arXiv subjects

Qing Sun

Publications and source records attributed to Qing Sun.

At least 37 records · Page 2Linked to original sources

arXiv2021

A General Approach to High Efficiency Perovskite Solar Cells by Any Antisolvent

Deposition of perovskite thin films by antisolvent engineering is one of the most common methods employed in perovskite photovoltaics research. Herein, we report on a general method that allows the fabrication of highly efficient perovskite solar cells by any antisolvent via the manipulation of the antisolvent application rate. Through a detailed structural, compositional and microstructural characterization of perovskite layers fabricated by 14 different antisolvents, we identify two key factors that influence the quality of the perovskite active layer: the solubility of the organic precursors in the antisolvent and its miscibility with the host solvent(s) of the perovskite precursor solution. Depending on these two factors, each antisolvent can be utilized to produce high performance devices reaching power conversion efficiencies (PCEs) that exceed 21%. Moreover, we demonstrate that by employing the optimal antisolvent application procedure, highly efficient solar cells can be fabricated from a broad range of precursor stoichiometries, with either a significant excess or deficiency of organic iodides.

cond-mat.mtrl-sci↗

arXiv2020

An Empirical Investigation Towards Efficient Multi-Domain Language Model Pre-training

Pre-training large language models has become a standard in the natural language processing community. Such models are pre-trained on generic data (e.g. BookCorpus and English Wikipedia) and often fine-tuned on tasks in the same domain. However, in order to achieve state-of-the-art performance on out of domain tasks such as clinical named entity recognition and relation extraction, additional in domain pre-training is required. In practice, staged multi-domain pre-training presents performance deterioration in the form of catastrophic forgetting (CF) when evaluated on a generic benchmark such as GLUE. In this paper we conduct an empirical investigation into known methods to mitigate CF. We find that elastic weight consolidation provides best overall scores yielding only a 0.33% drop in performance across seven generic tasks while remaining competitive in bio-medical tasks. Furthermore, we explore gradient and latent clustering based data selection techniques to improve coverage when using elastic weight consolidation and experience replay methods.

cs.CL↗

arXiv2020

Learn to Talk via Proactive Knowledge Transfer

Knowledge Transfer has been applied in solving a wide variety of problems. For example, knowledge can be transferred between tasks (e.g., learning to handle novel situations by leveraging prior knowledge) or between agents (e.g., learning from others without direct experience). Without loss of generality, we relate knowledge transfer to KL-divergence minimization, i.e., matching the (belief) distributions of learners and teachers. The equivalence gives us a new perspective in understanding variants of the KL-divergence by looking at how learners structure their interaction with teachers in order to acquire knowledge. In this paper, we provide an in-depth analysis of KL-divergence minimization in Forward and Backward orders, which shows that learners are reinforced via on-policy learning in Backward. In contrast, learners are supervised in Forward. Moreover, our analysis is gradient-based, so it can be generalized to arbitrary tasks and help to decide which order to minimize given the property of the task. By replacing Forward with Backward in Knowledge Distillation, we observed +0.7-1.1 BLEU gains on the WMT'17 De-En and IWSLT'15 Th-En machine translation tasks.

cs.LG↗

arXiv2020

Efficient and Stable PbS Quantum Dot Solar Cells by Triple-Cation Perovskite Passivation

Solution-processed quantum dots (QDs) have a high potential for fabricating low cost, flexible and large-scale solar energy harvesting devices. It has recently been demonstrated that hybrid devices employing a single monovalent cation perovskite solution for PbS QD surface passivation exhibit enhanced photovoltaic performance when compared to standard ligand passivation. Herein we demonstrate that the use of a triple cation Cs0.05(MA0.17FA0.83)0.95Pb(I0.9Br0.1)3 perovskite composition for surface passivation of the quantum dots results in highly efficient solar cells, which maintain 96 % of their initial performance after 1200h shelf storage. We confirm perovskite shell formation around the PbS nanocrystals by a range of spectroscopic techniques as well as high-resolution transmission electron microscopy. We find that the triple cation shell results in a favorable energetic alignment to the core of the dot, resulting in reduced recombination due to charge confinement without limiting transport in the active layer. Consequently, photovoltaic devices fabricated via a single-step film deposition reached a maximum AM1.5G power conversion efficiency of 11.3 % surpassing most previous reports of PbS solar cells employing perovskite passivation.

physics.app-ph↗

arXiv2020

Dipolar Hole-Blocking Layers for Inverted Perovskite Solar Cells: Effects of Aggregation and Electron Transport Levels

Herein, we report on the synthesis and investigation of two triazino-isoquinoline tetrafluoroborate electrolytes as hole-blocking layers in methylammonium triiodide perovskite photovoltaic devices with fullerene electron extraction layer. We find that increasing the thickness of the dipolar hole-blocking layer results in a gradual increase in the open-circuit voltage suggesting that aggregation of the molecules can enhance the dipole induced by the layer. This finding is confirmed by theoretical calculations demonstrating that while both molecules exhibit a similar dipole moment in their isolated state, this dipole is significantly enhanced when they aggregate. Ultra-violet photoemission spectroscopy measurements show that both derivatives exhibit a high ionisation potential of 7 eV, in agreement with their effective hole-blocking nature demonstrated by the devices. However, each of the molecules shows a different electron affinity due to the increased conjugation of one of the derivatives. While the change in electron transport level between the two derivatives is as high as 0.3 eV, the difference in the open-circuit voltage of both types of devices is negligible, suggesting that the electron transport level plays only a minor role in determining the open-circuit voltage of the device. Numerical device simulations confirm that the increase in built-in potential, arising from the high dipole of the electrolyte layer, compensates for the non-ideal energetic alignment of the charge transport levels, resulting in high VOC for a range of electron transport levels. Our study demonstrates that the application of small molecule electrolytes as hole-blocking layer in inverted architecture perovskite solar cells is a powerful tool to enhance the open-circuit voltage of the devices and provides useful guidelines for designing future generations of such compounds.

physics.app-ph↗

arXiv2020

Exotic Vortex States with Discrete Rotational Symmetry in Atomic Fermi Gases with Spin-Orbital-Angular-Momentum Coupling

We investigate the superfluidity of a two-component Fermi gas with spin-orbital-angular-momentum coupling (SOAMC). Due to the intricate interplay of SOAMC, two-photon detuning and atom-atom interaction, a family of vortex ground states emerge in a broad parameter regime of the phase diagram, in contrast to the usual case where an external rotation or magnetic field is generally required. More strikingly, an unprecedented vortex state, which breaks the continuous rotational symmetry to a discrete one spontaneously, is predicted to occur. The underlying physics are elucidated and verified by numerical simulations. The unique density distributions of the predicted vortex states enable a direct observation in experiment.

cond-mat.quant-gas↗

arXiv2020

Variational Problems for Tree Roots and Branches

This paper studies two classes of variational problems introduced in [7], related to the optimal shapes of tree roots and branches. Given a measure $μ$ describing the distribution of leaves, a sunlight functional $§(μ)$ computes the total amount of light captured by the leaves. For a measure $μ$ describing the distribution of root hair cells, a harvest functional $\H(μ)$ computes the total amount of water and nutrients gathered by the roots. In both cases, we seek a measure $μ$ that maximizes these functionals subject to a rami?ed transportation cost, for transporting nutrients from the roots to the trunk or from the trunk to the leaves. Compared with [7], here we do not impose any a priori bound on the total mass of the optimal measure $μ$, and more careful a priori estimates are thus required. In the unconstrained optimization problem for branches, we prove that an optimal measure exists, with bounded support and bounded total mass. In the unconstrained problem for tree roots, we prove that an optimal measure exists, with bounded support but possibly unbounded total mass. The last section of the paper analyzes how the size of the optimal tree depends on the parameters defining the various functionals.

math.OC↗

arXiv2019

Localization driven superradiant instability

The prominent Dicke superradiant phase arises from coupling an ensemble of atoms to cavity optical field when external optical pumping exceeds a threshold strength. Here we report a prediction of the superrandiant instability driven by Anderson localization, realized with a hybrid system of Dicke and Aubry-Andre (DAA) model for bosons trapped in a one-dimensional (1D) quasiperiodic optical lattice and coupled to a cavity. Our central finding is that for bosons condensed in localized phase given by the DAA model, the resonant superradiant scattering is induced, for which the critical optical pumping of superradiant phase transition approaches zero, giving an instability driven by Anderson localization. The superradiant phase for the DAA model with or without a mobility edge is investigated, showing that the localization driven superradiant instability is in sharp contrast to the superradiance as widely observed for Bose condensate in extended states, and should be insensitive to temperature of the system. This study unveils an insightful effect of localization on the Dicke superradiance, and is well accessible based on the current experiments.

cond-mat.quant-gas↗

arXiv2019

Synergistic Effect of One- and Two-dimensional Connected Coral-like Li$_{6.25}$Al$_{0.25}$La$_3$Zr$_2$O$_{12}$ in PEO-Based Composite Solid State Electrolyte

As one of the most promising next-generation energy storage device, lithium metal batteries have been extensively investigated. However, the poor safety issue and undesired lithium dendrites growth hinder the development of lithium metal batteries. The application of solid state electrolytes has attracted increasing attention as they can solve the safety issue and partly inhibit the growth of lithium dendrites. Polyethylene oxide (PEO)-based electrolytes are very promising due to their enhanced safety and excellent flexibility. However, PEO-based electrolytes suffer from low ionic conductivity at room temperature and can't effectively inhibit lithium dendrites at high temperature due to the intrinsic semi-crystalline properties and poor mechanical strength. In this work, a novel coral-like Li6.25Al0.25La3Zr2O12 (LALZO) is synthesized to use as an active ceramic filler in PEO. The PEO with LALZO coral (PLC) exhibits increased ionic conductivity and mechanical strength, which leads to the uniform deposition/stripping of lithium metal. The Li symmetric cells with PLC cycle for 1500 h without short circuit at 50 centidegree. The assembled LiFePO4/PLC/Li batteries display excellent cycling stability at both 60 and 50 centidegree. This work reveals that the electrochemical properties of organic and inorganic composite electrolyte can be effectively improved by tuning the microstructure of the filler, such as the coral-like LALZO architecture.

physics.chem-ph↗

arXiv2019

Enhancing the Open-Circuit Voltage of Perovskite Solar Cells by up to 120 mV using π-Extended Phosphoniumfluorene Electrolytes as Hole Blocking Layers

Four π-extended phosphoniumfluorene electrolytes (π-PFEs) are introduced as hole-blocking layers (HBL) in inverted architecture planar perovskite solar cells (PVSCs) with the structure of ITO/PEDOT:PSS/MAPbI3/PCBM/HBL/Ag. The deep-lying highest occupied molecular orbital (HOMO) energy level of the π-PFEs effectively blocks holes, decreasing contact recombination. We demonstrate that the incorporation of π-PFEs introduces a dipole moment at the PCBM/Ag interface, resulting in a significant enhancement of the built-in potential of the device. This enhancement results in an increase in the open-circuit voltage of the device by up to 120 mV, when compared to the commonly used bathocuproine HBL. The results are confirmed both experimentally and by numerical simulation. Our work demonstrates that interfacial engineering of the transport layer/contact interface by small molecule electrolytes is a promising route to suppress non-radiative recombination in perovskite devices and compensate for a non-ideal energetic alignment at the hole-transport layer/perovskite interface.

physics.app-ph↗

arXiv2019

Weighted Irrigation Plans

We model an irrigation network where lower branches must be thicker in order to support the weight of the higher ones. This leads to a countable family of ODEs, one for each branch, that must be solved by backward induction. Having introduced conditions that guarantee the existence and uniqueness of solutions, our main result establishes the lower semicontinuity of the corresponding cost functional, w.r.t. pointwise convergence of the irrigation plans. In turn, this yields the existence of an optimal irrigation plan, in the presence of these additional weights.

math.OC↗

arXiv2019

Effect of Ion Migration Induced Electrode Degradation on the Operational Stability of Perovskite Solar Cells

Perovskite-based solar cells are promising due to their rapidly improving efficiencies, but suffer from instability issues. Recently it has been claimed that one of the key contributors to the instability of perovskite solar cells is ion migration induced electrode degradation, which can be avoided by incorporating inorganic hole blocking layers (HBL) in the device architecture. In this work, we investigate the operational environmental stability of methylammonium lead iodide (MAPbI3) perovskite solar cells that contain either an inorganic or organic HBL, with only the former effectively blocking ions from migrating to the metal electrode. This is confirmed by X-ray photoemission spectroscopy measured on electrodes of degraded devices, where only electrodes of devices with an organic HBL show a significant iodine signal. Despite this, we show that when these devices are degraded under realistic operational conditions (i.e. constant illumination in a variety of atmospheric conditions), both types of devices exhibit nearly identical degradation behavior. These results demonstrate that contrary to prior suggestions, ion-induced electrode degradation is not the dominant factor in perovskite environmental instability under operational conditions.

physics.app-ph↗

arXiv2019

Larkin-Ovchinikov superfluidity in time-reversal symmetric bilayer Fermi gases

Larkin-Ovchinnikov (LO) state which combines the superfluidity and spatial periodicity of pairing order parameter and exhibits the supersolid properties has been attracting intense attention in both condensed matter physics and ultracold atoms. Conventionally, realization of LO state from an intrinsic s-wave interacting system necessitates to break the time-reversal (TR) and sometimes spatial-inversion (SI) symmetries. Here we report a novel prediction that the LO state can be realized in a TR and SI symmetric system representing a bilayer Fermi gas subjected to a laserassisted interlayer tunneling. We show that the intralayer s-wave atomic interaction acts effectively like a p-wave interaction in the pseudospin space. This provides distinctive pairing effects in the present system with pseudspin spin-orbit coupling, and leads to a spontaneous density-modulation of the pairing order predicted in a very broad parameter regime. Unlike the conventional schemes, our results do not rely on the spin imbalance or external Zeeman fields, showing a highly feasible way to observe the long-sought-after LO superfluid phase using the laser-assisted bilayer Fermi gases.

cond-mat.quant-gas↗

arXiv2018

Diverse Beam Search: Decoding Diverse Solutions from Neural Sequence Models

Neural sequence models are widely used to model time-series data. Equally ubiquitous is the usage of beam search (BS) as an approximate inference algorithm to decode output sequences from these models. BS explores the search space in a greedy left-right fashion retaining only the top-B candidates - resulting in sequences that differ only slightly from each other. Producing lists of nearly identical sequences is not only computationally wasteful but also typically fails to capture the inherent ambiguity of complex AI tasks. To overcome this problem, we propose Diverse Beam Search (DBS), an alternative to BS that decodes a list of diverse outputs by optimizing for a diversity-augmented objective. We observe that our method finds better top-1 solutions by controlling for the exploration and exploitation of the search space - implying that DBS is a better search algorithm. Moreover, these gains are achieved with minimal computational or memory over- head as compared to beam search. To demonstrate the broad applicability of our method, we present results on image captioning, machine translation and visual question generation using both standard quantitative metrics and qualitative human studies. Further, we study the role of diversity for image-grounded language generation tasks as the complexity of the image changes. We observe that our method consistently outperforms BS and previously proposed techniques for diverse decoding from neural sequence models.

cs.AI↗

arXiv2018

On the Optimal Shape of Tree Roots and Branches

This paper introduces two classes of variational problems, determining optimal shapes for tree roots and branches. Given a measure $μ$, describing the distribution of leaves, we introduce a sunlight functional $§(μ)$ computing the total amount of light captured by the leaves. On the other hand, given a measure $μ$ describing the distribution of root hair cells, we consider a harvest functional $\H(μ)$ computing the total amount of water and nutrients gathered by the roots. In both cases, we seek to maximize these functionals subject to a ramified transportation cost, for transporting nutrients from the roots to the trunk and from the trunk to the leaves. The main results establish various properties of these functionals, and the existence of optimal distributions. In particular, we prove the upper semicontinuity of $§$ and $\H$, together with a priori estimates on the support of optimal distributions.

math.OC↗

arXiv2017

An oscillating Casimir potential between two impurities in a spin-orbit coupled Bose-Einstein condensate

We study the Casimir potential between two impurities immersed in a spin-orbit coupled BoseEinstein condensate (BEC) with plane-wave order. We find that, by exchanging the virtual phonons/excitations, a remarkable anisotropic oscillating potential with both positive and negative parts can be induced between the impurities, with the period of the oscillation depending on the spin-orbit coupling strength. As a consequence, this would inevitably lead to a non-central Casimir force, which can be tuned by varying the strength of spin-orbit coupling . These results are elucidated for BECs with one-dimensional Raman-induced and two-dimensional Rashba-type SOC.

cond-mat.quant-gas↗

arXiv2017

Fulde-Ferrell-Larkin-Ovchinnikov state to topological superfluidity transition in bilayer spin-orbit coupled degenerate Fermi gas

Recently a scheme has been proposed for generating the 2D Rashba-type spin-orbit coupling (SOC) for ultracold atomic bosons in a bilayer geometry [S.-W. Su et al, Phys. Rev. A \textbf{93}, 053630 (2016)]. Here we investigate the superfluidity properties of a degenerate Fermi gas affected by the SOC in such a bilayer system. We demonstrate that a Fulde-Ferrell-Larkin-Ovchinnikov (FFLO) state appears in the regime of small to moderate atom-light coupling. In contrast to the ordinary SOC, the FFLO state emerges in the bilayer system without adding any external fields or spin polarization. As the atom-light coupling increases, the system can transit from the FFLO state to a topological superfluid state. These findings are also confirmed by the BdG simulations with a weak harmonic trap added.

cond-mat.quant-gas↗

arXiv2017

Bidirectional Beam Search: Forward-Backward Inference in Neural Sequence Models for Fill-in-the-Blank Image Captioning

We develop the first approximate inference algorithm for 1-Best (and M-Best) decoding in bidirectional neural sequence models by extending Beam Search (BS) to reason about both forward and backward time dependencies. Beam Search (BS) is a widely used approximate inference algorithm for decoding sequences from unidirectional neural sequence models. Interestingly, approximate inference in bidirectional models remains an open problem, despite their significant advantage in modeling information from both the past and future. To enable the use of bidirectional models, we present Bidirectional Beam Search (BiBS), an efficient algorithm for approximate bidirectional inference.To evaluate our method and as an interesting problem in its own right, we introduce a novel Fill-in-the-Blank Image Captioning task which requires reasoning about both past and future sentence structure to reconstruct sensible image descriptions. We use this task as well as the Visual Madlibs dataset to demonstrate the effectiveness of our approach, consistently outperforming all baseline methods.

cs.CV↗