Search

The performance of chemical mechanical polishing (CMP) was determined by the morphological characteristics of the abrasive particle. In this study, the shape of the aggregate of fumed silica was confirmed by a method of characterizing the morphology of the particles using transmission electron microscopy (TEM). Each aggregate was classified into four types—linear, branched, ellipsoidal, and spheroidal—and the shape distribution of a fumed-silica slurry was investigated with different specific surface areas. The CMP performance of fumed silica slurries at 90 m2/g with numerous linear-shaped aggregates and 400 m2/g with the highest portion of spheroidal aggregates were compared to that of colloidal silica slurry. A model for the effects of bumpy abrasives on CMP explained the reason for this result by using penetration depth and MRR with various numbers of contact bumps. Furthermore, a mixture of spherical and non-spherical abrasives in a ratio of 4:1 exhibited an improved CMP performance.

Various methods such as transmission electron microscopy, scanning electron microscopy, dynamic light scattering (DLS), and scanning mobility particle sizer were utilized to measure size distributions for sub-100 nm colloidal silica. Compared to electron micrographs, hydrodynamic diameter of sample is dependent on its concentration from ~60 nm to ~90 nm and it is difficult to measure the poly dispersed nanoparticles and analyze its own properties due to limitation of DLS method, whereas, nano nebulizer (NN)-SMPS method distinguished two different size nanoparticles (51 nm and 82 nm) perfectly and is considerably stable to alternations of concentration and dispersity.

One of the main factors affecting the production yield of semiconductor circuits is contamination by fine particles. Therefore, the wafer cleaning process has become important to the semiconductor manufacturing process. The best method for a cleaning evaluation is to use the actual wafers in the manufacturing process, but where a general cleaning evaluation is required, contaminated wafers cannot be obtained from this source. This study proposes a separate cleaning evaluation wafer as a method for the semiconductor cleaning evaluation. We here describe 300-mm wafers that had nanoparticles of silica (SiO2) and polystyrene latex spheres (PSLs) deposited in multiple designated areas by changing the electric field from 2 kV to 11.2 kV to make spots with diameter from 20 to 50 mm. A numerical analysis was performed to predict the particle trajectory during deposition, and the ensuing experimental results were in good agreement with the simulation, with an average error of 4.2%. Next, some actual wafers were cleaned to demonstrate possible cross-contaminations on the 300-mm wafer, and good effectiveness was achieved. These results provide useful data for optimizing cleaning conditions to improve semiconductor cleaning processes.

Megasonic and jet spray cleaning in single wafer process have been used for many years for removing particulate contaminants from wafer surface. These methods were alternately used during manufacturing process. As the design rule of novel devices shrinks to the nanometer scale, particle contaminations caused significant drop in product yield. To increase the particle removal efficiency (PRE), we optimized the cleaning methods for each material surface. These optimized methods were evaluated for aluminum surface cleaning in terms of the PRE and pattern damage. We observed that jet spray cleaning is more effective than megasonic cleaning process for aluminum surface layer. Similarly the PRE in jet spray is high as compared with megasonic due to the effect of CO2 dissolved DI water.

Semiconductor industry faces a continuous challenge to increase the production yield by eliminating defect sources. In this paper, generation of typical defect during dichlorosilane-based WSi2 deposition process is systematically evaluated. The root cause analysis using tape sampling test and parts analysis revealed that the source of silicon particles during the particle generation is due to Vapor-Liquid-Solid(VLS) mechanism. The particle generation according to gas flow rate, susceptor material, susceptor structure were studied systematically. As a result, we found that silicon component particle generation could be suppressed by AlN Coating susceptor with revised structure.

The purpose of this paper is to investigate the effect of diamond grit during chemical mechanical polishing (CMP) process. CMP is the most important for planarization of wafer surface in semiconductor manufacturing. During polishing process, the wafer surface and the pad are contacted with the slurry. There are various contact area, which are wafer-pad aperture and conditioner-pad aperture, and so on. Between diamond grits and pad, those contacts can affect wafer polishing process. Conditioner of CMP dresses pad surface to maintain CMP pad surface constant. Otherwise, CMP pad will be degraded out quickly after several wafer process. Shape of diamond pattern and their effect on pad property were reported, but there are no researchers related novel diamond conditioner. In this study, we experiment CMP process by using different conditioners and analyze wafer polishing property as conditions.

Megasonic cleaning process as a wet cleaning process is routinely used in the semiconductor industry for the removal of contaminant particles from wafer surfaces. The wafer surfaces can be cleaned effectively by choosing the proper chemicals, acoustic pressure, and the frequency of acoustic field. In the present study, we propose the design improvement of a megasonic waveguide to minimize wafer damage from the bubble cavitation. The conventional direct-type waveguide is compared with the newly designed and developed indirect-type waveguide in terms of its performance in minimizing the damage of wafers having 70 nm poly-Si patterns.

Megasonic cleaning process is routinely used in the semiconductor industry for removal of contaminant particles from wafer surfaces. Effective cleaning is achieved through proper choice of chemical solutions, transducer power density and frequency of the acoustic field. Two principal mechanisms, namely acoustic streaming and acoustic cavitation, are considered to be responsible for the removal of particles from a contaminated surface. In this study, we designed two megasonic waveguides, indirect type and direct type, for a comparative study of cleaning wafers with 70 nm poly-Si pattern. We observed that the indirect type waveguide has more uniform megasonic energy which is transferred to the wafer. Also bubble size was smaller than conventional direct type. The conventional direct type waveguide has better in energy transfer, but can cause structural damages on the wafer.