Letters to the Editors alcohols in a bubble column using various spargers. In some cases and when porous sintered plates were employed Kiinig et al. observed extremely low dispersion coefficients up to gas velocities of about 6cms-'.
The findings were explained qualitatively with the variation of the bubble diameter. If the bubbles are of intermediate size, i.e. l-3mm in diameter homogeneous bubbly flow prevails. In this flow regime, the bubbles are assumed to be of very uniform size and rise velocity and dispersion in small.
As the SRC reactor was operated at low gas velocity it might be. possible that such a special flow pattern with extraordinary small dispersion as described by KGnig el al. may occur in the SRC unit. However, another phenomenon should also be taken into account as a possibility to explain the discrepancy reported by Panvelker et al. This is the fact that the tracer compound used is not inert but undergoes chemical reaction. It is well known that halogenated condensed aromates can easily be dehalogenated. For instance, 9bromophenanthrene can be hydrodebromated at room temperature and atmospheric hydrogen pressure in the presence of a nickel catalyst [7]. Though a nickel catalyst might not be present in the liquefaction unit it can be taken for granted that under SRC operating conditions the intrinsic tracer Br-82 is no more present in the form of bromophenanthrene but is deliberated therefrom as HBr. Desorption of HBr into the gas phase can not be excluded. Hence, the radioactive tracer may also appear in the gas phase. The measured response curves are therefore a result of various phenomena as, for instance, chemical reaction, interphase mass transfer, dispersion of gas and liquid phase. It is obvious that misleading conclusions will be obtained if the above phenomena are neglected. In particular, the response curves measured by Panvelker et al. must not be interpreted as a result of liquid phase dispersion alone.