ABSTRACT: There is considerable interest in the fate of alkali elements in combustion (eg., Na or K that may be present in the fuel (‘biomass', or pulverized coal), or in the ingested air (eg., as sea-salt in a marine environment)). In the presence of excess sulfur (often present in the fuel), and oxygen (as in fuel-lean combustion) there is a strong thermodynamic tendency to form corrosive molten salts (eg.,Na2SO4(l) or K2SO4(l)), especially at cooled immersed solid surfaces (heat exchanger tubes or turbine blades) or containment surfaces (see, eg. Schofield(2005a)). Over the last 2 decades the authors have each made complementary measurements of Me2SO4 (Me = Li, K, Na, Rb, Cs) deposition from well-characterized seeded- atmospheric pressure pre-mixed laboratory gaseous flames (see, eg., Liang and Rosner(1987) and Steinberg and Schofield(1996)), and one of us has also recently reported HgSO4 deposition rates; Schofield(2005b)). Under the conditions of these particular experiments (chlorine-free) the dominant metal-carrier molecules are the metal hydroxides (MeOH(g) and/or the metal atoms themselves (Me(g))---ie., the indicated sulfates measured are therefore formed by heterogeneous reactions.

The present study initiates an overdue series of theoretical/computational investigations into the mechanism and rates of deposition of trace metals (ca. 25 ppmv, with 75 ppmv SO2)) from propane- and hydrogen-fueled flames based on the abovementioned recent data-sets. In particular, we inquire here if the relative rates of metal (sulfate) deposition reported in Schofield(2005b); Fig. 1 for Me = Li, K, Na, Rb, Cs and Hg at Tw/T(gas) = 0.256 can be quantitatively understood in terms of the relative rates of species transport to the cooled target surface. The principal mechanisms of species transport considered are convection, Fickian diffusion, and Soret diffusion in the prevailing local temperature gradients. Fortunately, adequate estimates of the required molecular species Fick diffusivities and Soret factors (Rosner, et al.(2000)) can be made using Chapman-Enskog ideal gas kinetic theory. Moreover, at the prevailing trace metal concentration levels and boundary layer residence times it is reasonable in the present cases to first invoke/examine the success of ‘chemically-frozen' (laminar) boundary layer (CFBL) theory (Rosner, et al.(1979)) with local thermochemical equilibrium imposed at the gas/condensate interface. Our preliminary results indicate that, provided one includes the augmenting effects of Soret transport (especially for the heavier elements Me = Rb, Cs and Hg), reasonable estimates can be made not only of the relative deposition rates shown in Schofield (2005b, Fig. 1), but also their absolute magnitudes (ca. 20 micro-moles/m2s). This encourages a closer investigation of deposition rate data obtained at higher surface temperatures, including those close to the threshold surface temperatures above which such deposits were not observed. Prepared for presentation at AIChE Annual Meeting, Nov 2006, San Francisco, CA. Liang, B. and Rosner, D.E.,(1987) AIChE J, vol 33(12) 1937-1948 Rosner, D.E. , Chen, B.K., Fryburg G.C. and. Kohl F.J. , (1979) Comb. Sci. Tech. 20, 87-106 Rosner, D.E.(2000) , Transport Processes in Chemically Reacting Flow Systems, DOVER Publications, N Y Rosner, D.E., Israel, R.A. and La Mantia, B., Comb. & Flame, vol. 123, 547-560(2000) Schofield, K.(2005a), Energy and Fuels (ACS) vol 19, 1898-1905 Schofield, K.(2005b), Proc. Comb Inst.(Elsevier) vol 30, 1263-1271 Schofield, K.(2004), Chem Phys Letters(Elsevier), vol 386, 65-69 Steinberg, M. and Schofield, K. (1996), Proc. 26th Sympos. (Int) on Combustion, Comb Inst. (Pittsburgh PA) 1835-1843