首 頁
滾動信息 更多 >>
npj Computational Materials獲得第一個SCI影響因... (2018-09-07)
英文刊《npj Computational Materials(計算材料學... (2017-05-15)
開放的數據庫 (2015-12-17)
上海硅酸鹽所與Nature出版集團聯合創辦的中國首本n... (2015-11-27)
Nature Publishing Group and Shanghai Institute ... (2015-09-24)
快捷服務
最新文章 研究綜述
過刊瀏覽 作者須知
期刊編輯 審稿須知
相關鏈接
· 在線投稿
會議信息
友情鏈接
  中國科學院上海硅酸鹽研究所
  無機材料學報
  OQMD數據庫
mm.jpg
期刊介紹
  《npj 計算材料學》是在線出版、完全開放獲取的國際學術期刊。發表結合計算模擬與設計的材料學一流的研究成果。本刊由中國科學院上海硅酸鹽研究所與英國自然出版集團(Nature Publishing Group,NPG)以伙伴關系合作出版。
  主編為陳龍慶博士,美國賓州大學材料科學與工程系、工程科學與力學系、數學系的杰出教授。
  共同主編為陳立東研究員,中國科學院上海硅酸鹽研究所研究員高性能陶瓷與超微結構國家重點實驗室主任。
  辦刊目的與報道范圍
  《npj 計算材料學》是在線出版、完全開放獲取的國際...
【查看詳細】
近期文章 更多 >>

Dodecagonal bilayer graphene quasicrystal and itsapproximants (雙層石墨烯準晶及其周期性近似結構)
Guodong Yu, Zewen Wu, Zhen Zhan, Mikhail I. Katsnelson & Shengjun Yuan
npj Computational Materials 5:122(2019)
doi:s41524-019-0258-0
Published online:12 December 2019

Abstract| Full Text | PDF OPEN

摘要:雙層石墨烯準晶具有十二重旋轉對稱性,但由于缺乏空間平移對稱性很多基于能帶理論的研究方法已不再適用。本文運用大尺度的緊束縛近似方法研究了由超過千萬個原子組成的雙層石墨烯準晶,并且提出了一系列周期性近似結構。這些周期性近似結構可以準確還原出雙層石墨烯準晶的電學和光學性質。通過對由2702個原子組成的周期性近似結構進行能帶反折疊,本文得到了雙層石墨烯準晶的有效能帶結構。計算結果表明,除了原有的單層石墨烯狄拉克點外,在兩個單層石墨烯的布里淵區內部出現了大量新的狄拉克點,同時在M點處出現了能隙。這些特性都和近期的實驗結果一致。此外,雙層石墨烯準晶的準晶特性在其朗道能級和光學性質中也有所體現。本文的研究結果表明,組成準晶的兩層石墨烯晶格常數匹配度是決定其周期性近似結構準確度的最重要因素。本文提出的周期性近似結構可以直接應用于其它的六角蜂窩層狀材料組成的準晶,并且其設計理念可以推廣到任意一種層狀準周期體系 

Abstract:Dodecagonal bilayer graphene quasicrystal has 12-fold rotational orderbut lacks translational symmetry which prevents theapplication of band theory. In this paper, we study the electronic and optical properties of graphene quasicrystal with large-scaletight-binding calculations involving more than ten million atoms. We propose a series of periodic approximants which reproduceaccurately theproperties of quasicrystal within a finite unit cell. By utilizing the band-unfolding method on the smallestapproximant with only 2702 atoms, the effective bandstructure of graphene quasicrystal is derived. The features, such as theemergence of new Dirac points (especially the mirrored ones), the band gap at M point and theFermi velocity are all in agreementwith recent experiments. The properties ofquasicrystal states are identified in the Landau level spectrum and opticalexcitations.Importantly, our results show that the lattice mismatch is the dominant factor determining the accuracy of layered approximants.The proposedapproximants can be used directly for other layered materials in honeycomb lattice,and the design principles can beapplied for any quasi-periodic incommensurate structures.

Editorial Summary

Approximants: Making band theory apply to graphene quasicrystal周期性近似結構:讓能帶論適用于雙層石墨烯準晶

扭轉的雙層石墨烯具有豐富的電子結構和物理性質。當扭轉的角度為30度時,雙層石墨烯組成具有十二重旋轉對稱性的準晶結構。但由于準晶缺乏空間平移對稱性,很多基于能帶理論的研究方法很難直接應用,這使得對近期實驗上制備的雙層石墨烯準晶進行理論研究充滿挑戰。最近武漢大學袁聲軍教授團隊,運用該小組發展的針對復雜量子體系的緊束縛傳播方法(TBPM),研究了由超過千萬個原子組成的雙層石墨烯準晶,并且提出了一系列周期性近似結構。這些周期性近似結構可以準確還原出雙層石墨烯準晶的電學和光學性質。其中最小的一個精確周期性近似結構包含2702個原子,通過對其能帶結構進行反折疊運算,得到了雙層石墨烯準晶的有效能帶結構。計算結果表明,除了原有的單層石墨烯狄拉克點外,在兩個單層石墨烯的布里淵區內部出現了大量新的狄拉克點,同時在M點處出現了能隙。這些特性都和近期的實驗結果一致。此外,雙層石墨烯準晶的準晶特性在其朗道能級和光學性質中也有所體現。本文的研究結果表明,組成準晶的兩層石墨烯晶格常數匹配度是決定其周期性近似結構準確度的最重要因素。本文提出的周期性近似結構可以直接應用于其它的六角蜂窩層狀材料組成的準晶,并且其設計理念可以推廣到任意一種層狀準周期體系

Twisted bilayer graphene has rich electronic structures and physical properties. When the twisted angle equals to 30-degree, twisted bilayer graphene forms a quasicrystal with 12-fold rotational order, which has been fabricated in recent experiments.Due to the absence of the translational symmetry, the band theory doesn’t apply to the quasicrystals, which makes it a big challenge to study theoretically twisted bilayer graphene quasicyrstal .Recently, a team led byProf. Shengjun Yuan from School of Physics and Technology,Wuhan University, China,studied the electronic structures and physical properties of graphene quasicrystal by means of the large-scale tight-binding propagation method using their home-made simulation package Tipsi.They proposed a series of approximants, which can reproduce electronic and optical properties of bilayergraphene quasicrystal accurately. The effective band structure obtained by unfolding the band structure of the smallest approximant with 2702 atoms is in agreement with recent experimental results. This study paved the way for applying band theory to graphene quasicrystal, and the design principle of the approximant can be applied for any quasi-periodic layered incommensurate structures.

Continuous strengthening in nanotwinned diamond (納米孿晶金剛石的持續強化)
Bin Wen, Bo Xu, Yanbin Wang, Guoying Gao, Xiang-Feng Zhou, Zhisheng Zhao, and Yongjun Tian
npj Computational Materials 5:117(2019)
doi:s41524-019-0256-2
Published online:05 December 2019

Abstract| Full Text | PDF OPEN

摘要:隨著晶粒尺寸和孿晶厚度的減小,納米晶和納米孿晶金屬的強度在增大。但當晶粒尺寸和孿晶厚度低于某一臨界值時,其強度在降低,這種強度隨晶粒尺度或孿晶厚度減小而降低的行為被稱為反霍爾-佩奇效應。在納米立方氮化硼和納米金剛石中,也能觀察到反霍爾-佩奇效應。然而,令人驚訝的是,隨著孿晶厚度的減小,甚至減小至幾個納米時,納米孿晶立方氮化硼和納米孿晶金剛石的硬度仍在持續增加,表明這些共價材料不存在反霍爾-佩奇效應。造成這種持續硬化行為的物理機制目前仍存在很大的爭議。本文利用分子動力學和第一性原理計算方法,研究了納米晶和納米孿晶金剛石的強化機理。對于納米金剛石,拖拉面位錯運動與晶界原子相關的運動之間的競爭導致了反霍爾-佩奇效應。對于納米孿晶金剛石,雖然拖拉面位錯運動仍然占主導,但由于沿孿晶界面的滑移在能量上并不比沿其他滑移面的滑移更具優勢,所有孿晶界面不容易遷移,結構比較穩定。因此,隨孿晶厚度的減小,孿晶界面仍能起到阻礙位錯運動的作用,從而抑制了反霍爾-佩奇效應的產生。該研究結果為納米共價材料的硬化機理提供了一個新的見解 

Abstract:Strengths of nanograined (ng) and nanotwinned (nt) metals increase with decreasing grain size and twin thickness, respectively, until reaching a critical value, below which strength decreases. This behavior is known as the reverse Hall–Petch effect (RHPE), which has also been observed in nanograined cubic boron nitride (cBN) and diamond. Surprisingly, however, hardness of nt-cBN and nt-diamond increases continuously with decreasing twin thickness down to several nanometers, suggesting the absence of RHPE in these covalent materials. The mechanism responsible for such a behavior remains controversial. Here we investigate the strengthening mechanisms in ng- and nt-diamond using molecular dynamics and first-principles calculations. For ng-diamond, the competition between shuffle-set dislocation (SSD) and grain boundary atom motions gives rise to RHPE. For nt-diamond, SSDs remain dominant but their slips along twin boundaries energetically show no advantage over those along other slip planes. Twin domains are locked and mechanically stable, resisting SSD propagation and inhibiting RHPE. These findings provide new insights into the hardening mechanism of nanotwinned covalent materials.

Editorial Summary

Continuous strengthening in nanotwinned diamond納米孿晶金剛石的持續強化

本研究提出了一種納米晶和納米孿晶金剛石強化的位錯運動機理,可以合理解釋納米孿晶共價材料的持續硬化行為。燕山大學田永君院士領導的團隊,與美國芝加哥大學的王雁賓教授合作。利用分子動力學和第一性原理計算方法,研究了納米晶和納米孿晶金剛石的強化機理。對于納米金剛石,拖拉面位錯運動與晶界原子相關的運動之間的競爭導致了反霍爾-佩奇效應。對于納米孿晶金剛石,雖然拖拉面位錯運動仍然占主導,但由于沿孿晶界面的滑移在能量上并不比沿其他滑移面的滑移更具優勢,所以孿晶界面不容易遷移,結構比較穩定。因此,隨孿晶厚度的減小,孿晶界面仍能起到阻礙位錯運動的作用,從而抑制了反霍爾-佩奇效應的產生。特別地,該研究將材料化學健與位錯運動模式建立了關聯,初步揭示了化學健的性質(例如方向性等)對位錯運動行為的影響規律。該研究結果為材料力學性能的微觀機理研究提供了一個新的見解

In this study, a dislocation-related mechanism for nanograined (ng) and nanotwinned (nt) diamond is proposed, which can reasonably explain the continuous hardening behavior of nt covalent materials. A team led by academician Yongjun Tian from Yanshan University, cooperating with Yanbin Wang from the University of Chicago, developed a continuous strengthening mechanism for nt-diamond.By using molecular dynamics and first principles calculation method, they found that, for ng-diamond, the competition between shuffle-set dislocation (SSD) and grain boundary atom motions gives rise to reverse Hall–Petch effect RHPE. For nt-diamond, SSDs remain dominant but their slips along twin boundaries energetically show no advantage over those along other slip planes. Twin domains are locked and mechanically stable, resisting SSD propagation and inhibiting RHPE. In particular, the study established a relationship between the chemical bond and dislocation motion, and preliminarily revealed the influence of the chemical bond properties (such as directionality, etc.) on dislocation motion behavior. The results provide a new insight for the study of the micro mechanism of mechanical properties of materials.

Beyond the RPA and GW methods with adiabatic xc-kernels for accurate ground state and quasiparticle energies (精確基態和準粒子能量描述的非絕熱xc-內核RPAGW計算方法)
Thomas OlsenChristopher E. PatrickJefferson E. BatesAdrienn Ruzsinszky & Kristian S. Thygesen
npj Computational Materials 5:106(2019)
doi:s41524-019-0242-8
Published online:11 November 2019

Abstract| Full Text | PDF OPEN

摘要:本文綜述了在絕熱連接漲落耗散理論和Hedin方程的框架內,用于基態能量和準粒子激發的從頭算絕熱交換-關聯(xc)泛函的理論與應用。首先介紹了各種源自均質電子氣中的不同交換關聯泛函,此后重點介紹了一類特定的重整化絕熱泛函,特別是rALDA(重整化絕熱局域密度近似)和rAPBE(重整化絕熱PBE廣義梯度近似)。與隨機相位近似(RPA)相比,這些泛函極大地改進了電子短程關聯的描述,從而顯著提高了關聯能。這種效應大大減少對誤差消除的依賴,而這些消除差在RPA中是必不可少的。同時,該效應也系統地提高了共價鍵能,同時保留了RPA中對色散相互作用的良好描述。對于準粒子能量,交換關聯泛函考慮了GW自能中缺失的頂點修正。在這種情況下,本研究顯示,短程關聯主要校正了絕對能帶的位置,而對帶隙的影響較小,這與GW已知的良好性能是一致的。重整化的交換關聯泛函提供了RPAGW方法的嚴格擴展,在幾乎沒有增加計算成本的情況下,準確性得到了明顯的改進 

Abstract:We review the theory and application of adiabatic exchange–correlation (xc)-kernels for ab initio calculations of ground state energies and quasiparticle excitations within the frameworks of the adiabatic connection fluctuation dissipation theorem and Hedin’s equations, respectively. Various different xc-kernels, which are all rooted in the homogeneous electron gas, are introduced but hereafter we focus on the specific class of renormalized adiabatic kernels, in particular the rALDA (renormalized adiabatic local density approximation) and rAPBE (renormalized adiabatic Perdew–Burke–Ernzerhof. The kernels drastically improve the description of short-range correlations as compared to the random phase approximation (RPA), resulting in significantly better correlation energies. This effect greatly reduces the reliance on error cancellations, which is essential in RPA, and systematically improves covalent bond energies while preserving the good performance of the RPA for dispersive interactions. For quasiparticle energies, the xc-kernels account for vertex corrections that are missing in the GW self-energy. In this context, we show that the short-range correlations mainly correct the absolute band positions while the band gap is less affected in agreement with the known good performance of GW for the latter. The renormalized xc-kernels offer a rigorous extension of the RPA and GW methods with clear improvements in terms of accuracy at little extra computational cost.

Editorial Summary

Ground state and quasiparticle energies: Accurate description精確基態和準粒子能量的精確描述

本文綜述的重點是含時密度泛函理論(TDDFT)中的靜態非局部交換關聯泛函所描述的RPAGW方法中的物理理論、實現途徑和意義。來自丹麥技術大學物理系計算原子尺度材料設計(CAMD)的Thomas Olsen教授帶領的團隊,并沒有詳述RPAGW方法本身,該方法的綜述已在另一篇綜述中闡明。在“理論”一節中,他們分別介紹了基于絕熱連接漲落耗散定理和Hedin方程的基態和QP能量計算的基本理論。作者為HEG引入了幾個非局域的交換關聯泛函,并描述了從(半)局域交換關聯泛函構造非局域交換關聯泛函的重整化過程。在“實現”一節中,作者描述了非局域交換關聯泛函的數值實現,包括將HEG泛函推廣到非均勻密度體系的不同策略、倒空間格點和基集收斂性等方面。在“結果”一節中,作者提供了一系列計算結果,用以說明交換關聯泛函對總能量和QP能帶結構的影響和重要性。具體而言,作者評估了重整化絕熱局域密度近似(rALDA)和重整化絕熱廣義梯度近似(rAPBE)在固體結構參數、共價固體和共價分子的原子化能、氧化物形成能、vdW鍵、靜態關聯的原子二聚體的解離、表面和化學吸附能、結構相變,以及塊體和二維半導體的QP能等方面的性能。最后,對全文進行了總結和展望

This review focuses on the theory, implementation, and implications of physics beyond the RPA and GW methods as described by static non-local xc-kernels from TDDFT. A team led by Prof. Thomas Olsen from the Computational Atomic-Scale Materials Design (CAMD), Department of Physics, Technical University of Denmark, did not dwell on the RPA and GW methods themselves but refer the interested reader to one of the existing reviews on these topics. The review is organized as follows. In section “Theory,” they present the basic theory of ground state and QP energy calculations based on the adiabatic connection fluctuation dissipation theorem and Hedin’s equations, respectively. They introduce several non-local xc-kernels for the HEG and describe a renormalization procedure for constructing non-local xc-kernels from (semi)local xc-functionals. In section “Implementation,” they describe the numerical implementation of non-local xc-kernels including different strategies for generalizing HEG kernels to inhomogeneous densities and some aspects of kk-point and basis set convergence. In section “Results,” they present a series of results serving to illustrate the effect and importance of the xc-kernels for both total energies and QP band structures. Specifically, authors assess the performance of the renormalized adiabatic local density approximation (rALDA) and renormalized adiabatic Perdew–Burke–Ernzerhof (rAPBE) xc-kernels for structural parameters of solids, atomization energies of covalently bonded solids and molecules, oxide formation energies, vdW bonding, dissociation of statically correlated atomic dimers, surface and chemisorption energies, structural phase transitions, and QP energies of bulk and two-dimensional semiconductors. Finally, they conclusions and outlook the whole content of the review.

Identification of advanced spin-driven thermoelectric materials via interpretable machine learning (通過可解釋性機器學習識別自旋驅動的先進熱電材料
Yuma IwasakiRyohto SawadaValentin StanevMasahiko IshidaAkihiro KiriharaYasutomo OmoriHiroko SomeyaIchiro TakeuchiEiji Saitoh & Shinichi Yorozu
npj Computational Materials 5:103(2019)
doi:s41524-019-0241-9
Published online:30 October 2019

Abstract| Full Text | PDF OPEN

摘要:機器學習正成為科學發現的重要工具。機器學習方法在新材料開發領域尤為表現突出,通過發現更好的新型功能材料來實現領域革新。要將機器學習應用于實際的材料開發中,科學家和機器學習工具之間必須緊密協作。但是,到目前為止,機器學習算法中的許多黑箱特征阻礙了這種相互協作。科學家通常很難從材料科學和物理學的角度來解釋這些數據驅動的模型。本研究通過一種可進行物理解釋的機器學習方法(FAB / HMEs:分解的漸進貝葉斯推理專家分層混合)來展示具有異常能斯特效應的自旋驅動的熱電材料的開發。基于材料科學和物理學的先驗知識,我們能夠從可解釋的機器學習數據中提取一些令人驚訝的相關性和關于自旋驅動的熱電材料的全新認知。在此指導下,本研究進行了實際的材料合成和表征,從而確定了一種新型的自旋驅動,且性能優異的熱電材料 

Abstract:Machine learning is becoming a valuable tool for scientific discovery. Particularly attractive is the application of machine learning methods to the field of materials development, which enables innovations by discovering new and better functional materials. To apply machine learning to actual materials development, close collaboration between scientists and machine learning tools is necessary. However, such collaboration has been so far impeded by the black box nature of many machine learning algorithms. It is often difficult for scientists to interpret the data-driven models from the viewpoint of material science and physics. Here, we demonstrate the development of spin-driven thermoelectric materials with anomalous Nernst effect by using an interpretable machine learning method called factorized asymptotic Bayesian inference hierarchical mixture of experts (FAB/HMEs). Based on prior knowledge of material science and physics, we were able to extract from the interpretable machine learning some surprising correlations and new knowledge about spin-driven thermoelectric materials. Guided by this, we carried out an actual material synthesis that led to the identification of a novel spin-driven thermoelectric material. This material shows the largest thermopower to date.

Editorial Summary

Identification of advanced spin-driven thermoelectric materials via interpretable machine learning探尋先進熱電材料:可解釋性機器學習

該研究證實了一種可解釋性的機器學習建模的工具,即分解的漸進貝葉斯推理專家分層混合(FAB / HME)。日本NEC公司中央研究實驗室的巖崎由馬教授領導的團隊,使用了最先進的可解釋機器學習模型,成功地展示了FAB / HMEs與材料科學家之間的協同工作和材料的研發。由于模型具有較高的預測能力和可解釋性,材料科學家可以從這些模型中獲得對新材料開發有用的新知識。通過可解釋模型得到的規律,作者成功地開發了一種自旋驅動的熱電材料,其熱電動勢SSTE大于目前的熱電材料。此外,他們從數據驅動模型中發現的新見解也可以使他們更全面地理解自旋驅動熱電材料背后的機理。因此,可解釋性的機器學習模型不僅可以幫助全新材料的開發,而且可以指導相關理論的研究

A utility of interpretable machine learning modeling, called factorized asymptotic Bayesian inference hierarchical mixture of experts (FAB/HMEs), in material development process is demonstrated. A team led by Prof. Yuma Iwasaki from the Central Research Laboratories, NEC Corporation, Japan, used state-of-the-art interpretable machine learning modeling to have shown a successful actual material development and synergy between the FAB/HMEs and the materials scientists. Because of model’s high predictive power and interpretability, materials scientists can obtain from such models non-trivial knowledge useful for novel materials development. Guided by the surprising correlation discovered by the interpretable model, the authors have succeeded in developing a spin-driven thermoelectric material, whose thermopower SSTE is larger than that of the current generation of thermoelectric materials. In addition, the novel insight they found from the data-driven model can lead to a more comprehensive understanding of the mechanism of emerging STE phenomena. Thus, the interpretable machine learning can help not only in the development of novel materials, but also in guiding the theoretical studies.

Absolute band alignment at semiconductor-water interfaces using explicit and implicit descriptions for liquid water (使用液態水的顯式和隱式描述在半導體-水界面處進行絕對帶對準)
Nicolas G. HormannZhendong GuoFrancesco AmbrosioOliviero AndreussiAlfredo Pasquarello & Nicola Marzari
npj Computational Materials 5:100(2019)
doi:s41524-019-0238-4
Published online:11 October 2019

Abstract| Full Text | PDF OPEN

摘要:量子力學模擬包含液體環境的作用,與固-液界面的表征息息相關,這對于各種器件的設計至關重要。本研究經過嚴格而系統的探索,通過一系列混合顯式/隱式模型與基于密度泛函理論的顯式原子模擬的對比,研究了水溶液中半導體的能帶對準。我們發現,如果第一層溶劑化殼被顯式處理,就可以獲得一致的結果。有趣的是,第一層的顯式水分子層僅與原始界面相關,即沒有解離吸附水的表面,暗示了在使用量子力學鍵進行描述時,界面飽和性的重要性。以顯式和隱式水相對于真空的平均靜電勢為基準,我們提供了絕對值校準,發現最大差異僅為~ 0.1~0.2 V。此外,隱式參考電勢相對于真空的固有偏移為-0.33 V,這是因為在隱式模型中,不存在顯式水界面的情況。這些結果為使用最小顯式/隱式模型精確模擬固-液界面鋪平了道路 

Abstract:Quantum mechanical simulations that include the effects of the liquid environment are highly relevant for the characterization of solid-liquid interfaces, which is crucial for the design of a wide range of devices. In this work we present a rigorous and systematic study of the band alignment of semiconductors in aqueous solutions by contrasting a range of hybrid explicit/implicit models against explicit atomistic simulations based on density-functional theory. We find that consistent results are obtained provided that the first solvation shell is treated explicitly. Interestingly, the first molecular layer of explicit water is only relevant for the pristine surfaces without dissociatively adsorbed water, hinting at the importance of saturating the surface with quantum mechanical bonds. By referencing the averaged electrostatic potentials of explicit and implicit water against vacuum, we provide absolute alignments, finding maximal differences of only  0.1–0.2 V. Furthermore, the implicit reference potential is shown to exhibit an intrinsic offset of -0.33 V with respect to vacuum, which is traced back to the absence of an explicit water surface in the implicit model. These results pave the way for accurate simulations of solid-liquid interfaces using minimalistic explicit/implicit models.

Editorial Summary

Absolute band alignment at semiconductor-water interfaces using explicit and implicit descriptions for liquid water液態水的描述:關乎半導體-水界面的帶對準

該研究測試了一些標準方法以使用隱式溶劑化模型精確模擬半導體的能帶對準,并以顯式模擬作對照對以下水中半導體進行了模型的驗證:具有分子吸附水的金紅石r-TiO2CdS、具有解離吸附水的GaN和同時具有這兩種情況的銳鈦礦α-TiO2GaAsGaP來自瑞士洛桑聯邦理工學院材料理論和模擬(THEOS)和國家材料計算設計與發現中心(MARVEL)的Nicolas G.Hormann領導的團隊發現,如果使用第一層水分子作顯式模擬,即可用自洽連續介質(SCCS)隱式水模型在不同材料和界面端以0.1-0.2 V的精度確定半導體-水界面處靜電勢的絕對校準。另一方面,不包含顯式水的模擬很可能具有1 V或更高的帶對準誤差。先前發現所有上述半導體端點均為穩定或亞穩定狀態,但作者指出,為了在隱式模型中準確地捕獲界面勢下降,必須明確地描述第一個溶劑化殼層。此外,他們發現,因隱式模型中不存在顯式水表面,隱式模型主體中的靜電勢比SCCS溶劑化的真空度降低約0.33 eV。該研究結果為在隱式環境中電化學固液界面的精確模擬提供了指導,并對界面水層的作用和性質以及界面水層可以被連續介質模型描述的程度有了新的認識。

Protocols to simulate accurately the band alignment of semiconductors leveraging implicit solvation models are tested and validated against explicit simulations for semiconductors in water, including (rutile) r-TiO2 and CdS with molecularly adsorbed water, GaN with dissociatively adsorbed water and (anatase) a-TiO2, GaAs and GaP with either of these two possibilities. A team led by Prof. Nicolas G. Hormann from the Theory and Simulation of Materials (THEOS), and National Centre for Computational Design and Discovery of Novel Materials (MARVEL), école Polytechnique Fédérale de Lausanne, Switzerland, revealed that the absolute alignment of electrostatic potentials at semiconductor-water interfaces can be determined with an accuracy of 0.1-0.2 V across different materials and interface terminations using the self-consistent continuum salvation (SCCS) implicit water model provided that the first water layer is simulated explicitly. On the other hand, simulations without inclusion of explicit water might very likely exhibit alignment errors of 1 V or more. All the terminations of above semiconductors were previously found to be stable or meta-stable, but the authors showed that it is necessary to describe the first solvation shell explicitly in order to capture interfacial potential drops accurately within an implicit model. Furthermore, they found that the electrostatic potential in the bulk of the implicit model is approximately 0.33 eV below the vacuum level for SCCS solvation, which can be explained by the absence of an explicit water surface in implicit models. Their research provides a guideline for accurate simulations of electrochemical solid-liquid interfaces in implicit environments and new insights in the effects and properties of interfacial water layers and in the extent that they can be described by continuum models.

A comparative study using state-of-the-art electronic structure theories on solid hydrogen phases under high pressures(使用最新的電子結構理論對高壓下固態氫相的比較研究)
Ke Liao,Xin-Zheng Li , Ali Alavi & Andreas Gruneis
npj Computational Materials 5:110(2019)
doi:s41524-019-0243-7
Published online:14 November 2019

Abstract| Full Text | PDF OPEN

摘要:使用理論方法確定維格納-亨廷頓過渡非常具有挑戰性。盡管近幾十年來現代從頭算理論取得了重大進展,但預先確定高壓下固體氫的原子結構和性質是高壓物理學的一個長期存在的問題,在行星和材料科學中具有深遠的意義。氫的壓力-溫度-相圖的確定,需要滿足極端條件,還需要對固體氫的組成電子和原子核分別作精確的量子力學處理,因而對于實驗和理論來說都很有挑戰性。本研究確切地證明了耦合簇理論可以作為計算高效的理論工具,來高精度地預測固體氫相。我們使用第一性原理計算,研究了100~450 GPa壓力下的固態氫相,將計算出的靜態晶格焓與擴散蒙特卡羅計算結果、密度泛函理論計算結果作了比較。本研究包括C2/c-24P211/c-24在內的最穩定相的耦合簇理論結果,與使用擴散蒙特卡洛法獲得的結果很好地相互吻合,只是Cmca-4例外,因該相被預測為很不穩定相。我們討論了幾個所用方法的適用范圍,以及它們在有效、互補的理論工具方面的作用,以了解和解決高壓下固態氫相的老大難問題。 

Abstract:Determining the Wigner–Huntington transition using theoretical methods is extremely challenging. Despite the significant advancements of modern ab initio theories in the past decades, the preidentifying the atomic structure and properties of solid hydrogen under high pressures is a long-standing problem of high-pressure physics with far-reaching significance in planetary and materials science. Determining the pressure-temperature phase diagram of hydrogen is challenging for experiment and theory due to the extreme conditions and the required accuracy in the quantum mechanical treatment of the constituent electrons and nuclei, respectively. Here, we demonstrate explicitly that coupled cluster theory can serve as a computationally efficient theoretical tool to predict solid hydrogen phases with high accuracy.We present a first principles study of solid hydrogen phases at pressures ranging from 100 to 450 GPa. The computed static lattice enthalpies are compared to state-of-the-art diffusion Monte Carlo results and density functional theory calculations.Our coupled cluster theory results for the most stable phases including C2/c-24 and P211/c-24 are in good agreement with those obtained using diffusion Monte Carlo, with the exception of Cmca-4, which is predicted to be significantly less stable.We discuss the scope of the employed methods and how they can contribute as efficient and complementary theoretical tools to solve the long-standing puzzle of understanding solid hydrogen phases at high pressures.

Editorial Summary

Solid hydrogen phases: electronic structure theories (使用最新的電子結構理論對高壓下固態氫相的比較研究 還要做金屬氫嗎?

本研究證明,量子化學波函數理論有望作為一種有效而準確的工具,用于研究固體氫的高壓相。特別地,來自德國馬克斯·普朗克固體研究所的Ke Liao和奧地利維也納工業大學理論物理研究所的AndreasGruneis共同領導的團隊,發現耦合簇理論在采用新近開發的技術后,可在計算成本和準確性之間取得良好的權衡,這些技術可以有效地模擬周期系統的熱力學極限。他們注意到,這些有限的尺寸校正為許多從頭算研究鋪平了道路,包括分子-表面相互作用的預測、碳和氮化硼同素異形體的壓力-溫度-相圖。他們的研究表明,對于沒有強相關性的固態系統,耦合簇方法的精度與擴散蒙特卡羅方法接近。而且,耦合簇方法已在模型氫系統中,以各種更準確的方法作了基準測試,顯示了在弱相關情況下該方法的高度精確性。此外,在這項針對小型系統的研究中,作者用全構象相互作用量子蒙特卡洛方法,驗證了耦合簇方法的有效性

Quantum chemical wave function theories which hold the promise to serve as an efficient and accurate tool for the investigation of high-pressure phases of solid hydrogen are demonstrated. A team co-led by Ke Liao from the Max Planck Institute for Solid State Research, Germany, and Andreas Gruneis from the Institute for Theoretical Physics, Vienna University of Technology, Austria, particularly found that coupled cluster theory achieves a good trade-off between computational cost and accuracy when employing recently developed techniques that allow for simulating the thermodynamic limit of periodic systems in an efficient manner. They note that these finite size corrections have paved the way for a number of ab initio studies, including predictions of molecule–surface interactions and pressure–temperature phase diagrams of carbon and boron nitride allotropes. The studies referred to above have demonstrated that coupled cluster methods achieve a similar level of accuracy as DMC for solid state systems that are not strongly correlated. Moreover, coupled cluster methods have been benchmarked against various more accurate methods in model hydrogen systems, showing the high accuracy of the methods in weakly correlated situations. Furthermore, the authors employ full configuration interaction Quantum Monte Carlo in their work for small systems to examine the validity of the coupled cluster method.

Topological phases in pyrochlore thallium niobateTl2Nb2O6+x(燒綠石結構鈮酸鉈Tl2Nb2O6+x中的拓撲相)
Wei Zhang,Kaifa Luo, Zhendong Chen,Ziming Zhu, Rui Yu, Chen Fang &Hongming Weng
npj Computational Materials 5:105(2019)
doi:s41524-019-0245-5
Published online:1 November 2019

Abstract| Full Text | PDF OPEN

摘要:新型拓撲電子材料的發現帶來了發現新物理的機會。迄今,很多拓撲材料已經由理論提出并被實驗驗證。可惜的是,拓撲氧化物的存在仍舊缺少令人信服的實驗證據。這是因為氧元素的強氧化作用通常會導致離子晶體的形成,使能帶反轉不太可能出現。另外,在單一材料中實現不同拓撲電子態非常困難,但這對于探索拓撲相變又是急需的。本研究運用第一性原理計算和對稱性分析方法,提出在燒綠石結構的Tl2Nb2O6+x0x 1.0)中,氧固溶體在實驗上的連續可調會導致多種不同的拓撲電子態。通過改變氧含量和/或調節晶體對稱性,可以得到拓撲絕緣體、狄拉克半金屬和三重簡并點半金屬等。當x = 1時,體系是半金屬,其費米能級處具有二次型色散接觸點。在平面內不同應變的作用下,其會轉變為Dirac半金屬或者拓撲絕緣體。當x = 0.5時,Tl2Nb2O6.5體系空間反演對稱性的自然破缺導致了三重簡并點的產生。當x 0時,Tl2Nb2O6成為窄能隙的普通絕緣體。這些由氧的固溶體驅動的拓撲相變是獨特的、合乎物理的,其本質是源于T1+Tl3+的價態改變。這種拓撲氧化物將有望用于電子關聯引起的拓撲電子態的研究及其潛在應用   

Abstract: discovery of new topological electronic materials brings a chance to uncover new physics. Up to now, many materials have been theoretically proposed and experimentally proved to host different kinds of topological states.Unfortunately, there is little convincing experimental evidence for the existence of topological oxides. The reason is that oxidation of oxygen leads to ionic crystal in general and makes band inversion unlikely. In addition, the realization of different topological states in a single material is quite difficult, but strongly needed for exploring topological phase transitions. In this work, using first-principles calculations and symmetry analysis, we propose that the experimentally tunable continuous solid solution of oxygen in pyrochlore Tl2Nb2O6+x (0x1.0) leads to various topological states. Topological insulator, Dirac semimetal, and triply degenerate nodal point semimetal can be realized in it via changing the oxygen content and/or tuning the crystalline symmetries. When x=1, it is a semimetal with quadratic band touching point at Fermi level. It transits into a Dirac semimetal or a topological insulator depending on the in-plane strain. When x=0.5, the inversion symmetry is spontaneously broken in Tl2Nb2O6.5, leading to triply degenerate nodal points. When x=0, Tl2Nb2O6 becomes a trivial insulator with a narrow band gap. These topological phase transitions driven by solid solution of oxygen are unique and physically plausible due to the variation of valence state of Tl+ and Tl3+. This topological oxide will be promising for studying correlation induced topological states and potential applications. 

Editorial Summary

New topological oxidepromotes the diversity of topological electronic states福建師大張薇聯手翁紅明、余睿:一窺拓撲電子態萬花筒

燒綠石結構的Tl2Nb2O6+x0x 1.0)是氧的固溶體,且具有氧含量在實驗上連續可調的特性。本研究報道了此特性使得Tl2Nb2O6+x展現出獨特的拓撲相變。福建師范大學的張薇副教授與中國科學院物理研究所的翁紅明研究員、武漢大學的余睿教授等合作,運用第一性原理計算和對稱性分析方法,提出通過調控氧含量及/或晶體對稱性,可在Tl2Nb2O6+x體系中實現多種不同的拓撲電子態:包括拓撲絕緣體、狄拉克半金屬和三重簡并點半金屬等。T1價態由+1轉變成+3導致了這些拓撲相變的產生。該氧化物系列拓撲性質的研究將有助于關聯拓撲電子態的探索及其潛在應用

The pyrochlore Tl2Nb2O6+x (0 x 1.0) is a solid solution of oxygen, whose oxygen content iscontinuously tunable in experiments. This study reported that the solid solution of oxygen led to unique topological phase transitions. Associate Professor Wei Zhang of Fujian Normal University, in collaboration with Professor Hongming Weng at the Institute of Physics, Chinese Academy of Sciences, and Professor Rui Yu from Wuhan University, using first-principles calculations and symmetry analysis, proposedthat a variety of topological electronic states could be achieved in the Tl2Nb2O6+x system,by tuning oxygen content and/or crystal symmetry. These topological electronic states include topological insulator, Dirac semimetal, and triply degenerate nodal point semimetal. The topological phase transitions are due to the T1 valence transition from +1 to +3. The study of the topological oxide will be promising for the exploration of the correlated topological electronic states and potential applications.

Ternary superconducting cophosphorus hydrides stabilized via lithium (通過鋰摻雜穩定的三元超導磷氫化合物)
Ziji Shao, Defang Duan, Yanbin Ma, Hongyu Yu, Hao Song, Hui Xie, Da Li, Fubo Tian, Bingbing Liu & Tian Cui
npj Computational Materials 5:104(2019)
doi:s41524-019-0244-6
Published online:30 October 2019

Abstract| Full Text | PDF OPEN

摘要:受高壓下硫化氫200 K超導電性和磷化氫復雜結構相變的啟發,本研究將第一性原理計算與晶體結構預測技術相結合,探索高壓下Li-P-H三元化合物的穩定結構,以期發現新型高溫超導體。研究發現在P-H體系中通過摻雜鋰元素可以使得磷氫化合物在高壓下穩定,在100300 GPa的壓力范圍內,發現了四個穩定存在的化合物LiPH3LiPH4LiPH6LiPH7。尤其值得關注的是,理論預測具有對稱性的原子相LiPH6是潛在的高溫超導體,在200 GPa時超導轉變溫度為150-167 K,并且隨著壓力的增加超導轉變溫度降低。除了-LiPH6是離子型原子相外,其余三個穩定存在的三元氫化物中都包含PH共價框架,而離子鋰分布在框架周圍。本研究工作還提出了合成三元鋰磷氫化物的一種可能路徑:LiP + H2LiPHn,為將來的高壓合成實驗提供了有益和明確的指導,并為三元超導氫化物的進一步探索提供了參考   

Abstract:Inspired by the diverse properties of sulfur hydrides and phosphorus hydrides, we combine first-principles calculations with structure prediction to search for stable structures of LiPH ternary compounds at high pressures with the aim of finding novel superconductors. It is found that phosphorus hydrides can be stabilized under pressure via additional doped lithium. Four stable stoichiometries LiPH3, LiPH4, LiPH6, and LiPH7 are uncovered in the pressure range of 100–300 GPa. Notably, we find an atomic LiPH6 with symmetry which is predicted to be a potential high-temperature superconductor with a Tc value of 150–167K at 200GPa and the Tc decreases upon compression. All the predicted stable ternary hydrides contain the P–H covalent frameworks with ionic lithium staying beside, but not for –LiPH6. We proposed a possible synthesis route for ternary lithium phosphorus hydrides: LiP+H2→LiPHn, which could provide helpful and clear guidance to further experimental studies. Our work may provide some advice on further investigations on ternary superconductive hydrides at high pressure. 

Editorial Summary

Conventional high-temperature superconductors:  lithium-doped phosphorus hydrides吉林大學預測傳統高溫超導體:鋰摻雜磷氫化合物

該研究工作預測了一種新型的三元傳統高溫超導體LiPH6,該材料在200萬大氣壓的高壓下超導轉變溫度可達167 K,遠高于液氮溫度(77 K)。來自吉林大學物理學院的崔田教授/段德芳副教授課題組,通過第一性原理計算結合晶體結構搜索技術,預測了一系列高壓下穩定存在的三元氫化物LiPH3LiPH4LiPH6LiPH7,這類材料是由Li摻雜二元磷氫化合物形成的。母體氫化物P-H體系在高壓下不穩定,會分解為單質PH2,而引入額外的電子可以穩定P-H體系,在高壓下形成穩定的三元化合物。其中超導性質最為突出的是LiPH6,在200萬大氣壓的壓力下超導轉變溫度可達167 K。該研究工作還提出了合成三元鋰磷氫化合物的一種可能路徑:LiP + H2→LiPHn,為高壓下三元氫化物超導材料的深入研究提供了重要參考

In this study, a new conventional high temperature superconductor LiPH6 is predicted. The critical temperature Tof this material is calculated to be 167 K at 200 GPa, which is much higher than liquid nitrogen temperature (77 K). The research group of Professor Tian Cui/ Associate Professor Defang Duan from the College of Physics of Jilin University, predicted a series of stable ternary hydrides LiPH3, LiPH4, LiPHand LiPH7 under high pressure by means of first-principles calculations and structure prediction. These materials can be formed through doping lithium in binary phosphorus hydrides. The host P-H system is unstable under high pressure and will decompose into elemental P and H2. The introduction of additional electrons stabilizes the P-H system and form a series of stable ternary compounds at high pressure. Notably, an atomic LiPH6 with symmetry which is predicted to be a potential high-temperature superconductor with aTvalue of 150–167K at 200GPa. In addition, the research also proposed a possible synthesis route for ternary lithium phosphorus hydrides: LiP+H2→LiPHn, which could provide helpful and clear guidance to further experimental studies. This work may provide some advice on further investigations on ternary superconductive hydrides at high pressure.

Tensorial stress-strain fields and large elastoplasticity as well as friction in diamond anvil cell up to 400GPa (承受400 GPa的金剛石砧座單元:拉伸應力-應變場、大彈性塑性和摩擦)
Valery I. LevitasMehdi Kamrani & Biao Feng
npj Computational Materials 5:94(2019)
doi:s41524-019-0234-8
Published online:1 October 2019

Abstract| Full Text | PDF OPEN

摘要:在極端壓力下金剛石砧座單元中的各種現象(斷裂、相變和化學反應)都受到應力和塑性應變張量所有分量場的強烈影響。然而,它們卻無法測量。本研究提出一種實驗-理論-計算耦合的測量方法,允許(使用已公開的實驗數據)優化、校準和驗證高達400 GPa的鎢(W)和金剛石彈塑性行為模型和接觸摩擦模型,并重建W和金剛石中所有應力分量和大塑性應變張量。盡管人們已普遍接受了應變引起的各向異性、應變硬化和路徑依賴的可塑性,但本研究表明,大塑性應變后的W表現為各向同性、完美塑性和獨立于路徑的完美塑性表面。此外,即使在如此大的場梯度中,也發現了彈塑性特性的尺度非依賴性。本研究結果為定量極端應力科學和創紀錄高壓打開了新的機遇   

Abstract:Various phenomena (fracture, phase transformations, and chemical reactions) studied under extreme pressures in diamond anvil cell are strongly affected by fields of all components of stress and plastic strain tensors. However, they could not be measured. Here, we suggest a coupled experimental theoretical computational approach that allowed us (using published experimental data) to refine, calibrate, and verify models for elastoplastic behavior and contact friction for tungsten (W) and diamond up to 400GPa and reconstruct fields of all components of stress and large plastic strain tensors in W and diamond. Despite the generally accepted strain-induced anisotropy, strain hardening, and path-dependent plasticity, here we showed that W after large plastic strains behaves as isotropic and perfectly plastic with path-independent surface of perfect plasticity. Moreover, scale-independence of elastoplastic properties is found even for such large field gradients. Obtained results open opportunities for quantitative extreme stress science and reaching record high pressures. 

Editorial Summary

Diamond anvil cell: tensorial stress-strain fields, large elastoplasticity and friction攬瓷器活的金剛砧:需要承受多少委屈?

靜態高壓研究中,兆伏壓力是由金剛砧座單元中的兩顆金剛石對稀薄樣品的壓縮而產生的。但金剛砧所受的應力和塑性應變張量卻無法測量。該研究提出了一種新穎的實驗-理論-計算耦合的測試方法,這一方法可使用已知實驗數據來構建模型,從而提取有關彈塑性和摩擦方式的完整信息,以及金剛砧單元(diamond anvil cells)中的材料在受極端壓力時的各種復雜張量場的完整信息。來自美國愛荷華州立大學航空航天工程和機械工程系的Valery I. Levitas教授團隊,優化、校準和驗證了任意樣品的彈塑性行為模型及鎢(W)樣品可高達400 GPa的接觸摩擦,重建了W和金剛石中的應力和大塑性應變張量的所有分量場。該研究除了獲取屈服強度和摩擦的壓力依賴性定量信息、高階彈性常數的定量信息外,還揭示了極大應變和壓力下,彈塑性行為的一些共同特性:1)盡管人們已普遍認定應變會導致各向異性、應變硬化和路徑相關的塑性,但大塑性應變之后的W卻表現為各向同性,且未表現出應變硬化和路徑依賴。2)盡管樣品厚度達μm級、應力可達5 GPa /μm和塑性應變梯度巨大,但樣品的彈塑性特性卻具有非尺度依賴性。這兩種性能都極大地簡化了可塑性理論和極端條件下的測量。該研究在塑性方面的發現還暗示了在這種極端載荷下塑性摩擦的重要特性:塑性摩擦與塑性應變、塑性應變路徑和尺寸毫無關聯

A novel coupled experimental theoretical computational approach is suggested, which allowed to extract complete information about elastoplastic properties and friction rules, as well as all complex tensorial fields using known experimental data for materials compressed in a diamond anvil cells under extreme pressure. A team led by Prof. Valery I. Levitas from the Departments of Aerospace Engineering and Mechanical Engineering, Iowa State University, USA, refined, calibrated, and verified models for elastoplastic behavior of a sample and contact friction for W up to 400GPa and reconstruct fields of all components of stress and large plastic strain tensors in W and diamond. In addition to quantitative information on the pressure dependence of the yield strength and friction, as well as higher-order elastic constants, they justifed some general unique properties of elastoplastic behavior under very large strains and pressures: (a) Despite the generally accepted strain-induced anisotropy, strain hardening, and path-dependent plasticity, W after large plastic strains behaves isotropically and does not exhibit strain hardening and path dependence. (b) Despite the μm-sized sample thickness and huge stress (5GPa/μm) and plastic strain gradients, scale-independence of elastoplastic properties is found. Both of these properties drastically simplify plasticity theory and measurements under extreme conditions. Their finding for plasticity also implies important properties for plastic friction under such extreme loading: Plastic friction is plastic strain-, plastic strain path-, and scale-independent.

Interpretable deep learning for guided microstructure-property explorations in photovoltaics (基于可解釋性深度學習探索光伏引導性微結構-性能)
Balaji Sesha Sarath PokuriSambuddha GhosalApurva KokateSoumik Sarkar & Baskar Ganapathysubramanian
npj Computational Materials 5:95(2019)
doi:s41524-019-0231-y
Published online:1 October 2019

Abstract| Full Text | PDF OPEN

摘要:微觀結構決定了薄膜有機半導體膜的光伏性能。微觀結構與宏觀性能之間的構效關系通常呈高度非線性化,其評估成本很高,因此微觀結構的優化具有相當大的挑戰性。本研究展示了一種數據驅動的方法,使用深度卷積神經網絡將微觀結構映射到光伏性能。我們用兩個關鍵指標來表征這一方法:可泛化性(是否學到了一個合理的映射?)和可解釋性(能否產生有意義的微觀結構特征來幫助其預測?)。同時具又這兩個關鍵指標的替代模型對于后續的設計和開發是非常有價值的。我們通過使用替代模型來說明這一點,該模型用于手動探索(驗證已知領域的洞察力)和微觀結構的自動優化。我們認為這樣的方法將可廣泛應用于各種對微結構敏感的設計問題   

Abstract:The microstructure determines the photovoltaic performance of a thin film organic semiconductor film. The relationship between microstructure and performance is usually highly non-linear and expensive to evaluate, thus making microstructure optimization challenging. Here, we show a data-driven approach for mapping the microstructure to photovoltaic performance using deep convolutional neural networks. We characterize this approach in terms of two critical metrics, its generalizability (has it learnt a reasonable map?), and its intepretability (can it produce meaningful microstructure characteristics that influence its prediction?). A surrogate model that exhibits these two features of generalizability and intepretability is particularly useful for subsequent design exploration. We illustrate this by using the surrogate model for both manual exploration (that verifies known domain insight) as well as automated microstructure optimization. We envision such approaches to be widely applicable to a wide variety of microstructure-sensitive design problems. 

Editorial Summary

Interpretable deep learning: microstructure-property explorations名字與火車一樣長的教授等:微結構-大性能的玩法

該研究利用卷積神經網絡(CNN)的多功能性,將薄膜有機光伏(OPV)的活性層形態映射到性能指標——短路電流Jsc。來自美國愛荷華州立大學機械工程系的Soumik Sarkar教授和Baskar Ganapathysubramanian(這個姓恰好20字母)教授等,訓練了一種形態分類器,可將OPV形態和短路電流關聯到一起。他們測試了幾種深度和寬度不同的學習框架,這些學習框架可以從給定的一組形態及其標簽中學習,最后的結果發現該學習框架具有很高的準確性和F1得分。為了區分這些同樣表現良好的模型并對它們進行排序,作者使用了兩個額外的附加度量。首先是揭示習得的結構-性質關系的概括性。他們確定了網絡架構,可以用可用的數據集概括地圖,并根據“將未看到的形態投射到已學習過的分布上”的能力進行量化,并做出了良好的預測。其次是可解釋性。這是理解工程系統行為的相當重要的指標。他們引入了一種稱為DLSP(結構屬性查詢深度學習)的方法,用于從數據中學習結構-性能之間的構效關系。作者認為這種方法可廣泛應用于各種對微結構敏感的設計問題

Versatility of Convolutional Neural Networks (CNN) is reported to map the active layer morphology of thin film organic photovoltaics (OPV) to a performance metric, the short-circuit current Jsc. A team co-led by Prof. Soumik Sarkar and Prof. Baskar Ganapathysubramanian from the Department of Mechanical Engineering, Iowa State University, USA, trained a morphology classifier that maps an OPVmorphology to a short-circuit current.They tested several architectures (of varying depth and width) that can learn from a given set of morphologies and their labels, and demonstrated very high accuracy, and F1 score. To distinguish and rank order between these equally well performing models, the authors used two additional measures. The first is generalizability to reveal the learnt structure-property relationship. They identified network architectures that can generalize the map with the available dataset and quantify in terms of the ability to ‘project the unseen’ morphology onto the learnt distribution and made good predictions.The second is interpretability, a fairlyimportant metric for understanding the behavior of engineered systems.They introduced an approach called DLSP (Deep Learning for Structure Property interrogation) for learning the structure-property relationship from data. The authorsenvision such approaches to be widely applicable to a wide variety of microstructure-sensitive design problems.

版權所有 © 中國科學院上海硅酸鹽研究所
地址:上海市長寧區定西路1295號 郵政編碼:200050
心水论坛高手榜