Plasma polymerization in CHF3 glow discharges was investigated in a bell jar type RIE system. The rate of polymerization in A/min, Ret , was studied as a function of dc self-bias in the range -200 to -530V. It Was found that at the lower values of dc self-bias, there was considerable enhancement in Rpt on the powered electrode compared to that on the grounded electrode. Rpt decreased monotonically on the powered, and increased monotonically on the grounded electrode as the dc self-bias increased. At the higher values of dc self-bias within our range of investigation, polymer formation was suppressed on the powered electrode, Rpt eventually falling to zero. XPS analysis indicated a lower fluorine content for polymer deposited on the powered electrode. A model is proposed for the plasma polymerization mechanism and some of its implications investigated using Monte Carlo simulations.Plasma polymerization plays an important role in many dry etch processes. By forming selectively on the sidewalls of etch features, formation on horizontal surfaces being prevented by the energetic ion flux from the plasma, polymers can block lateral etching and thereby produce anisotropic profiles. Further, the polymer films sometimes exhibit substrate specificity, forming on certain substrates and not on others. This phenomenon helps achieve high selectivity, and is exploited, for example, in the etching of oxide over silicon. Polymer is not deposited on the oxide, and etching proceeds unhindered. The moment the underlying silicon is reached, a thin polymer film begins to form, inhibiting further etching and thereby enhancing selectivity (2, 3). Excessive polymer formation can be deleterious, necessitating additional process steps for polymer removal.An understanding of plasma polymerization in the etch environment is necessary to fully comprehend and realistically model the overall etch process. It was this that motivated the work presented here. We studied plasma polymerization in CHF3 glow discharges because: (i) this is a discharge in which significant polymer formation does occur (4, 5) and (ii) CHF3 (with some amount of O2 as an additive) discharges are routinely employed in oxide etching.The mechanisms involved in glow discharge polymerization remain obscure. The electric breakdown of a gas results in the generation of numerous atomic, free radical, and molecular species. Positive and negative ions and electrons are also generated. Fluxes of all these diverse species are incident onto substrates placed in the discharge. Some of the charged particles, especially positive ions, often carry a great deal of kinetic energy when they strike the substrate, complicating the kinetics of surface chemical processes. The overall mechanism is very intricate and complex, and the work described in this paper is an attempt to unravel some of it. The effect of the energetic ion flux on the plasma polymerization process is especially emphasized. ExperimentalPlasma polymerization in CHF3 glow discharges was carried out in a bell jar-type reactor s...

No abstract

The etching of CHF 3 plasma polymer in fluorine containing electrical discharges was studied. The fluorine sources were SF 6' CF 4 , and mixtures of the two. For discharges in SF 6 and mixtures of SF6 and CF 4 , a good correlation was obtained between the etch rate and the atomic fluorine concentration measured using actinometry. For CF 4 , the etch rate was found to be much higher than that predicted from this correlation. This is attributed to the energetic ion bombardment of the polymer surface in the CF 4 discharge. X-ray photoelectron spectroscopy analysis of the etched polymer surface shows an increased fluorine content, but the F:C ratio was independent of the etching conditions. The implications of the results for the kinetics of fluorocarbon plasma polymerization are discussed.

Specific contact resistivity measurements have been performed on Ohmic contacts formed by etching in a CHF3/O2 reactive ion etching plasma, to determine the extent to which this process affects contact resistance. Contacts between n+ (Ns≊1020 cm−3) silicon and Ti were used. The low intrinsic specific contact resistivity of this system assures that the experimentally measured resistivity will depend on the contact etch process. Also, deep junction depths (xj≊1.0 μm) were used to minimize the sensitivity of the measurements to possible changes in sheet resistance due to overetching. The resistivity data, which was extracted via a two-dimensional analytical technique, show a strong dependence on the use of a postetch NF3/Ar clean-up plasma, as well as on a forming-gas anneal. However, it also shows little dependence on the amount of O2 flowing in the plasma. Resistivity measurements range from (9±5)×10−9 up to (2.4±0.6)×10−5 Ω cm2, depending on whether the contacts received a NF3 clean-up treatment and a forming-gas anneal. Surface analysis (x-ray photoelectron spectroscopy) of the Si surface shows a layer of fluorocarbon polymer, and a layer of silicon containing atomic F, C, and O.

A Abstract IntroductionThe discharge that is struck during plasma etching is a complex environment containing a myriad of chemical species. Some of these species can undergo polymerization reactions leading to the formation of polymeric residues on surfaces exposed to the discharge. These residues play an important role in the etch process. They may affect subsequent processing steps and may require additional steps for their removal. Often, they are responsible for obtaining good selectivities. Finally, they may act as inhibitor layers, reducing the etch rate. Lai