본문

서브메뉴

Scaling in Lithium Niobite: Synaptic Devices for Neuromorphic Computing
Scaling in Lithium Niobite: Synaptic Devices for Neuromorphic Computing
Scaling in Lithium Niobite: Synaptic Devices for Neuromorphic Computing

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202105524
ISBN  
9798263350581
DDC  
300
저자명  
Mccrone, Timothy.
서명/저자  
Scaling in Lithium Niobite: Synaptic Devices for Neuromorphic Computing
발행사항  
[Sl] : Georgia Institute of Technology, 2024
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2024
형태사항  
131 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
주기사항  
Advisor: Doolittle, W. Alan.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2024.
초록/해제  
요약Due to the end of Dennard scaling, problems with the Von Neumann bottleneck, and the unique challenges posed by the volume of data generated on the internet neuromorphic computing has emerged an energy efficient computing method. While standard neuromorphic solutions use matrix math to churn through tables of numbers, in the last several years dedicated hardware has emerged that has optimized the overhead these computations incur in digital systems. Research into memristors has emerged as an alternative method to create neuromorphic systems capable of implementations of synaptic inference more closely matching the computations made in the brain than faux digital or analog systems. Instead of a series of amplifiers, transistors, and resistors necessary for digital and analog systems memristive systems can implement a biomimetic computational paradigm within several two terminal devices. In the variety of memristive systems LiNbO2 stands out due to its already proven large change in resistivity, low power programmability, dynamic and static control of resistivity ranges and temporal response, ability to create volatile, non-volatile and even mixed volatility responses and the ability to implement inductive analogues via ionic momentum. In hopes of unlocking the lowest power neuromorphic implementation to date, LiNbO2 has been investigated as a promising class of memristor devices. Prior investigations were limited to optical lithography scale devices yet compared very well to nanoscale devices from other technologies. This thesis explored methods to a) produce suitable LiNbO2 films allowing scaling to nanometer length scales and b) LiNbO2 devices were scaled past optical lithography lengths resulting in state-ofthe-art breakthroughs in power efficiency.CHAPTER 1 focuses on why neuromorphic computing has become a topic of interest and the need for more biomimetic architectures. It shows how a variety of VLSI silicon systems have successfully implemented a more biomimetic style of neuromorphic computing with results that indicate the approach can improve the power use and accuracy of neural nets.CHAPTER 2 introduces the different types of memristors in the field and discusses how these systems change their resistivity. CHAPTER 3 then discusses in depth LiNbO2, the focus of this dissertation. It discusses how this material system can use delithiation as a resistance modulation mechanism that can allow it to change its resistivity by over 4 orders of magnitude, the methods of fabrication and the recent history of the exploration of the memristive properties of LiNbO2. CHAPTER 4 begins the primary body of work of this thesis by discussing Electron Beam Lithography (EBL) and the unique chemistry and materials challenges encountered in scaling LiNbO2 to the nanoscale. CHAPTER 5 demonstrates the devices created using ebeam lithography and uncovers how at the nanoscale these systems may be lithium extraction limited. It also finds that the initial low power programming that sparked this investigation continues to scale down to 150 nm. CHAPTER 6 discusses the experiments performed with LPEE and found that despite a promising lattice match to LiNbO2, SiC may not be a suitable substrate due to chemical reactions with the LPEE environment. However, the discovery of a novel interfacial NbC thin film could warrant further exploration. CHAPTER 7 then concludes this dissertation and discusses how future work could improve LiNbO2 devices. One of the major objectives delivered in this project was the scaling of LiNbO2 devices to the nanoscale to better compare its properties to other state of the art material systems. It is found that scaled LiNbO2 memristors outperform all other published memristors in terms of the crucial sensitivity figure of merit R/V.
일반주제명  
Monte Carlo simulation
일반주제명  
Neurons
일반주제명  
Technological change
일반주제명  
Brain research
일반주제명  
Carbon
일반주제명  
Titanium
일반주제명  
Molybdenum
일반주제명  
Synapses
일반주제명  
Neural networks
일반주제명  
Molecular beam epitaxy
일반주제명  
Potassium
일반주제명  
Nickel
일반주제명  
Large language models
일반주제명  
Thin films
일반주제명  
Lithium
일반주제명  
Scanning electron microscopy
일반주제명  
Analytical chemistry
일반주제명  
Condensed matter physics
일반주제명  
Materials science
일반주제명  
Neurosciences
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2024        us                              c    eng  d
■001000017360430
■00520260202105524
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798263350581
■035    ▼a(MiAaPQ)AAI32309730
■035    ▼a(MiAaPQ)GeorgiaTech77763
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a300
■1001  ▼aMccrone,  Timothy.
■24510▼aScaling  in  Lithium  Niobite:  Synaptic  Devices  for  Neuromorphic  Computing
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2024
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2024
■300    ▼a131  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  B.
■500    ▼aAdvisor:  Doolittle,  W.  Alan.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2024.
■520    ▼aDue  to  the  end  of  Dennard  scaling,  problems  with  the  Von  Neumann  bottleneck,  and  the  unique  challenges  posed  by  the  volume  of  data  generated  on  the  internet  neuromorphic  computing  has  emerged  an  energy  efficient  computing  method.  While  standard  neuromorphic  solutions  use  matrix  math  to  churn  through  tables  of  numbers,  in  the  last  several  years  dedicated  hardware  has  emerged  that  has  optimized  the  overhead  these  computations  incur  in  digital  systems.  Research  into  memristors  has  emerged  as  an  alternative  method  to  create  neuromorphic  systems  capable  of  implementations  of  synaptic  inference  more  closely  matching  the  computations  made  in  the  brain  than  faux  digital  or  analog  systems.  Instead  of  a  series  of  amplifiers,  transistors,  and  resistors  necessary  for  digital  and  analog  systems  memristive  systems  can  implement  a  biomimetic  computational  paradigm  within  several  two  terminal  devices.  In  the  variety  of  memristive  systems  LiNbO2  stands  out  due  to  its  already  proven  large  change  in  resistivity,  low  power  programmability,  dynamic  and  static  control  of  resistivity  ranges  and  temporal  response,  ability  to  create  volatile,  non-volatile  and  even  mixed  volatility  responses  and  the  ability  to  implement  inductive  analogues  via  ionic  momentum.  In  hopes  of  unlocking  the  lowest  power  neuromorphic  implementation  to  date,  LiNbO2  has  been  investigated  as  a  promising  class  of  memristor  devices.  Prior  investigations  were  limited  to  optical  lithography  scale  devices  yet  compared  very  well  to  nanoscale  devices  from  other  technologies.  This  thesis  explored  methods  to  a)  produce  suitable  LiNbO2  films  allowing  scaling  to  nanometer  length  scales  and  b)  LiNbO2  devices  were  scaled  past  optical  lithography  lengths  resulting  in  state-ofthe-art  breakthroughs  in  power  efficiency.CHAPTER  1  focuses  on  why  neuromorphic  computing  has  become  a  topic  of  interest  and  the  need  for  more  biomimetic  architectures.  It  shows  how  a  variety  of  VLSI  silicon  systems  have  successfully  implemented  a  more  biomimetic  style  of  neuromorphic  computing  with  results  that  indicate  the  approach  can  improve  the  power  use  and  accuracy  of  neural  nets.CHAPTER  2  introduces  the  different  types  of  memristors  in  the  field  and  discusses  how  these  systems  change  their  resistivity.  CHAPTER  3  then  discusses  in  depth  LiNbO2,  the  focus  of  this  dissertation.  It  discusses  how  this  material  system  can  use  delithiation  as  a  resistance  modulation  mechanism  that  can  allow  it  to  change  its  resistivity  by  over  4  orders  of  magnitude,  the  methods  of  fabrication  and  the  recent  history  of  the  exploration  of  the  memristive  properties  of  LiNbO2.  CHAPTER  4  begins  the  primary  body  of  work  of  this  thesis  by  discussing  Electron  Beam  Lithography  (EBL)  and  the  unique  chemistry  and  materials  challenges  encountered  in  scaling  LiNbO2  to  the  nanoscale.  CHAPTER  5  demonstrates  the  devices  created  using  ebeam  lithography  and  uncovers  how  at  the  nanoscale  these  systems  may  be  lithium  extraction  limited.  It  also  finds  that  the  initial  low  power  programming  that  sparked  this  investigation  continues  to  scale  down  to  150  nm.  CHAPTER  6  discusses  the  experiments  performed  with  LPEE  and  found  that  despite  a  promising  lattice  match  to  LiNbO2,  SiC  may  not  be  a  suitable  substrate  due  to  chemical  reactions  with  the  LPEE  environment.  However,  the  discovery  of  a  novel  interfacial  NbC  thin  film  could  warrant  further  exploration.  CHAPTER  7  then  concludes  this  dissertation  and  discusses  how  future  work  could  improve  LiNbO2  devices.  One  of  the  major  objectives  delivered  in  this  project  was  the  scaling  of  LiNbO2  devices  to  the  nanoscale  to  better  compare  its  properties  to  other  state  of  the  art  material  systems.  It  is  found  that  scaled  LiNbO2  memristors  outperform  all  other  published  memristors  in  terms  of  the  crucial  sensitivity  figure  of  merit  R/V.
■590    ▼aSchool  code:  0078.
■650  4▼aMonte  Carlo  simulation
■650  4▼aNeurons
■650  4▼aTechnological  change
■650  4▼aBrain  research
■650  4▼aCarbon
■650  4▼aTitanium
■650  4▼aMolybdenum
■650  4▼aSynapses
■650  4▼aNeural  networks
■650  4▼aMolecular  beam  epitaxy
■650  4▼aPotassium
■650  4▼aNickel
■650  4▼aLarge  language  models
■650  4▼aThin  films
■650  4▼aLithium
■650  4▼aScanning  electron  microscopy
■650  4▼aAnalytical  chemistry
■650  4▼aCondensed  matter  physics
■650  4▼aMaterials  science
■650  4▼aNeurosciences
■690    ▼a0800
■690    ▼a0486
■690    ▼a0611
■690    ▼a0794
■690    ▼a0317
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05B.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2024
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360430▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF17218 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.