Study on Cu Cleaning Efficacy Depending on Initial Contamination Method
- PDF / 1,054,012 Bytes
- 7 Pages / 414.72 x 648 pts Page_size
- 71 Downloads / 234 Views
INTRODUCTION
With down scaling of ultra large scale integrated (ULSI) devices, metal contaminationfree Si surface is essential to realize the high performance since it has a fatal effect on device characteristics.' Among them, noble metals such Cu, Ag and Au with high oxidationreduction (redox) potential form chemical bond with Si substrate and are hardly removed from Si surface. 2,3 Moreover, Cu is the metal impurity to be contaminated most easily in device manufacturing processes such as implantation, dry etch and ashing process.4 In order to retain the metal contamination free Si surface, we must understand accurately the mechanism for the adsorption and desorption of metal contaminants on Si surface and design the optimal cleaning method. Up to now, the cleaning efficiency of a certain chemical solution for metal has been generally known to be identical with no consideration for initial contamination conditions such as initial contamination concentration and contamination solutions. However, the expected metal contamination level ranges over different levels and there are many chemical solutions that can be expected as a metal contamination source. For this reason, various studies for metal adsorption and desorption on Si surface have been reported. In this work, we investigated the dependence of cleaning efficiency of typical dissoluble chemical solutions such as 0 3 -UPW, SPM and HPM for Cu on the pre-contamination method with various contamination solutions and different initial concentrations and 259
Mat. Res. Soc. Symp. Proc. Vol. 477 01997 Materials Research Society
RESULTS AND DISCUSSION
The stability and corrosion characteristics of Cu in Cu-water system depend on pH and redox potential vs. NHE (normal hydrogen electrode as a reference electrode). In Fig. I, we summarized the dissoution behaviors of Cu in above mentioned typical dissoluble chemical solutions used in this study from pH and redox potential values in Table I and potential-pH diagram. However, according to our previous research, Cu on Si surface can not be dissolved 23 in UPW and DHF having a lower redox potential than 0.75 V and 0.85 V respectively. Therefore, we can expect that the contaminated Cu on Si surface is fundamentally removed by cleaning in typical dissoluble chemical solutions such as 0 3-UPW, SPM and HPM. In Fig.2, it is shown that the cleaning efficiency of typical dissoluble chemical solutions for Cu on Si surface has a dependence on contamination solution. The hydrophobic Si wafers contaminated by means of dipping into UPW and DHF spiked with 10ppb CuCl2 for 3min. were cleaned by typical dissoluble chemical solutions such as 0 3 -UPW, SPM and HPM for 10 min. followed by a post-UPW rinse for 10min.. The cleaning efficiencies for Cu are almost same among these chemical solutions regardless of contamination solutions. Cotarary to our expectation mentioned in Fig. 1, after cleaning, the residual Cu concentration on Si surface has a variation of about one order showing a dependence on contamination solution. The reason why
Data Loading...
