无颗粒晶圆清洗干燥技术

COMPLETE elimination of all possible impurities from semiconductor processing environments is very  important for realizing ULSI devices of submicrometer to  lower submicrometer feature sizes [l]. It is obvious that  residual contaminants on wafer surfaces should be remarkably decreased with an accompanying increase in  pattern densities. Because the particulate contamination  on wafer surfaces mainly cause device failures, particlefree wet chemical processes are required for obtaining  high yield and reliability for semiconductor fabrication.  Generally, wet chemical processes can be classified as  cleaning and/or etching processes using ultrapure water  and chemicals and a drying process to remove water from  wafer surfaces. 

Wafer cleaning experiments were performed on artificially contaminated wafers. Polystyrene latex spheres  having a diameter of 0.43 pm and silica latex spheres having a diameter of 0.5 pm were employed as organic and  inorganic artificial particulate contaminants, respectively.  Silicon wafers used for tests were 3-inches in diameter  and ( 1, 0, 0) oriented p-type (6-8 Q-cm) or n-type (3-5  Q-cm). Latex suspensions containing 7 X lo9-7 x 10"  particles/ml were diluted to lo3- lo4 particles/ml, and  dropped onto the wafer surfaces, and then spun dry. By  this procedure, wafers each having 500-2000 particles on  its surface were prepared for cleaning tests.

The wafer surfaces were observed using a scanning  electron microscope (SEM). Particle counts on wafers  were evaluated using an Aeronca WIS- 100 wafer inspection system. The WIS-100 classifies defect sizes to particles (over 0.5 pm) and a so-called haze (below 0.5 pm).  Before the experiments, the WIS- 100 was calibrated with  a relative standard wafer.

Immersion time in each of the cleaning solutions is 10  min except for the dilute-HF solution. The time for immersion in dilute-HF is 1 min. After immersion into the  solutions, the wafers were rinsed for 10 min in ultrapure  water and then spun dry. Chemicals used in these tests are  all of electronic grade and low particle levels [5]-[7]. The  particle counts of'those over 0.5-pm diameter size in the  chemicals were below 20 counts/ml or less. The quality  of ultrapure water was 18.24 Ma-cm in resistivity, 5 ppb  in total organic carbon (TOC), 2-3 ppb in silica concentration, and 0.3 count/ml in particle counts of those over  0.2-pm diameter size [8]. All equipment used for the  wafer cleaning and rinsing were made of either TeflonPFA or quartz glass. The cleaning and rinsing processes  were manually performed in a wet station.

Fig1

Fig. 1 shows a schematic diagram of the newly developed wafer drying system. This system consists of an IPA  heating and chilling system, a wafer transport system, an  IPA liquid delivery system, and a fan filter unit. The IPA  liquid in the chamber is indirectly heated by the bottom  heater. The evaporated IPA vapor is condensed by the  chiller placed at the top inside of the chamber. Condensed  IPA liquid is recycled. To make this system particle-generation free, all materials in the chamber in contact with  IPA vapor and/or liquid such as the vaporizing chamber  and the chiller were made of 316L stainless steel. All internal surfaces were precisely buffed by a mechanical polishing and a chemical treatment [9]. In the IPA delivery  system, IPA liquid is fed from the stainless steel vessel  through the Teflon membrane filter having pore size of  0.1 pm and then sent to the chamber. To avoid the particle  generation from the tubes, the electro-polished 3 16L  stainless steel tubes were employed. The wafer transport  system is automatically operated in obedience to the sequences. To prevent particle contamination during wafer  transport, the moving portions of the system are kept away  from the wafers as far as possible. Clean air is passed  through an ULPA filter and is streamed with the vertical  and laminar flow to keep the high quality of cleanliness  in wafer transport spaces.

Wafers which have diameters of 3, 4, and 5 in were  cleaned with chemicals and ultrapure water as described  above. After rinsing, the wafers were immersed into the  vapor zone for 5 min. During the wafer drying operation,  the wafers were set above the receiver in the vapor zone  and any condensed IPA containing water was immediately taken out from the chamber through the receiver.  After drying, the wafers were taken out from the chamber  using the automatic transportation robot.

Comparison of Particle Removal Eficiency: Fig. 2  shows the particle removal efficiency of the five cleaning  solutions for polystyrene latex spheres. The H2S04-H202  and the NH40H-H202 solutions with the low NH40H  content were found to have a high removal efficiency for  both organic particles and haze. The other cleaning solutions were also effective for removing particles. It  should be noticed that only the NH40H-H202 solution of  the standard ratio reported by Kern and Puotinen [2]  showed an increase in haze formation. This is due to an  irregular etching of wafer surfaces during cleaning as described later.