WO2023097643A1 - 可编程理性设计纳米晶体材料的方法 - Google Patents

可编程理性设计纳米晶体材料的方法 Download PDF

Info

Publication number
WO2023097643A1
WO2023097643A1 PCT/CN2021/135297 CN2021135297W WO2023097643A1 WO 2023097643 A1 WO2023097643 A1 WO 2023097643A1 CN 2021135297 W CN2021135297 W CN 2021135297W WO 2023097643 A1 WO2023097643 A1 WO 2023097643A1
Authority
WO
WIPO (PCT)
Prior art keywords
module
experimental
nanocrystal
reaction
preparation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2021/135297
Other languages
English (en)
French (fr)
Inventor
赵海涛
喻学锋
陈薇
陈子健
张雪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Priority to PCT/CN2021/135297 priority Critical patent/WO2023097643A1/zh
Publication of WO2023097643A1 publication Critical patent/WO2023097643A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/20Design optimisation, verification or simulation

Definitions

  • the invention belongs to the field of new material preparation technology and digital manufacturing technology, and in particular relates to the research on the fusion mechanism of automatic preparation of robot process at the macro-scale level and the regulation and control of crystal growth by morphology control agents at the atomic scale level, so as to realize the automation and controllability of robot-assisted nanocrystal morphology control synthesis.
  • the purpose of the present invention is to provide a digital controllable synthetic robot auxiliary equipment of the dynamic morphology control mechanism of nanocrystals, which realizes the preparation of digital and programmable robot-assisted nanocrystals.
  • the growth kinetics of the nanocrystals was studied, and the growth kinetics model of nanocrystals was constructed, and the digital, automatic and controllable synthesis of nanocrystals with controllable morphology was realized.
  • One aspect of the present invention provides a digital and programmable robotic auxiliary equipment for nanocrystal preparation
  • the equipment includes a nanocrystal preparation reaction execution module, a nanocrystal growth solution information acquisition module, and a data storage and calculation module
  • the preparation reaction execution module includes components for performing the preparation reaction
  • the nanocrystal growth liquid information collection module includes components for obtaining nanocrystal optical absorption topography features or color features
  • the data storage and calculation module includes components for experimental data, Experimental results, storage components for collecting data, and computing modules for controlling the execution module of the nanocrystal preparation reaction and the information collection module of the nanocrystal reaction liquid;
  • the calculation module includes a literature method digitization module, an experiment scheme editing module, an experiment scheme execution control module, an experiment data analysis module and an experiment result display module;
  • the document method digitization module realizes the digitization of the preparation scheme in the document by extracting the reaction raw material types, raw material consumption and reaction conditions in the document method; the digitized data is stored in the storage unit and can be called by the calculation module;
  • the experimental scheme editing module can edit the scheme to be tested, can call the digital preparation scheme of the literature method digitization module or call the experimental scheme calculated by the experimental data analysis module after modeling; where the scheme to be tested is edited as the experimental raw material for editing Type, concentration of experimental raw materials, volume of experimental raw materials, order of adding different raw materials, experimental reaction temperature, reaction time, type of experimental result collection;
  • the experimental program execution control module can control the nanocrystal preparation reaction execution module and the nanocrystal reaction liquid information collection module to realize the experimental program set by the experimental program editing module;
  • the experimental data analysis module can call the optical absorption data obtained by the nanocrystal growth solution information collection module and the corresponding experimental plan information, and learn and analyze the called information through the method of machine learning;
  • the experimental result display module can visually display the model established by the machine learning obtained by the experimental data analysis module.
  • the preparation reaction execution module includes components that automatically perform liquid extraction, liquid separation, mixing, vibration, and temperature rise, and the components that perform liquid extraction and liquid separation are selected from pipetting robotic arms, walking robots, six-axis robotic arms, 8 Channel pipette gun, 96-well plate, heating furnace, temperature-raising components are selected from temperature-controlled heating muffle furnace, and vibration components are selected from 96-well plate mixing vibrator.
  • the nanocrystal reaction solution information collection module includes at least one of a color ultra-sensitive high-definition camera system, an ultraviolet-visible-near-infrared absorption spectrometer, a transmission electron microscope, and an X-ray diffractometer.
  • Another aspect of the present invention provides an automatic preparation method for preparing nanocrystals.
  • the preparation method includes the preparation using the digital and programmable robot auxiliary equipment for nanocrystal preparation described in the present invention.
  • Another aspect of the present invention provides a modeling method of nanocrystal growth kinetics, which includes the following steps:
  • the nanocrystal morphology characteristic data described in S1) is selected from any one of UV-visible-near-infrared absorption spectrum data, OD LSPR , LSPR value, and RGB color data of the reaction solution.
  • time point of data collection in S1) is from the start of the reaction until the OD LSPR of the reaction no longer increases.
  • the method of machine learning in S2) is selected from the SISSO algorithm.
  • nano crystals are selected from nano gold crystals.
  • Another aspect of the present invention provides a method for controllable preparation of nanocrystals, the method includes using the model established by the above modeling method to predict the experimental conditions obtained, and preparing by the above equipment.
  • the invention adopts robot assistance, and based on the nanocrystal dynamic shape control mechanism, carries out the digital controllable synthesis of the nanocrystal, and realizes the automatic synthesis of the nanocrystal with a specified shape.
  • the invention pioneered the design of fully automatic synthesis equipment for nanocrystals. Based on the combination of big data and experimental data, the kinetic control method for preparing nanocrystals with a specified shape was obtained through machine learning, and the automatic preparation components were designed to automate the synthesis process. synthesis.
  • Figure 1 is a 3D schematic diagram of the overall layout of the platform, including a walking robot, an automatic liquid suction (dispensing) module, a color ultra-sensitive high-definition camera system, an intelligent material storage warehouse, and an electric valve control pull-open muffle furnace.
  • Figure 2 is a schematic diagram and code of the experimental scheme extracted from the literature.
  • Figure 3 is based on the experimental program design and executable program in Figure 2.
  • Figure 4 is a schematic diagram of the collection, automatic extraction and analysis of high-throughput experimental results.
  • Figure 5 is a schematic diagram of high-throughput experimental machine learning fitting model and visualization.
  • Fig. 6 is the optical absorption spectrum of the gold nanorod solution at 0-360 minutes under three experimental conditions.
  • Fig. 7(A) is the curve of the absorbance OD LSPR corresponding to the maximum value of LSPR on the absorption spectrum as a function of time t under three experimental conditions.
  • (B) is the curve of the color (RGB) value of the reaction solution changing with time t under three experimental conditions.
  • (C) ln(OD[Au + ])) calculated from the absorption spectra OD LSPR and OD Lmax has a linear relationship with time t, indicating that the growth of gold nanoparticles conforms to the first-order reaction kinetics.
  • a digital and programmable robot auxiliary device for nanocrystal preparation includes a nanocrystal preparation reaction execution module, a nanocrystal reaction liquid information collection module, and a data storage and calculation module;
  • the preparation reaction execution module includes components for performing the preparation reaction;
  • the nanocrystal reaction liquid information collection module includes components for obtaining nanocrystal morphology or color characteristics;
  • the data storage and calculation module includes components for experimental data, experimental results,
  • the calculation module includes a literature method digitization module, an experiment scheme editing module, an experiment scheme execution control module, an experiment data analysis module and an experiment result display module;
  • the digitization module of the literature method realizes the digitization of the preparation scheme in the literature by extracting the types of reaction raw materials, the amount of raw materials, and the reaction conditions in the literature method; the digitized data is stored in the storage unit and can be called by the calculation module; The schematic diagram and code of the experimental scheme are shown in Figure 2
  • the preparation reaction execution module includes components that automatically perform liquid extraction, liquid separation, mixing, vibration, and heating.
  • the nanocrystal reaction solution information collection module includes at least one of a color ultra-sensitive high-definition camera system, an ultraviolet-visible light-near-infrared absorption spectrometer, a transmission electron microscope, and an X-ray diffractometer.
  • the experimental scheme editing module can edit the scheme to be tested, can call the digital preparation scheme of the literature method digitization module or call the experimental scheme calculated by the experimental data analysis module after modeling; where the scheme to be tested is edited as the experimental raw material for editing Type, concentration of experimental raw materials, volume of experimental raw materials, order of adding different raw materials, experimental reaction temperature, reaction time, type of experimental result collection;
  • the experimental program execution control module can control the nanocrystal preparation reaction execution module and the nanocrystal reaction liquid information acquisition module to realize the experimental program set by the experimental program editing module; in a specific embodiment, the experimental program based on Fig. 2
  • the design and executable program are shown in Figure 3.
  • the experimental data analysis module can call the data obtained by the nanocrystal reaction liquid information collection module and the corresponding experimental plan information, and learn and analyze the called information through the method of machine learning;
  • the experimental result display module can visually display the model established by the machine learning obtained by the experimental data analysis module.
  • a schematic diagram of a high-throughput experimental machine learning fitting model is shown in FIG. 5 .
  • experimental condition 1 10 ⁇ L of AgNO 3 , 0.1M CTAB and 0.2M HCl, 0.27M and 0.4M experiments.
  • Experimental condition 2 10 ⁇ L of AgNO 3 , 0.1M CTAB and 0.27M HCl were used for the reaction.
  • Experimental condition 3 Use 12 ⁇ L of AgNO 3 , 0.12M CTAB and 0.42M HCl to react, wherein the concentration of AgNO 3 is 0.01M, the volume of CTAB solution is 1 mL, the volume of HCl is 20 ⁇ L, and the remaining experimental drugs are added 50 ⁇ L of 0.1M HAuCl4 solution, 8 ⁇ L of 0.1M ascorbic acid solution and 2.4 ⁇ L of pre-prepared gold seed solution. And set the ultraviolet-visible light-near-infrared optical absorption spectrum measurement within 0min-360min respectively, the experimental results are shown in Figure 6. In addition, the color transformation during the growth of gold nanocrystals was continuously tested by a color ultra-sensitive high-definition camera.
  • t represents the reaction time
  • OD LSPR represents the absorbance value corresponding to the peak value of LSPR at the test time
  • OD Lmax represents the maximum absorbance value after the reaction is completed.
  • R represents the color R value during the test
  • Rmin represents the minimum R value when the reaction is completed
  • the R value gradually decreases as the gold nanoparticles grow.
  • the squares, circles and triangles represent different reaction conditions, namely the experiment with C(HCl) of 0.2M in the square group, 0.27M in the triangle group and 0.42M in the circle group.
  • a programmable kinetic model was then established using the machine learning SISSO algorithm.
  • the expression of the growth kinetic model is shown in Table 1.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Investigating Or Analysing Materials By Optical Means (AREA)

Abstract

本发明涉及一种可编程理性设计纳米晶体材料的方法。具体公开了,全流程的可编程理性基于机器人辅助设备包括纳米晶体制备反应执行模块、纳米晶体反应液信息采集模块以及数据存储和计算模块。通过该设备实现了纳米晶体的生长动力学研究,建立了生长动力学模型以及相关数据库,通过机器学习实现了可编程理性设计方法。本发明的设备全自动化,平行性和准确性更高,本发明验证的不同实验条件下都能实现可靠的结果。

Description

可编程理性设计纳米晶体材料的方法 技术领域
本发明属于新型材料制备技术与数字制造技术领域,尤其涉及研究宏观尺度层面机器人流程自动化制备和原子尺度层面形貌控制剂调控晶体生长的融合机制,实现机器人辅助纳米晶形貌控制的自动化可控合成。
背景技术
近年来,国内外十分关注数字驱动材料创新的研究,发展材料与人工智能技术的交叉融合的优势越来越明显:效率和精度在提升[1-3];对计算资源的依赖在降低[4,5];机器学习算法和自训练能力在加强[6,7];学习尺度和数据库在拓宽[8,9];功能材料逆向开发在逐渐成熟[1,10]。然而,可靠的实验数据和特定性质的描述符(Descriptor)限制了材料领域人工智能的发展,目前以第一性原理计算数据与机器学习相结合(辅以实验)的研究为主[11-15],机器人/流程自动化与人工智能相结合的材料数字化和智能化制备研究方兴未艾[16-20]。
流程自动化的有机材料制备是当前研究的热点:Li、Ballmer等人[18]研究了14种不同的有机小分子全自动制备,为更普遍的有机小分子自动制备明确了可行的路线图;Cronin等人[19]开发了一种机器学习驱动的有机合成机器人,可以在进行少量实验后预测可能出现的反应。在生物材料方面,何凯等人[21]研发了应用于合成生物学的自动化实验平台;Ager等人[22]结合自动化和计算机技术开发了用于生物催化剂性能优化的工作站。在无机材料研究方面,Burger和Cooper等人[20]提出使用移动式机器人,筛选用于水制氢的光催化剂,注重多种催化剂配方评价实验而非新型材料制备环节。
然而,上述研究基本思路主要集中于有机材料和生物材料流程自动化和移动机器人材料性能评价,采用机器人/机器手臂系统辅助对于纳米晶体的动力学研究鲜有报道。
1.Zunger,A.,Inverse design in search of materials with target functionalities.Nature Reviews Chemistry,2018.2(4):p.0121.
2.Hautier,G.,A.Jain,and S.P.Ong,From the computer to the laboratory:materials discovery and design using first-principles calculations.Journal of Materials Science,2012.47(21):p.7317-7340.
3.Liu,Y.,et al.,Materials discovery and design using machine learning.Journal of Materiomics,2017.3(3):p.159-177.
4.Chen,C.,et al.,A Critical Review of Machine Learning of Energy Materials.Advanced  Energy Materials,2020.10(8):p.1903242.
5.Schleder,G.R.,C.M.Acosta,and A.Fazzio,Exploring Two-Dimensional Materials Thermodynamic Stability via Machine Learning.ACS Applied Materials&Interfaces,2020.12(18):p.20149-20157.
6.Jablonka,K.M.,et al.,Big-Data Science in Porous Materials:Materials Genomics and Machine Learning.Chemical Reviews,2020.120(16):p.8066-8129.
7.Ward,L.,et al.,A general-purpose machine learning framework for predicting properties of inorganic materials.npj Computational Materials,2016.2(1):p.16028.
8.Tabor,D.P.,et al.,Accelerating the discovery of materials for clean energy in the era of smart automation.Nature Reviews Materials,2018.3(5):p.5-20.
9.Chen,C.,et al.,Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals.Chemistry of Materials,2019.31(9):p.3564-3572.
10.Sanchez-Lengeling,B.and A.Aspuru-Guzik,Inverse molecular design using machine learning:Generative models for matter engineering.Science,2018.361(6400):p.360.
11.Bruix,A.,et al.,First-principles-based multiscale modelling of heterogeneous catalysis.Nature Catalysis,2019.2(8):p.659-670.
12.Zhong,M.,et al.,Accelerated discovery of CO2electrocatalysts using active machine learning.Nature,2020.581(7807):p.178-183.
13.Seh,Z.W.,et al.,Combining theory and experiment in electrocatalysis:Insights into materials design.Science,2017.355(6321):p.eaad4998.
14.Ma,X.-Y.,et al.,Large family of two-dimensional ferroelectric metals discovered via machine learning.Science Bulletin,2021.66(3):p.233-242.
15.Lu,S.,et al.,Accelerated discovery of stable lead-free hybrid organic-inorganic perovskites via machine learning.Nature Communications,2018.9(1):p.3405.
16.Perera,D.,et al.,A platform for automated nanomole-scale reaction screening and micromole-scale synthesis in flow.Science,2018.359(6374):p.429-434.
17.Mehr,S.H.M.,et al.,A universal system for digitization and automatic execution of the chemical synthesis literature.Science,2020.370(6512):p.101-108.
18.Li,J.,et al.,Synthesis of many different types of organic small molecules using one automated process.Science,2015.347(6227):p.1221-1226.
19.Granda,J.M.,et al.,Controlling an organic synthesis robot with machine learning to search for new reactivity.Nature,2018.559(7714):p.377-381.
20.Burger,B.,et al.,A mobile robotic chemist.Nature,2020.583(7815):p.237-241.
21.何凯,陈书,赵文亮,方海涛.应用于合成生物学的自动化铸造平台系统集成技术:中国, 111482212A,2020-08-04.审中-公开.
22.Ager,D.,et al.,Method and apparatus for optimization of high-throughput screening and enhancement of biocatalyst performance.2002,Google Patents.
发明内容
本发明的目的是提供一种纳米晶动力学形貌控制机制的数字化可控合成的机器人辅助设备,实现了数字化和可编程化机器人辅助纳米晶体的制备,同时通过该机器人辅助设备对纳米晶体的生长动力学进行研究,构建了纳米晶体生长动力学模型,实现了形貌可控合成的纳米晶体的数字化自动化可控合成。
本发明一个方面提供了一种用于纳米晶体制备的数字化和可编程化机器人辅助设备,所述设备包括纳米晶体制备反应执行模块、纳米晶体生长溶液信息采集模块以及数据存储和计算模块;所述制备反应执行模块包括用于执行制备反应的部件;所述纳米晶体生长液信息采集模块包括获取纳米晶体光学吸收形貌特征或颜色特征的部件;所述数据存储和计算模块包括用于实验数据、实验结果、采集数据的存储部件以及用于控制纳米晶体制备反应执行模块和纳米晶体反应液信息采集模块的计算模块;
所述计算模块中具有文献方法数字化模块、实验方案编辑模块、实验方案执行控制模块、实验数据分析模块以及实验结果展示模块;
所述文献方法数字化模块通过提取文献方法中的反应原料种类、原料用量、反应条件实现将文献中的制备方案数字化;数字化后的数据存储在存储部件中,并能够被计算模块调用;
所述实验方案编辑模块能够对待实验的方案进行编辑、能够调用文献方法数字化模块数字化的制备方案或调用实验数据分析模块建模后计算获得的实验方案;其中对待实验的方案进行编辑为编辑实验原料种类、实验原料的浓度、实验原料的体积、不同原料的添加顺序、实验反应温度、反应时间、实验结果采集的种类;
所述实验方案执行控制模块能够控制纳米晶体制备反应执行模块、纳米晶体反应液信息采集模块实现实验方案编辑模块设定的待实验的方案;
所述实验数据分析模块能够调用纳米晶体生长溶液信息采集模块获得的光学吸收数据以及对应的实验方案信息,通过机器学习的方法对调用的信息进行学习和分析;
所述实验结果展示模块能够通过可视化地展示实验数据分析模块获得的机器学习建立的模型。
进一步的,所述制备反应执行模块包括自动化执行取液、分液、混合、振动、升温的部件,执行取液、分液的部件选自移液机器手臂、行走机器人、六轴机器手、8通道移液枪、 96孔板、加热炉,升温的部件选自控温加热马弗炉,振动的部件选自96孔板混匀振动器。
进一步的,所述纳米晶体反应液信息采集模块包括颜色超敏感高清照相系统、紫外-可见光-近红外吸收光谱仪、透射电镜、X射线衍射仪中的至少一种。
本发明另一个方面提供了一种自动化制备纳米晶的制备方法,所述制备方法包括在采用本发明所述的用于纳米晶体制备的数字化和可编程化机器人辅助设备进行制备。
本发明再一个方面提供一种纳米晶体生长动力学的建模方法,其包括以下步骤:
S1)收集制备纳米晶体的实验条件和结果数据,包括不同时间点的纳米晶体形貌特征数据;
S2)通过机器学习的方法建立反应实验与纳米晶形貌特征的模型。
进一步地,S1)中所述的纳米晶体形貌特征数据选自紫外-可见光-近红外吸收光谱数据、OD LSPR、LSPR值、反应液RGB颜色数据中的任意一种。
进一步地,S1)中数据采集的时间点从反应开始至反应的OD LSPR不再升高。
进一步地,S2)中机器学习的方法选自SISSO算法。
进一步地,所述的纳米晶体选自纳米金晶体。
进一步地,建模方法通过本发明上述设备试验获得。
本发明再一个方面提供了一种纳米晶体的可控制备方法,所述方法包括采用上述建模方法建立的模型预测获得的实验条件,并通过上述设备制备。
有益效果
本发明采用机器人辅助,基于纳米晶动力学形貌控制机制,进行纳米晶的数字化可控合成,实现了指定形貌纳米晶的全自动合成。
本发明开拓性地设计了纳米晶全自动合成设备,基于大数据和实验数据的结合,通过机器学习获得制备指定形貌纳米晶的动力学控制方法,并通过设计的自动制备组件进行了自动化的合成。
附图说明
图1为平台整体布局3D示意图,包含了行走机器人、自动吸(分)液模块、颜色超敏感高清照相系统、智能物料存储库、电动气门控制拉开式马弗炉。
图2为从文献中提取实验方案示意图和代码。
图3基于图2的实验程序设计和可执行程序。
图4为高通量实验结果采集、自动化提取和分析的示意图。
图5为高通量实验机器学习拟合模型和可视化的示意图。
图6为三种实验条件下金纳米棒溶液在0-360分钟的光学吸收光谱。
图7(A)为三种实验条件下吸收光谱上LSPR最大值对应的吸光度OD LSPR随时间t变化的曲线。(B)为三种实验条件下反应溶液颜色(RGB)数值随时间t变化的曲线。(C)根据吸收光谱OD LSPR和OD Lmax计算的ln(OD[Au +]))与时间t呈线性关系,显示纳米金生长符合一级反应动力学规律。(D)根据颜色变化R(测量时刻的R值)和R min(反应完成时的R值)计算的ln(R-R min)与时间t呈线性关系,显示用颜色变化描述纳米金动力学生长符合一级反应动力学规律。在图B-D中,方块、圆形和三角形图标分别表示不同的反应条件,即方块组C(HCl)为0.2M,三角组0.27M和圆形组0.4M的实验。
具体实施方式
为了使本发明的上述目的、特征和优点能够更加明显易懂,下面对本发明的具体实施方式做详细的说明,但不能理解为对本发明的可实施范围的限定。
如图1所示,提供了用于纳米晶体制备的数字化和可编程化机器人辅助设备,所述设备包括纳米晶体制备反应执行模块、纳米晶体反应液信息采集模块以及数据存储和计算模块;所述制备反应执行模块包括用于执行制备反应的部件;所述纳米晶体反应液信息采集模块包括获取纳米晶体形貌或颜色特征的部件;所述数据存储和计算模块包括用于实验数据、实验结果、采集数据的存储部件以及用于控制纳米晶体制备反应执行模块和纳米晶体反应液信息采集模块的计算模块;
所述计算模块中具有文献方法数字化模块、实验方案编辑模块、实验方案执行控制模块、实验数据分析模块以及实验结果展示模块;
所述文献方法数字化模块通过提取文献方法中的反应原料种类、原料用量、反应条件实现将文献中的制备方案数字化;数字化后的数据存储在存储部件中,并能够被计算模块调用;文献中提取实验方案示意图和代码见图2
所述制备反应执行模块包括自动化执行取液、分液、混合、振动、升温的部件,执行取液、分液的部件选自移液机器手臂、行走机器人、六轴机器手、8通道移液枪、96孔板、加热炉,升温的部件选自加热炉,振动的部件选自振动器。
所述纳米晶体反应液信息采集模块包括颜色超敏感高清照相系统、紫外-可见光-近红外吸收光谱仪、透射电镜、X射线衍射仪中的至少一种。
所述实验方案编辑模块能够对待实验的方案进行编辑、能够调用文献方法数字化模块数 字化的制备方案或调用实验数据分析模块建模后计算获得的实验方案;其中对待实验的方案进行编辑为编辑实验原料种类、实验原料的浓度、实验原料的体积、不同原料的添加顺序、实验反应温度、反应时间、实验结果采集的种类;
所述实验方案执行控制模块能够控制纳米晶体制备反应执行模块、纳米晶体反应液信息采集模块实现实验方案编辑模块设定的待实验的方案;在一个具体的实施方案中,基于图2的实验程序设计和可执行程序见图3。
所述实验数据分析模块能够调用纳米晶体反应液信息采集模块获得的数据以及对应的实验方案信息,通过机器学习的方法对调用的信息进行学习和分析;
所述实验结果展示模块能够通过可视化地展示实验数据分析模块获得的机器学习建立的模型。在一个具体的实施方案中,高通量实验机器学习拟合模型示意图见图5。
进一步采用本发明的数字化和可编程化机器人辅助设备进行纳米金生长动力学研究
对纳米金生长过程产物的光学吸收浓度变化进行了实时检测,产物浓度变化的吸收光谱结果如图6所示。
首先对待实施的方案进行编程,设定实验条件,实验条件共有3种,分别为实验条件1:10μL的AgNO 3、0.1M CTAB和0.2M HCl、0.27M和0.4M的实验。实验条件2:采用10μL的AgNO 3、0.1M CTAB和0.27M HCl的条件进行反应。实验条件3:采用12μL的AgNO 3、0.12M CTAB和0.42M HCl的条件进行反应,其中AgNO 3的浓度为0.01M,CTAB溶液的体积为1mL,HCl的体积为20μL,添加的其余的实验药品为50μL浓度为0.1M的HAuCl4溶液,8μL浓度为0.1M的抗坏血酸溶液和2.4μL预先制备的金种子溶液。并设定在0min-360min内分别进行紫外-可见光-近红外光学吸收谱测定,实验结果如图6所示。此外,通过颜色超敏高高清相机对金纳米晶体生长过程中的颜色变换进行连续测试。
进一步对实验结果进行分析,以上述结果中OD LSPR对时间作图,结果见图7A所示,对颜色超敏感相机连续监控的反应溶液颜色变化对时间作图,如图7B所示。。进一步研究不同条件的纳米金的反应生长动力学,其中图7C,7D显示了以金溶液光学吸收浓度的对数ln(OD[Au +]))或者溶液颜色变化的对数,跟生长时间呈现出线性关系,即说明该反应为一阶反应。并且,在上述三个条件下,盐酸浓度越低,体系的斜率绝对值越大,说明纳米金晶体生长越快。
其中,t代表反应时间,OD LSPR代表测试时刻LSPR峰值对应的吸光度值,OD Lmax代表反应完成最大的吸光度值,随着纳米金的生长,OD LSPR逐渐升高。R代表测试时候的颜色R值,Rmin代表反应完成时最小的R值,随着纳米金的生长,R值逐渐降低。。在图B-D中, 方块、圆形和三角形图标分别表示不同的反应条件,即方块组C(HCl)为0.2M,三角组0.27M和圆形组0.42M的实验。
随后采用机器学习SISSO算法,建立了可编程动力学模型。生长动力学模型表达式如表1所示。
表1.机器学习计算棒状纳米金晶体的生长动力学模型。
Figure PCTCN2021135297-appb-000001
可以看出通过本发明的实验装置,能够实现纳米晶体制备的自动化,进而可以更容易的获得纳米晶生长动力学的大量数据,平行性和准确性更高,本发明验证的不同实验条件下都能实现可靠的结果,可以看到三组模型的方差均大于0.99。
通过建立的生长动力学模型结果,可以调入实验方案编辑模块,对于后续的实验方案进行实验条件的预测。
通过对上述形貌、反应条件、时间等的动力学数据库进行数据预处理、离群数据清理等措施,优化数据集结构,筛选特征描述,调用机器学习SISSO算法,提升了计算效率和准确性,实现机器学习动力学参数与形貌的内在规律;研究融合数学模型-高通量计算-动力学参数-AI新生产数据等分布式数据库,实现纳米晶数学模型、数据库和AI算法等全流程的可编程理性设计。

Claims (10)

  1. 一种用于纳米晶体制备的数字化和可编程化机器人辅助设备,其特征在于,所述设备包括纳米晶体制备反应执行模块、纳米晶体反应液信息采集模块以及数据存储和计算模块;所述制备反应执行模块包括用于执行制备反应的部件;所述纳米晶体反应液信息采集模块包括获取纳米晶体形貌或颜色特征的部件;所述数据存储和计算模块包括用于实验数据、实验结果、采集数据的存储部件以及用于控制纳米晶体制备反应执行模块和纳米晶体反应液信息采集模块的计算模块;
    所述计算模块中具有文献方法数字化模块、实验方案编辑模块、实验方案执行控制模块、实验数据分析模块以及实验结果展示模块;
    所述文献方法数字化模块通过提取文献方法中的反应原料种类、原料用量、反应条件实现将文献中的制备方案数字化;数字化后的数据存储在存储部件中,并能够被计算模块调用;
    所述实验方案编辑模块能够对待实验的方案进行编辑、能够调用文献方法数字化模块数字化的制备方案或调用实验数据分析模块建模后计算获得的实验方案;其中对待实验的方案进行编辑为编辑实验原料种类、实验原料的浓度、实验原料的体积、不同原料的添加顺序、实验反应温度、反应时间、实验结果采集的种类;
    所述实验方案执行控制模块能够控制纳米晶体制备反应执行模块、纳米晶体反应液信息采集模块实现实验方案编辑模块设定的待实验的方案;
    所述实验数据分析模块能够调用纳米晶体反应液信息采集模块获得的数据以及对应的实验方案信息,通过机器学习的方法对调用的信息进行学习和分析;
    所述实验结果展示模块能够通过可视化地展示实验数据分析模块获得的机器学习建立的模型。
  2. 根据权利要求1所述的数字化和可编程化机器人辅助设备,其特征在于,制备反应执行模块包括自动化执行取液、分液、混合、振动、升温的部件,执行取液、分液的部件选自移液机器手臂、行走机器人、六轴机器手、8通道移液枪、96孔板、加热炉,升温的部件选自加热炉,振动的部件选自振动器。
  3. 根据权利要求1所述的数字化和可编程化机器人辅助设备,其特征在于,所述纳米晶体反应液信息采集模块包括颜色超敏感高清照相系统、紫外-可见光-近红外吸收光谱仪、透射电镜、X射线衍射仪中的至少一种。
  4. 一种自动化制备纳米晶的制备方法,其特征在于,所述制备方法包括在采用权利要求 1-3任一项所述的用于纳米晶体制备的数字化和可编程化机器人辅助设备进行制备。
  5. 一种纳米晶体生长动力学的建模方法,其特征在于,其包括以下步骤:
    S1)收集制备纳米晶体的数据,包括不同时间点的纳米晶体形貌特征数据;
    S2)通过机器学习的方法建立反应实验与纳米晶形貌特征的模型。
  6. 根据权利要求5所述的建模方法,其特征在于,S1)中所述的纳米晶体形貌特征数据选自紫外-可见光-近红外吸收光谱数据、OD LSPR、LSPR值、反应液RGB颜色数据中的任意一种。
  7. 根据权利要求5所述的建模方法,其特征在于,S2)中机器学习的方法选自SISSO算法。
  8. 根据权利要求5所述的建模方法,其特征在于,S1)中数据采集的时间点从反应开始至反应结束时OD LSPR不再升高。
  9. 根据权利要求5所述的建模方法,其特征在于,所述的纳米晶体选自纳米金晶体。
  10. 一种纳米晶体的可控制备方法,其特征他在于,所述可控制备方法包括采用权利要求5-9任一项所述的建模方法建立的模型预测获得的实验条件,并通过权利要求1-3任一项所述的设备制备。
PCT/CN2021/135297 2021-12-03 2021-12-03 可编程理性设计纳米晶体材料的方法 Ceased WO2023097643A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/135297 WO2023097643A1 (zh) 2021-12-03 2021-12-03 可编程理性设计纳米晶体材料的方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/135297 WO2023097643A1 (zh) 2021-12-03 2021-12-03 可编程理性设计纳米晶体材料的方法

Publications (1)

Publication Number Publication Date
WO2023097643A1 true WO2023097643A1 (zh) 2023-06-08

Family

ID=86611337

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/135297 Ceased WO2023097643A1 (zh) 2021-12-03 2021-12-03 可编程理性设计纳米晶体材料的方法

Country Status (1)

Country Link
WO (1) WO2023097643A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110793982A (zh) * 2019-11-21 2020-02-14 山东建筑大学 一种纳米晶化动力学过程的高能x射线表征方法
WO2020161440A1 (fr) * 2019-02-05 2020-08-13 Centre National De La Recherche Scientifique Nouveau procédé de synthèse de nanocristaux hydrosolubles et système automatisé
CN113699592A (zh) * 2020-05-22 2021-11-26 天津理工大学 一种利用机器学习辅助生长大尺寸lbo晶体的方法及其应用

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020161440A1 (fr) * 2019-02-05 2020-08-13 Centre National De La Recherche Scientifique Nouveau procédé de synthèse de nanocristaux hydrosolubles et système automatisé
CN110793982A (zh) * 2019-11-21 2020-02-14 山东建筑大学 一种纳米晶化动力学过程的高能x射线表征方法
CN113699592A (zh) * 2020-05-22 2021-11-26 天津理工大学 一种利用机器学习辅助生长大尺寸lbo晶体的方法及其应用

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
HUANG CHUAN, SONG XIAOYAN, WEI JUN, HAN QINGCHAO: "Computer Simulation and Experimental Study of Thermal Stability of Nanocrystalline Metals", CHINESE JOURNAL OF STEREOLOGY AND IMAGE ANALYSIS, vol. 13, no. 3, 30 September 2008 (2008-09-30), pages 193 - 199, XP093069201 *

Similar Documents

Publication Publication Date Title
CN116724340B (zh) 电子显微镜数据的人工智能(ai)辅助分析
Elsherbiny et al. Rapid grapevine health diagnosis based on digital imaging and deep learning
CN105588827B (zh) 活体单细胞拉曼分析平台数字控制系统和方法
Li et al. Synchrotron radiation data-driven artificial intelligence approaches in materials discovery
CN118456126B (zh) 基于机器学习的圆柱桶内表面熔覆层抛光路径优化方法
EP4361944A1 (en) Monitoring of biofabrication processes
WO2025107933A1 (zh) 细胞培养液代谢物浓度确定方法、装置、设备和介质
Zhang et al. Afm imaging defect detection and classification using deep learning
Aziz et al. Digital Twins in Smart Manufacturing
CN114329906A (zh) 纳米晶体材料数学模型、数据库和ai算法全流程的可编程理性设计方法
WO2023097643A1 (zh) 可编程理性设计纳米晶体材料的方法
CN116445391B (zh) 一种基于微纳机器人操作的完整单细胞显微切割捕获方法
Yazid et al. Deep learning-empowered robot vision for efficient robotic grasp detection and defect elimination in industry 4.0
WO2023097664A1 (zh) 机器人辅助纳米晶自动表征加速材料的智能化逆向设计
WO2021180982A1 (de) Verfahren und vorrichtung zur untersuchung von mikroskopproben unter verwendung optischer marker
Pandey et al. Reinforcement learning based carbon nanotube growth automation
CN114357857B (zh) 机器人辅助数字化可控合成纳米晶体形貌的方法
Guo et al. From algorithms to applications: a comprehensive review of machine learning in computational materials science
CN114357712B (zh) 机器人辅助纳米晶自动表征加速材料的智能化逆向设计
CN108795741A (zh) 藻类观测与藻种智能识别系统
Liu et al. Dynamic control framework for automated particle transport based on optically induced dielectrophoresis
Alexandar et al. AI in Nanomaterials: Enhancing the Efficiency of the Applications through Advanced Intelligence
CN118379733B (zh) 一种针对r显带染色体的切割方法
CN121438307A (zh) 一种基于计算机视觉的面向钴/镍基高温合金激光熔覆过程熔池自动识别以及预测方法与系统
CN120833331B (zh) 基于生物标志物的智能化海洋环境污染预警方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21966070

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21966070

Country of ref document: EP

Kind code of ref document: A1