ICMCTF1998 Session H3: Chemical Mechanical Polishing
Time Period ThA Sessions | Abstract Timeline | Topic H Sessions | Time Periods | Topics | ICMCTF1998 Schedule
Start | Invited? | Item |
---|---|---|
1:30 PM | Invited |
H3-1 Advanced Applications of Orbital CMP Technology
K. Holland (Ipec Planar) |
2:10 PM |
H3-3 Non-Contact Post-CMP Cleaning
A. Busnana (Clarkson University) Current post-CMP cleans are contact cleaning techniques. Contact cleaning may cause scratching, and brush shedding which occur under many operating conditions causes additional defects. There is a need to minimize contact post-CMP clean process. Megasonic cleaning provides the best complement or alternative to brush clean. Non-contact cleaning of polished wafers (using silica-based slurry) is studied using DI water and/or basic chemistry. The removal of slurry from dipped and polished thermal oxide wafers is studied. The effectiveness of the megasonic cleaning process in post-CMP cleaning has been demonstrated. A non-contact (megasonic) cleaning process for the removal of slurry from dipped and polished wafers has been developed. Complete particle removal (100%) was achieved using megasonics with DI water or DI water with 1% NH4OH using wafer dipped in STI silica slurry. The optimum conditions for megasonic cleaning power, temperature and time were determined for the removal of the STI silica slurry (using DI water with 1% NH4OH). Up to 99% particle removal (of STI silica slurry) from polished wafers was accomplished using non-contact megasonic cleaning with 1% ammonia. |
|
2:30 PM |
H3-4 Assessment of Post-CMP Cleaning Mechanisms Using Statistically-Designed Experiments
G. Zhang, F. Dai, A. Zutshi (IPEC); T. Bibby (IPEC Planar); S.P. Beaudoin (Arizona State University) Post-CMP wafer cleaning has become a key step in successful CMP processing. The cleanup strategy includes post polishing buffing, wet wafer transfer, and wet wafer cleaning. Typically post-CMP cleaning uses a combination of double-sided brush scrubbing, megasonic cleaning, and spin-rinse drying to remove polishing residues from wafer surfaces. This paper investigates the importance of these processes by evaluating IPEC Planar's Avanti 9000 post-CMP cleaner, and infers the mechanisms of particle removal during these processes. Previously polished and cleaned W sheet wafers were used for this study. These substrates were dipped in diluted CMP slurry before being cleaned by the Avanti 9000 with the megasonic tank off-line. Diluted SC-1 and DI water were used as the cleaning solutions in the brush scrubbing and spin-rinse drying units, respectively. Our results show that the number of brush cycles, the brush pressure, and the brush rotating speed are key parameters in the removal of particles with diameters greater than 0.21 microns from the wafer. Two factor interactions between: 1) the number of brush cycles and the brush pressure, and 2) the number of brush cycles and the brush rotating speed, were also important to the process performance. Our results also show that the extent of particle removal is relatively independent of the operation of the spin-rinse dryer over the range studied. Thus, after some minimum level of SRD speed and number of cycles (in our studies 500 rpm and 1), operating the SRD for longer does not improve cleaning effectiveness. This suggests that particle removal occurs only during the brushing process, and that the SRD serves only to transport previously dislodged particles off the wafers and to facilitate drying, not to actually remove bound particles from the wafers. Based on these results we will present physical and theoretical arguments for the mechanisms of particle transport and removal during post-CMP cleaning. |
|
2:50 PM |
H3-5 A Modified Multi-Chemicals Spray Cleaning for Post STI CMP Cleaning Application
Y.L. Wang, T.H. Kuan, Y.K. Lin, W.T. Tseng, S.A. Wu (Taiwan Semiconductor Manufacturing Co., Ltd., ROC) As device geometries have continued to shrink and circuit complexity has continued to grow, device isolation has become a major factor limiting circuit density. Shollow trench isolation technology (STI) has been developed and applied to deep-sub-micro process. Poly Si or CVD Si2N4 or SiO2 were deposited on trench and planarized by CMP. However, the wafer surface after processing CMP is contaminated with particles and metallic impurities in slurry. The silica particles will defect VLSI patterns, and metallic impurities even if it is a bit of quanity, will induce many crystal defects in Si wafers during furnace processing. Therefore, the post CMP cleaning very important for STI process. However, the wafers for poly Si surface is hydrophobic, siO2 surface is hydrophilic and Si3N4 films is easy to charge up. The defect is difficult to remove them by conventional cleaning technique. In this study, we propose to use a modify multi-chemicals cleaning for post STI CMP clean. We use the modified heated (75-90) °C) ammonia/peroxide mix (APM) clean with 5 sec NH4OH pre-soak time in order to pre-wetting the polished surface and then the NH4OH helps to remove slurry particles from the wafer surface through including a negative zeta potential between the silica particles and polished surface. There is a clear advantage of etching a thin layer of hydrous oxide with dilute HF, since the films had been damaged during CMP process. The metallic ions can be "trapped" in the damage film and can be very difficult to remove without a light etching. Also, the hydrochloric/peroxide mix (HPM) clean with high temperature (75-80 °C) can help to remove all the metallic ions. This cleaning method of improving the solubility is the formation of soluble complexes. |
|
3:30 PM | Invited |
H3-7 Summary of Recent Developments in CMP Modelling
S.R. Runnels (Southwest Research) |
4:10 PM |
H3-9 Chemical-Mechanical Polishig for Current-in-Plane Giant Magnetoresistance Device Integration
R. Gutmann, Y.Z. Hu, T.P. Chow (Rensselaer Polytechnic Institute); W. Witcraft (Honeywell, Inc.) Current-in-plane (CIP) giant magnetoresistance (GMR) device integrations with Si Ics requires an atomically-smooth dielectric surface for the conformal deposition of GMR multilayers. A chemical-mechanical polishing (CMP) process demonstrates tht PVD silicon nitride films can be polished with less than 0.2 nms surface roughness with a wide process window. For this application, CMP of silicon nitride si a more robust and manufacturable process than CMP of PTEOS oxide. The CMP processes will be described along with alternative post-CMP cleaning results. |
|
4:30 PM |
H3-10 Material Characteristics and Chemical Mechanical Polishing of Aluminum Alloy Thin Films.
T.H. Kuan, Y.L. Wang, Y.K. Lin (Taiwan Semiconductor Manufacturing Co., Ltd., ROC); S.A. Wu (Taiwan Semiconductor Manufactuirng Co., Ltd., ROC) An experimental slurry for chemical-mechanical polish (CMP) of aluminum alloy thin films is formulated and its polish performance evaluated. De-agglomerated α-Al2O3 particles are added as the abrasive. Hydrogen peroxide is added into the slurry as the main oxidizer while phosphoric acid acts as the reducer to dissolve oxidized Al. The addition of citric acid improves the pH stability of the slurry significantly. During metal CMP, the metal film surface reacts with the slurry chemicals, resulting in corrosion or dissolution reactions. In this study, Al-CMP characteristics are evaluated on Al alloy thin films of different compositions including, pure Al, Al-1.0%Si-0.5%Cu, Al-1%Cu, Al-0.5%Cu and Al-1%Si. Characteristics of the films such as native oxide, alloy content and grain size and their effects on the CMP removal rates will be assessed. Corrosion potentials of each film will also be measured to investigate the corrosion susceptibilities of these films and their correlation with Al CMP results will be discussed. Slurry chemistry is also a critical issue in the Al CMP process. The slurry pH values and the concentrations of hydrogen peroxide are varied to investigate their influences on Al CMP removal rate. Static etching experiment is designed to simulate the chemical environment that the films are subjected to during the CMP process. These results are consistent with those obtained from the CMP experiments, suggesting that alloying compositions influence the polishing behavior to a great extent, and chemical reactions play a dominant role in the Al CMP process. |
|
4:50 PM |
H3-11 Copper CMP for Damascene Patterned Cu-Polymer Interconnects
R. Gutmann, D.T. Price (Rensselaer Polytechnic Institute) High performance IC interconnects consist of copper metallization and low-k dielectrics. In this work alternative dual damascene patterning techniqures have been evaluated with parylene and benzocyclobutene (BCB) two-level test structures. The effect of alternative barrier layers, polish stops and CMP process parameters will be presented. Specific contact resistance of 2-3 x 10E-9 ohm-cm2 have been achieved over 125 mm diameter wafers, comparable to that achieved previously with oxide interlevel dielectrics (ILDs) and copper metallization. |