Drc 2010 D.J.Pawlik
- 1. Sub-Micron InGaAs Esaki Diodes With Record High Peak Current DensityD. Pawlik, M. Barth, P. Thomas,S. Kurinec, S. RommelS. Mookerjea, D. Mohata, S. DattaS. Cohen, D. RitterThis work is partially supported by NSF (ECCS-0725760)Device Research Conference: June 22, 2010
- 2. Esaki Tunnel Diode (ETD)+p++JZin++JPKey Metrics Peak Current: JP = IP/Area
- 3. Valley Current: JV = IV/Area
- 4. Peak-to-Valley current ratio: PVCRJV-VPVV
- 5. ApplicationsTunneling FET (TFET) Gated “p-i-n” Esaki diode
- 6. Operated in Zener direction
- 7. High JZ needed for drive current
- 8. Scaled to deep sub-micron dimensionsSourceP++iGateMultijunction Solar CellsTunneling SRAMBJT & Contactsn++Tunnel JunctionDrainTunnel Diodeswebsite:www.sustainability.rit.edu/nanopower/E. Yalon, et Al., DRC 2010
- 9. Scaling D. J. Pawlik, et al., ISDRS 2009Max. PVCR improved with area
- 10. What about current density?InGaAs Tunnel DiodesTD1TD2MBE Grown at IQE
- 11. Higher level of doping than TD2
- 12. Contact layers: reduced doping
- 13. MOMBE Grown at Technion, Haifa, Israel
- 14. Lower Doping than TD1
- 15. Contact layers; high dopingJTD1 > JTD2RTD1 > RTD2Predictions:Fabrication ProcessMetal 1
- 16. Defined by E-Beam lithography
- 17. Au/Zn/Au (20nm/20nm/160nm)
- 18. Mesa Isolation Etch
- 19. Citric Acid : H2O2 (20:1)
- 20. ~80 nm deep
- 21. ~20 nm below the junction200 nm200 nm
- 22. Device Area CharacterizationMask Defined Radius = 100nmExtensive SEM analysis needed
- 23. Critical for accurate analysisFabrication (Continued)(1) Metal and Mesa(2) BCB Etch back(3) Contact padsFinal StructureMetal ContactsBCBJunction40 mmSubstrate
- 24. Electrical CharacterizationGround Connection: Large area ETD as virtual ground
- 25. ~25,200 mm2
- 26. Current Scales with area
- 27. PVCRTD2 > PVCRTD1
- 28. JTD1 > JTD2Preliminary Results
- 29. Area Scaling & Current DensityArea scaling of IPJP HistogramRecord high current density !!Current scales linearly with area
- 30. Deviations from trendline are likely due to uncertainties in area
- 31. Devices measured: 200+ each
- 32. Gaussian fit used to determine JP
- 33. As predicted JTD1 > JTD2Area Scaling & Current Density
- 34. Current Density vs. N*
- 35. Series Resistance (RS)~0.3 V2.5xRS introduces voltage shift
- 36. Effects: Apparent (false) reduction in JZ
- 37. Smaller currents result in smaller shifts
- 38. Smaller junctions have are effected by RS lessSeries Resistance: Constant y = m * x + bTypes of RS: Constant (RC)Provides linear change in VP(Meas) with respect to
- 39. RC is the slope of VP vs. IP(2) Area dependent (rArea)
- 40. Series Resistance: Area DependantTD2: I-V CurvesConstant Current CurvesSpecific Resistivity y = m * x + b For a constant current (IC);
- 41. RC constant V shift
- 42. rA variable V shift y m * x
- 43. ConclusionExtensive SEM measurements needed to characterize junction areas.
- 44. Junctions as small as ~0.015 mm2(R = 69 nm) were fabricated.
- 45. Degradation of JP and JZ were not directly observed
- 46. Sub-micron junctions needed to minimize RSeries effects
- 48. Highest reported JP of 9.75 mA/mm2 for any tunnel diode was measured