화학공학소재연구정보센터
Journal of Vacuum Science & Technology A, Vol.16, No.6, 3259-3265, 1998
Surface reaction mechanisms of trifluoracetylacetone on clean and pre-oxidized Ni(110) : An example where etching chemistry does not follow volatility trends
Ni contamination is an increasing problem in the fabrication of thin microelectronic films. In previous work, we examined nickel removal using oxygen and 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (Hhfac) or 2,4-pentanedione (Hacac) and oxygen. In both cases, Ni was etched cleanly by the precursors under net oxidizing conditions. Decomposition pathways were prevalent, however,under reducing conditions. In the results reported in this article, we used temperature-programmed desorption to examine the use of another beta-diketone, 1,1,1-trifluoro-2,4-pentanedione, and oxygen in the same process. Ni was etched on the oxidized surface by the desorption of Ni(CF3COCHCOCH3)(2). between 260 and 430 K. Etching ceased as decomposition products formed. The structures of decomposition products on both surfaces indicates the promotion of beta-bond scission by Ni, predominantly at C-CF3, C-CH3 and C-CH/C-CH2. In the hexafluoropentanedione and pentanedione studies, beta scission was also the decomposition pathway. However, with trifluoropentanedione, the pathways produced many more species than with either the, unfluorinated or the fully fluorinated analog. This observation was unexpected. Another unexpected observation was that some maximum desorption rate temperatures do not follow expectations based on volatility : Instead, Ni(CF3COCHCOCH3)(2) desorbed at nearly the same temperature as Ni(CH3COCHCOCH3)(2) and much lower than Ni(CF3COCHCOCF3)(2). Another trend that does not follow predictions based on volatility is the desorption of the intact precursor from the monolayer on the pre-oxidized surface : Molecular CH3COCH2COCH3 desorbed at a lower temperature than either the CF3COCH2COCF or CH3COCH2COCH3. At present, it is unclear why the results disagree with the volatility trends.