Surface Loss Probabilities

The surface loss probability of a species of interest can be determined using the cavity technique as described in Sect. 11.3.1. So far, cavity probes have been applied in low-temperature plasma experiments in the laboratory and in the fusion experiments JET and ASDEX Upgrade. 

Another consideration concerns the profiles obtained for the different precursor gases. We expect the same radicals to be present in the plasmas of different gases. For example, C2H5 will certainly be produced in an ethane plasma by simple dissociation, but it will also be produced in a methane discharge due to various gas phase reactions. Therefore, it seems reasonable to model the profiles from all cavity positions and for all precursor gases with superpositions of one common set of single-/ profiles. The question that arises 

The set of values of ft leading to the best overall agreement is ft = 0.8, ft2 = 0.35, and / 3 10 2. For the calculation ft2 10 3 was chosen, nevertheless the normalized profiles for all f3 10-2 cannot be distinguished. Furthermore, one exception was made for the C2H2 profiles ft[ = 0.9 was chosen because it improves the fit considerably in contrast to all other profiles, where ft = 0.8 leads to a better fit [38]. 

The question arises why there are only three discrete values of ft although there are much more than just three radical species anticipated in the plasmas. The answer is twofold  

Two cavity probes were installed in the divertor of JET in the period from August f999 till November 2001. One was installed below the septum and the other in the inner module of the divertor. The analysis of these cavities was published by M. Mayer in 2003 [51]. Thick deposits of up to 14 and 45 pm were found on the outside of the inner module and septum cavity, respectively. The corresponding maximum thicknesses inside the cavity at the position opposite to the entrance slit are 15 and 10 pm. Ion-beam analysis of the layers showed that they consist mainly of D and C with a D/C ratio of about 1. Interestingly, the morphology and density of the two samples is very different. While the layer from the inner module is smooth with a density typical of a C H layers, the one from the septum is very porous and has about half the density only. The reason for this obvious difference remains unclear. A fit of the deposition profiles in the cavities leads to the conclusion that the layers are mainly deposited from species with a high surface loss probability (fH 0.9) and a minor contribution of species with a low surface loss probability (f3 10-2). New cavity probes were installed inside the JET divertor in November 2001. They are planed to be removed in 2004. 

---

A review is given on the physical and chemical reactions that occur if atomic hydrogen, hydrocarbon radicals, and low-energy ions interact with carbonaceous surfaces. In a first set of experiments the surface loss probabilities of different hydrocarbon radicals are determined in low-temperature plasmas using the cavity technique. The following values were determined / (C2H) = 0.90 0.05, / (C2H3) = 0.35 0.15, and / (CH3, C2H5) < KT2. 

This article is organized as follows. In the next section, the properties of a C H layers are shortly summarized. The following section presents the experimental methods and set-ups applied in our experiments. Section 11.4.1 summarizes the knowledge about surface loss probabilities of different hydrocarbon radicals. The remainder of the article is dedicated to a review of the results from our particle-beam experiment MAJESTIX. 

With respect to deposition processes the most important of the mentioned quantities is the sticking coefficient s. However, its determination requires a rather high experimental effort. An experiment to measure s of selected species will be described in the next section. To get easier access to a figure of merit for the overall reactivity of a radical with surfaces often the surface loss probability... 

An elegant yet technically simple method to determine surface loss probabilities is the cavity technique [30,36-38] A cavity with a small entrance slit (see Fig. 11.2) or a different well-defined geometry - [39-41] is exposed to a flux of reactive species. The transport of the particles is studied via the cross-sectional film thickness profiles. The dimensions of the geometry are chosen much smaller than the mean free path of the neutral radical species so that gas phase collisions are negligible. Then the normalized profiles depend on the surface loss probability (3 only. If the total fluence of particles into the slit is not known no conclusions can be drawn concerning the sticking coefficient except s < / . 

Fig. 11.2. Principle of a cavity probe Reactive particles enter the cavity through the slit and deposit films on the inner surfaces. Their transport and thus the resulting film thickness profiles are determined by the surface loss probability...

The value of / is derived via comparison with model calculations. Model calculations can be carried out with Monte Carlo methods The particles start at random positions outside the cavity with a randomly chosen direction. The angular distribution of the particle directions does, however, not necessarily have to be uniform. Each particle is followed if it enters the slit and as long as it is inside the cavity. Upon each wall collision a fraction s of the particle sticks to the wall and a fraction r = 1 — / is re-emitted with a cosine distribution with respect to the surface normal. When only a negligible part of the particle is left, e. g. 10-3, the next particle is started outside the cavity. In general, the trajectories of more than 106 particles have to be calculated to reach good statistics. For convenience, s = / is chosen in the calculation (this is equivalent to 7 = 0). As said before, the normalized profiles depend on the surface loss probability (3 only, so that this choice has no influence on the profile. 

It was emphasized in the previous sections that the cavity probe technique is sensitive to the surface loss probability of radicals rather than their sticking probability. The sticking probability is only accessible by experiments which allow to directly compare the rate of sticking to the flux of incoming species. The ability of the MAJESTIX experiment to perform such a task was described in Sect. 11.3.2. As a first example, we shall review our results on the sticking coefficient of methyl radicals on a-C H surfaces. 

A. von Keudell, C. Hopf, T. Schwarz-Selinger, W. Jacob Surface loss probabilities of hydrocarbon radicals on amorphous hydrogenated carbon film surfaces Consequences for the formation of re-deposited layers in fusion experiments. Nuclear Fusion 39, 1451 (1999)... 

C. Hopf, K. Letourneur, W. Jacob, T. Schwarz-Selinger, A. von Keudefi Surface loss probabilities of the dominant neutral precursors for film growth in methane and acetylene discharges. Appl. Phys. Lett. 74, 3800 (1999)... 

If the amount of metal removal by erosion is significant the surface will probably be continually active. Metal loss will be the additive effect of erosion and active corrosion. Sometimes the erosion rate is higher than that of active corrosion. The material selection judgment can then disregard coirosion and proceed on the basis of erosion resistance provided the corrosion rates of aetive surfaces of the alloys considered are not much different. As an example of magnitudes, a good high-chromium iron may lose metal from erosion only a tenth as fast as do the usual stainless steels. 

Heterogeneous uptake on surfaces has also been documented for various free radicals (DeMore et al., 1994). Table 3 shows values of the gas/surface reaction probabilities (y) of the species assumed to undergo loss to aerosol surface in the model. Only the species where a reaction probability has been measured at a reasonable boundary layer temperature (i.e. >273 K) and on a suitable surface for the marine boundary layer (NaCl(s) or liquid water) have been included. Unless stated otherwise, values for uptake onto NaCl(s), the most likely aerosol surface in the MBL (Gras and Ayers, 1983), have been used. Where reaction probabilities are unavailable mass accommodation coefficients (a) have been used instead. The experimental values of the reaction probability are expected to be smaller than or equal to the mass accommodation coefficients because a is just the probability that a molecule is taken up on the particle surface, while y takes into account the uptake, the gas phase diffusion and the reaction with other species in the particle (Ravishankara, 1997). 

This surface reaction would involve a change in the cyclopentadienyl hapaticity prior to SiO-H activation by the metallic species. Loss of cyclopentadiene by reductive elimination would then occur to provide an allyl palladium(ll) surface species, probably stabilized by a silanol group, in which the oxygen atom acts as a 2e donating ligand. However, when the temperature is raised significant carbon contamination has been evidenced by TPD and TPO experiments. These results are consistent with the absence of further SiO-H activation to eliminate propene [57]. 

If this explanation is correct, then one also has to consider the subsequent slow decline in activity (figure 2). One possible explanation involves the deposition of more vanadium, slowly decreasing the amount of molybdenum at the surface. More probably, the slow decline reflects slow loss of surface area by infilling of pores in the deposit, rather than in the catalyst. The initial deposition of carbon should create a porous deposit on the catalyst via the deposition of large crystallites of coke and of metal sulphides [35]. Slow infilling of these pores via metals deposition should lead to slow loss of surface area and of activity. 

Direct photochemical degradation of 1,1,1-trichloroethane in the troposphere is not expected to be an important fate process, because there is no chromophore for absorption of ultraviolet light (>290 nm) found in sunlight at tropospheric altitudes (Hubrich and Stuhl 1980 VanLaethem-Meuree et al. 1979). A laboratory experiment performed in sealed Pyrex ampules showed loss of 1,1,1-trichloroethane in 2 weeks under the influence of sunlight however, catalysis by the Pyrex surface was probably responsible for the enhanced reactivity (Buchardt and Manscher 1978). 

Once the electron does arrive near the emitting surface, the probability of its actual escape into the vacuum is the final limitation on yield. Any barrier present at the surface is the primary impediment, but we have partially considered this in describing and in implying that after generation the electron must be transported to the surface without loss of energy below xh( for metals or... 

---