-
TorchDriveEnv: A Reinforcement Learning Benchmark for Autonomous Driving with Reactive, Realistic, and Diverse Non-Playable Characters
Authors:
Jonathan Wilder Lavington,
Ke Zhang,
Vasileios Lioutas,
Matthew Niedoba,
Yunpeng Liu,
Dylan Green,
Saeid Naderiparizi,
Xiaoxuan Liang,
Setareh Dabiri,
Adam Ścibior,
Berend Zwartsenberg,
Frank Wood
Abstract:
The training, testing, and deployment, of autonomous vehicles requires realistic and efficient simulators. Moreover, because of the high variability between different problems presented in different autonomous systems, these simulators need to be easy to use, and easy to modify. To address these problems we introduce TorchDriveSim and its benchmark extension TorchDriveEnv. TorchDriveEnv is a light…
▽ More
The training, testing, and deployment, of autonomous vehicles requires realistic and efficient simulators. Moreover, because of the high variability between different problems presented in different autonomous systems, these simulators need to be easy to use, and easy to modify. To address these problems we introduce TorchDriveSim and its benchmark extension TorchDriveEnv. TorchDriveEnv is a lightweight reinforcement learning benchmark programmed entirely in Python, which can be modified to test a number of different factors in learned vehicle behavior, including the effect of varying kinematic models, agent types, and traffic control patterns. Most importantly unlike many replay based simulation approaches, TorchDriveEnv is fully integrated with a state of the art behavioral simulation API. This allows users to train and evaluate driving models alongside data driven Non-Playable Characters (NPC) whose initializations and driving behavior are reactive, realistic, and diverse. We illustrate the efficiency and simplicity of TorchDriveEnv by evaluating common reinforcement learning baselines in both training and validation environments. Our experiments show that TorchDriveEnv is easy to use, but difficult to solve.
△ Less
Submitted 7 May, 2024;
originally announced May 2024.
-
Semantically Consistent Video Inpainting with Conditional Diffusion Models
Authors:
Dylan Green,
William Harvey,
Saeid Naderiparizi,
Matthew Niedoba,
Yunpeng Liu,
Xiaoxuan Liang,
Jonathan Lavington,
Ke Zhang,
Vasileios Lioutas,
Setareh Dabiri,
Adam Scibior,
Berend Zwartsenberg,
Frank Wood
Abstract:
Current state-of-the-art methods for video inpainting typically rely on optical flow or attention-based approaches to inpaint masked regions by propagating visual information across frames. While such approaches have led to significant progress on standard benchmarks, they struggle with tasks that require the synthesis of novel content that is not present in other frames. In this paper we reframe…
▽ More
Current state-of-the-art methods for video inpainting typically rely on optical flow or attention-based approaches to inpaint masked regions by propagating visual information across frames. While such approaches have led to significant progress on standard benchmarks, they struggle with tasks that require the synthesis of novel content that is not present in other frames. In this paper we reframe video inpainting as a conditional generative modeling problem and present a framework for solving such problems with conditional video diffusion models. We highlight the advantages of using a generative approach for this task, showing that our method is capable of generating diverse, high-quality inpaintings and synthesizing new content that is spatially, temporally, and semantically consistent with the provided context.
△ Less
Submitted 30 April, 2024;
originally announced May 2024.
-
Nearest Neighbour Score Estimators for Diffusion Generative Models
Authors:
Matthew Niedoba,
Dylan Green,
Saeid Naderiparizi,
Vasileios Lioutas,
Jonathan Wilder Lavington,
Xiaoxuan Liang,
Yunpeng Liu,
Ke Zhang,
Setareh Dabiri,
Adam Ścibior,
Berend Zwartsenberg,
Frank Wood
Abstract:
Score function estimation is the cornerstone of both training and sampling from diffusion generative models. Despite this fact, the most commonly used estimators are either biased neural network approximations or high variance Monte Carlo estimators based on the conditional score. We introduce a novel nearest neighbour score function estimator which utilizes multiple samples from the training set…
▽ More
Score function estimation is the cornerstone of both training and sampling from diffusion generative models. Despite this fact, the most commonly used estimators are either biased neural network approximations or high variance Monte Carlo estimators based on the conditional score. We introduce a novel nearest neighbour score function estimator which utilizes multiple samples from the training set to dramatically decrease estimator variance. We leverage our low variance estimator in two compelling applications. Training consistency models with our estimator, we report a significant increase in both convergence speed and sample quality. In diffusion models, we show that our estimator can replace a learned network for probability-flow ODE integration, opening promising new avenues of future research.
△ Less
Submitted 16 July, 2024; v1 submitted 12 February, 2024;
originally announced February 2024.
-
A Diffusion-Model of Joint Interactive Navigation
Authors:
Matthew Niedoba,
Jonathan Wilder Lavington,
Yunpeng Liu,
Vasileios Lioutas,
Justice Sefas,
Xiaoxuan Liang,
Dylan Green,
Setareh Dabiri,
Berend Zwartsenberg,
Adam Scibior,
Frank Wood
Abstract:
Simulation of autonomous vehicle systems requires that simulated traffic participants exhibit diverse and realistic behaviors. The use of prerecorded real-world traffic scenarios in simulation ensures realism but the rarity of safety critical events makes large scale collection of driving scenarios expensive. In this paper, we present DJINN - a diffusion based method of generating traffic scenario…
▽ More
Simulation of autonomous vehicle systems requires that simulated traffic participants exhibit diverse and realistic behaviors. The use of prerecorded real-world traffic scenarios in simulation ensures realism but the rarity of safety critical events makes large scale collection of driving scenarios expensive. In this paper, we present DJINN - a diffusion based method of generating traffic scenarios. Our approach jointly diffuses the trajectories of all agents, conditioned on a flexible set of state observations from the past, present, or future. On popular trajectory forecasting datasets, we report state of the art performance on joint trajectory metrics. In addition, we demonstrate how DJINN flexibly enables direct test-time sampling from a variety of valuable conditional distributions including goal-based sampling, behavior-class sampling, and scenario editing.
△ Less
Submitted 24 October, 2023; v1 submitted 21 September, 2023;
originally announced September 2023.
-
Nature of the current-induced insulator-to-metal transition in Ca$_2$RuO$_4$ as revealed by transport-ARPES
Authors:
Cissy T Suen,
Igor Marković,
Marta Zonno,
Niclas Heinsdorf,
Sergey Zhdanovich,
Na-Hyun Jo,
Michael Schmid,
Philipp Hansmann,
Pascal Puphal,
Katrin Fürsich,
Valentin Zimmerman,
Steef Smit,
Christine Au-Yeung,
Berend Zwartsenberg,
Maximilian Krautloher,
Ilya S Elfimov,
Roland Koch,
Sergey Gorovikov,
Chris Jozwiak,
Aaron Bostwick,
Marcel Franz,
Eli Rotenberg,
Bernhard Keimer,
Andrea Damascelli
Abstract:
The Mott insulator Ca$_2$RuO$_4$ exhibits a rare insulator-to-metal transition (IMT) induced by DC current. While structural changes associated with this transition have been tracked by neutron diffraction, Raman scattering, and x-ray spectroscopy, work on elucidating the response of the electronic degrees of freedom is still in progress. Here we unveil the current-induced modifications of the ele…
▽ More
The Mott insulator Ca$_2$RuO$_4$ exhibits a rare insulator-to-metal transition (IMT) induced by DC current. While structural changes associated with this transition have been tracked by neutron diffraction, Raman scattering, and x-ray spectroscopy, work on elucidating the response of the electronic degrees of freedom is still in progress. Here we unveil the current-induced modifications of the electronic states of Ca$_2$RuO$_4$ by employing angle-resolved photoemission spectroscopy (ARPES) in conjunction with four-probe transport. Two main effects emerge: a clear reduction of the Mott gap and a modification in the dispersion of the Ru-bands. The changes in dispersion occur exclusively along the $XM$ high-symmetry direction, parallel to the $b$-axis where the greatest in-plane lattice change occurs. These experimental observations, together with dynamical mean-field theory (DMFT) calculations simulated from the current-induced structural distortions, indicate the intimate interplay of lattice and orbital-dependent electronic response in the current-driven IMT. Furthermore, based on a free energy analysis, we demonstrate that the current-induced phase, albeit thermodynamically equivalent, is electronically distinct from the high-temperature zero-current metallic phase. Our results provide insight into the elusive nature of the current-induced IMT of Ca$_2$RuO$_4$ and advance the challenging, yet powerful, technique of transport-ARPES.
△ Less
Submitted 6 July, 2024; v1 submitted 10 August, 2023;
originally announced August 2023.
-
Don't be so negative! Score-based Generative Modeling with Oracle-assisted Guidance
Authors:
Saeid Naderiparizi,
Xiaoxuan Liang,
Berend Zwartsenberg,
Frank Wood
Abstract:
The maximum likelihood principle advocates parameter estimation via optimization of the data likelihood function. Models estimated in this way can exhibit a variety of generalization characteristics dictated by, e.g. architecture, parameterization, and optimization bias. This work addresses model learning in a setting where there further exists side-information in the form of an oracle that can la…
▽ More
The maximum likelihood principle advocates parameter estimation via optimization of the data likelihood function. Models estimated in this way can exhibit a variety of generalization characteristics dictated by, e.g. architecture, parameterization, and optimization bias. This work addresses model learning in a setting where there further exists side-information in the form of an oracle that can label samples as being outside the support of the true data generating distribution. Specifically we develop a new denoising diffusion probabilistic modeling (DDPM) methodology, Gen-neG, that leverages this additional side-information. Our approach builds on generative adversarial networks (GANs) and discriminator guidance in diffusion models to guide the generation process towards the positive support region indicated by the oracle. We empirically establish the utility of Gen-neG in applications including collision avoidance in self-driving simulators and safety-guarded human motion generation.
△ Less
Submitted 31 July, 2023;
originally announced July 2023.
-
Realistically distributing object placements in synthetic training data improves the performance of vision-based object detection models
Authors:
Setareh Dabiri,
Vasileios Lioutas,
Berend Zwartsenberg,
Yunpeng Liu,
Matthew Niedoba,
Xiaoxuan Liang,
Dylan Green,
Justice Sefas,
Jonathan Wilder Lavington,
Frank Wood,
Adam Scibior
Abstract:
When training object detection models on synthetic data, it is important to make the distribution of synthetic data as close as possible to the distribution of real data. We investigate specifically the impact of object placement distribution, keeping all other aspects of synthetic data fixed. Our experiment, training a 3D vehicle detection model in CARLA and testing on KITTI, demonstrates a subst…
▽ More
When training object detection models on synthetic data, it is important to make the distribution of synthetic data as close as possible to the distribution of real data. We investigate specifically the impact of object placement distribution, keeping all other aspects of synthetic data fixed. Our experiment, training a 3D vehicle detection model in CARLA and testing on KITTI, demonstrates a substantial improvement resulting from improving the object placement distribution.
△ Less
Submitted 23 May, 2023;
originally announced May 2023.
-
Video Killed the HD-Map: Predicting Multi-Agent Behavior Directly From Aerial Images
Authors:
Yunpeng Liu,
Vasileios Lioutas,
Jonathan Wilder Lavington,
Matthew Niedoba,
Justice Sefas,
Setareh Dabiri,
Dylan Green,
Xiaoxuan Liang,
Berend Zwartsenberg,
Adam Ścibior,
Frank Wood
Abstract:
The development of algorithms that learn multi-agent behavioral models using human demonstrations has led to increasingly realistic simulations in the field of autonomous driving. In general, such models learn to jointly predict trajectories for all controlled agents by exploiting road context information such as drivable lanes obtained from manually annotated high-definition (HD) maps. Recent stu…
▽ More
The development of algorithms that learn multi-agent behavioral models using human demonstrations has led to increasingly realistic simulations in the field of autonomous driving. In general, such models learn to jointly predict trajectories for all controlled agents by exploiting road context information such as drivable lanes obtained from manually annotated high-definition (HD) maps. Recent studies show that these models can greatly benefit from increasing the amount of human data available for training. However, the manual annotation of HD maps which is necessary for every new location puts a bottleneck on efficiently scaling up human traffic datasets. We propose an aerial image-based map (AIM) representation that requires minimal annotation and provides rich road context information for traffic agents like pedestrians and vehicles. We evaluate multi-agent trajectory prediction using the AIM by incorporating it into a differentiable driving simulator as an image-texture-based differentiable rendering module. Our results demonstrate competitive multi-agent trajectory prediction performance especially for pedestrians in the scene when using our AIM representation as compared to models trained with rasterized HD maps.
△ Less
Submitted 19 September, 2023; v1 submitted 19 May, 2023;
originally announced May 2023.
-
Conditional Permutation Invariant Flows
Authors:
Berend Zwartsenberg,
Adam Ścibior,
Matthew Niedoba,
Vasileios Lioutas,
Yunpeng Liu,
Justice Sefas,
Setareh Dabiri,
Jonathan Wilder Lavington,
Trevor Campbell,
Frank Wood
Abstract:
We present a novel, conditional generative probabilistic model of set-valued data with a tractable log density. This model is a continuous normalizing flow governed by permutation equivariant dynamics. These dynamics are driven by a learnable per-set-element term and pairwise interactions, both parametrized by deep neural networks. We illustrate the utility of this model via applications including…
▽ More
We present a novel, conditional generative probabilistic model of set-valued data with a tractable log density. This model is a continuous normalizing flow governed by permutation equivariant dynamics. These dynamics are driven by a learnable per-set-element term and pairwise interactions, both parametrized by deep neural networks. We illustrate the utility of this model via applications including (1) complex traffic scene generation conditioned on visually specified map information, and (2) object bounding box generation conditioned directly on images. We train our model by maximizing the expected likelihood of labeled conditional data under our flow, with the aid of a penalty that ensures the dynamics are smooth and hence efficiently solvable. Our method significantly outperforms non-permutation invariant baselines in terms of log likelihood and domain-specific metrics (offroad, collision, and combined infractions), yielding realistic samples that are difficult to distinguish from real data.
△ Less
Submitted 17 June, 2022;
originally announced June 2022.
-
Critic Sequential Monte Carlo
Authors:
Vasileios Lioutas,
Jonathan Wilder Lavington,
Justice Sefas,
Matthew Niedoba,
Yunpeng Liu,
Berend Zwartsenberg,
Setareh Dabiri,
Frank Wood,
Adam Scibior
Abstract:
We introduce CriticSMC, a new algorithm for planning as inference built from a composition of sequential Monte Carlo with learned Soft-Q function heuristic factors. These heuristic factors, obtained from parametric approximations of the marginal likelihood ahead, more effectively guide SMC towards the desired target distribution, which is particularly helpful for planning in environments with hard…
▽ More
We introduce CriticSMC, a new algorithm for planning as inference built from a composition of sequential Monte Carlo with learned Soft-Q function heuristic factors. These heuristic factors, obtained from parametric approximations of the marginal likelihood ahead, more effectively guide SMC towards the desired target distribution, which is particularly helpful for planning in environments with hard constraints placed sparsely in time. Compared with previous work, we modify the placement of such heuristic factors, which allows us to cheaply propose and evaluate large numbers of putative action particles, greatly increasing inference and planning efficiency. CriticSMC is compatible with informative priors, whose density function need not be known, and can be used as a model-free control algorithm. Our experiments on collision avoidance in a high-dimensional simulated driving task show that CriticSMC significantly reduces collision rates at a low computational cost while maintaining realism and diversity of driving behaviors across vehicles and environment scenarios.
△ Less
Submitted 21 January, 2023; v1 submitted 30 May, 2022;
originally announced May 2022.
-
Constraints on the two-dimensional pseudo-spin 1/2 Mott insulator description of Sr$_2$IrO$_4$
Authors:
Berend Zwartsenberg,
Ryan P. Day,
Elia Razzoli,
Matteo Michiardi,
Mengxing Na,
Guoren Zhang,
Jonathan D. Denlinger,
Ivana Vobornik,
Chiara Bigi,
Bumjoon Kim,
Ilya S. Elfimov,
Eva Pavarini,
Andrea Damascelli
Abstract:
Sr$_{2}$IrO$_{4}$ has often been described via a simple, one-band pseudo-spin 1/2 model, subject to electron-electron interactions, on a square lattice, fostering analogies with cuprate superconductors, believed to be well described by a similar model. In this work we argue - based on a detailed study of the low-energy electronic structure by circularly polarized spin and angle-resolved photoemiss…
▽ More
Sr$_{2}$IrO$_{4}$ has often been described via a simple, one-band pseudo-spin 1/2 model, subject to electron-electron interactions, on a square lattice, fostering analogies with cuprate superconductors, believed to be well described by a similar model. In this work we argue - based on a detailed study of the low-energy electronic structure by circularly polarized spin and angle-resolved photoemission spectroscopy combined with dynamical mean-field theory calculations - that a pseudo-spin 1/2 model fails to capture the full complexity of the system. We show instead that a realistic multi-band Hubbard Hamiltonian, accounting for the full correlated $t_{2g}$ manifold, provides a detailed description of the interplay between spin-orbital entanglement and electron-electron interactions, and yields quantitative agreement with experiments. Our analysis establishes that the $j_{3/2}$ states make up a substantial percentage of the low energy spectral weight, i.e. approximately 74% as determined from the integration of the $j$-resolved spectral function in the $0$ to $-1.64$ eV energy range. The results in our work are not only of relevance to iridium based materials, but more generally to the study of multi-orbital materials with closely spaced energy scales.
△ Less
Submitted 25 May, 2022;
originally announced May 2022.
-
The Three-Dimensional Electronic Structure of LiFeAs: Strong-coupling Superconductivity and Topology in the Iron Pnictides
Authors:
Ryan P. Day,
MengXing Na,
Manuel Zingl,
Berend Zwartsenberg,
Matteo Michiardi,
Giorgio Levy,
Michael Schneider,
Doug Wong,
Pinder Dosanjh,
Tor M. Pedersen,
Sergey Gorovikov,
Shun Chi,
Ruixing Liang,
Walter N. Hardy,
Douglas A. Bonn,
Sergey Zhdanovich,
Ilya S. Elfimov,
Andrea Damascelli
Abstract:
Amongst the iron-based superconductors, LiFeAs is unrivalled in the simplicity of its crystal structure and phase diagram. However, our understanding of this canonical compound suffers from conflict between mutually incompatible descriptions of the material's electronic structure, as derived from contradictory interpretations of the photoemission record. Here, we explore the challenge of interpret…
▽ More
Amongst the iron-based superconductors, LiFeAs is unrivalled in the simplicity of its crystal structure and phase diagram. However, our understanding of this canonical compound suffers from conflict between mutually incompatible descriptions of the material's electronic structure, as derived from contradictory interpretations of the photoemission record. Here, we explore the challenge of interpretation in such experiments. By combining comprehensive photon energy- and polarization- dependent angle-resolved photoemission spectroscopy (ARPES) measurements with numerical simulations, we establish the providence of several contradictions in the present understanding of this and related materials. We identify a confluence of surface-related issues which have precluded unambiguous identification of both the number and dimensionality of the Fermi surface sheets. Ultimately, we arrive at a scenario which supports indications of topologically non-trivial states, while also being incompatible with superconductivity as a spin-fluctuation driven Fermi surface instability.
△ Less
Submitted 27 September, 2021;
originally announced September 2021.
-
Establishing non-thermal regimes in pump-probe electron-relaxation dynamics
Authors:
MengXing Na,
Fabio Boschini,
Arthur K. Mills,
Matteo Michiardi,
Ryan P. Day,
Berend Zwartsenberg,
Giorgio Levy,
Sergey Zhdanovich,
Alexander F. Kemper,
David J. Jones,
Andrea Damascelli
Abstract:
Time- and angle-resolved photoemission spectroscopy (TR-ARPES) accesses the electronic structure of solids under optical excitation, and is a powerful technique for studying the coupling between electrons and collective modes. One approach to infer electron-boson coupling is through the relaxation dynamics of optically-excited electrons, and the characteristic timescales of energy redistribution.…
▽ More
Time- and angle-resolved photoemission spectroscopy (TR-ARPES) accesses the electronic structure of solids under optical excitation, and is a powerful technique for studying the coupling between electrons and collective modes. One approach to infer electron-boson coupling is through the relaxation dynamics of optically-excited electrons, and the characteristic timescales of energy redistribution. A common description of electron relaxation dynamics is through the effective electronic temperature. Such a description requires that thermodynamic quantities are well-defined, an assumption that is generally violated at early delays. Additionally, precise estimation of the non-thermal window -- within which effective temperature models may not be applied -- is challenging. We perform TR-ARPES on graphite and show that Boltzmann rate equations can be used to calculate the time-dependent electronic occupation function, and reproduce experimental features given by non-thermal electron occupation. Using this model, we define a quantitative measure of non-thermal electron occupation and use it to define distinct phases of electron relaxation in the fluence-delay phase space. More generally, this approach can be used to inform the non-thermal-to-thermal crossover in pump-probe experiments.
△ Less
Submitted 10 September, 2020;
originally announced September 2020.
-
Probabilistic Surrogate Networks for Simulators with Unbounded Randomness
Authors:
Andreas Munk,
Berend Zwartsenberg,
Adam Ścibior,
Atılım Güneş Baydin,
Andrew Stewart,
Goran Fernlund,
Anoush Poursartip,
Frank Wood
Abstract:
We present a framework for automatically structuring and training fast, approximate, deep neural surrogates of stochastic simulators. Unlike traditional approaches to surrogate modeling, our surrogates retain the interpretable structure and control flow of the reference simulator. Our surrogates target stochastic simulators where the number of random variables itself can be stochastic and potentia…
▽ More
We present a framework for automatically structuring and training fast, approximate, deep neural surrogates of stochastic simulators. Unlike traditional approaches to surrogate modeling, our surrogates retain the interpretable structure and control flow of the reference simulator. Our surrogates target stochastic simulators where the number of random variables itself can be stochastic and potentially unbounded. Our framework further enables an automatic replacement of the reference simulator with the surrogate when undertaking amortized inference. The fidelity and speed of our surrogates allow for both faster stochastic simulation and accurate and substantially faster posterior inference. Using an illustrative yet non-trivial example we show our surrogates' ability to accurately model a probabilistic program with an unbounded number of random variables. We then proceed with an example that shows our surrogates are able to accurately model a complex structure like an unbounded stack in a program synthesis example. We further demonstrate how our surrogate modeling technique makes amortized inference in complex black-box simulators an order of magnitude faster. Specifically, we do simulator-based materials quality testing, inferring safety-critical latent internal temperature profiles of composite materials undergoing curing.
△ Less
Submitted 20 January, 2023; v1 submitted 25 October, 2019;
originally announced October 2019.
-
Computational Framework for Angle-Resolved Photoemission Spectroscopy
Authors:
Ryan P. Day,
Berend Zwartsenberg,
Ilya S. Elfimov,
Andrea Damascelli
Abstract:
We have developed the numerical software package $chinook$, designed for the simulation of photoemission matrix elements. This quantity encodes a depth of information regarding the orbital structure of the underlying wavefunctions from which photoemission occurs. Extraction of this information is often nontrivial, owing to the influence of the experimental geometry and photoelectron interference,…
▽ More
We have developed the numerical software package $chinook$, designed for the simulation of photoemission matrix elements. This quantity encodes a depth of information regarding the orbital structure of the underlying wavefunctions from which photoemission occurs. Extraction of this information is often nontrivial, owing to the influence of the experimental geometry and photoelectron interference, precluding straightforward solutions. The $chinook$ code has been designed to simulate and predict the ARPES intensity measured for arbitrary experimental configuration, including photon-energy, polarization and spin-projection, as well as consideration of both surface-projected slab and bulk models. This framework then facilitates an efficient interpretation of the photoemission experiment, allowing for a deeper understanding of the electronic structure in addition to the design of new experiments which leverage the matrix element effects towards the objective of selective photoemission from states of particular interest.
△ Less
Submitted 30 September, 2019; v1 submitted 25 September, 2019;
originally announced September 2019.
-
Spin-orbit-controlled metal-insulator transition in Sr$_2$IrO$_4$
Authors:
Berend Zwartsenberg,
Ryan P. Day,
Elia Razzoli,
Matteo Michiardi,
Nan Xu,
Ming Shi,
Jonathan D. Denlinger,
Guixin Cao,
Stuart Calder,
Kentaro Ueda,
Joel Bertinshaw,
Hidenori Takagi,
Bumjoon Kim,
Ilya S. Elfimov,
Andrea Damascelli
Abstract:
In the context of correlated insulators, where electron-electron interactions (U) drive the localization of charge carriers, the metal-insulator transition (MIT) is described as either bandwidth (BC) or filling (FC) controlled. Motivated by the challenge of the insulating phase in Sr$_2$IrO$_4$, a new class of correlated insulators has been proposed, in which spin-orbit coupling (SOC) is believed…
▽ More
In the context of correlated insulators, where electron-electron interactions (U) drive the localization of charge carriers, the metal-insulator transition (MIT) is described as either bandwidth (BC) or filling (FC) controlled. Motivated by the challenge of the insulating phase in Sr$_2$IrO$_4$, a new class of correlated insulators has been proposed, in which spin-orbit coupling (SOC) is believed to renormalize the bandwidth of the half-filled $j_{\mathrm{eff}} = 1/2$ doublet, allowing a modest U to induce a charge-localized phase. Although this framework has been tacitly assumed, a thorough characterization of the ground state has been elusive. Furthermore, direct evidence for the role of SOC in stabilizing the insulating state has not been established, since previous attempts at revealing the role of SOC have been hindered by concurrently occurring changes to the filling. We overcome this challenge by employing multiple substituents that introduce well defined changes to the signatures of SOC and carrier concentration in the electronic structure, as well as a new methodology that allows us to monitor SOC directly. Specifically, we study Sr$_2$Ir$_{1-x}$T$_x$O$_4$ (T = Ru, Rh) by angle-resolved photoemission spectroscopy (ARPES), combined with ab-initio and supercell tight-binding calculations. This allows us to distinguish relativistic and filling effects, thereby establishing conclusively the central role of SOC in stabilizing the insulating state of Sr$_2$IrO$_4$. Most importantly, we estimate the critical value for spin-orbit coupling in this system to be $λ_c = 0.42 \pm 0.01$ eV, and provide the first demonstration of a spin-orbit-controlled MIT.
△ Less
Submitted 30 March, 2020; v1 submitted 1 March, 2019;
originally announced March 2019.
-
Emergence of pseudogap from short-range spin-correlations in electron doped cuprates
Authors:
F. Boschini,
M. Zonno,
E. Razzoli,
R. P. Day,
M. Michiardi,
B. Zwartsenberg,
P. Nigge,
M. Schneider,
E. H. da Silva Neto,
A. Erb,
S. Zhdanovich,
A. K. Mills,
G. Levy,
C. Giannetti,
D. J. Jones,
A. Damascelli
Abstract:
Electron interactions are pivotal for defining the electronic structure of quantum materials. In particular, the strong electron Coulomb repulsion is considered the keystone for describing the emergence of exotic and/or ordered phases of quantum matter as disparate as high-temperature superconductivity and charge- or magnetic-order. However, a comprehensive understanding of fundamental electronic…
▽ More
Electron interactions are pivotal for defining the electronic structure of quantum materials. In particular, the strong electron Coulomb repulsion is considered the keystone for describing the emergence of exotic and/or ordered phases of quantum matter as disparate as high-temperature superconductivity and charge- or magnetic-order. However, a comprehensive understanding of fundamental electronic properties of quantum materials is often complicated by the appearance of an enigmatic partial suppression of low-energy electronic states, known as the pseudogap. Here we take advantage of ultrafast angle-resolved photoemission spectroscopy to unveil the temperature evolution of the low-energy density of states in the electron-doped cuprate Nd$_{\text{2-x}}$Ce$_{\text{x}}$CuO$_{\text{4}}$, an emblematic system where the pseudogap intertwines with magnetic degrees of freedom. By photoexciting the electronic system across the pseudogap onset temperature T*, we report the direct relation between the momentum-resolved pseudogap spectral features and the spin-correlation length with an unprecedented sensitivity. This transient approach, corroborated by mean field model calculations, allows us to establish the pseudogap in electron-doped cuprates as a precursor to the incipient antiferromagnetic order even when long-range antiferromagnetic correlations are not established, as in the case of optimal doping.
△ Less
Submitted 17 August, 2019; v1 submitted 18 December, 2018;
originally announced December 2018.
-
Role of matrix elements in the time-resolved photoemission signal
Authors:
F. Boschini,
D. Bugini,
M. Zonno,
M. Michiardi,
R. P. Day,
E. Razzoli,
B. Zwartsenberg,
E. H. da Silva Neto,
S. dal Conte,
S. K. Kushwaha,
R. J. Cava,
S. Zhdanovich,
A. K. Mills,
G. Levy,
E. Carpene,
C. Dallera,
C. Giannetti,
D. J. Jones,
G. Cerullo,
A. Damascelli
Abstract:
Time- and angle-resolved photoemission spectroscopy accesses the ultrafast evolution of quasiparticles and many-body interactions in solid-state systems. However, the momentum- and energy-resolved transient photoemission intensity may not be unambiguously related to the intrinsic relaxation dynamics of photoexcited electrons. In fact, interpretation of the time-dependent photoemission signal can b…
▽ More
Time- and angle-resolved photoemission spectroscopy accesses the ultrafast evolution of quasiparticles and many-body interactions in solid-state systems. However, the momentum- and energy-resolved transient photoemission intensity may not be unambiguously related to the intrinsic relaxation dynamics of photoexcited electrons. In fact, interpretation of the time-dependent photoemission signal can be affected by the transient evolution of both the one-electron removal spectral function as well as the photoemission dipole matrix elements. Here we investigate the topological insulator Bi$_{1.1}$Sb$_{0.9}$Te$_2$S to demonstrate, by means of a careful probe-polarization study, the transient contribution of matrix elements to the time-resolved photoemission signal.
△ Less
Submitted 15 October, 2018;
originally announced October 2018.
-
Stable Weyl points, trivial surface states and particle-hole compensation in WP2
Authors:
E. Razzoli,
B. Zwartsenberg,
M. Michiardi,
F. Boschini,
R. P. Day,
I. S. Elfimov,
J. D. Denlinger,
V. Süß,
C. Felser,
A. Damascelli
Abstract:
A possible connection between extremely large magneto-resistance and the presence of Weyl points has garnered much attention in the study of topological semimetals. Exploration of these concepts in transition metal phosphide WP2 has been complicated by conflicting experimental reports. Here we combine angle-resolved photoemission spectroscopy (ARPES) and density functional theory (DFT) calculation…
▽ More
A possible connection between extremely large magneto-resistance and the presence of Weyl points has garnered much attention in the study of topological semimetals. Exploration of these concepts in transition metal phosphide WP2 has been complicated by conflicting experimental reports. Here we combine angle-resolved photoemission spectroscopy (ARPES) and density functional theory (DFT) calculations to disentangle surface and bulk contributions to the ARPES intensity, the superposition of which has plagued the determination of the electronic structure in WP2. Our results show that while the hole- and electron-like Fermi surface sheets originating from surface states have different areas, the bulk-band structure of WP2 is electron-hole-compensated in agreement with DFT. Furthermore, the detailed band structure is compatible with the presence of at least 4 temperature-independent Weyl points, confirming the topological nature of WP2 and its stability against lattice distortions.
△ Less
Submitted 11 January, 2018;
originally announced January 2018.
-
Influence of Spin Orbit Coupling in the Iron-Based Superconductors
Authors:
R. P. Day,
G. Levy,
M. Michiardi,
B. Zwartsenberg,
M. Zonno,
F. Ji,
E. Razzoli,
F. Boschini,
S. Chi,
R. Liang,
P. K. Das,
I. Vobornik,
J. Fujii,
D. A. Bonn,
W. N. Hardy,
I. S. Elfimov,
A. Damascelli
Abstract:
We report on the influence of spin-orbit coupling (SOC) in the Fe-based superconductors (FeSCs) via application of circularly-polarized spin and angle-resolved photoemission spectroscopy. We combine this technique in representative members of both the Fe-pnictides and Fe-chalcogenides with ab initio density functional theory and tight-binding calculations to establish an ubiquitous modification of…
▽ More
We report on the influence of spin-orbit coupling (SOC) in the Fe-based superconductors (FeSCs) via application of circularly-polarized spin and angle-resolved photoemission spectroscopy. We combine this technique in representative members of both the Fe-pnictides and Fe-chalcogenides with ab initio density functional theory and tight-binding calculations to establish an ubiquitous modification of the electronic structure in these materials imbued by SOC. The influence of SOC is found to be concentrated on the hole pockets where the superconducting gap is generally found to be largest. This result contests descriptions of superconductivity in these materials in terms of pure spin-singlet eigenstates, raising questions regarding the possible pairing mechanisms and role of SOC therein.
△ Less
Submitted 20 July, 2018; v1 submitted 17 November, 2017;
originally announced November 2017.
-
Collapse of superconductivity in cuprates via ultrafast quenching of phase coherence
Authors:
F. Boschini,
E. H. da Silva Neto,
E. Razzoli,
M. Zonno,
S. Peli,
R. P. Day,
M. Michiardi,
M. Schneider,
B. Zwartsenberg,
P. Nigge,
R. D. Zhong,
J. Schneeloch,
G. D. Gu,
S. Zhdanovich,
A. K. Mills,
G. Levy,
D. J. Jones,
C. Giannetti,
A. Damascelli
Abstract:
The possibility of driving phase transitions in low-density condensates through the loss of phase coherence alone has far-reaching implications for the study of quantum phases of matter. This has inspired the development of tools to control and explore the collective properties of condensate phases via phase fluctuations. Electrically-gated oxide interfaces, ultracold Fermi atoms, and cuprate supe…
▽ More
The possibility of driving phase transitions in low-density condensates through the loss of phase coherence alone has far-reaching implications for the study of quantum phases of matter. This has inspired the development of tools to control and explore the collective properties of condensate phases via phase fluctuations. Electrically-gated oxide interfaces, ultracold Fermi atoms, and cuprate superconductors, which are characterized by an intrinsically small phase-stiffness, are paradigmatic examples where these tools are having a dramatic impact. Here we use light pulses shorter than the internal thermalization time to drive and probe the phase fragility of the Bi$_2$Sr$_2$CaCu$_2$O$_{8+δ}$ cuprate superconductor, completely melting the superconducting condensate without affecting the pairing strength. The resulting ultrafast dynamics of phase fluctuations and charge excitations are captured and disentangled by time-resolved photoemission spectroscopy. This work demonstrates the dominant role of phase coherence in the superconductor-to-normal state phase transition and offers a benchmark for non-equilibrium spectroscopic investigations of the cuprate phase diagram.
△ Less
Submitted 2 April, 2018; v1 submitted 7 July, 2017;
originally announced July 2017.
-
Comparative study of rare earth hexaborides using high resolution angle-resolved photoemission
Authors:
Shyama V. Ramankutty,
Nick de Jong,
Ying-Kai Huang,
Berend Zwartsenberg,
Freek Massee,
Tran V. Bay,
Mark S. Golden,
Emmanouil Frantzeskakis
Abstract:
Strong electron correlations in rare earth hexaborides can give rise to a variety of interesting phenomena like ferromagnetism, Kondo hybridization, mixed valence, superconductivity and possibly topological characteristics. The theoretical prediction of topological properties in SmB$_{6}$ and YbB$_{6}$, has rekindled the scientific interest in the rare earth hexaborides, and high-resolution ARPES…
▽ More
Strong electron correlations in rare earth hexaborides can give rise to a variety of interesting phenomena like ferromagnetism, Kondo hybridization, mixed valence, superconductivity and possibly topological characteristics. The theoretical prediction of topological properties in SmB$_{6}$ and YbB$_{6}$, has rekindled the scientific interest in the rare earth hexaborides, and high-resolution ARPES has been playing a major role in the debate. The electronic band structure of the hexaborides contains the key to understand the origin of the different phenomena observed, and much can be learned by comparing the experimental data from different rare earth hexaborides. We have performed high-resolution ARPES on the (001) surfaces of YbB$_{6}$, CeB$_{6}$ and SmB$_{6}$. On the most basic level, the data show that the differences in the valence of the rare earth element are reflected in the experimental electronic band structure primarily as a rigid shift of the energy position of the metal 5$\textit{d}$ states with respect to the Fermi level. Although the overall shape of the $\textit{d}$-derived Fermi surface contours remains the same, we report differences in the dimensionality of these states between the compounds studied. Moreover, the spectroscopic fingerprint of the 4$\textit{f}$ states also reveals considerable differences that are related to their coherence and the strength of the $\textit{d}$-$\textit{f}$ hybridization. For the SmB$_6$ case, we use ARPES in combination with STM imaging and electron diffraction to reveal time dependent changes in the structural symmetry of the highly debated SmB$_{6}$(001) surface. All in all, our study highlights the suitability of electron spectroscopies like high-resolution ARPES to provide links between electronic structure and function in complex and correlated materials such as the rare earth hexaborides.
△ Less
Submitted 16 May, 2016; v1 submitted 19 June, 2015;
originally announced June 2015.
-
Angle-resolved and core-level photoemission study of interfacing the topological insulator Bi1.5Sb0.5Te1.7Se1.3 with Ag, Nb and Fe
Authors:
N. de Jong,
E. Frantzeskakis,
B. Zwartsenberg,
Y. K. Huang,
D. Wu,
P. Hlawenka,
J. Sánchez-Barriga,
A. Varykhalov,
E. van Heumen,
M. S. Golden
Abstract:
Interfaces between a bulk-insulating topological insulator (TI) and metallic adatoms have been studied using high-resolution, angle-resolved and core-level photoemission. Fe, Nb and Ag were evaporated onto Bi1.5Sb0.5Te1.7Se1.3 (BSTS) surfaces both at room temperature and 38K. The coverage- and temperature-dependence of the adsorption and interfacial formation process have been investigated, highli…
▽ More
Interfaces between a bulk-insulating topological insulator (TI) and metallic adatoms have been studied using high-resolution, angle-resolved and core-level photoemission. Fe, Nb and Ag were evaporated onto Bi1.5Sb0.5Te1.7Se1.3 (BSTS) surfaces both at room temperature and 38K. The coverage- and temperature-dependence of the adsorption and interfacial formation process have been investigated, highlighting the effects of the overlayer growth on the occupied electronic structure of the TI. For all coverages at room temperature and for those equivalent to less than 0.1 monolayer at low temperature all three metals lead to a downward shift of the TI's bands with respect to the Fermi level. At room temperature Ag appears to intercalate efficiently into the van der Waals gap of BSTS, accompanied by low-level substitution of the Te/Se atoms of the termination layer of the crystal. This Te/Se substitution with silver increases significantly for low temperature adsorption, and can even dominate the electrostatic environment of the Bi/Sb atoms in the BSTS near-surface region. On the other hand, Fe and Nb evaporants remain close to the termination layer of the crystal. On room temperature deposition, they initially substitute isoelectronically for Bi as a function of coverage, before substituting for Te/Se atoms. For low temperature deposition, Fe and Nb are too immobile for substitution processes and show a behaviour consistent with clustering on the surface. For both Ag and Fe/Nb, these differing adsorption pathways leads to the qualitatively similar and remarkable behavior for low temperature deposition that the chemical potential first moves upward (n-type dopant behavior) and then downward (p-type behavior) on increasing coverage.
△ Less
Submitted 27 August, 2015; v1 submitted 28 April, 2015;
originally announced April 2015.
-
Dirac states with knobs on: interplay of external parameters and the surface electronic properties of 3D topological insulators
Authors:
Emmanouil Frantzeskakis,
Nick de Jong,
Berend Zwartsenberg,
Tran V. Bay,
Ying Kai Huang,
Shyama V. Ramankutty,
Alona Tytarenko,
Dong Wu,
Yu Pan,
Shira Hollanders,
Milan Radovic,
Nicholas C. Plumb,
Nan Xu,
Ming Shi,
Cosmin Lupulescu,
Tiberiu Arion,
Ruslan Ovsyannikov,
Andrei Varykhalov,
Wolfgang Eberhardt,
Anne de Visser,
Erik van Heumen,
Mark S. Golden
Abstract:
Topological insulators are a novel materials platform with high applications potential in fields ranging from spintronics to quantum computation. In the ongoing scientific effort to demonstrate controlled manipulation of their electronic structure by external means, stoichiometric variation and surface decoration are two effective approaches that have been followed. In ARPES experiments, both appr…
▽ More
Topological insulators are a novel materials platform with high applications potential in fields ranging from spintronics to quantum computation. In the ongoing scientific effort to demonstrate controlled manipulation of their electronic structure by external means, stoichiometric variation and surface decoration are two effective approaches that have been followed. In ARPES experiments, both approaches are seen to lead to electronic band structure changes. Such approaches result in variations of the energy position of bulk and surface-related features and the creation of two-dimensional electron gases.The data presented here demonstrate that a third manipulation handle is accessible by utilizing the amount of illumination a topological insulator surface has been exposed to under typical experimental ARPES conditions. Our results show that this new, third, knob acts on an equal footing with stoichiometry and surface decoration as a modifier of the electronic band structure, and that it is in continuous competition with the latter. The data clearly point towards surface photovoltage and photo-induced desorption as the physical phenomena behind modifications of the electronic band structure under exposure to high-flux photons. We show that the interplay of these phenomena can minimize and even eliminate the adsorbate-related surface band bending on typical binary, ternary and quaternary Bi-based topological insulators. Including the influence of the sample temperature, these data set up a framework for the external control of the electronic band structure in topological insulator compounds in an ARPES setting. Four external knobs are available: bulk stoichiometry, surface decoration, temperature and photon exposure. These knobs can be used in conjunction to tune the band energies near the surface and consequently influence the topological properties of the relevant electronic states.
△ Less
Submitted 17 April, 2015;
originally announced April 2015.
-
Micro-metric electronic patterning of a topological band structure using a photon beam
Authors:
Emmanouil Frantzeskakis,
Nick de Jong,
Berend Zwartsenberg,
Yingkai Huang,
Tran V. Bay,
Pieter Pronk,
Erik van Heumen,
Dong Wu,
Yu Pan,
Milan Radovic,
Nicholas C. Plumb,
Nan Xu,
Ming Shi,
Anne de Visser,
Mark S. Golden
Abstract:
In an ideal 3D topological insulator (TI), the bulk is insulating and the surface conducting due to the existence of metallic states that are localized on the surface; these are the topological surface states. Quaternary Bi-based compounds of Bi$_{2-x}$Sb$_{x}$Te$_{3-y}$Se$_{y}$ with finely-tuned bulk stoichiometries are good candidates for realizing ideal 3D TI behavior due to their bulk insulati…
▽ More
In an ideal 3D topological insulator (TI), the bulk is insulating and the surface conducting due to the existence of metallic states that are localized on the surface; these are the topological surface states. Quaternary Bi-based compounds of Bi$_{2-x}$Sb$_{x}$Te$_{3-y}$Se$_{y}$ with finely-tuned bulk stoichiometries are good candidates for realizing ideal 3D TI behavior due to their bulk insulating character. However, despite its insulating bulk in transport experiments, the surface region of Bi$_{2-x}$Sb$_{x}$Te$_{3-y}$Se$_{y}$ crystals cleaved in ultrahigh vacuum also exhibits occupied states originating from the bulk conduction band. This is due to adsorbate-induced downward band-bending, a phenomenon known from other Bi-based 3D TIs. Here we show, using angle-resolved photoemission, how an EUV light beam of moderate flux can be used to exclude these topologically trivial states from the Fermi level of Bi$_{1.46}$Sb$_{0.54}$Te$_{1.7}$Se$_{1.3}$ single crystals, thereby re-establishing the purely topological character of the low lying electronic states of the system. We furthermore prove that this process is highly local in nature in this bulk-insulating TI, and are thus able to imprint structures in the spatial energy landscape at the surface. We illustrate this by `writing' micron-sized letters in the Dirac point energy of the system.
△ Less
Submitted 27 November, 2015; v1 submitted 2 December, 2014;
originally announced December 2014.
-
Low carrier concentration crystals of the topological insulator Bi$_{2-x}$Sb$_{x}$Te$_{3-y}$Se$_{y}$: a magnetotransport study
Authors:
Y. Pan,
D. Wu,
J. R. Angevaare,
H. Luigjes,
E. Frantzeskakis,
N. de Jong,
E. van Heumen,
T. V. Bay,
B. Zwartsenberg,
Y. K. Huang,
M. Snelder,
A. Brinkman,
M. S. Golden,
A. de Visser
Abstract:
In 3D topological insulators achieving a genuine bulk-insulating state is an important research topic. Recently, the material system (Bi,Sb)$_{2}$(Te,Se)$_{3}$ (BSTS) has been proposed as a topological insulator with high resistivity and a low carrier concentration (Ren \textit{et al.} \cite{Ren2011}). Here we present a study to further refine the bulk-insulating properties of BSTS. We have synthe…
▽ More
In 3D topological insulators achieving a genuine bulk-insulating state is an important research topic. Recently, the material system (Bi,Sb)$_{2}$(Te,Se)$_{3}$ (BSTS) has been proposed as a topological insulator with high resistivity and a low carrier concentration (Ren \textit{et al.} \cite{Ren2011}). Here we present a study to further refine the bulk-insulating properties of BSTS. We have synthesized Bi$_{2-x}$Sb${_x}$Te$_{3-y}$Se$_{y}$ single crystals with compositions around $x = 0.5$ and $y = 1.3$. Resistance and Hall effect measurements show high resistivity and record low bulk carrier density for the composition Bi$_{1.46}$Sb$_{0.54}$Te$_{1.7}$Se$_{1.3}$. The analysis of the resistance measured for crystals with different thicknesses within a parallel resistor model shows that the surface contribution to the electrical transport amounts to 97% when the sample thickness is reduced to $1 μ$m. The magnetoconductance of exfoliated BSTS nanoflakes shows 2D weak antilocalization with $α\simeq -1$ as expected for transport dominated by topological surface states.
△ Less
Submitted 7 May, 2015; v1 submitted 10 November, 2014;
originally announced November 2014.
-
Kondo hybridisation and the origin of metallic states at the (001) surface of SmB6
Authors:
E. Frantzeskakis,
N. de Jong,
B. Zwartsenberg,
Y. K. Huang,
Y. Pan,
X. Zhang,
J. X. Zhang,
F. X. Zhang,
L. H. Bao,
O. Tegus,
A. Varykhalov,
A. de Visser,
M. S. Golden
Abstract:
SmB6, a well-known Kondo insulator, has been proposed to be an ideal topological insulator with states of topological character located in a clean, bulk electronic gap, namely the Kondo hybridisation gap. Seeing as the Kondo gap arises from many body electronic correlations, this would place SmB6 at the head of a new material class: topological Kondo insulators. Here, for the first time, we show t…
▽ More
SmB6, a well-known Kondo insulator, has been proposed to be an ideal topological insulator with states of topological character located in a clean, bulk electronic gap, namely the Kondo hybridisation gap. Seeing as the Kondo gap arises from many body electronic correlations, this would place SmB6 at the head of a new material class: topological Kondo insulators. Here, for the first time, we show that the k-space characteristics of the Kondo hybridisation process is the key to unravelling the origin of the two types of metallic states observed directly by ARPES in the electronic band structure of SmB6(001). One group of these states is essentially of bulk origin, and cuts the Fermi level due to the position of the chemical potential 20 meV above the lowest lying 5d-4f hybridisation zone. The other metallic state is more enigmatic, being weak in intensity, but represents a good candidate for a topological surface state. However, before this claim can be substantiated by an unequivocal measurement of its massless dispersion relation, our data raises the bar in terms of the ARPES resolution required, as we show there to be a strong renormalisation of the hybridisation gaps by a factor 2-3 compared to theory, following from the knowledge of the true position of the chemical potential and a careful comparison with the predictions from recent LDA+Gutzwiler calculations. All in all, these key pieces of evidence act as triangulation markers, providing a detailed description of the electronic landscape in SmB6, pointing the way for future, ultrahigh resolution ARPES experiments to achieve a direct measurement of the Dirac cones in the first topological Kondo insulator.
△ Less
Submitted 13 December, 2013; v1 submitted 1 August, 2013;
originally announced August 2013.
