The kinetics of wood degradation and its respective components have been and still are obtained by measuring the rate of weight loss of the sample as a function of time and temperature. The most common technique for this investigation is thermogravimetric analysis (TGA). TGA involves continuous weighing and recording of data obtained from a sample, heated at either constant temperature or a fixed heating rate, enclosed in a furnace (e. g. 1,2, 3, 4, 5, 6, 9, 13, 15, 18, 20, 40, 41, 42, 43, 44, 45, 46, 55]. These experiments are normally carried out under vacuum or in a nitrogen atmosphere at both low temperatures and heating rates. Some workers have used steam as the gaseous environment in their experimental system [45, 46, 56]. As there is no provision for the collection of volatile pyrolysis products, the TGA data are normally used to derive overall kinetic expressions.

The global thermal degradation process can be described by a simplistic reaction scheme as shown in equation M.2.

biomass —> char + volatiles M.2

and the rate of the above reaction is then described in the form of a first order Arrhenius type rate law as shown in equation M.3.

dW = — A exp (-E ) (W — Щ) M.3 dt RT .

where: W: residual weight fraction,

Wf: final weight,

A: pre-exponential (frequency) factor,

E: activation energy,

R: Universal gas constant.

T: reactor temperature

If the sample is heated at a constant rate M, then dW = — A M.4

dt M exp (- E) (W — Wf) where: M = dT

RT dt

Sometimes the weight term is replaced by a density term. It is assumed here that no shrinkage occurs during char formation. Table 2.2 shows some selected kinetic parameters for overall reaction rate expressions.

The Arrhenius kinetic parameters, the activation energy (E) and the pre-exponential (frequency) factor (A) are derived by obtaining best fit curves through the experimental data and solving the Arrhenius rate law, using a least squares method. A comparison of the different kinetic parameter estimation methods is made by Vovelle et al. which highlights the variability in the values obtained (57). Most of these values have been obtained from weight loss data.

Kinetic parameters estimated by the researchers given in Table 2.2 show a wide variability in the values, even with similar feedstocks. Some of the variability in the parameters may be accounted for by the neglect of temperature variation of the sample during heat up and the use of the steady state temperature as the overall reaction temperature. Some of the kinetic modelling has also be performed with large biomass samples where the effects of mass and heat transfer cannot be neglected. This is evidenced by Salazar who gives different value for the pre­exponential factor and the activation energy for two differing cylinders of eucalyptus (68). Other researchers have used two or more consecutive steps of zero and first order reactions to describe the pyrolytic degradation of the biomass (13, 58, 59, 60, 61,62,63,64).

Bilbao et al. have taken into account the influence of heating rate on the kinetic parameters and have therefore studied a range of heating rates from 1.25 C°/min to 80C7min and the effects on the reaction order (2, 3, 4, 5, 6). Other discrepancies arise from too simplistic modelling and the presence of impurities or ash which may influence the decomposition kinetics. Varhegyi et al. have investigated the effect of NaCI, FeS04 and ZnCl2 on the pyrolysis of Avicel cellulose under different

conditions with four different modelling approaches (55). To predict the product distribution, stepwise models should be applied.