Influence of entrapped gas on initiation of explosion in liquids and solids By A. Yoffe Laboratory for the Physics and Chemistry of Rubbing Solids, Department of Physical Chemistry, Cambridge ( Communicated by F. P. Bowden, F.—Received 17 January 1949) [Plate 15] Further evidence has been obtained for the view that the initiation of explosion by impact in liquids is due to the compression and adiabatic heating of trapped gas. It is shown that the sensitivity of an explosive is very dependent upon the pressure ratio. Flat impact experi ments on nitroglycerine spread as a ring on a flat anvil show that the explosion efficiency, which is high when the initial gas pressure is 1 atm., is reduced to zero when the initial gas pressure is c. 30 atm. A high explosion efficiency is still observed when the initial air pressure is less than 10~6 mm., and it is suggested that under these conditions the initiation is due to the compression of the nitroglycerine vapour itself present at a pressure of c. 10“8 mm. It is further suggested that the explosion begins in the vapour phase.
Entrapped gas and the initiation of explosions 375 could be varied between 1 and 100 atm. The final temperature T2 will depend on the initial gas pressure pv If explosion requires a definite high temperature T% to be reached, then initiation will be the more difficult the greater the initial gas pressure. The arrangement used was as follows. Nitroglycerine was spread as a ring on a flat steel anvil inside a chamber which was then filled with air or nitrogen from a gas cylinder. When the gas pressure had reached a required value, the ex plosive was hit with a flat steel hammer. A sketch of the apparatus is given in figure 2.
378 A. Yoffe Impact experiments in low pressure atmospheres of different gases When gases having a value for y lower than air are included in the gas space, there is a fall in explosion efficiency. Values have been obtained for the explosion efficiency over a fairly wide range of initial gas pressures. The nitroglycerine was spread as a ring and pure samples of the gases introduced into the apparatus described in the previous section.
Entrapped gas and the initiation of explosions 379 The adiabatic compression of saturated vapour need not always result in con densation. The condition for no liquefaction to occur is (Roberts 1943) ^uq •~T + d T <()‘^ This relation is satisfied by carbon tetrachloride but not by w-pentane or methyl nitrate. However, if one takes the value y 1*08 for w-pentane, the estimated final initial gas pressure, p1 (mm.) Figure 5. Explosion efficiency for different initial pressures of air, pentane and ethylene. Mass of striker 210 g. Height of fall 30 cm.
380 A. Yoffe The value of y appropriate to the sudden compression of a gas and defined by equation (1), is not the value of y at room temperature and pressure found from measurements of the speed of sound, etc. Even for the permanent gases, y varies with pressure and temperature (see International critical tables, first edition) and during a compression y depends on the rate of pressure rise (Lewis & von Elbe 1939). This variation has been ascribed to the failure of the kinetic energy to share itself out among all the modes of motion.
Entrapped gas and the initiation of explosions 381 explosive vapour has not been closely examined. Preliminary experiments by Chamberlain, Gray & Walsh (1947) demonstrated the ease with which methyl nitrate vapour, diluted ten times with nitrogen, ignites, and confirm the likelihood of such inflammation.