Investigation of fluorocarbon plasma deposition from for use as passivation during deep silicon etching
Schematic of the Bosch anisotropic deep silicon etching process. Etch and deposition steps are cycled in order to produce an anisotropic feature profile during deep silicon etching. The -based deposition step deposits a thin polymer layer which, due to the difference in lateral vs vertical etch rates, protects the feature sidewalls during the -based etch step. The etch∕deposition cycle is repeated to achieve the required Si depth. The length of the etch and deposition steps determines the degree of scalloping on the feature sidewalls.
Cross-sectional schematic of the Surface Technology Systems High Rate Advanced Silicon Etch tool used in this work. The system consists of an inductively coupled plasma source with an wafer chuck specially designed to minimize across wafer stress and allow for through-wafer etches. An ISA model 550 optical emission spectrometer was mounted on the system to collect plasma emissions via the quartz window at the top of the chamber.
Deposition rate as a function of pressure (a) and power (b). Deposition rate generally decreases as a function of increasing pressure, while it increases with increasing ICP power. Both trends are expected and are discussed later in this work.
Refractive index as a function of pressure and power. Refractive index increases linearly with pressure, while the response to increased power is more complex, with the refractive index only decreasing when power is increased from . Despite the large changes in deposition rate observed in Figs. 3(a) and 3(b), the range of refractive indices covered by these conditions is quite small (1.373–1.384), indicating that the film composition does not appear to vary significantly over these process conditions. Carbon XPS of these films confirms these results.
Deposition rate as a function of the F:C ratio of the deposited films. The data falls into two groups: low F:C ratio and high F:C ratio , with corresponding refractive indices (1.40–1.42 and 1.365–1.385, respectively). Within each group, the deposition rate can vary quite widely as a result of changes in the process parameters. A specific combination of parameters is needed to move between one group and the other; in this case, a specific coupling of pressure and power settings.
Comparison of XPS spectra for representative films from each compositional group (see Table III for ). One film is dominated by structures (high F:C film), while the other has a more equal distribution of structures (low F:C film). The high F:C ratio film occurs when there is a combination of low pressure and high power , whereas the low F:C ratio film is generated by high pressure, low power plasmas .
Typical optical emission spectra (OES) for (a) low pressure, high power plasmas, and (b) high pressure, low power plasmas of . High pressure∕low power plasmas favor primary or secondary dissociation products (e.g., ), whereas low pressure∕high power plasmas favor highly fragmented dissociation products . Contrary to expectations, correlating the OES spectra with their deposited films shows that the -rich plasma produces a -poor film, while the highly fragmented-species plasma produces a -rich film.
XPS spectra for films deposited at 1, 4, and into the deposition step of the Bosch etch∕deposition cycle. The deposition pressure and power conditions fall within the high F:C ratio film regime, and, as expected, produce a high F:C ratio film, as evidenced by the -dominated XPS spectra. These spectra also capture some of the time-dependent behavior of the deposition process (flow changes, gradual bulk film formation, etc.).
(a) Etch rate of each of the films from the deposition DOE as a function of the film F:C ratio. The low F:C ratio films are more etch resistant than the high F:C ratio films, with a generally linear increase in etch rate as the F:C ratio increases. (b) Ratio of the etch and deposition rates as a function of the film F:C ratio. A low ratio is desired for optimal Bosch cycling, and thus, although the low F:C ratio films are more etch resistant, their slow deposition rates offset that advantage. For the films analyzed in this work, the optimum passivant film appears to have a F:C ratio of 1.45.
Film deposition process matrix for half-fractional factorial DOE. Note that the “Middle” condition is the nominal operating condition. All DOE parameters are centered on these conditions except for power, where a slightly wider power range is used.
Statistically significant responses of depositon rate and F:C ratio to the five parameters analyzed by the deposition DOE. The experimental matrix captured two-factor interactions and the statistically significant interactions are identified in this table.
Percentages of , , CF, and for representative films (Fig. 6) from the two composition regimes as found via the deposition DOE and determined by decomposition of high resolution XPS spectra. The F:C ratio is determined from the integrated areas of the high resolution C and F peaks.
Percentages of , , CF, and for both the Bosch cycle film and a bulk film deposited at the same process conditions as determined by decomposition of high resolution XPS spectra. The film composition is nearly identical to that of the equivalent bulk film, indicating that the deposition DOE results can be applied to the Bosch cycle process with minimal adjustments.
Article metrics loading...
Full text loading...