Development of Selective SiGe Etchants

Transcription

1 Development of Selective Ge Etchants Glenn Westwood and Rebecca Duffy Avantor, 1013 US Highway 202/206, Bldg. JR-1, Bridgewater, NJ

2 Introduction Goals Develop Ge etchants with high selectivity to,, and 3 N 4. Target etchants for Ge with <30% Ge. Target Gate-All-Around (GAA) applications. Outline of Presentation Applications Approaches to Ge/ Etch Development Etchant Development work Inhibitor identification, Clouding Performance Aging effect, Bath life Ge Thickness Effects Etch rates and interfacial Ge content. Conclusions and Future work 2

3 Applications Gate All Around (GAA) Transistors Option 1: Cavity Etch Partial Release Option 2: Full Release 3 N 4 Ge Ge Ge Substrate Selective Ge Etch 3 N 4 Ge Ge Ge Substrate 3 N 4 Substrate 1. Alternating deposition of and Ge layers 2. Partial or full release of the Ge layers 3. nanosheets or nanowires remain 3

4 Approaches to Selectively Etching Ge (20%) vs. Ge Ge Ge Pure Etch Selectivity Selective Oxidation Ge > Oxidation Ge Effective at >>Ge with TMAH-based chemistries. Avantor efforts unsuccessful to date. Achieved Ge with highly alkaline chemistries. Oxide Removal Ge Ge Limiting step is surface oxidation. Proven approach for Ge> etch. and 3 N 4 compatibility is inconceivable. Selective Oxide Removal Ge Ge Oxide > Oxide Ge Limiting step - oxide removal. Necessary to achieve and 3 N 4 compatibility. 4

5 HF/Acetic Acid/H2O2 Selective Ge Etchant Peroxyacids RCO 2 H+H 2 O 2 RCO 3 H+H 2 O Ge RCO 3 H Selective Oxidation of Ge vs. Slow to form (Days) Variation in performance Ge HF - Oxide Removal Not selective for oxide removal Poor oxide and nitride selectivity Ge Very high Ge etch rates achieved, but reported Ge: Etch selectivity of ~20:1 is still not ideal. 5

6 Avantor Approach to Developing New Ge Etchant Peroxyacids RO 2 H+H 2 O 2 RO 3 H+H 2 O HF - Oxide Removal Ge Non-Acetic Acid based peroxyacid Selected to minimize formation time Enhanced Ge : Selectivity Ge Corrosion Inhibitors selected to slow down oxide and nitride etch vs. Ge Ge Formulation Evaluated >95% Water Hydrofluoric Acid Ge/ selectivity enhancer Chelator Etch Inhibitor Added Mixed 5:1 H2O2 6

7 Etch Inhibitor Screening First Attempt Formulation Evaluated >95% Water Hydrofluoric Acid Ge/ selectivity enhancer Chelator Etch Inhibitor Added Mixed 5:1 H 2 O 2 Results No clear winners. Many improved p selectivity. Cationic surfactants (A1-A3) - worse selectivity. Picked 4 select additives for further testing. Etch Inhibitors p None A A B B B C C D D D D E E E E E F

8 Etch Inhibitor Screening Second Attempt Formulation Evaluated >95% Water Reduced Hydrofluoric Acid Ge/ selectivity enhancer Chelator Increased Etch Inhibitor Mixed 5:1 H 2 O 2 Etch Inhibitors Etch Rate p Ge (20%) (Å/min.) Ge (20%) / p Etch Selectivity B D E F Results Formulation w/ F1 Increasing ph dramatically increases Ge etch rate. p etch rates also increase with ph. 8

9 RL-1 and RL-1B Cloudy Formulations Formulation XM-529-V3 + Etch Inhibitor 3 N 4 (Å/min.) p Ge (20%) (Å/min.) Ge (20%) / p Etch Selectivity RL-1* F Rl-1B* F1 + B * Mixed 5:1 H 2 O 2 at 35 C All high selectivity etchants in initial stages of this study were cloudy. Cloudiness precipitation of: Etch Inhibitor + Ge/ Selectivity Enhancer 9

10 Formulations without Clouding Prevented clouding by: 1. Reducing Ge/ Selectivity Enhancer RL-3 RL-1 RL-4 2. Modified Etch Inhibitor RL-4 Best Etch solution to date. 3. Comparable selectivity, but lower Ge etch rates. Formulation XM-529-V3 + Etch Inhibitor 3 N 4 (Å/min.) p Ge (20%) (Å/min.) Ge (20%) / p Etch Selectivity RL-3* F RL-4* F * Mixed 5:1 H 2 O 2 at 35 C 10

11 RL-4: Bath Performance Data RL-4 was mixed with H 2 O 2 (5:1) and tested after days of storage. ~50% increase in Ge etch rates. Less increase over time than acetic acid based blends. Stable 5 RL-4 : 1 H 2 O 2 at 35 C, 650 rpm. RL-4 bath life stable out to 7 hrs. 11

12 Dependence on Etch Times RL-4 : H 2 O 2 at 35 C, 650 rpm. Etched Ge wafers for different etch times. The longer the etch time, the lower the etch rate. Ge etch rate does not decrease with bath life, instead the Ge (20%) etch is slowing because the Ge thickness is decreasing. 12

13 Etch Time Effect (XPS) Ge Etch Depth (Å) Ge ER (Normalized) % Ge Longer etch times result in decreased Ge and GeO 2 at the interface. Etch rates decrease with thinning of Ge films because of reduced Ge content Possible explanations Uneven starting Ge content in Ge films. Interface becomes depleted in Ge due to the etch Ge>. XPS collected by Ryan Thorpe, Rutgers K-Alpha XPS Facility 13

14 Additional XPS Results C, N, O Spectra Carbon Only adventitious carbon identified No nitrogen detected No indication that components from chemistry remain on the Ge XPS collected by Ryan Thorpe, Rutgers K-Alpha XPS Facility 14

15 Conclusions Ge Etchant Development Achieving high Ge: selectivity is not difficult. High Ge: selectivity without clouding is a challenge. Best Solution RL-4 Identified etch inhibitor that doesn t reduce Ge etch rates. s for partial release not full release. Stable bath life and stable when storing mixed with H 2 O 2. Ge etch rate slows with increased Ge removal due to decrease in Ge at interface. Formulation p Ge (20%) (Å/min.) Ge (20%) / p Etch Selectivity RL

16 Next Steps Improvements based on cloudy formulations High Ge etch rate and selectivities Clear solution separated from cloudiness. Use a cloudy solution p Test Application on devices Related GAA etch solutions /Ge, /Ge, and Ge/Ge. III-V release chemistries InGaAs/GaAs Ge (20%) (Å/min.) Ge (20%) / p Etch Selectivity < p Ge (20%) (Å/min.) Ge (20%) / p Etch Selectivity