FIG. 4. 
posal operation in May 1956. After two 
men had been killed and a third lost an 
arm while attempting to re-crate scrap Zr 
for removal, it was decided to burn re- 
maining scrap cautiously in outdoor storage 
Flare-up of Zr scrap during dis- 
area 
components, and the initial pressure. 
With initial mixtures at optimum pro- 
portions and atmospheric pressure, 
maximum explosion pressures are 
roughly 0-150 psig. 
Effect of particle size. Qualita- 
tively, the ease of ignition and explosion 
pressures attainable with metal dusts 
decrease rapidly with increasing parti- 
cle size. The maximum particle size 
that will just support combustion fol- 
lowing dispersion in air has not been 
adequately investigated. The amount 
of energy required to initiate-explosion 
of fine metal powders dispersed in air is 
often surprisingly low (6). In some 
cases the energy involved in dispersing 
the powder is sufficient to initiate spon- 
taneous explosions. 
Oxides Reduce Pyrophoricity 
Although pyrophoric metals burn in 
oxygen, the pyrophoricity of many 
metals, e.g., U (8), can often be dra- 
matically reduced using oxygen. As 
an example, finely divided iron powder, 
prepared by hydrogen reduction of iron 
formate, will often ignite when first 
exposed to air. If, however, after pre- 
paring such powders, the hydrogen is 
replaced with an inert gas into which 
small amounts of dry oxygen are slowly 
fed while the furnace is being cooled to 
room temperatures, an iron powder is 
obtained that may be handled in air 
without ignition. A similar commer- 
cial process reduces pyrophoricity of 
highly porous Ti and Zr sponge pro- 
duced by the Kroll process. 
The ability of oxide coatings to re- 
duce apparent metal pyrophoricity 
varies considerably among different 
122 
metals and even among specimens of 
the same metal. Both the physical 
and chemical properties of oxide coat- 
ings may change markedly with in- 
creasing temperatures. Some metals 
such as Ti and Zr are capable of dis- 
solving large quantities of surface 
oxides when heated above phase trans- 
formation temperatures, while other 
metals such as ‘Al and U do not absorb 
significant quantities of surface oxides 
even when the metal is heated to its 
melting point. 
Several investigators, while admit- 
ting that it does not make sense from 
a thermodynamic viewpoint, have 
found by test that Zr powders contain- 
ing up to 25% oxide are more readily 
ignited than similar powders relatively 
free of oxides. The role played by sur- 
face oxides in the metal pyrophoricity 
picture is obviously important, com- 
plex, often erratic, and confusing. 
Much information is available in the 
chemical literature on the oxidation of 
metals by oxygen, air, and water. 
Much is also known about the effects 
of oxide films on metal oxidation rates. 
The amount of heat liberated from 
combustion of specific metals can be 
accurately predicted, as can many fac- 
tors influencing the probability of a 
reaction (such as temperature, pres- 
sure, etc.). 
Despite this profusion of informa- 
tion, very little is known about the 
specific sequence of physical and chemi- 
cal steps occurring during metal burn- 
ing. Even less is known about reac- 
tion-rate variables during intermediate 
steps preceding and during metal com- 
bustion. Thisis particularly true when 
moisture and metal contaminants are 
involved. 
The study of many unusual pyro- 
phoric-metal incidents (12, 13), indi- 
cates apparently several independent 
mechanisms may contribute to metal 
pyrophoricity, and that surface oxides, 
stress, moisture, and contaminants are 
major facets deserving evaluation. 
Moisture Increases Pyrophoricity 
Metal pyrophoricity is strongly in- 
fluenced by moisture. Metals such as 
Na and K can react violently with 
water (9), as is well known. In the 
case of U, the high-temperature reac- 
tion with steam is much more violent 
than with oxygen (10). 
Studies have indicated that the reac- 
tion between some metals (e.g., pow- 
dered Al or Mg) and water is theoreti- 
cally capable of producing slightly more 
explosive energy than nitroglycerin or 
TNT. Russian scientists reported (11) 
attainment of close-to-theoretical 
energy releases during tests aimed at 
preliminary evaluation of the use 
of water-metal reactions as high- 
explosives. 
Many methods are known for induc- 
ing water-metal explosions (12). A few 
violent water-metal explosions have 
accidentally occurred in industry (nota- 
bly during are melting of Ti in water- 
cooled crucibles), However, the rarity 
of such occurrences suggests that explo- 
sions of this type are only attainable 
under a very narrow set of conditions. 
That they can occur at all, however, 
justifies current AEC research in this 
area, 
Two recent reports (12, 18) provide 
strong circumstantial evidence that 
pyrophoricity of U and Zr can some- 
times be very materially increased by 
prior quiescent exposure to moisture at 
room temperature. Research is cur- 
rently under way to determine the 
mechanism by which metal surfaces are 
pyrophorically influenced by water. 
Stress Increases Pyrophoricity 
There exists fragmentary but im- 
pressive evidence that pyrophoricity is 
increased when the metal is under 
stress. This stress can be physical or 
chemical. Uranium lathe chips, for 
example, appear to be very much more 
susceptible to spontaneous ignition 
than annealed metal of the same dimen- 
sions in sheet form. Several incidents 
(14) suggest a connection between 
stress and metal-surface explosions in 
nitric acid. Sudden release of elastic 
Combustion of Metals: Solids vs 
Powders 
Powder-to- 
solid ratio of 
Property compared property* 
Combustion surface area 13,300 
Metal mass involved 0.524 
Total-surface: total-mass 
ratio 25,400 
Fraction of total surface 
area exposed to external 
surroundings 0.000118 
* Based on cubic packing of spherical par- 
ticles lu in dia. 
