Estimation of Flammability Limits of Selected Fluorocarbons with F(sub 2) and CIF(sub3) Page: 44 of 78

                                
available, similar experimental data are listed. Most of the experimental values, however, are
from binary mixtures, so there is no guarantee that they would be the minimum concentration in
a ternary system. When N2 is the diluent, the shape of the flammability envelopes typically is
such that the minimum fuel or oxidizer concentrations lie in the vicinity of dilute mixtures.
When the higher heat capacity UF6 is the diluent, however, the minimum concentrations usually
occur on the no-diluent axis.
Table 3. Minimum flammable concentrations
Oxidizer                         Fuel
Expt     Det.   Equilibr.       Expt     Det.  Equilibr..
dataa   model    model          dataa   Model   Model
F2   CFC-114      N2     11.0%     5.2%    7.3%           6.0%     3.4%    4.1%
F2     c-C4F8     N2     22.9%     5.1%    7.8%           7.0%     1.7%    2.1%
F2     C4F1o      N2               6.7%    8.9%                    2.3%    2.9%
CIF3  CFC-114      N2               3.6%    6.5%                    4.0%    6.2%
CIF3    c-C4F8     N2              7.4%     7.1%          11.0%     2.3%    3.0%
CIF3    C4F1o      N2              11.1%    8.5%                    3.5%    4.5%
F2   CFC-114     UFs     17.5%    11.9%   11.8%           8.0%     5.5%    4.2%
F2     c-C4F8    UFs     22.9%    14.6%   16.3%           7.0%     2.8%    2.1%
F2     C4F10     UFs             24.6%    17.4%                    4.2%    2.9%
CIF3  CFC-114     UFs               9.5%   10.0%                   13.9%   11.7%
CIF3    c-C4F8    UFs             20.2%    13.9%          11.0%     6.9%    6.3%
CIF3    C4F,o     UF6              33.5%   16.8%                   10.1%    8.9%
a Experimental data in bold italic are from binary (fuel-oxidizer) mixtures only and are directly comparable to minimum flammable
concentrations only if these occur at the 0% diluent boundary (more likely true of UF,-diluted systems than N2-diluted systems).
3.3 DISCUSSION
To reiterate, flammability limit predictions are purely empirical. No real theoretical justification
exists for either prediction scheme, though some precedent has been established for use of
similar empirical schemes in other chemical systems [10, 11]. The models used here are based
on correlations and calibrated to the data set shown, so a certain degree of agreement should be
expected for the chemical systems used for calibration. The reliability of predictions made for
other similar systems is derived solely from the range of systems for which the predictions
appear reasonable and on the chemical similarity of the systems in question. Other systems for
which we would expect good agreement would be those involving the same fuels and oxidizers
but a different diluent, or a mixture of diluents. Similarly, a mixture of fuels or oxidizers from
the current set would probably give reliable predictions, but some modification of the models
would be needed to handle such mixtures. Similar compounds - for example CFC-115, iso-
C4Fo, C5F12, or c-C4F80 (if one could find appropriate thermodynamic data) - would also likely
yield flammability boundary predictions that are about as accurate as those presented here. The
more dissimilar the compounds and chemistry become, however, the less confidence one would

31