Isoconversional Methodology - Isoconversional Kinetics of Thermally Stimulated Processes - page 30

E
RT

α

ln() β = Const1052 − . ,

(2.11)

i

α , i

where E ʱ is estimated as a slope of the linear plot ln( ʲ i ) against 1/ T ʱ,i . Repeating
the procedure for a set of different ʱ 's gives rise to a dependence of E ʱ on ʱ . Note
that the integral methods can be also applied to the differential type of data that
would have to be integrated numerically. Unlike numerical differentiation, integra-
tion does not amplify experimental noise that makes integral methods well suitable
of either type of data.
It does not appear accidental that all three methods (i.e., Friedman, Ozawa, and
Flynn and Wall) were proposed by workers who studied decomposition of complex
polymeric materials, i.e., the processes for which finding an adequate kinetic model
is more than challenging. It should be noted that the first applications of the meth-
ods immediately brought to light the issue of variable activation energy. Friedman
[ 13 ] observed a variation in E ʱ (Fig. 2.3 ) for decomposition of cured phenolic resin.
Ozawa [ 14 ] found that E ʱ varied for decomposition of both Nylon 6 and CaC 2 O 4
(Fig. 2.4 ). By using simulated data for competing and independent parallel reac-
tions, Flynn [ 16 ] linked a variation in E ʱ to the activation energies of the individual
steps. This link has been explored systematically by Elder [ 20 - 23 ] and Dowdy [ 24 ,
25 ] for competing or independent reactions and by Vyazovkin for reactions compli-
cated by diffusion [ 26 ], as well as for consecutive [ 27 ] and reversible reactions [ 28 ].
The studies have concluded that analysis of the E ʱ can be used for obtaining some

Fig. 2.3 The activation ener-
gies determined by Friedman
for the thermal degradation of
phenolic plastic. (Reproduced
from Friedman [ 13 ] with
permission of Wiley)