東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁 到查詢結果 [ null ]

切換: 標籤 | MARC模式 | ISBD

Advancing Sustainable Energy Materia...

Anwar, Husna,

FindBook

Google Book

Amazon

博客來

Advancing Sustainable Energy Materials: Accelerated Discovery Techniques /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Advancing Sustainable Energy Materials: Accelerated Discovery Techniques // Husna Anwar.
作者:	Anwar, Husna,
面頁冊數:	1 electronic resource (154 pages)
附註:	Source: Dissertations Abstracts International, Volume: 86-05, Section: B.
Contained By:	Dissertations Abstracts International86-05B.
標題:	Physical chemistry. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31488231
ISBN:	9798342751148

Advancing Sustainable Energy Materials: Accelerated Discovery Techniques /
Anwar, Husna,

Advancing Sustainable Energy Materials: Accelerated Discovery Techniques /Husna Anwar. - 1 electronic resource (154 pages)

Source: Dissertations Abstracts International, Volume: 86-05, Section: B.

Most electronics since the advent of integrated circuits in the 1950s are silicon-based. We need readily-manufactured, and thus affordable, materials processing options to meet the burgeoning needs of the fourth energy transition. Recent advances in computing techniques allow us to explore new chemical spaces, but translating from theoretically-predicted structures to physically-realized molecules is laborious. High-throughput experimental techniques and machine learning methods can track trends and test more parameters than can do humans alone. In this thesis, I use these techniques to derive insight into new materials, chemical pathways, and optimization strategies for photovoltaics and LEDs.First, I screen a database of sputtered thin-film characterization data to predict optical absorption, conductivity, bandgap, and charge carrier mobility. I use three ML algorithms to accelerate this process 5x-10x over a full-time experimentalist (FTE) and achieve a 3x lower MAE and a 1000x improvement on compute time over first principles calculations. The models are validated by doping an important hole transport material, NiOx, and improving conductivity 2x over pristine NiOx.Next, I create a high-throughput workflow to find new spacer molecules for reduced-dimensional perovskites to produce blue emission. I build an automated characterization station and combine it with a drop-setting robot that allows me to try out 54 spacer ligands via vapor-assisted antisolvent crystallization 120x faster than can an individual experimentalist. I cluster the data and use Tanimoto similarity and predict 18 new molecules and synthesize one, (3,4-difluorobenzyl lead bromide) that produces a 10% PLQY. The remaining 17 candidates, 100,000 crystal images, and 16,128 photoluminescence spectra are compiled in the first ever dataset of perovskite emission spectra.Lastly, I use in-situ Grazing Incidence Wide Angle Scattering (GIWAXS) to study the formation of thin-film Cs1−xRbxPbI1.75Br1.25 (x= 0.05-0.20). I observe that a threshold of x=0.15 appears to create a stable cubic perovskite that degrades into the ẟ-phase after prolonged annealing. The resulting dataset comprises over 10,000 time-dependent GIWAXS images and integrated diffraction spectra and can be used by the perovskite community to study the crystallization dynamics of this perovskite.Broadly, this work demonstrates how domain knowledge, AI-assisted analytics, and autonomous systems can reduce decision latency, improve mechanistic understanding, and accelerate the discovery of materials for optoelectronics.

English

ISBN: 9798342751148Subjects--Topical Terms:

1981412
Physical chemistry.
Subjects--Index Terms:

Automation

Advancing Sustainable Energy Materials: Accelerated Discovery Techniques /
LDR:03999nmm a22004573i 4500 001 2430650
005 20250527055723.5
006 m o d
007 cr|nu||||||||
008 261011s2024 miu||||||m |||||||eng d
020 $a 9798342751148
035 $a (MiAaPQD)AAI31488231
035 $a AAI31488231
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Anwar, Husna, $e author. $0 (orcid)0000-0002-4027-284X $3 3819076
245 1 0 $a Advancing Sustainable Energy Materials: Accelerated Discovery Techniques / $c Husna Anwar.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (154 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-05, Section: B.
500 $a Advisors: Sargent, Edward H.
502 $b Ph.D. $c University of Toronto (Canada) $d 2024.
520 $a Most electronics since the advent of integrated circuits in the 1950s are silicon-based. We need readily-manufactured, and thus affordable, materials processing options to meet the burgeoning needs of the fourth energy transition. Recent advances in computing techniques allow us to explore new chemical spaces, but translating from theoretically-predicted structures to physically-realized molecules is laborious. High-throughput experimental techniques and machine learning methods can track trends and test more parameters than can do humans alone. In this thesis, I use these techniques to derive insight into new materials, chemical pathways, and optimization strategies for photovoltaics and LEDs.First, I screen a database of sputtered thin-film characterization data to predict optical absorption, conductivity, bandgap, and charge carrier mobility. I use three ML algorithms to accelerate this process 5x-10x over a full-time experimentalist (FTE) and achieve a 3x lower MAE and a 1000x improvement on compute time over first principles calculations. The models are validated by doping an important hole transport material, NiOx, and improving conductivity 2x over pristine NiOx.Next, I create a high-throughput workflow to find new spacer molecules for reduced-dimensional perovskites to produce blue emission. I build an automated characterization station and combine it with a drop-setting robot that allows me to try out 54 spacer ligands via vapor-assisted antisolvent crystallization 120x faster than can an individual experimentalist. I cluster the data and use Tanimoto similarity and predict 18 new molecules and synthesize one, (3,4-difluorobenzyl lead bromide) that produces a 10% PLQY. The remaining 17 candidates, 100,000 crystal images, and 16,128 photoluminescence spectra are compiled in the first ever dataset of perovskite emission spectra.Lastly, I use in-situ Grazing Incidence Wide Angle Scattering (GIWAXS) to study the formation of thin-film Cs1−xRbxPbI1.75Br1.25 (x= 0.05-0.20). I observe that a threshold of x=0.15 appears to create a stable cubic perovskite that degrades into the ẟ-phase after prolonged annealing. The resulting dataset comprises over 10,000 time-dependent GIWAXS images and integrated diffraction spectra and can be used by the perovskite community to study the crystallization dynamics of this perovskite.Broadly, this work demonstrates how domain knowledge, AI-assisted analytics, and autonomous systems can reduce decision latency, improve mechanistic understanding, and accelerate the discovery of materials for optoelectronics.
546 $a English
590 $a School code: 0779
650 4 $a Physical chemistry. $3 1981412
650 4 $a Energy. $3 876794
650 4 $a Materials science. $3 543314
650 4 $a Engineering. $3 586835
653 $a Automation
653 $a Discovery
653 $a LEDs
653 $a Machine learning
653 $a Photovoltaics
653 $a Renewable energy
690 $a 0794
690 $a 0494
690 $a 0791
690 $a 0537
710 2 $a University of Toronto (Canada). $b Electrical and Computer Engineering. $e degree granting institution. $3 3816307
720 1 $a Sargent, Edward H. $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-05B.
790 $a 0779
791 $a Ph.D.
792 $a 2024
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31488231