Planarization and Regrowth of Self-Aligned Ohmic Contacts on InGaAs

Transcription

1 MBE 2008, Vancouver, B.C. Planarization and Regrowth of Self-Aligned Ohmic Contacts on InGaAs Mark Wistey, Greg Burek, Uttam Singisetti, Austin Nelson, Brian Thibeault, Joël Cagnon, Susanne Stemmer, Arthur Gossard, Mark Rodwell University of California, Santa Barbara Seth Bank University of Texas-Austin

2 Outline Motivation: Can InGaAs MOSFETs beat Si? Contact resistance is key: Rc ~ 1 / f 2 Planarization & Etchback Self-aligned Regrown Low Resistance Contacts and MEE InGaP Etch Stop Layer Conclusions 2

3 In Order to Beat Silicon... Source n+ Regrowth Raccess Rc{ Gate High-k Channel Barrier Substrate Drain Rsd = 180 Ω-µm Target... Strategy... InxGa1-xAs channel Ids = 6 ma/µm & low gate leakage Lg = 22 nm, low SD leakage High gate barrier: MOSFET High back barrier: AlGaAs Rc = 1 Ω-µm 2 (10-8 Ω-cm 2 ) Raccess = 10 Ω-µm Rc ~ 1 / f 2 Regrow S/D epitaxially.* Backfill channel recess etch.* *Major challenges for MBE. Source: Rodwell IPRM

4 Surface Cleaning for Regrowth Encapsulate all gate metals Surface clean: Recess etch UV ozone 30min 1:10 HCl:H2O for 1 min, DI rinse UHV bake 200 C H clean or thermally desorb oxide No extended defects, low contact resistance: 1.3 Ω-µm 2 Channel HRTEM InGaAs n+ blanket regrowth Interface HAADF-STEM` Interface Resistance (!) TLM Measurement T=100µm n-3.6e19 cm -3 W=25µm InGaAs n+ 2 nm Rc=1.3 Ω-µm Gap (µm) 4

5 Planarization & Etchback Goal: self-aligned, selective area contacts 2. O2 Ash or Developer Polymer Regrowth Channel Regrowth Channel 3. Mo & InGaAs Etch 4. Strip resist Polymer Source Regrowth Channel Regrowth Channel Drain 5

6 Planarization & Etchback Goal: self-aligned, selective area contacts 1. Spin on thick polymer Polymer 2. O2 Ash or Developer Polymer Regrowth Channel Regrowth Channel 3. Mo & InGaAs Etch 4. Strip resist Polymer Source Regrowth Channel Regrowth Channel Drain 5

7 Planarization & Etchback Goal: self-aligned, selective area contacts 1. Spin on thick polymer Polymer 2. O2 Ash or Developer Polymer Regrowth Channel Regrowth Channel 3. Mo & InGaAs Etch 4. Strip resist Polymer Source Regrowth Channel Regrowth Channel Drain 5

8 Planarization & Etchback Goal: self-aligned, selective area contacts 1. Spin on thick polymer Polymer 2. O2 Ash or Developer Polymer Regrowth Channel Regrowth Channel 3. Mo & InGaAs Etch 4. Strip resist Polymer Source Regrowth Channel Regrowth Channel Drain 5

9 Choice of Polymers PMGI / MIF developer BCB / CF4+O2 SPR510 / O2 plasma 1µm Lost in Mo etch Lost in Mo etch (Gate missing) Difficult to process Smooth & uniform Spin: 4kRPM 30s. Bake: 90 C 1min, 110 C 1min. Photoresist (SPR-510): Easy to process Withstands wet & dry etches Smooth & uniform... 6

10 Low Power Plasma Gives Smooth Etchback Etchback in high energy O2 plasma: Extreme roughness (micromasking) Scum Etchback in low energy O2 plasma (ICP) or UV ozone: Clean, smooth surfaces, no scum After etchback After Mo etch and PR strip: No Mo Mo 7

11 Planarization: Repeatable and Easy Thickness by naked eye Purple/Blue 300nm Yellow 200nm Light Blue 100nm Burnt/Black <70nm High yield esp. for <1µm Mo+InGaAs SPR-510A SiO 2 Cr 8

12 MBE Regrowth: Bad at any Temperature? Conditions: 0.5 µm/hr, V/III=35 Low growth temperature (<400 C): Smooth in far field Gap near gate (shadowing) No contact to channel! Gate 200nm Gap Source-Drain Regrowth s e r Channel SEMs: Uttam Singisetti 9

13 MBE Regrowth: Bad at any Temperature? Conditions: 0.5 µm/hr, V/III=35 Low growth temperature (<400 C): Smooth in far field Gap near gate (shadowing) No contact to channel! Gate 200nm Gap Source-Drain Regrowth s e r Channel Gate High growth temperature (>490 C): Selective/preferential epi on InGaAs No gaps near gate Rough far field High resistance Source-Drain Regrowth Regrowth: 50nm InGaAs:Si, 5nm InAs:Si. Si=8E19/cm3, 20nm Mo. SEMs: Uttam Singisetti 9

14 Gap-free Regrowth by MEE Migration-Enhanced Epitaxy (MEE) conditions: C (pyrometer) As flux constant ~ 1x10-6 Torr: V/III~3, not interrupted. 0.5nm InGaAs:Si pulses (3.7 sec), sec As soak RHEED: 4x2 ==> 1x2 ==> 4x2 with each pulse. SEM Cross Section SEM Side View (Oblique) Top of gate dummy gate dummy gate InGaAs Regrowth Original Interface InGaAs Regrowth SEM: Greg Burek SEM: Uttam Singisetti No gaps Smooth surfaces. High Si activation (4x10 19 cm -3 ). Quasi-selective: no growth on sidewalls 10

15 Rough Regrowth on Thin InP Etch Stop Layer Conversion of 2-4nm InP to InAs Strain relaxation As P InGaAs InP etch stop (2-4 nm) InAlAs barrier As 11

16 Rough Regrowth on Thin InP Etch Stop Layer InP regrowth RHEED InP regrowth SEM SEM: U. Singisetti Conversion of 2-4nm InP to InAs Strain relaxation As P InGaAs InP etch stop (2-4 nm) InAlAs barrier InAs InGaAs InAlAs barrier InAs As 11

17 Regrowth on InGaP Replace InP with InGaP Converts to InGaAs (good!) Strain compensation As P InGaAs InGaP Converted from InGaP InGaAs InGaAs InGaP InGaP InAlAs barrier InGaP regrowth RHEED InGaP regrowth SEM 2 12

18 Summary Surface clean before regrowth: UV ozone, 10% HCl, then H clean or thermal desorb Lowest resistance regrown contacts: Rc = 1.3 Ω-µm 2 (1.3x10-8 Ω cm 2 ) Planarization by photoresist: simple & repeatable No lithography needed MBE + Planarization = Self-Aligned Regrowth Gap-free regrowth (n=4x10 19 cm -3 ) by MEE above 490 C InGaP etch stop prevents relaxation before regrowth 13

19 Acknowledgements Chris Palmstrøm and Erdem Arkun (now at UCSB) IBM Yorktown: Yanning Sun, Edward Kiewra, Devendra Sadana SRC 14

20 Additional Slides 15

21 Transmission Line Method contact Length (L) contact Conducting layer R(L) = (R SH /W)*(L+2L t ) R(0) = 2R c = 2L t *R SH /W 2L t r c = R c *A c r c = R SH *L t 2 2R c Slide courtesy Adam Crook, 2007

22 Transmission Line Method I 2L t 2R c Width Gap V 4-Point Probe Resistance measurements Resistance (!) TLM Measurement 100µm n-3.6e19 cm -3 W=25µm Rc=1.3x10-8 Ω-cm Gap (µm) + Image courtesy Adam Crook, 2007

23 Contact Resistance: In-situ Mo Contacts Step Ω-µm 2 In-situ Mo n+ Regrowth 0.5 Ω-µm 2 Step 1 In-situ Mo n+ InGaAs SI InP n+ Regrowth n+ InGaAs SI InP In-situ Mo n+ Regrowth In-situ Mo InGaAs-InGaAs regrown interface resistance < 1 Ω -µm 2 on unprocessed surfaces. Regrown interfaces comparable with 0.5 Ω-µm 2 from continuous epitaxy. TLMs by U. Singisetti 18

24 19 SRC Nonclassical CMOS Research Center MOSFET Process Flow Detail M. Rodwell, SRC Review, 2008 non-selective area S/D regrowth PR PR selective area S/D regrowth

25 Process Flow: Gate Deposition High-k first on pristine channel. Tall gate stack. Litho. Selective etches to channel. NID InGaAs Channel InP etch stop InAlAs barrier InP substrate 20

26 Process Flow: Gate Deposition High-k first on pristine channel. Cr Tall gate stack. Litho. Selective etches to channel. s e r NID InGaAs Channel InP etch stop InAlAs barrier InP substrate 20

27 Process Flow: Gate Deposition High-k first on pristine channel. Cr Tall gate stack. Litho. Selective etches to channel. Critical etch process: Stop on channel with no damage s e r NID InGaAs Channel InP etch stop InAlAs barrier InP substrate 20

28 Gate Stack: Multiple Layers & Selective Etches Key: stop etch before reaching dielectric, then gentle low-power etch to stop on dielectric Rodwell IPRM

29 Process Flow: Sidewalls & Recess Etch SiNx or sidewalls Encapsulate gate metals Controlled recess etch Slow facet planes Not needed for depletion-mode FETs, SiNx s e r Channel InP etch stop InAlAs barrier InP substrate 22

30 Regrowth Interface Resistances: Measured Data SRC Nonclassical CMOS Research Center When tested individually in separate experiments: In-situ Mo Contact ρ c =1 Ω - μm 2 25 nm regrown InGaAs R sh =70 Ω/sq InGaAs-InGaAs re-growth resistance < 1 Ω - μm 2. InGaAs-InP re-growth resistance = 6 Ω - μm 2 (on thick InP).