: In chapter 2, a relationship holding on an adiabatic self-heating process of zero-order is derived. When a material heats spontaneously at T_s K under zero-order adiabatic condition, Eq.1 holds generally. If we assume that the heating rate dT_s/dt remains effectively constant within a minute temperature range near T_s, Eq.1 can be integrated to yield Eq.2, where Δt min is the time for sample temperature T_s to rise by a given minute temperature increment ΔT_s deg from the initial temperature T_s, and c cal/g·deg, ρ g/cm³ and ΔH cal/mole are the specific heat, density and molar heat of reaction of the material, respectively. Thus both the apparent activation energy E and the apparent frequency factor A of the very slow self-heating reaction can be determined from the gradient (Eq.3) and intercept (Eq.4) of the experimental linear plot of log Δt against the reciprocal of T_s.
    In chapter 3, a relationship between Eq.2 and Frank-Kamenetskii's critical condition for thermal explosion (Eq.5) is discussed, where δ_c (dimensionless), λ cal/cm·min·deg and γ cm are critical parameter for explosion, thermal conductivity and the radius of the sample package, respectively. Eq.2 is concerned with the adiabatic case, whereas Eq.5 the nonisothermal. So far as the heating process is concerned, however, both equations deal with the similar slow heating process in the subcritical state prior to ignition. Thus an assumption may be thought legitimate that ΔH, E and A appearing in Eq.5 are equivalent to those of the heating decomposition reaction under adiabatic conditions. Eq.5 can be rewritten to give Eq.6. Substitution of Eq.3 and 4 into Eq.6 leads to Eq.7. As thermal conductivity λ is expressed in terms of thermal diffusivity a as in Eq.8, finally we obtain Eq.9. It is worth while to mention that the actual value of ΔH, E, A, λ and cρ of the material is not necessary, so far as the calculation of the critical ignition temperature T_c is concerned. It is only necessary to measure α for calculation of T_c, except for recording the adiabatic self-heating behavior at the initial slow heating stages of the material.
    In chapter 4, T_c of BPO in a commercial package of 10 kg is calculated. The adiabatic heating processes of BPO powder were recorded at temperatures ranging from 75 to 85 °C. Then the logarithm of the time Δt for T_s to rise by 1.25 deg from each T_s was plotted against the reciprocal of T_s according to Eq.2. The coefficients a and b were determined by the least-squares method to be 1.3739 × 10⁴ and -36.3477, respectively. From these values E and A were obtained as follows : E = 62.87 kcal/mole, A = 8.75 × 10³⁰ mole/cm³ ·min. This E value is near to an E value of 58 kcal/mole, which was reported on BPO based on the induction period⁶⁾. The shape of the package was assumed to be an infinite cylinder (δ_c = 2) of 26 cm diameter, according to the dimension of it (27 cm × 27 cm × 40 cm). Thermal diffusivity of BPO at ρ = 0.5 g/cm³ was measured by the Xe flash method to be 5.76 × 10^-2 cm² /min⁵⁾. Molar heat of reaction of BPO was taken from Fine's data⁸⁾. On substituting these values into Eq.9, T_c is determined to be 79 °C. This value is fairly near to 82 °C, which is the value of SADT on solid BPO⁷⁾.
    In chapter 5, a self-ignition accident of magnetite (Fe3O4 ) and its heating behavior are described. The weathering effect which is characteristic of most solid materials having oxidative heating properties was observed in this material (Fig.2). Thus, Fe3O4 granules show a vigorous tendency to heat due to oxidation immediately after it was taken out of the container ; it becomes, however, less active to oxidation, the longer it is exposed to air. Similarly the heating profile is also steeper at the initial stages in each run, but it shows a tendency to saturate gradually at the latter stages. This effect was verified by X-ray micro-diffractometry to be attributed to a fact that oxidation occurs predominantly near the surface of granules.
    In chapter 6, self-ignition accidents of dry spray mist of some oxidatively polymerizing paint and its heating behavior are described. Accidents of this type were experienced often in winter rather than in summer, contrary to common sense. Hydroperoxide (hpo) is probably present in the resin substrate, as the dry mist was found to heat even in nitrogen atmosphere (Fig.3 and 4). Linoleic acid/cotton wool mixture which had reserved for 6 days in a refrigerator at 7 °C showed a more vigorous heating tendency and a greater peroxide value than that reserved in an oven at 35 °C, other conditions being equal (Fig.6). Thus such situations were revealed that in winter hpo is, though slowly, formed in the resin substrate while the mist sticks to the wall, but the hpo is not consumed for polymerization due to lower ambient temperatures and is gradually accumulated in the dry mist ; whereas in summer it is actively consumed for polymerization, so that the paint mist gets to lose the tendency to heat spontaneously.
    In chapter 7, a comparative study of the adiabatic oxidative heating, rates of a series of unsaturated fatty acids is described. Times for T_s to rise by 2.5 deg from T_s at temperatures from 40 to 50 °C were measured for arachidonic acid (A), linolenic acid (B) and linoleic acid (C) in the form of mixture with cotton wool, respectively. The order of heating rate was A > B > C, in the ratio of 9.1 : 3.2 : 1 (Table 2 and Fig.8), However, Holman et al. reported that the relative rates of oxidation based on weight gain at 37 °C of acid esters mentioned above were 4.8 : 2.4 : 1, respectively⁹⁾. This apparent inconsistency is probably due to the reason that the heating process proceeds on a mechanism which is not necessarily parallel with up-take of oxygen.
    In conclusion the apparatus was found to be effective, not only for recording the profile of the adiabatic self-heating process, but also for estimating T_c of various chemicals.