ever, it has occurred in air at room tem- 
peratures in the case of Pu, U, and Th 
with initial metal temperatures ranging 
from that of dry ice to red heat. Wit- 
nesses generally agree that these inci- 
dents are accompanied by slow but 
noticeable- flaking off of the metal sur- 
face with sporadic emanation of local- 
ized sparking. 
From these observations it is appar- 
ent that combustion of abnormally 
pyrophoric massive metals includes a 
“built-in” mechanism for converting 
a solid metal surface into an extremely 
fine, brittle, and chemically reactive 
powder, and that this mechanism is 
accelerated during auto-oxidation of 
the powder. Since ignition of massive 
metal has occurred at quite low tem- 
peratures, combustion in air of the sur- 
face powder must involve a highly 
exothermic reaction that more than 
offsets the heat loss from the metal sur- 
face. Thus, a basic mechanism con- 
tributing to ultimate burning of mas- 
sive metal must involve both physical 
aspects (e.g., stress, brittleness, cracks, 
voids, etc.) and chemical aspects 
(initiating and secondary combustion). 
Powder Combustion 
The large difference in observed pyro- 
phoricity of metal powders and massive 
metal proceeds from differences in sec- 
ondary factors influencing combustion 
efficiency, rather than from any basic 
difference in the reaction mechanisms 
involved. To illustrate the magnitude 
of these secondary ‘‘efficiency”’ factors, 
identical volumes of solid metal and 
powder are compared in the table 
on p. 96. 
The table shows that metal powders 
possess a much greater surface area on 
which to generate heat, have a much 
smaller mass of metal in which to dissi- 
pate this heat, and have a much smaller 
percentage of total surface area avail- 
able for loss of heat by radiation to ex- 
ternal surroundings. All of these fac- 
tors combine to increase the probability 
of metal combustion. 
Effect of mass. Unlike solid metal 
(in which heat can be internally trans- 
ferred by the relatively efficient process 
of direct heat conduction), escape of 
any heat from within metal powders 
must largely take place by radiation 
through multiple insulating layers of 
air. Thus, accumulation of heat gen- 
erated within metal powders by slow 
oxidation becomes increasingly prob- 
able with increasing size of the powder 
120 
Unusual Metal Fires and Explosions 
The following unusual metal fires or explosions that occurred in the U. S. atomic 
energy program have for the most part been taken from a much more comprehensive 
list of such incidents (/2, /3). 
It should be noted that such incidents have been relatively infrequent in the large- 
scale handling of pyrophoric metals over a period of years in the atomic energy 
rogram. 
y In reviewing such incidents, it is of interest to note that increased pyrophoricity 
can apparently be “built into” a metal at the time of its initial manufacture or may 
be subsequently acquired as a surface phenomenon. 
Zirconium Incidents 
4 Up to May 1955 no serious fires had been encountered during storage of scrap-Zr 
turnings, chips, plates, rods, etc. Such scrap had been stored (pending contemplated 
future recovery) in segregated open-top bins as shown in Figs. 1 and 2. Several 
days after a heavy rain, a fire of unknown origin took place in one of the bins with 
flames extending 100 ft into the air. Shortly afterwards, contents of other (but not 
necessarily adjoining) bins suddenly and intermittently flared up. Material in all bins 
soon became involved and 159,000 Ib of Zr were consumed. The heat was sufficiently 
intense to crack windows and ignite wood located over 150 feet away. Particles of 
burning Zr were carried over one-quarter of a mile through the air. [This took 
place at Bettis laboratory, Pittsburgh.] 
4 About 5 years ago some water-wet scrap Zr powder in wooden barrels was placed 
in outside storage pending development of scrap-recovery processes. During the next 
several years, a few minor spontaneous fires broke out in this material. In January 
1956, the material in several deteriorated wooden barrels was wet with water and 
repackaged in steel drums. 
In May 1956, employees working in the area noted that one of the steel drums 
lying on its side contained a black material “similar to carbon dust.” Just exactly 
what happened there is uncertain, but a spontaneous explosion occurred accompanied 
by streaks of red fire with black smoke extending 100 ft into the air. A pronounced 
concussion wave was noted and sound of the blast heard several miles away. Two 
employees were killed, one having been blown 80 ft through the air, and a third lost 
an arm. 
The drum contained Zr, probably in the form of a fine powder. 
Using extensive precautions, the remaining drums of scrap Zr were subsequently 
burned. During this operation, one of the drums exploded in a manner shown in 
Fig. 4. [This incident took place at the Y-12 area, Oak Ridge.] 
4 Two men died and two others were seriously iajured in 1954 in a spontaneous 
explosion initiated during removal of the friction-top lid from a polythene-bag-lined, 
one-gallon metal can containing Zr powder 16% wet with water. A ball of flame 
enveloped the entire area, accompanied by a definite concussion wave. 
4 A 2-lb sample of carbon-tetrachloride-moistened powdered Zr was placed in a 
glass flask, vacuum applied, and the flask very gently heated with a Bunsen burner. 
The Zr suddenly began to heat up and detonated with a blinding flash. The explo- 
sion was attributed to a small amount of water. 
Uranium Incidents 
{In June 1956, three flat U plates (44 X 3 X 14 in.), made by powder metallurgical 
means, were observed to have “blown up” to approximately 3-in.-diameter rods. 
The “rods” were removed from the building awaiting disposal. The following morn- 
ing, one of the plates exploded and “took off like a rocket,” hitting a tree 30 ft 
away. The remaining plates were pierced with bullets but no explosions occurred. 
It was the opinion of company personnel that the plates had not been pressed to full 
density and that there was an incomplete decomposition of the hydride in the U 
powder, causing a build-up of internal hydrogen pressure. 
qIm January 1955, an attempt was made to roll two 1,000-lb U slabs to 0.01-in.- 
thick strip. After initial heating to 1,150° F in a lithium-carbonate—potassium- 
carbonate bath, several 30% reductions were made by rolling. It was observed that 
heavy work passes had caused overheating. The strip, then % in. thick, was cooled 
to 1,200° F before entering the four finishing stands. The strip again excessively 
heated during the next three reductions and became so ductile on entering the fourth 
stand that it pulled into two parts and a cobble resulted. The strip at this stage was 
cherry red, but by the time it had been removed to the mill floor it was observed to 
increase in temperature to a white heat followed by melting and burning. 
In February 1956, a technician was attempting to roll a plate consisting of 
Zr-clad U, which, in turn, was clad in a low-carbon-steel jacket. During preheating, 
the furnace temperature control (which had been set to 1,450° F) failed, allowing the 
temperature to rise to 1,800° F. During subsequent rolling, molten Fe-Zr eutectic 
alloy within the steel jacket was forced to one end of the strip where it burst into 
flames as it sprayed out over an area approximately 10 ft wide, 10 ft high, and 25 ft 
long. One employee was seriously injured. 
{In the early program for the large-scale manufacture of metallic U, fine powder 
was allowed to collect under roughly 25 ft of water. At approximately one-month 
intervals, and without prior warning, a geyser about 30 ft high would suddenly de- 
velop over the powder and then immediately subside. 
qA series of cases is known in which massive pieces of metallic U, Pu, and Th 
have displayed unusual pyrophoricity, e.g., spontaneously igniting at room tempera- 
ture. Spontaneous fires in U chips are, however, much more common and im one 
briquettes that had been in outside storage for several weeks. 
drum contents were normal and at approximately room temperature, he was warned 
by an operator to stand back. A few seconds later, a flame shot to a height of abou 
