strain can, under suitable conditions, 
cause adiabatic heating of metallic Ti 
fragments to their ignition temperature 
in air (16). 
That pyrophoricity is increased by 
metal brittleness would be expected. 
For example, brittle alloys are used in 
lighter flints. 
Na-K and Ti-Zr alloys are more 
pyrophoric and have lower melting 
points than the pure constituent metals. 
This suggests that metal contaminants 
that lower melting points can, in some 
cases, add to metal pyrophoricity. 
Sheet Zr has ignited while under 
nitric acid when the sheet accidentally 
touched the graphite container. This 
suggests that surface contaminant oc- 
clusions can (in part through electro- 
lytic effects) substantially alter metal 
pyrophoricity. 
Significance of Hydrogen 
The metal fire and explosion vagaries 
studied covered a wide range of tem- 
peratures, metal purities, atmospheres, 
surface-to-mass-ratios, etc. The only 
factor that appeared to be commonly 
involved in all of these incidents was 
hydrogen, i.e., all of the metals involved 
had either been made by hydride de- 
composition (and could therefore be 
suspected of retaining some hydrogen) 
or had at some time been exposed to 
water in some form. 
The evidence of the significance of 
hydrogen was often dramatic. Thus, 
U chips often ignited in moist air but 
seemed nonpyrophoric in dry ir. 
Metal spontaneously ignited (some- 
times even in the absence of air) after 
prolonged storage under moist argon or 
water. Hydrogen concentrations on 
machined U surfaces sometimes ranged 
up to 50 times the published room tem- 
perature solubility of hydrogen in U. 
Brittle, pyrophoric uranium hydride 
has been identified following U expo- 
sure to water, etc. 
While it is known that hydrogen on 
and below the metal surface is not the 
sole cause of pyrophoric anomalies, it is 
reasonably certain that it is a major 
common contributing cause. The vari- 
ous mechanisms by which hydrogen de- 
rived from water can enter and migrate 
within a metal (16, 18) are of less im- 
mediate interest than the ultimate 
effect of such hydrogen on pyrophor- 
icity and the fate of the oxygen left 
after hydrogen removal from water. 
It is thought that the principal ef- 
fects of hydrogen on metal pyrophor- 
icity inelude increasing internal stress 
(17, 18), promoting metal surface frag- 
mentation (in part through effects on 
brittleness, accentuation of grain 
boundary defects, and through subsur- 
face gas evolution under heat), and re- 
duction of surface-oxide protection of 
the metal. 
Moesel’s theory. According to a 
theory proposed by F. C. Moesel, three 
possible reactions occur: (a) metal plus 
moisture form metal hydride and hy- 
drogen peroxide; (b) metal plus oxygen, 
in the presence of moisture, form metal 
oxide and additional peroxide; (c) 
metal plus hydrogen peroxide form 
metal peroxide or at least associate 
with each other by physical adsorption 
or absorption, which makes this reac- 
tion imminent. A metal fire in the 
absence of air would be the reaction of 
metal peroxide (or metal and hydrogen 
peroxide) with adjacent metal to form 
metal oxide with a liberation of heat. 
With a sufficient buildup in metal 
peroxide concentration, this reaction 
could be explosive. In the presence of 
air, the dispersion resulting from the 
primary explosion would result in a 
secondary explosion involving the oxi- 
dation of not only the earlier formed 
metal hydride, but also of the residual 
metal in the dispersed particles. 
The general theory involved in spon- 
taneous ignition of metal powders has 
been previously discussed. However, 
a number of incidents show that moist 
metal powders can, under some condi- 
tions, literally explode (either in the 
presence or absence of air), instead of 
simply igniting and quiescently burn- 
ing. Mixed hydride and peroxide on 
the metal surface is one of several possi- 
ble high-energy sources capable of 
spontaneously and very rapidly raising 
powder temperatures to a high degree. 
In the discussion of massive-metal 
pyrophoricity theories, several mecha- 
nisms were covered relative to frag- 
mentizing of heated metal surfaces. 
The same principles are thought to be 
even more applicable to metal powders. 
Finally, the Russian experimenters 
obtained high-energy water-metal ex- 
plosions under conditions comparable 
to hot fragmentized powdered metal 
(11). Thus, both theory and experi- 
ence justify considerable precautions 
when handling or heating moist, finely 
divided powders with the degree of con- 
cern increasing with increasing quanti- 
ties and decreasing average particle 
size. 
Experience and experiments have 
shown that when water contacts molten 
pyrophoric metals, rapid generation of 
steam is, under quite narrow condi- 
tions, occasionally accompanied by 
violently explosive water-metal reac- 
tions. In the case of high-melting- 
point pyrophoric metals, experiments 
have shown that when water-metal ex- 
plosions are attained, only a small per- 
centage of the molten metal reacts with 
water. From this fact it can be as- 
sumed that the critical conditions are 
disrupted during the explosion. It is 
thought that during that very brief 
interval in which the very-high-tem- 
perature liquid metal can remain in 
contact with water, a combination of 
mechanical and exothermic chemical 
forces cause rapid spewing of finely 
divided and exceedingly hot metal par- 
ticles into water (12). The basic 
mechanism involved in liquid metal- 
water explosions is viewed as being very 
similar to, but more severe than, the 
moist-metal-powder explosions. 
BIBLIOGRAPHY 
. Nuc.eonics 14, No. 6, 24 (1956) 
. Nucueonics 14, No. 7, R12 (1956) 
. H. C. Anderson, L. H. Belz. Factors con- 
trolling the combustion of zirconium powders, 
J. Electrochem. Soc. 100, 240 (1953) 
4. W. R. DeHollander, ‘‘Spontaneous Com- 
bustion of Zirconium,.’’ Memorandum to 
D. F. Hayes (1956) (Included as Appendix C 
to Ref. 13) 
6. B. Kopleman, N. B. Compton. Spontaneous 
combustion of metal powders, Metal Progr. 
61, 77 (1953) 
6. I. Hartmann, J. Nagy, M. Jacobson, ‘‘ Explo- 
sive characteristics of titanium, zirconium, 
thorium, uranium, and their hydrides’’ 
(Bureau of Mines Report of Investigations 
4835, 1951) 
7. E. W. Cousins, P. E. Cotton. Design of 
closed vessels to withstand internal explosions, 
Chem, Eng. 58, No. 8,133 (1951) 
8. U. S. Patents Nos. 2,667,668 and 2,667,669 
(1954) 
9. ‘‘Liquids Metals Handbook, Sodium-Nak, 
Supplement”’ (Superintendent of Documents, 
Government Printing Office, Washington 25, 
D. C., 1955) 
10. J. Katz, E. Rabinowitch, ‘‘The Chemistry of 
Uranium,"’ National Nuclear Energy Series 
VIII-5 (McGraw-Hill Book Co., Inc., New 
York, 1951) 
11. Zhurnal prikladnoy khimii, 19, No. 4, p. 371 
(1946) 
12. R.B. Smith, The fire properties of uranium. 
TID-8011, (1956) 
13. R. B. Smith. Zirconium fire and explosion 
hazard evaluation. TID-5635 (1956) 
14. ‘‘Recent Titanium-Nitric Acid Explosions"’ 
(AEC Accident and Fire Prevention Informa- 
tion, September 13, 1954) 
15. ‘‘Handbook on Titanium Metal,"’ 7th ed., 
p. 26 (Titanium Metals Corp. of America, 
New York, 1953) 
16. D. 8. Gibbs, H. J. Svec, J. Am. Chem. Soc. 
76, 6,052 (1953) 
17. S. S. Sidhu. Neutron diffraction studies of 
Hafnium-Hydrogen and Titanium-Hydregen 
Systems,'’ ANL-5501 (1956) 
18. T. P. Smith, L. W. Eastwood, D. J. Carney, 
C. E. Sims, ‘Gases in Metals'' (American 
Society for Metals 1953) 
19, Explosions involving pickling of zirconium 
uranium alloys, WAPD-84 (1953) 
wth 
123 
