Surface Characteristics, Etching Behaviors and Chemicalmechanical Polishing of Aluminum Alloy thin Films
- PDF / 445,542 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 16 Downloads / 234 Views
microstructures
and
electrochemical
nature
of metals.
Among the metals for
interconnection, Al, despite its dominance, receives disproportionately little attention for its CMIP process development. At a time when the implementation of Cu-based metallization still remains virtually at the R&D level, extension of Al metallizations down to 0.25 ýLrm design rule and below has been demonstrated, owing to the development of Al damascene process [1]. Such progress makes it urgently necessary to develop a feasible Al CMII process. In our previous investigations, we found that the Al CMP polish rate depends heavily on the pH and oxidizer concentration of slurry used [2]. Besides, the addition of minimum amount of Si and Cu into Al brings significant changes to Al CMII behaviors [3]. As part of our continuing research, the present study is targeted at elucidating the chemical mechanisms during Al CMI and relating them to the material characteristics of Al thin films. Static etch rate data for Al thin films of various compositions is compared with dynamic CMP polish rate data gained under the same chemical environment. Using H202-based etchant/slurry as a reference, the chemical erosion and removal mechanisms involved during etching and CMP are proposed. EXPERIMENTAL Nominally pure Al (99.995%), Al-l.0%Si, Al-0.5%Cu, Al-l.O%Si-0.5%Cu, and Al-l.0%Cu thin films are sputter-deposited onto oxidized 150 mm Si wafers to a thickness of about I ýsm. These samples are cut into 1 cm 2 pieces for etching experiments. Hydrogen peroxide is added into the etching solution as the main oxidizer whose concentration is varied from 0 to 10 % by volume. 5% 125 Mat. Res. Soc. Symp. Proc. Vol. 477 0 1997 Materials Research Society
phosphoric acid is used as the pH buffer solutions. Adjustment in pH was achieved by adding variable amount of KOH into the solution. Cut samples are immersed in the above etchant for 1000, 3600, and 10000 seconds. After etching, samples are cleaned in D.I. water and blow dried. Thickness of Al thin films is determined from sheet resistance measurements and confirmed subsequently by scanning electron microscopy (SEM). An EG&G Princeton Model 273A potentiostat/galvanostat is used for electrochemical (Tafel) measurements [4]. A scan rate of I mV/s is used on Al samples of approximately 1 cm 2 in area exposed to the etchant with 2.5 vol % H202 and a pH of 2. Quantitative ESCA analyses are performed to all Al films. Relative native oxide thickness is calculated by comparing the relative intensities of the oxide and metallic Al 2p peaks based on the following formulae [5]: dox(Ak) = 24 In (1.4 Io/Im +1)
(1)
where I, and Im are the intensities (i.e. peak areas) of the oxide and metal photoelectron peaks respectively, which can be determined from the computer curve fit of experimental Al 2p peak envelopes. All spectra were acquired with Mg Kot X-ray source and all binding energies were referenced to C Is line at 285.0 eV. The precision for binding energy measurements was ± 0.1 eV. Roughness is determined by Atomic For
Data Loading...
