In this thesis, the wafer environment contamination control and the plasma process of silicide shallow junction were the main topics. The influences of AMC and plasma process to device performance were studied and the new improvement methods were developed.
In chapter 2, a clean bench with different filter modules such as the NEUROFINE PTFE filter, the glass-fiber ULPA filter, and a combination of chemical filters with both of these was set up in the cleanroom. All elements in the air or on the wafer were analyzed through air sampling and wafer sampling inside the bench, to determine clearly the ability of each filter module to suppress AMC. The practical effects of AMC on device performance were also investigated by actually exposing a wafer to air in an especially controlled clean bench before gate oxidation.
In chapter 3, the performance of two ULPA filter modules - the PTFE and the glass-fiber was investigated. PTFE fiber is considered to be resistant to acid vapor corrosion and it is a good material from which to fabricate ULPA filters. The source of AMC and the effects of AMC on devices exposed in the clean bench in a HF vapor environment were examined to elucidate the impact of filter material. The results indicated that the PTFE ULPA filter can provide a better cleanroom environment than a commercial glass fiber ULPA filter.
In chapter 4, the optimum process condition of NiSi on different silicon substrates and the relation between sheet resistance and linewidth were investigated. In addition, the thermal stability of NiSi processed with second rapid thermal annealing (RTA) was also studied.
In chapter 5, a new selective liquid-phase deposition (S_LPD) process was proposed to overcome the RIE damage problems. S_LPD process is the selective deposition technology
which deposits silicon dioxide on silicide surface against photoresist. This technology was conduct under non-plasma environment and hence it will not induce the surface damaged and contaminated layer. The experimental results indicate that the new S_LPD process indeed has the superior of suppressing plasma damage on devices. Therefore, the new S_LPD process may become the candidate of non-plasma process for future advanced device manufacturing.
Finally, conclusions and recommendations for further research were summarized in Chapter 6.
TABLE 1-1
ITRS 2003 Technology Requirements
Year of Production 2003 2004 2005 2006 2007 2008 2009
Technology Node hp90 hp65
MPU/ASIC ½ Pitch (nm) 107 90 80 70 65 57 50
MPU Printed Gate Length (nm) 65 53 45 40 35 32 28
MPU Physical Gate Length (nm) 45 37 32 28 25 22 20
Equivalent physical oxide thickness
for MPU/ASIC Tox (nm) 1.3 1.2 1.1 1.0 0.9 0.8 0.8
Drain extension Xj (nm) 24.8 20.4 17.6 15.4 13.8 8.8 8.0
Maximum drain extension sheet
resistance (PMOS)(Ω /sq) 545 663 767 833 884 1739 1800
Maximum drain extension sheet
resistance (NMOS) (Ω/sq) 255 310 358 389 412 811 840
Contact Xj (nm) 49.5 40.7 35.2 30.8 27.5 NA NA
Maximum silicon consumption (nm) 24.8 20.4 17.6 15.4 13.8 13.2 12
Silicide thickness (nm) 25 20 21 19 17 16 14
Contact silicide sheet Rs (Ω/sq) 6.5 7.9 7.5 8.6 9.6 10.0 11.1
Contact maximum resistivity (Ω-cm2) 1.93E-07 1.62E-07 1.44E-07 1.20E-07 1.05E-07 0.87E-07 0.72E-07 Airborne Molecular Contaminants in gas phase (pptM))
Lithography—bases (as amine,
amide, and NH3) 750 750 750 <750 <750 <750 <750
Gate—metals (as Cu, E=2×10–5) 0.15 0.1 0.1 0.07 <0.07 <0.07 <0.07 Gate—organics
(as molecular weight to 250,
E=1×10–3) 80 70 60 60 50 50 50
Dopants <10 <10 <10 <10 <10 <10 <10
Airborne Molecular Contaminants, Surface Deposition Limits (for Si Witness Wafer, 24-hour Exposure to Closed FOUP, Pod, Mini-environment or Air)
SMC organics on wafers, ASTM
1982–99, ng/cm2 4 2 2 2 2 2 2
Front-end processes, bare Si, total dopants added to 24-hour witness
wafer, atoms/cm2 <2E12 <2E12 <2E12 1E12 1E12 1E12 1E12
Front-end processes, bare Si, total metals added to witness wafer,
atoms/cm2 <2E10 <2E10 <2E10 <2E10 <2E10 <2E10 <2E10 SMC-Surface Molecular Condensable
C
HAPTER2
Investigation of Airborne Molecular Contamination in Cleanroom and Its Effects on Device Performance
2.1INTRODUCTION
The cleanness of a cleanroom environment must be tightly controlled to achieve high-yield and high-performance ULSI manufacturing [12]-[15]. Recently, new developments of material and process have helped to shrink device geometry. The advanced contamination control of cleanrooms has also been indispensable in this regard [16]. So far, only particle contamination has been intensively studied [17], and recently developed filters, such as HEPA and ULPA, can already sufficiently suppress particles [18]. However, the controllability of airborne molecular contamination (AMC) by present cleanroom technologies has not yet been validated because AMC is a kind of atom or molecular-level gas-phase contamination like organic and inorganic ones. As the minimum feature size of devices is continuously scaled down to far below 0.1µm, AMC will gradually become crucial in ULSI manufacturing. In the near future, especially in the nanodevice era, contamination by organic compounds, inorganic ions, and trace doping impurities [19] in cleanroom air may also dominate the characteristics, reliability, and even yield of devices. Technologies for eliminating both particles and AMC from manufacturing environment must be considered to completely solve this problem.
To date, several investigations of AMC, and its effects on device performance, have been
In this work, a clean bench was specially equipped with different filter modules such as the NEUROFINE PTFE filter, the glass-fiber ULPA filter, and a combination of chemical filters with both of these. All elements in the air or on the wafer were analyzed through air sampling and wafer sampling inside the bench, to determine clearly the ability of each filter module to suppress AMC. The practical effects of AMC on device performance were also investigated by actually exposing a wafer to air in an especially controlled clean bench before gate oxidation.