Dirac-Band Materials 2D Group-IV monolayers Bi2Se3 3D-Topological Insulator Carrier Transport- Ballistic and Acoustic Phonons (Chapter 6) Graphene Heavier Group-IV Devices - Graphene electro-optic transistor (Chapter 3) Devices - Spin-Separator and Filter (Chapter 4 and 5) Carrier Transport- Contact Effects (Chapter 7) Device - Interconnects (Chapter 8) Carrier Transport- Defects (Chapter 8) Carrier Transport- Longitudinal & Vertical Transport (Chapter 9) Devices – Resonance Devices (Chapter 9)

Decision-Making Process

Minimum required characteristics Objective Find beyond-CMOS Materials and Devices for High-Speed Low-Power Computing Minimum required characteristics (i) Good Gate Control (ii) Large Mobility  Large Current  Faster Switching (iii) Turn-Off Capability  Low –Leakage Short-Channel Effect in CMOS  Poor Gate Control Multi-Gate Devices  + Better gate Control _ Complex Process + Strong Quantization CMOS  Velocity Saturation Dirac band Materials Large Mobility Long mean free path High thermal conductivity Graphene 2D Material Can We Solve It ? Chemical Modification ? Confinement ?  Poor Mobility [3]

Edge Effects & Transition Width Chapter 3 Performance Evaluation of Electro-Optic Effect based Graphene Transistor How to Switch-Off Graphene Transistor effectively (large ION/IOFF ratio) without strong confinement and chemical modification to open the band-gap ? Can We Solve It ? A new scheme to switch-off by creating “Virtual Band-Gap”  Allows us to relax “iii” constraint for now (iv) Non-Volatile  Achieve Zero Standby Power  Controllable Spin-Based Properties + ION/IOFF > 104 + SS < 60 mV/dec _ ION ~ 0.1 mA/μm (required > 1mA/μm) Partially Achieved Next Step of The Problem Edge Effects & Transition Width Considerable Leakage & Performance Degradation CMOS  Volatile Graphene ? Is there better or similar Dirac-band material which can give good spin control + good mobility + good electrostatic control + turn-off (directly or via unconventional schemes)? Weak SOC & Magnetic Doping  + Good for spin transport (e.g. spin interconnects) _ Bad for our targeted type of Devices  poor spin control Can We Solve It ?

Heavier Group-IV 2D Materials: Tsai W F, Huang C Y, Chang T R, Lin H, Jeng H T and Bansil A 2013 Gated silicene as a tunable source of nearly 100% spin-polarized electrons Nat Commun 4 1500 Theoretically predicted Good Bulk Spin Control in Y-shape Group-IV three-terminal devices Heavier Group-IV 2D Materials: + Larger SOC than Graphene + Similar Dirac bands ++ Controllable Bands via E field + Silicene and Germanene could be CMOS process compatible _ Slightly lower mobility than Graphene + Better Possibility  Abandon Graphene line of thought Partially Achieved + Good Spin-Separation on both edges _ Only edge spin-transport  no bulk _ Fano interference on edges  Poor conductance _ E field may change materials’ phase : QSH to BI Chapter 4 Y-Shape Spin-Separator for two-dimensional Group-IV How seriously do these negatives affect carrier transport in heavier Group IV materials ? Can they be circumvent ? Can We Solve It ?

+ Not Achieved + No Fano interference  Conductance Improved _ Still only edge spin-transport  Bulk significantly degrades spin information (Spin-Filter device) _ E field phase transition especially strong for Silicene & Germanene. Can We Solve It ? Chapter 5 Effect of Phase Transition on Quantum Transport in Group-IV Two-Dimensional U-shape Device Process Issues in making heavier 2D Group-IV materials Not Achieved Bi2Se3 3D-Topological Insulator: + 2D Dirac surface states with spin-momentum locking + Time-Reversal Symmetry prohibits backscattering  Expected : robust transport & high-mobility + Bulk Band-gap ~ 0.3 eV  exploit large energy window for spin-polarized surface transport Is there better material which has least ? 2D material like high velocity Dirac-bands Instead of edge, spin-transport should be supported on entire surface to have large spin-polarized current Large bulk-band gap to provide sufficiently large energy window to exploit Dirac surface states Can 3D-TI furnish all 4 requirements? Better Possibility  Abandon 2D Group-IV line of thought  Transition to 3D-TI +

Partially Achieved Is the 3D-TI really worth it ? But why so many different resistance trends (w.r.t. temperature) and yet everybody claimed to be demonstrating topological surface dominated transport? How strong is phonon scattering ? From State of Art-Literature in 2012-13: Can 3D-TI furnish all 4 requirements? Answer: Could Be. Very Promising Material. Note: Good electronic device needs Good transport characteristic + Provide explanation for all the resistance observations until then (> 2 yrs.)  Fundamental insights in 3D-TI transport _ Phonon scattering could be strong BUT conclusion can only be drawn when information of Fermi-Level, bias, thickness and channel length is available.  Inconclusive to accept or reject the material  Decision had to be based on application and design constraints. Chapter 6 Carrier Transport in Bi2Se3 Topological Insulator Slab Partially Achieved What about contacts and spin-flow ? 1. Device is not just channel in isolation. How do contacts affect transport ? 2. How good is TI for carrying and controlling spin-information? (spin injected via FM contacts) Chapter 7 Contact Effects in thin 3D-Topological Insulators: How does the current flow ?

+ Provide explanation for another experimental data (FM contacts) + Show current redistributes to flow on topological surfaces + Contacts very important  may result in seemingly unanticipated results _ Poor spin-info at high-temperature  Pristine Bi2Se3 3D-TI not appropriate for spintronics. BUT: How about magnetic doping  Hotly researched track of TI  Too broad with too many possibilities  Outside the purview of this thesis Chapter 7 Contact Effects in thin 3D-Topological Insulators: How does the current flow ? Seems “fine” for charge transport so far, how about suspected devices ? Partially Achieved Interconnects to replace Cu ? (based on ITRS 2011 roadmap) Band-Alignment modulation based resonant devices ? (based on IEEE TED, 60 (2013) 951-957) Chapter 8 Evaluation of mobility in thin Bi2Se3 Topological Insulator for prospects of Local Electrical Interconnects Chapter 9 Effect of Band-Alignment Operation on Carrier Transport in Bi2Se3 Topological Insulator

Partially Achieved Partially Achieved + High mobility @ Low Temp. _ Poor Mobility @ High Temp. _ Low DOS near Dirac-point (Topological Surface States)  At least thin Bi2Se3 3D-TI seemingly not good for interconnects. ? Scope still for even larger bulk band-gap material with weak phonon scattering at room temperature Chapter 8 Evaluation of mobility in thin Bi2Se3 Topological Insulator for prospects of Local Electrical Interconnects Partially Achieved Conclusion 1. 2D Group-IV material do not seem very promising for our problem 2. 3D-TI  Inconclusive (still May or May Not) 3. Need 3D-TI with weak electron-phonon coupling and large bulk band-gap. Note: For such material defects may become important (Defects cannot back-scatter but still scatter) 4. Bi2Se3 may not be suitable for our problem. Deeper into carrier transport in 3D-TI: * Effect of Defects, Vacancies * Real-Space Transport on GP-GPU * Mode-matching & Simultaneous transport in longitudinal and vertical direction Partially Achieved + Band-Alignment modulation  Resonant devices _ ION/IOFF very weak (lateral transport) _ Vertical transport  assumptions in IEEE TED 2013 invalidated  even weaker resonance to total elimination ? Lateral transport based design could be improved with the unconventional switching scheme investigated in Chapter 3 Chapter 9 Effect of Band-Alignment Operation on Carrier Transport in Bi2Se3 Topological Insulator