There is now convincing theoretical and experimental evidence, assembled and discussed here, for the remarkable insensitivity of the orientation-averaged thermophoretic properties of aggregated particles to aggregate size and structure (morphology), as well as the nature of gas/surface scattering. Indeed, theoretical consideration of straight chains, and uniformly “packed” quasi-spherical agglomerates, as well as recent experimental data on soot aggregate transport in/from laminar flames at atmospheric pressure, indicates that the orientation-averaged thermophoretic diffusivity, (α_TD)_n, of an aggregate containing N primary particles is usually within about 8% of the (α_TD)_i,-value for a single “primary” sphere in the free-molecule regime and within about 21 % in the continuum limit. Among other things, this implies that, especially in the free-molecule regime, thermophoretically-dominated transport rates can be adequately predicted without a detailed knowledge of the size and morphology (-distribution) of the aggregated particles or the nature of gas/particle surface scattering, which is definitely not the case for particle transport by Brownian diffusion or inertial drift (see, e.g., Rosner, 1991). This result also implies that thermophoretic particle sampling from “low pressure” flames (Dobbins and Megaridis, 1987) does not itself introduce a significant bias in the relative populations of various sampled aggregate sizes and morphologies. As a corollary, local gas temperature and sool volume fraction estimates based on mass transfer rates- or thermocouple response-methods (Eisner and Rosner, 1985, 1986) will be negligibly influenced by the inevitably uncertain sizes and morphologies of the prevailing soot agglomerates. Since the optical properties (e.g., effective cross-sections for light-scattering and extinction) of soot aggregates are now known to be size- and structure-.?™v»'w (Mackowski. 1987, 1988, Dobbins, Santoro, and Semerjian, 1991. Dobbins and Megaridis, 1991)) we anticipate that the drag vs. thermal force “compensation” effects that produce (α_TD)-values insensitive to aggregate si:e (N), and structure, and the nature of gas molecide/surface scattering will find important R&D applications for many systems in which agglomeration occurs. It is also concluded that thermophoretic means would not be useful to rapidly separate various asymmetric particle morphologies unless orientation-averaging is suppressed, perhaps using external fields (E, B).

Dispersion relations and threshold values for the onset of buoyancy–thermocapillary instability are given for the case of prescribed heat flux on the boundaries of a liquid layer heated from below when the upper boundary is a deformable surface open to the ambient air. The nonlinear evolution equations of this free surface under various circumstances are also provided (without and with buoyancy, for microgravity or standard ground conditions).