서브메뉴
검색
Design and Optimization of Integrated Plasmonic Devices for High-Performance Computing
Design and Optimization of Integrated Plasmonic Devices for High-Performance Computing
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202105546
- ISBN
- 9798263393861
- DDC
- 690
- 서명/저자
- Design and Optimization of Integrated Plasmonic Devices for High-Performance Computing
- 발행사항
- [Sl] : Georgia Institute of Technology, 2023
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2023
- 형태사항
- 136 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-05, Section: A.
- 주기사항
- Advisor: Naeemi, Azad.
- 학위논문주기
- Thesis (Ph.D.)--Georgia Institute of Technology, 2023.
- 초록/해제
- 요약Plasmonics has opened a wealth of prospects in high-performance computing offering high throughput and dense integration capability. To realize plasmon-based computing, it is crucial to consider the interaction among the connected components and trade-offs among different performance parameters during the design process. However, research in this area has mainly focused on the design or demonstration of individual device components in isolation, and there has not been any systematic and comprehensive research effort considering the major aspects of a plasmonic computing system in a holistic approach. The doctoral thesis aims to design an integrated plasmonic computing system for high-end streaming server applications with optimized system-level performance, considering the trade-offs among footprint, throughput, latency, energy consumption, and thermal management.Augmenting complementary metal-oxide semiconductor (CMOS) processors with plasmonic computing modules in high-end servers offers the advantage of compact processor cores with exceptionally high clock rates. The high clock speed helps to reduce the massive number of cores needed in many streaming applications and, thereby, reducing the long interconnection wires that dominate the energy consumption in today's CMOS-based servers. The computing system designed in this thesis includes photonic to plasmonic mode converters to generate surface plasmon polariton waves, plasmonic phase modulators to encode data, a plasmonic logic gate for wave-computation, a plasmonic photodetector, and coupler between the WGs to convert the logic gate output to electrical signal. The system design and optimization also considers high-speed CMOS circuits for detecting the logic gate's output and driving the plasmonic phase modulator.The main element of the computing system is a plasmonic multi-functional logic gate that can perform non-Boolean majority and threshold logic operations along with Boolean operations. The gate operates on the interference of the incoming surface plasmon polariton (SPP) signal, resulting in multiple output levels. A multiplier structure is designed using the logic gate as a primitive, showing its application in arithmetic operations.To detect and differentiate different output levels of the logic gate, a plasmonic AlGe-Cu metal-semiconductor-metal detector is designed. Through numerical analysis, the performance of the Ge-based plasmonic detector is optimized considering the trade-off between responsivity and operation bandwidth. The designed detector offers a low dark current of a few nA, responsivity of 0.32A/W, and a bandwidth of ∼ 200GHz at only 100mV .To excite SPP for the computing system, photonic to plasmonic mode converters are designed numerically. Five different coupling approaches including the directional and end coupling of wire and Si slot photonic WGs with plasmonic metal-insulator-metal WGs have been explored. As the plasmonic logic gate is a multiple-input device, multiple mode converters and therefore, multiple Si photonic WGs are needed, which can largely penalize the total footprint due to significant mismatch between the pitch of the conventional Si WG and plasmonic MIM WG. As a result, the converter design and optimization include considerations such as the WG pitch, the total footprint along with the coupling efficiency. The design also considers arrangement of the photonic and plasmonic layer ensuring that there is negligible power coupling back to the photonic mode after SPP excitation. Simulation results show that the optimized mode conversion efficiency can be as high as 78%.To encode the input data for the plasmonic logic gate, a plasmonic phase shifter with nonlinear electro-optic slot material is analytically modeled using the coupled wave equation formalism for χ (2) nonlinear processes in lossy systems. The analytical model calculates the modulated and unmodulated signals at the output of the modulator. To find the design space of the phase shifter, such as its length, input optical power, and driving voltage, an integrated system consisting of the phase shifters, the plasmonic logic gate, the plasmonic detector, and the supporting CMOS circuits is designed.With the help of the phase shifter's analytical model and electromagnetic simulation of the connected devices, the BER of the integrated plasmonic devices is calculated. As Joule heating is common in plasmonic devices, and the nonlinear material of the phase shifter becomes unstable at ∼ 100oC, thermal analysis is performed using a 1-D thermal model to find the average temperature of the devices. A maximal but realistic BER of 10−3 and a driving voltage of 1V is taken as the limit. The length and driving voltage of the phase shifter are varied, and the input optical power is searched that maintains the target BER, and at the same time, does not heat the system to the extent of making the devices inoperable.
- 일반주제명
- Cooling
- 일반주제명
- Signal to noise ratio
- 일반주제명
- Wire
- 일반주제명
- Bandwidths
- 일반주제명
- Boolean
- 일반주제명
- Electric fields
- 일반주제명
- Circuits
- 일반주제명
- Supercomputers
- 일반주제명
- Design
- 일반주제명
- CMOS
- 일반주제명
- Transistors
- 일반주제명
- Optics
- 일반주제명
- Energy consumption
- 일반주제명
- Computer science
- 일반주제명
- Electrical engineering
- 일반주제명
- Materials science
- 일반주제명
- Electromagnetics
- 기본자료저록
- Dissertations Abstracts International. 87-05A.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2023 us c eng d■001000017360553
■00520260202105546
■006m o d
■007cr#unu||||||||
■020 ▼a9798263393861
■035 ▼a(MiAaPQ)AAI32315574
■035 ▼a(MiAaPQ)GeorgiaTech73120
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a690
■1001 ▼aNoor, Samantha Lubaba.
■24510▼aDesign and Optimization of Integrated Plasmonic Devices for High-Performance Computing
■260 ▼a[Sl]▼bGeorgia Institute of Technology▼c2023
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2023
■300 ▼a136 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-05, Section: A.
■500 ▼aAdvisor: Naeemi, Azad.
■5021 ▼aThesis (Ph.D.)--Georgia Institute of Technology, 2023.
■520 ▼aPlasmonics has opened a wealth of prospects in high-performance computing offering high throughput and dense integration capability. To realize plasmon-based computing, it is crucial to consider the interaction among the connected components and trade-offs among different performance parameters during the design process. However, research in this area has mainly focused on the design or demonstration of individual device components in isolation, and there has not been any systematic and comprehensive research effort considering the major aspects of a plasmonic computing system in a holistic approach. The doctoral thesis aims to design an integrated plasmonic computing system for high-end streaming server applications with optimized system-level performance, considering the trade-offs among footprint, throughput, latency, energy consumption, and thermal management.Augmenting complementary metal-oxide semiconductor (CMOS) processors with plasmonic computing modules in high-end servers offers the advantage of compact processor cores with exceptionally high clock rates. The high clock speed helps to reduce the massive number of cores needed in many streaming applications and, thereby, reducing the long interconnection wires that dominate the energy consumption in today's CMOS-based servers. The computing system designed in this thesis includes photonic to plasmonic mode converters to generate surface plasmon polariton waves, plasmonic phase modulators to encode data, a plasmonic logic gate for wave-computation, a plasmonic photodetector, and coupler between the WGs to convert the logic gate output to electrical signal. The system design and optimization also considers high-speed CMOS circuits for detecting the logic gate's output and driving the plasmonic phase modulator.The main element of the computing system is a plasmonic multi-functional logic gate that can perform non-Boolean majority and threshold logic operations along with Boolean operations. The gate operates on the interference of the incoming surface plasmon polariton (SPP) signal, resulting in multiple output levels. A multiplier structure is designed using the logic gate as a primitive, showing its application in arithmetic operations.To detect and differentiate different output levels of the logic gate, a plasmonic AlGe-Cu metal-semiconductor-metal detector is designed. Through numerical analysis, the performance of the Ge-based plasmonic detector is optimized considering the trade-off between responsivity and operation bandwidth. The designed detector offers a low dark current of a few nA, responsivity of 0.32A/W, and a bandwidth of ∼ 200GHz at only 100mV .To excite SPP for the computing system, photonic to plasmonic mode converters are designed numerically. Five different coupling approaches including the directional and end coupling of wire and Si slot photonic WGs with plasmonic metal-insulator-metal WGs have been explored. As the plasmonic logic gate is a multiple-input device, multiple mode converters and therefore, multiple Si photonic WGs are needed, which can largely penalize the total footprint due to significant mismatch between the pitch of the conventional Si WG and plasmonic MIM WG. As a result, the converter design and optimization include considerations such as the WG pitch, the total footprint along with the coupling efficiency. The design also considers arrangement of the photonic and plasmonic layer ensuring that there is negligible power coupling back to the photonic mode after SPP excitation. Simulation results show that the optimized mode conversion efficiency can be as high as 78%.To encode the input data for the plasmonic logic gate, a plasmonic phase shifter with nonlinear electro-optic slot material is analytically modeled using the coupled wave equation formalism for χ (2) nonlinear processes in lossy systems. The analytical model calculates the modulated and unmodulated signals at the output of the modulator. To find the design space of the phase shifter, such as its length, input optical power, and driving voltage, an integrated system consisting of the phase shifters, the plasmonic logic gate, the plasmonic detector, and the supporting CMOS circuits is designed.With the help of the phase shifter's analytical model and electromagnetic simulation of the connected devices, the BER of the integrated plasmonic devices is calculated. As Joule heating is common in plasmonic devices, and the nonlinear material of the phase shifter becomes unstable at ∼ 100oC, thermal analysis is performed using a 1-D thermal model to find the average temperature of the devices. A maximal but realistic BER of 10−3 and a driving voltage of 1V is taken as the limit. The length and driving voltage of the phase shifter are varied, and the input optical power is searched that maintains the target BER, and at the same time, does not heat the system to the extent of making the devices inoperable.
■590 ▼aSchool code: 0078.
■650 4▼aCooling
■650 4▼aSignal to noise ratio
■650 4▼aWire
■650 4▼aBandwidths
■650 4▼aBoolean
■650 4▼aElectric fields
■650 4▼aChemical vapor deposition
■650 4▼aCircuits
■650 4▼aSupercomputers
■650 4▼aDesign
■650 4▼aCMOS
■650 4▼aTransistors
■650 4▼aOptics
■650 4▼aEnergy consumption
■650 4▼aHigh performance computing
■650 4▼aComputer science
■650 4▼aElectrical engineering
■650 4▼aMaterials science
■650 4▼aElectromagnetics
■690 ▼a0389
■690 ▼a0752
■690 ▼a0984
■690 ▼a0544
■690 ▼a0794
■690 ▼a0607
■71020▼aGeorgia Institute of Technology.
■7730 ▼tDissertations Abstracts International▼g87-05A.
■790 ▼a0078
■791 ▼aPh.D.
■792 ▼a2023
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360553▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.


