2 outputs. We input light into our device one polarization at a time and measured the transmission efficiency into the correct output. This allowed us to verify that the device performs as designed. This is analogous to separating two channels of communication (for example, a video stream from PBS and another from Netflix). Previously such separation would have required time and power-consuming electronics or if photonics devices were used, they would have been much larger (so much harder to integrate onto a chip).

Figure 1: (a) Scanning-electron micrograph of fabricated polarization beamsplitter. Simulated intensity distribution at (b) TE polarization and (c) TM polarization showing the separation of the beams.

In the big picture, our research has the potential to maintain Moore's law for photonics. By enabling integrated photonics devices to be much smaller (in fact, close to their theoretical limits), we allow the integration of more devices in the same area (which increases functionality) and also enable the devices to communicate faster (since they are closer together; light has to travel shorter distances). Finally, by packing more devices into the same chip, one also exploits economies of scale to reduce the cost per chip (similar to what has happened in electronics). The practical impact for customers is that one can expect to drastically reduce power consumption and enable faster communications and computing. Data centers today consume over 2% of the total global electricity. Reducing power consumption in data centers and other electronics can go a long way to reduce our CO2 emissions and stem global climate change.

Our vision is to create a library of ultra-compact devices (including beamsplitters, but also other devices) that can then be all connected together in a variety of different ways to enable both optical computing and communications. The first devices were fabricated at a University. Next, we need to fabricate these in a standard process at a company, and then provide this library of devices to designers and hopefully, unleash their creativity. We believe that these devices will usher in unpredictable, but unbelievably exciting applications.

References :
[1] Bing Shen, Peng Wang, Randy Polson, Rajesh Menon, “An integrated-nanophotonic polarization beamsplitter with 2.4 × 2.4 μm² footprint”, Nature Photonics, 9, 378-382 (2015). Abstract.
[2] Bing Shen, Peng Wang, Randy Polson, Rajesh Menon, “Integrated metamaterials for efficient, compact free-space-to-waveguide coupling”, Optics Express, 22, 27175-27182 (2014). Abstract.
[3] Bing Shen, Randy Polson, Rajesh Menon, “Integrated digital metamaterials enables ultra-compact optical diodes”, Optics Express, 23, 10847-10855 (2015). Abstract.

Site-Dependent Evolution of Electrical Conductance from Tunneling to Atomic Point Contact

Howon Kim (left) and Yukio Hasegawa

Authors: Howon Kim and Yukio Hasegawa

Affiliation: The Institute for Solid State Physics, University of Tokyo, Japan.

In our recent work [1], the evolution of electrical conductance was investigated from tunneling to atomic point contact whose atomic geometry was precisely defined using scanning tunneling microscopy (STM). We found that the conductance evolution depends on the contact site; for instance, on-top, bridge, or hollow ('hexagonal close packed, hcp' and 'face centered cubic, fcc') site in the close-packed lattice of the substrate, indicating the importance of the atomic configuration in the conductance of the atomic junctions.

Electronic conduction through atomic-sized metal contacts is of fundamental interest as a transport mechanism though the ultimately squeezed conductor [2]. Several seminal phenomena, such as quantization and step-wise variation [3, 4] in the conductance, have been reported using a method called break junction in which the conductance is measured just before the breaking moment of a nanometer-width thin wire. Since atomic geometry of the point contact cannot be controlled and the measured conductance fluctuates at every breaking, therefore, the obtained conductances are usually analyzed in a statistical manner. Scanning tunneling microscopy (STM) has also been utilized for the study of atomic point contacts, in which the contact is formed by pushing the probe tip toward the sample surface. Because of plastic deformation of the tip by the contact formation, however, quantitatively reliable and reproducible measurements have been difficult.

Figure 1. schematic showing the atomic geometry of the atomic point contacts formed at an on-top site (left) and a 3-fold hollow site (right) of a close-packed surface.

Here, in our study, making most of the capability of the atomically resolved imaging of STM, we measured the conductance of the atomic point contact in an atomically controlled manner. We first positioned the probe tip on a specific site, for instance, on-top, bridge, or hollow (fcc and hcp) site, in the crystallographic lattice of the substrate surface (Fig. 1), and then measured the conductance while moving the tip toward the substrate from tunneling to contact regimes. It is found that the conductance evolution depends significantly on the contact site. When the contact is formed, the hollow site has the largest conductance, and among the two hollow sites the hcp site is more conductive than fcc. When the tip is pulled from the contact by just 20-30 pm, a crossover occurs and the conductance at on-top site becomes the largest.

Figure 2: electrical conductance measured from tunneling (Δz = -20 pm) to contact (Δz = -60 pm) regimes. The measured conductance G is normalized by the quantum conductance G₀ given by 2e²/h (~77.5 μS). For each conductance trace, 10 traces taken at the corresponding sites marked in the atomically-resolved STM image (inset) are averaged.

The traces of the electrical conductance measured from tunneling to contact at on-top, bridge, fcc, and hcp sites of the Pb(111) surface are shown in Fig. 2. For each plot, 10 traces obtained from the corresponding marked sites in the inset STM image are averaged. At the tip displacement Δz of -50 ~ -60 pm from the tunneling (Δz = 0), the atomic contact is formed as the conductance shows saturation around the quantum conductance G₀ given by 2e²/h (~77.5 μS). The contact conductance shows strong site dependence; the conductance at the hcp site is largest and more than 50 % larger than the one measured at the on-top site. Around Δz = -30 ~ -40 pm, that is, when the tip is located above the substrate by 20 ~ 30 pm from the contact, the plot indicates the largest conductance at on-top site.

Figure 3. Spatial mappings of the conductance at various tip displacements (upper left) topographic STM image (3.0 X 3.0 nm²) taken simultaneously with 64 X 64 conductance traces. (lower left) conductance mapping at Δz = -32pm, where the largest conductance at on-top site is enhanced (Lower right) conductance mapping at Δz = -60 pm, that is, the contact regime, where hollow site, particularly hcp site, has large conductance. (upper right) schematics explaining the site dependence of the conductance. The atoms on which the chemical interaction is exerted are marked red.

In order to spatially demonstrate the site dependence, we performed real-space mappings of the conductance in the on-top enhancement region and in the contact regime. The upper-left panel of Fig. 3 is an STM image showing the atomic contrast taken simultaneously with the conductance traces. At a tip displacement Δz of -32pm, the conductance mapping (lower-left of Fig. 3) exhibits bright contrast at the on-top site, similarly to that in the topographic image. As the conductance mapping at Δz = 0 does not have any contrast, the bright contrast indicates the conductance enhancement at the on-top site. On the other hand, the conductance mapping in the contact regime (lower right of Fig. 3, Δz = -60pm) has its contrast reversed from that of the topographic one, indicating a larger conductance at the hollow site than at the on-top site. These results clearly demonstrate that the point contact conductance is quite sensitive to the atomic configuration.

When the distance between the tip and substrate is reduced, the attractive chemical interaction is exerted between the surface and tip apex atoms. This interaction presumably opens up the conduction channel and contributes to the development of the conductance. Schematics in the upper right panel of Fig. 3 show how the chemical interaction works in the case of contacts formed at on-top and hollow sites. When the tip approaches from the tunneling regime, the attractive interaction is exerted first at the on-top site (the force-exerted atoms are marked red in the schematics) because the substrate atoms are closer at on-top site than at hollow sites, thus making the on-top conductance enhanced. In the contact regime, however, the attractive force becomes stronger at hollow sites because of the greater number of involved atoms than the on-top site. This is probably the reason why conductance becomes larger there at the contact. Obviously, theoretical studies [5], simultaneous measurements of force and conductance by atomic force microscopy [6], and/or conduction channel analysis of the atomic point contact [7] are needed to elucidate the observed conductance behaviors.

References :
[1] Howon Kim and Yukio Hasegawa, "Site-dependent evolution of electrical conductance from tunneling to atomic point contact", Physical Review Letters, 114, 206801 (2015). Abstract.
[2] Nicolás Agraı̈t, Alfredo Levy Yeyati, Jan M. van Ruitenbeek, "Quantum properties of atomic-sized conductors", Physics Reports, 377, 81 (2003). Abstract.
[3] J. M. Krans, J. M. van Ruitenbeek, V. V. Fisun, I. K. Yanson, L. J. de Jongh; "The signature of conductance quantization in metallic point contacts", Nature, 375, 767 (1995). Abstract.
[4] L. Olesen, E. Lægsgaard, I. Stensgaard, F. Besenbacher, J. Schiøtz, P. Stoltze, K. W. Jacobsen, J. K. Nørskov; "Quantized conductance in an atom-sized point contact", Physical Review Letters, 72, 2251 (1994). Abstract.
[5] Jose Manuel Blanco, Cesar González, Pavel Jelínek, José Ortega, Fernando Flores, Rubén Pérez, "First-principles simulations of STM images: From tunneling to the contact regime", Physical Review B, 70, 085405 (2004). Abstract.
[6] Yoshiaki Sugimoto, Keiichi Ueda, Masayuki Abe, Seizo Morita "Three-dimensional scanning force/tunneling spectroscopy at room temperature", Journal of Physics : Condensed Matter, 24, 084008 (2012). Abstract.
[7] Howon Kim and Yukio Hasegawa, "Site-dependent conduction channel transmission in atomic-scale superconducting junctions", arXiv:1506.05528 [cond-mat.mes-hall].