Electron-impact dissociation, rate

In general, the substrate temperature will remain unchanged, while pressure, power, and gas flow rates have to be adjusted so that the plasma chemistry is not affected significantly. Grill [117] conceptualizes plasma processing as two consecutive processes the formation of reactive species, and the mass transport of these species to surfaces to be processed. If the dissociation of precursor molecules can be described by a single electron collision process, the electron impact reaction rates depend only on the ratio of electric field to pressure, E/p, because the electron temperature is determined mainly by this ratio. 

High etch rates and selectivity can be achieved by judicious selection of feed gases to a plasma reactor. The atomic and radical species formed by electron impact dissociation depend largely on feed gas composition, and the intrinsic etch rates measured in the absence of a plasma (i.e., downstream etching) provide a useful indicator of chemical selectivity in the presence of a plasma. For example, the ratio of (100) silicon (34) to thermal oxide (Si02) (37) etching by F atoms is 41 1 at room temperature. As etch rates generally follow an Arrhenius type dependence on substrate temperature. 

Formulation of Equations. Discharge structure influences chemistry primarily through electron-impact dissociation and surface ion bombardment. To predict the rate of electron-impact dissociation, local electron number density and energy must be known. These quantities are obtained from equations for electron continuity and electron energy, respectively. 

The capillary plasma reactor consists of a Pyrex glass body and mounted electrodes which are not in direct contact with the gas flow in order to eliminate the influence of the cathode and anode region on CO2 decomposition. Analysis of downscaling effects on the plasma chemistry and discharge characteristics showed that the carbon dioxide conversion rate is mainly determined by electron impact dissociation and gas-phase reverse reactions in the capillary microreactor. The extremely high CO2 conversion rate was attributed to an increased current density rather than to surface reactions or an increased electric field. 

FIG. 15. Electron impact reaction rales as a function of the average electron energy. A I 1 mixture of SiHa and H2 was used, at a total pressure of 83 Pa. (a) Reaction rates for SiHa, (b) reaction rates for Si2H6 (dotted lines) and H2 (solid lines). Abbreviations are ion. ionization dis, dissociation vib, vibrational excitation att, attachment. See Table II for details and references. (Adapted from G. J. Nienhuis, Ph.D. Thesis. Universileit Utrecht. Utrecht, the Netherlands. 1998.)... 

The power dissipated at two different frequencies has been calculated for all reactions and compared with the energy loss to the walls. It is shown that at 65 MHz the fraction of power lost to the boundary decreases by a large amount compared to the situation at 13.56 MHz [224]. In contrast, the power dissipated by electron impact collision increases from nearly 47% to more than 71%, of which vibrational excitation increases by a factor of 2, dissociation increases by 45%, and ionization stays approximately the same, in agreement with the product of the ionization probability per electron, the electron density, and the ion flux, as shown before. The vibrational excitation energy thresholds (0.11 and 0.27 eV) are much smaller than the dissociation (8.3 eV) and ionization (13 eV) ones, and the vibrational excitation cross sections are large too. The reaction rate of processes with a low energy threshold therefore increases more than those with a high threshold. 

A second type of gas phase collision is that occurring between the various (heavy) species generated by electron impact reactions, as well as between these species and the unreacted gas-phase molecules (25,2d). Again, dissociation and ionization processes occur, but in addition, recombination and molecular rearrangements are prevalent. Similar rate expressions to that of Equation 2 can be written for these collisions (27). In this case, the concentration of each chemical species, along with the collision cross section, and the species energy distribution function must be known if k is to be calculated. Clearly, much of this information is presently unknown. 

We have seen that ZTRID can be successfully observed and interpreted for cluster ions. It is of interest to look as well at covalent molecular ions for new thermochemical information. The parent ion of tetraethylsilane illustrates these possibilities. The ion is formed in adequate abundance directly in the FTICR cell by electron impact, and the more abundant triethylsilyl ion is readily removed by ion ejection. Temperatures substantially above room temperature are needed to give measurable ZTRID rates. Figure 10 shows the low-pressure dissociation chemistry at 403 K. At this temperature, some water vapor outgasses in the cell and reacts with the tetraethylsilane parent ion to give the EtjSi(H20) ion, but this competing bimolecular reaction is well behaved and easily allowed for in the kinetic fitting. The parent ion undergoes the ZTRID process. 

For the conditions shown in Fig. 8, it was estimated that the electron density was 8x 10 cm" . This means that kj = 2.5 x 10" cm /s, a value in good agreement with measured rate coefficients for dissociation of small molecules by electron impact . The gas phase propagation rate coefficient kp was also found to be in very good agreement with values determined for conventional butadiene polymerization. The agreement of the adjusted parameter values with those measured independently lends further support to the validity of the proposed model of plasma polymerization. 

If, on the other hand, the excitation energy of the metastable is lower than the dissociation energy of the species AB, the latter can only be decomposed either by dissociative de-excitation of short-lived electronically excited states, or by ion-molecule reactions, or by electron impact. Much higher plasma energy values are necessary to produce a high rate of decomposition of the species AB in such systems28,44 (e.g. C/02,TiN/Cl2). 

Direct determination of non-Boltzmann distributions of the vibrational levels of the ground N2(X1S ) state has recently been performed by a new diagnostic technique, coherent anti-Stokes Raman spectroscopy21 with results consistent with calculated distributions. Kinetic data on N2 dissociation amenable to a comparison with theoretical predictions are scanty. Data from Ref.223 can however be quoted and are reported in Fig. 26. The dashed line has been calculated on the assumption that dissociation takes place via predissociated electronic states excited by direct electron impact. Observed dissociation rates are higher and a much better agreement has been claimed with dissociation rates calculated on the basis of a pure vibrational mechanism223 22b. ... 

---