独立晶片表面颗粒沉积的数值分析

时间:2023-05-09 08:54:00 浏览量:0

ABSTRACT. Numerical analysis was performed to characterize the particle deposition behavior on a horizontal freestanding wafer with thermophoretic effect under  the turbulent flow field. A low Reynolds number turbulent k-E model was used to  analyze the turbulent flow field around the water. The deposition mechanisms  considered were convection, Brownian and turbulent diffusion, sedimentation, and  thermophoresis. The averaged particle deposition velocities and their radial distributions for both the upper and lower surfaces of the wafer were calculated from the  particle concentration equation in an Eulerian frame of reference. When the wafer  is unheated, in the diffusion-controlleddeposition regime with particle size d, < 0.1  pm the averaged particle deposition velocity under the turbulent flow was about 1.3  times higher than the laminar flow case, and the local deposition velocity near the  center of the wafer was high equivalent to that near the edge. The particle  deposition on the lower surface was comparable to that on the upper surface. When  heated, the deposition-free zone, where the deposition velocity is lower than lo-'  cm/s, exists between 0.096 pm and 1.6 pm with AT of 10K, indicating shifting  behavior to larger size range compared with the laminar flow case. As for the local  deposition velocities, for dp < 0.05 pm, the deposition velocity is higher near the  center of the wafer than near the wafer edge, whereas for dp=2.0 pm the  deposition takes place mainly on the inside area of the wafer. The comparison of  the present numerical results with the experimental data by Ye et al. (1991) showed  reasonably good agreement. Finally, an approximate deposition velocity model was  suggested. The comparison of the model calculations with the present numerical  results and the experimental data of Opiolka et al. (1994) showed good agreement.  AEROSOL SCIENCE AND TECHNOLOGY 25: 141- 156 (1996).


Particle deposition on semiconductor wafers is an important problem in VLSI circuit  manufacturing. For this reason so far much  effort has been devoted to reducing the  contaminant levels in clean rooms and in  the VLSI manufacturing equipment. In order to understand and develop strategies  for minimizing particle deposition it is necessary to characterize the flow field around  the wafer and investigate various transport  mechanisms. Theoretical and experimental  works concerned with this subject before  1990 were summarized by Pui et al. (1990).  Since then several research works for reducing particle deposition onto a wafer by  using thermophoresis have been progressed  theoretically and experimentally (Opiolka  et al., 1990, 1994; Ye et al., 1991). Recently,  Fujii et al. (1989) and Pui et al. (1991)  reported the existence of flow fluctuations  in the vicinity of the wafer produced by  filter media and other obstructions in the  clean rooms; however, none of the existing  theoretical studies on particle deposition  onto a wafer is sufficiently detailed to allow  the understanding of particle transport and  deposition behaviors around the wafer under the turbulent flow field. Therefore, effects of turbulence on particle deposition  onto a wafer deserve examining.


The deposition mechanisms related to  particles are diffusion, convection, sedimentation, inertia, thermophoresis, electrostatic force, and turbulence. It is generally  known that there is not enough information  about the charge distribution of clean room  particles to evaluate the exact influence of  electrostatic forces on particle flux. In this  work, the particle deposition onto a horizontal freestanding wafer was studied numerically under the turbulent flow field  similar to that reported recently in the clean  room. In addition, sedimentation and thermophoresis were also considered. In order  to predict the turbulent flow field around  the wafer, a low Reynolds number k-s turbulence model was used in the analysis.  This model allows the facility on the treatment of the very thin particle concentration  boundary layer due to high Schmidt number of particles. Both average and local  deposition velocities were calculated and  discussed in detail. The averaged deposition velocity were compared with laminar  flow results, those of the available experimental data. Finally, an approximate deposition velocity model is suggested. The results of our suggested model are compared  with the present numerical results, those of  the existing model and the available experimental data.  


1

Fig1


The SIMPLE algorithm by Patankar (1980)  was modified to consider the particle transport including the migration due to the  external force fields as pointed out by  Stratmann and Whitby (1989). The velocity  and temperature fields of the fluid were  calculated first using the momentum and  energy conservation equations of the fluid  Eqs. 6-14, and these calculated values were  then used as input for the convective-diffusion equation Eq. 4. Finally, the related  particle deposition velocities were calculated for the upper and the lower sides of  the wafer.


The calculation grid system introduced  here consists of 175 (axial) X 104 (radial)  grid lines. The grid test was carried out for  the laminar flow including thermophoresis  and sedimentation before starting the main  calculation, an the calculated results for  AT = 10K, where AT = T, - To, were  shown in Figure 2 with the grid spacings  100, 25, 10, and 5 prn for the test. The  experimental results of Ye et al. (1991)  were also shown for comparison. It is noted  that the numerical results are very sensitive  to the grid spacing between the nearest  grid and the wall of the wafer. It can be  seen that for d, < 1 pm, the deposition  velocities for the grid spacing of 100 Fm  overestimate those of the 10 pm grid spacing largely, but in the cases of the grid  spacings of 10 and 5 pm, the deposition  velocities are the same. In the present work,  the first grids from the surface of the wafer  in all directions were located at the distance of 10 pm, and the other grids were  arranged to increase geometrically with geometric ratios less than 1.2.

文件下载请联系管理员: 400-876-8096