UO, diameter in inches 
- Natural : 
0.70 SGT eR od AN 
ro) 
O 
6 8 10 12 4 16 18 
Enrichment (wt % U#5) 
FIG. 4. Dependence of fn(xa) on UO> 
diameter and enrichment for 95% 
dense, solid oxide cylinders (60) 
unirradiated UO, at temperatures be- 
low 200° C. In the range applicable 
to operating fuel elements, i.e., above 
400° C, the difference is only 30%. 
However, a 30% uncertainty in the 
thermal conductivity results in a much 
larger uncertainty in calculated fuel- 
core temperatures, as is evident from 
Fig. 6. 
Recently, Eichenberg (32) measured 
the center temperature of an assembly 
of UOz pellets clad in stainless steel 
during irradiation in the Materials 
Testing Reactor. The pellets had been 
prepared from thermally decomposed 
uranyl nitrate by cold-pressing and 
sintering in hydrogen, the same process 
used for the Shippingport reactor fuel. 
Values of the effective thermal conduc- 
tivity required to produce an observed 
UO: central temperature of 150°-600° C 
were calculated and are shown in Fig. 2. 
The calculated results are about a fac- 
tor of three lower than the corrected 
Kingery data of Fig. 2 (28). 
Ross (27) has measured the thermal 
conductivity of stoichiometric UO, pel- 
lets at 60° C after irradiations up to an 
integrated thermal-neutron flux of 2 < 
10'°n/em*. The pellets were prepared 
by cold-pressing and sintering UO, 
powder prepared by the hydrogen re- 
duction of ammonium diuranate. The 
maximum central temperature in the 
pellets during the irradiation did not 
exceed 500° C. The measurements 
were made on a divided-bar type of 
comparative-heat-flow apparatus using 
Zircaloy-2 probes. Only a small de- 
crease in thermal conductivity, ~25%, 
was observed after an irradiation of 
9 X 107 n/em?. Longer irradiations 
produced no further significant-change. 
Additional measurements for shorter 
irradiations are under way to determine 
the irradiation level at which the de- 
crease occurs. 
There is a marked disagreement. be- 
tween Eichenberg’s results and those 
of Ross. It is possible that the thermal 
conductivity of UOz in a neutron flux 
is appreciably lower than postirradi- 
ation measurements indicate because of 
point defects, which might anneal out 
when the irradiation ended. Fuel-ele- 
ment tests described later do not sup- 
port such a hypothesis, however. 
Several laboratories have reported 
that a threefold decrease in thermal 
conductivity at 60° C can be produced 
in UO, pellets by increasing the O/U 
ratio above stoichiometric by 8% (8, 
15, 27). On the other hand, Thack- 
ray’s values for hot-pressed UOe.13 over 
the temperature range 100°-600° C (33) 
were, essentially the same as the cor- 
rected values of Kingery et al. (28) for 
stoichiometric UO... Measurements 
by Ross (27) are given in Fig. 3. The 
conductivity of 0.017 w/cm/°C deter- 
mined for the steam-sintered sample 
approaches the theoretical lower limit 
of 0.015 w/em/°C for the UOz lattice 
determined by Kingery (3, 32). It 
is not yet known whether a compar- 
able decrease in conductivity occurs 
with increased O/U ratio at higher 
temperatures. 
Some of the apparent discrepancies 
in thermal-conductivity measurements 
may be due to variations in the fabrica- 
tion technique. For example, the con- 
ductivity of stoichiometric UO: pellets, 
prepared by steam-sintering at 1,400° C 
followed by hydrogen-cooling, is 25% 
lower than for those that were also 
steam-sintered but annealed in hydro- 
gen before cooling. 
Much work remains to be done be- 
fore the thermal conductivity of oper- 
ating UO, fuel elements can be pre- 
dicted with certainty. 
Thermal expansion. The several 
published values for the linear thermai- 
expansion coefficient of UO» are in rea- 
sonable agreement (Table 1). Murray 
and Livey (33) reported that the bulk 
density and O/U ratio had little effect 
on the thermal-expansion coefficient. 
\Fabrication of UO; 
“ UO. compacts can be fabricated from 
a variety of starting powders by con- 
ventional ceramic processes such as 
cold-pressing, extrusion, — slip-casting 
and isostatic-pressing. Dense bodies 
can be produced from the compacts 
by sintering at 1,300-2,000° C. Other 
fabrication methods include rotary- 
swaging and loose compaction of pow- 
der in metal sheaths and hot-pressing 
in metal or graphite dies. 
These processes are detailed in the 
following discussion. 
Powder preparation. UO» powder 
can be prepared by the hydrogen reduc- 
tion of several compounds, e.g., UOs, 
UO; hydrates, U3;303, ammonium di- 
uranate, uranium oxalate and uranium 
peroxide, and by steam oxidation of 
uranium hydride or uranium metal. 
UO, powders prepared by different 
methods show wide variations in physi- 
cal characteristics such as surface area, 
particle density, porosity distribution 
and crystallite size (2, 36-39). Such 
variations have a marked effect on the 
sinterability of the powder as. illus- 
trated in Table 2. 
Pellets with densities above 10 gm/ 
cm* can be produced from the least sin- 
terable powder shown in Table 2, how- 
ever, if high compacting pressures and 
long sintering times are used. The first 
oxide charge for the Shippingport reac- 
tor, for example, was prepared by com- 
pacting UO». powder obtained from 
Mallinckrodt at 250,000 psi and sinter- 
ing for 8 hr in hydrogen at 1,675° C (41). 
The pellets varied in density from 10.2 
to 10.4 gm/em’. 
Powders of low sinterability can be 
activated by wet-ball-milling before hy- 
drogen reduction as shown in Table 2. 
A simpler, more economic method is to 
grind UO; powder to particle sizes less 
than 1 micron in a fluid-jet mill and 
then to reduce in hydrogen. Pellets 
with densities up to 10.7 gm/cm* have 
been produced from such powder after 
pressing at 40,000 psi and sintering for 
1 hr at 1,625° C (39). 
Many laboratories have investigated 
the effects of chemical additives to pro- 
mote more rapid sintering of UOs. 
Canadian workers have verified the 
many earlier reports on the effective- 
ness of a 0.1% TiOz addition and have 
found that 0.4% Nb; has an even 
larger effect (39, 42). 
The preparation of UO2 powder via 
the ammonium diuranate (ADU) route 
has been extensively investigated (10, 
39, 43). A reproducible ADU powder 
was made by the batch precipitation of 
ADU from urany] nitrate solution (ura- 
nium concentration of 100 gm/I at 
60° C) by the rapid addition of concen- 
trated, aqueous ammonium hydroxide 
untila pH of 9wasreached. The ADU 
was filtered, washed with water and 
oven-dried in air at 200° C. Higher 
uranium concentration, slower precipi- 
61 
