Copper CVD preparation in filling trenches under the iodine assistance

碩士 === 國立臺灣科技大學 === 化學工程系 === 92 === The theme of this thesis is to investigate the iodine effect in chemical vapor deposition of copper metal, which is a potential metallization technique in the back-end process in semiconductor manufacturing. The initial growth of copper, its surface reaction cons...

Full description

Bibliographic Details
Main Authors: JAN JEN TZE, 詹鎮澤
Other Authors: Dah-Shyang Tsai
Format: Others
Language:zh-TW
Published: 2004
Online Access:http://ndltd.ncl.edu.tw/handle/09596923946511854926
Description
Summary:碩士 === 國立臺灣科技大學 === 化學工程系 === 92 === The theme of this thesis is to investigate the iodine effect in chemical vapor deposition of copper metal, which is a potential metallization technique in the back-end process in semiconductor manufacturing. The initial growth of copper, its surface reaction constant, and its bottom-up trench filling capability on the TiN/SiO2/Si(100) substrate are studied in a cold-wall reactor, using Cu(hfac)(vtms) as reactant. The experiments on copper initial growth at 95C on TiN surface indicate that the initial growth rate is significantly different between the surface with and without iodine. The copper nuclei emerge earlier on the surface dosed with iodine. They increase in size and extend in horizontal direction so that they easily connect with each other to form a flat thin film. On the other hand, the copper nuclei on the TiN surface without iodine emerge after an induction time. They tend to grow in size in the vertical direction, and difficult to connect with each other to form a film. After 2 minutes of deposition, the surface dosed with iodine forms a copper layer, while the copper nuclei on the surface without iodine remain separate and isolated after 6 minutes. These differences in the morphological variation are summarized and compared in terms of the nucleus size, the surface fraction covered by copper, and the nucleus number density. The experiments on surface reaction constant are analyzed under the assumption of first order surface reaction is the rate controlling step in deposition. The deposition experiments are performed on a mini-chamber of confined space to eliminate the probable involvement of gas-phase reaction. The decrease of film thickness from the chamber entrance allows us to extract the surface reaction constant. Deposition rates on the surfaces with 4.2, 3.6, and 0.0% of iodine are measured. The activation energy of surface reaction is 8.6 kcal/mol (4.2% I), 9.6 kcal/mol (3.6% I), and 16.8 kcal/mol (0.0% I). The results indicate that the dosed iodine reduces the activation energy barrier and enhances the reaction constant, therefore, the deposition is increased. Experiments on the bottom-up trench filling show that the surface dosed with iodine exhibit such effect in a trench with specific opening at a proper substrate temperature. For instance, trenches with 2.675 and 1.725 m openings can be filled with copper at 105C. But the trench with 0.8 m opening filled with a keyhole at 105C. If the copper CVD is executed at 95C, deposition in the trenches with opening larger than 0.55 m exhibits bottom-up filling effect. Deposition in the trenches with opening less than 0.55 m displays a keyhole. If deposition is executed on the trenches without iodine, the deposition rate is very slow in the temperature range of 105-85C. The deposition rate is fast enough at 145C, but very little copper enters the trenches before the opening is closed by depositing copper.