,
L"
WY
4
.
A
.
#
.
.
..
.
UNCLASSIFIED
ORNL
T
Ti. -
.
er
''
12.
:
SY-
313
2
.
.
2

".
2
D
97
Olent photo gadiso.co....
DIVISION OF
Polot his and contare stor betur ting Technology (Provised
FIFLE. OF PAPER Recovery of Uranium from Spent Zirconium-Based Reactor
Fuels in the Oak Ridge Volatility Pilot Plants
Prosenbalion Time
77
20-letnutog
2716
Y AUTHORS
(underlino name of speaker)
Mambert
W. H. Carr
R. E. Brooksbank
S. Mann
R. P. Milford
(lial address only once U all authors
at name address)
Chemical Technology Division
Oak Ridge National Laboratory
Post Office Box X
Oak Ridge, Tennessee
Yes
No
No
Yes
Division American Chomist or Chomi.
Hambur? el Enginouri (u nov. pine
eleselfriction owed as bielosion,
physicist, ode.)
Yes Chemical Engineer
No ! Chemical Engineer
Chemical Engineer
Yes Chemica). Engineer
Windsid
Work done at
ABSTRACT. Length 200 words. Plewe type double spaced within the rulad ante Il you need more opace for 200 wondo, or more space
for special abocracle (oce column headed "Instruccions" for individual programs), please continue your address on a secondo plain shesh
L A molten-salt fluoride-volatility process for the recovery and purification of un-
burned uranium in spent nuclear fuels has been developed at Oak Ridge National Laboratory.
In this nonaqueous process, as applied to U-Zr alloy, the fuel 18 dissolved into molten
LiF-NaF-ZrF, at 650°C by hydrogen fluoride. Muorine 18 used to convert the UF in the
resulting salt to volatile UFg. This partially decontaminated UF 18 further purified.
in NaF and MgF, absorber beds and then recovered in cold traps.
1.
The process was studied in the Volatility Pilot Plant with enriched, highly burned
U-2r alloy fuel elements decayed for as short a time as six months. The uranium product
recovered during the pilot plant operation met requirements for return to production
channels. Decontamination from fission products was among the best ever reported; in
one run, the decontamination factor from 92r-No exceeded 5 x 1010. Uranium losses below
0.2% and the capability for sustained operation were demonstrated.
The process offers several advantages over aqueous methods for treating enriched
zirconium-based fuel. These include reduced process and waste volumes, and increased
nuclear safety. A disadvantage is the higher corrosion rate encountered in certain
steps.
A commercial plant for irradiated V-Zr alloy fuels can be designed on the basis of
present knowledge.
"Research sponsored by the U. S. Atomic Energy Commission under contract with the
Union Carbide Corporation.
MAIL ABSTRACT TO PERSON INDICATED IN PROJECTED DIVISION PROGRAMS
61-7
.26407
im
,
va
U
.
AIL
UN
N
1
.IT
L
".
..
.
PT.
1
:
24
alcul 12%:
It. S
The
Car:
7/16/64
Arr. 8/11/0r
Recovery of Uranium from Spent Zirconium-Based Reactor
Fuels in the Oak Ridge Volatility Pilot Planting
Nuclear fuels are never completely burnel in reactors. Hence,
Tor économical operetion, the unbumed fuel, or 118gionable reterial,
must be recovered from tļae spent fuel of most muclear reactors.
Unfortunately, this unburned fuel 18 intimately associated with 218nion
products, which are highly radioactive. Many of the fission products
are also neutron poisons, and absorb neutrons which would otherwise
be available for causing additional 1188ioning. Thus, the recovered
fissionable material must be sufficiently separated from reutron poisons
that it can custain a chain reaction, and be sufficiently free of
radioactive fission products that the fuel can be reconstituted in a
form suiteble for recycle.
There are many processes for recovering and purifying unturned
uranium from spent nuclear fuels. Most of these processes involve solvent
extraction from aqueous solutions. However, there are several nonaqueous
grocesses being developed, including the molten-sait fluoride-volatility
process developed at Oak Ridge National Laboratory.
Nilir 12.5 of
.
.
The molten-selt fluoride-volatility process is applicable to most
types of enriched uranium fuels. For example, the next paper will
discuss process development for uranium-aluminum alloy fuels. This paper
will be coceinca with the process as applied to uranium-zirconium alloy
.
ficls.
Within the limiterions of the allotted time, I shall describe the
process and give the results of processing fuel elements decayed as short
a tine as six months. This will be followed by a brief discussion of
<
A
ll
Es
-
-
K
.
ER
.
*
>
.
*
some of the advantages and disadventages of the proceso when compared
with aqueous processes. Finally, I hope that there will be time for
me to try to answer any questions you might have. .
The molten-salt fluoride-volatility process consists of fuel .
d18solution in molten salt by reaction with hydrogen fluorldc, uranium
volatilizetion by conversion of UFL to UFG with fluorine, and further
purificetion and recovery of the UF: The process is shown in a little
more detail on the first slide, which 13 a schematic flow-sheet of
Slide 1: ORNL-D:g. 63-4519 (slide 6 at '63 Info. Mtng).
.
I
osa
the ORNL Volatility Pilot Plant.
For dissolution, Zircaloy fuel elements weighing 42 kilograms,
coutaining half a kilogram of uranium, are lowered into the bottom section
Harshilin
of the wronta dissolver. This section 18 5 Inches in diameter and 9 feet
high. The upper section, 24 inches in diameter, 18 provided for
:
deentrainment.
i
Molten berren salt 1s cherged to the dissolver until the fuel
ELE
elements are coverca; normal liquid level is in the conical section. The
Si
barren salt used in a mixture of 37 1/2 - 37:/2 - 25 mole $ 11thium,
Bodlum, end zirconium fluorides, melting at about 600°C.
.
..
Aniyaxous !F is sparged through the salt, and reacts with the metals
in the fu2.. ciceats. Zirconium is converted to ZrF%, and uraniun to
UF2. The !! :crides are soluble in the sali. Dissolution is
continued until the fuel elements are dissolved, and the meit composition
is about 27 2,2 - 27 1/2 - 45 mole % lithium fluoride - sodium fluoride -
.
.
.
I
R
SKO
.
Fax
217
..
.
-
-
.
.
NOW
With
7
.
II.
I
#
-
gooi
'YE -
W
+
A
IR : - -
zirconium tetrafluoride. As dissolution proceeds, the melting point
of the galt Crops to about 450 °C; operating temperature 18
simultaneously lowered from 650 to 500°C.
3F for the dissolution reaction 18 supplied through a recycle
system, which includes facilities for initial charging, vaporization of
EF for dissolver feed, and condensation for recycle of unreacted HF in
the dissolver off-ges.
After dissolution is complete, the salt, containing UF, 18 trans-
ferred to the Iluorinator. This 16 inch vessel of low-carbon nickel
has a working capacity of about 65 liters in the lower portion; again,
the upper portion is for deentrainment. Fluorine, sparged through the
melt, converts the UF4 to UFgThe volatile UFG 18 carried out of the
fluorinator in the gas stream, leaving greater than 99.9% of the fission
.
.-
product activity in the salt. We'll get back to this weste salt la a
-
moment.
2
.
.
ST
BAL
47.
The 62s Leaving the fluorinator flows through the lower part of the
movable bed absorber, through the chemical trep, to the off-les system,
which incluacs caustic scrubbers. The movable bed absorber is filled
with sodium fluoride pellets. The bottom part is held at 100°C. At this
temperature, the sodium fluoride will sort meny flosion and corrosion
products, such cs zirconiuti, niobium, ruthenium, and chronium fluorides.
The la 0.70 of the fission products pass through this portion of
the 2004; the principai cotoninants passing through are telluriwa,
iodine, moldcruci, technetiur, anå neptunium.
arter passing through the 400° scdiun fluoride, the gas stream goes
throusa a portion of the bed held at 100°C. Tae sodium fluoride in this
.
.
-
LA
in
73
YAM
?
i
19
tice

19"
.
Jan
zono sorbs the uraülum, molybdomum, technetium, and neptunium fluorides.
"I'
-
-
-
Thus, the stream leaving the movable bed absorber consists of excess
fluorine, inert purge gases, tellurium, and lodine. This stream flows
through the chemical trap, a bed of sodium fluoride at room temperature,
to ensure against the loss of UFF, and into the off-ras system.
After fluorination and absorption are complete, the 100° Sollum
fluoride is heated to 150°C to strip off the molybdenum. I might
will look on
which sheirledi'
mention that we have rocortiso revised the procedure to avoid the necessity
for this molybdenum stripping. By absorbing at 150°C rather than at 100,
molybdenum "Le not, absorbea; the quantity of UFG not absorbed to very ***
small, ara 'is held in the chemical trap, from which it 18 easily
recoverable.
.
-
.
The next operation 18 desorption of the UF, Irom the sodiuna fluoride.
For this, the entire movable bed absorber (8 heated to 400°C with a
sufficient sweep of fluorine to displace the UF desorbed. Unfortunately,
technetium and neptunium are also desorbed. Hence, the stream is routed
throush a bed of magnesium fluoride pellets, where the technetium and
neptinium Iluorides are trapped. The UF, is then filtered through sintered
metal, and is frozen in a receiver. The product from several runs say
be composited in the larger cold traps to permit the more convenient
shipment in a standard size UF, cylinder.
I promised earlier to come back to the waste salt remaining in the
fluorinator. Aiter desorption is complete, part of the sodium fluoride in
the roveble bed absorber 18 discharged into the fluorinator by a hydraulic
piston. Fresh sodium fluoride, or material crom the chemical trap, is
then charsed at the top of the absorber. Thus, counter-current operation
3
14
*
1_
"
NG
A
is achieved.
T
.
LU
11
.
V
.
WI
.
2
"
L
IUNI
NONI
The sodium fluoride discharged from the absorber dissolves in
the salt in the fluorinator. A large plant, this weste salt could
be trensferred directly into a large underground tank for permanent
waste storege. On a development program, it is more convenient to
transfer the salt into a mild steel can, allow the salt to freeze,
and then bury the can.
Lights, please sol (dip
An extensive development program on uranium-zirconium alloy
fuel processing hes been completed at Oak Ridge National Laboratory.
This development program culminated in a series v? 5 runs in which
uranium was recovered from fuel elements decayed as short a time as 6
w
*
.
.
-
-
-
-
1
months. Since the lessons learned earlier are embodied in the final
.
.
.
.
..
5 runs, I'll report only on this series. It is interesting to note
that the series was shorter than planned; after the 5 runs, there was
enough data that we felt that the process had been adequately proven on
a pilot plant scale.
Throughout the series, operations were smooth, and the ease of
operation increased steadily throughout the series. In fact, plant
performance can be summarized merely by stating that the equiprent
functioneå 25 designed.
Rediation exposures were quite low. For the most active 3 month
period of the program, the average radiation dose received by the operating
crew was only likes of the maximum permissible, ani no individual's exposure
exceeded 100 mr in any week.
Daconcealntion iron fission products was amazingly higa, end in some
cases was betto: than obtainable even with 3 cycles of solvent extraction.
.
:
.
.
.
.
+
LE
WA
ct
E
'•
,
'
t
i
-
".
..
сенімімім алу »
wriлiкке». 4. ..
vitAlow
-
u.
a
ь
аммо... . .. . . . .
. . .
Decontaiantion factors for specific muclides are shown on the second
slide, which includes the decay time for fuel procossed in each run.
Slide 2: To be made from data in Table 2.2 of '64 CTD Annuel; some
rearrangins may be required for legib11ity on a slide. (Instead of
"Run No.", should we use "Run" and 1st, 2nd, etc?)
-
In this table, in cases where the final product was below the analytical
limit for the particular nuclide, the analytical llmit was used as the
analysis. I'd like to call your attention particum:ly to the zirconium-
niobium. DF of >5 x 1040 in the last run. To my knowledge, this is the
highest DF that has ever been reported on a comparable basis.
Even thougia there are some instances of a drop in DF, such as cesium
in the 3rå run, there is no general trend toward loss of decontamination
with additional runs. On the contrary, if there 18 any trend at all, it
is in the direction of improvement.
From a cherical stanapoint, the composited product was satisfactory
when it was returned to production channels, even though product analyses
from individual runs showed high impurity levels. On the rext slide, the
Slide 3: ORNL-Dwg. 63-4668 (slide 11, '63 Info. Mtng.) (See note on
Slide 2 re, "Run No.") (should it be "cationic" rether than "cation?")
1
A
.
.
cationic iurisics in the UF, products are listed for each rin. The
specifications for UF, returned to production channels, published in the
Federal kegister, set a molybdenum limit of 200 ppm and a total cationic
.
.
.
*********
**
ST

impurity linit of 300 ppm. Even though the molybdenum specification
was exceedea by 20% in 2 runs, and the total was too high in all rins
but the first, after the products from all 5 runs were transferred to
a single shipping cylinder, the UF, was acceptable.
You have probaüly noticed by now that none of the columns will
add up to the totels shown. These totals were obtained by including
the analyčical limits of detection for cations which were below the
-
-
-
•
-
?
limits.
The erratic molybdenum levels indicate that the molybdenum
removal procedure was not uniformly effective. We now believe that the
time at 150 °C, the fluorine sweep, or both, were inadequate, and that
sorbing the UE, at 250 °C rather than at 100° will solve this problem.
In all runs, technetium and neptunium contamination of the product
was satisfactorily low, even though the range was relatively high. Hence,
the magnesiu fluoride bed is adequate for trapping these two nuclides.
Uroniun recovery was quite satisfactory. Valid material balances for
individual mms are impossible because of the system bola-un, but all
ی امررر..
وره
1
- ترم
. هي ردة
.
materiel balance : 100 was achieved for the antire more lion-
recoveraine Cosses are sho:in on the next slide, which also includes
S-
fluorination results. You will note that uranium remaining in the
waste salt citez fluorination was uniformly low. For a corrercial
plant: ii rould be more economical to reduce the 2 hour fluorination period
ani dised Silly pre uranium.
I'm lost comme is the total nonre coʻrerable uraniu: loss in weight
This
comes toc rzu: remaining in the sait after fluorinction,
uranium lose to the waste salt by sodium fluoride discharge from the
movable bed cosoxter, and 'ranium lost to the caustic scrubber in the off-cas
...
.......
m.com
-
...
mo...
...
.

.
Slide 4: to include the following data:
Use one or
the other
(See slide 2 rote)
· Uranium in Salt, ppm
Before
After:
Fluorination Fluorization
Run
Run No.
R-7
1st
Nonrecoverable U LOSS
vt. &
0.10
0.02
0.15
0.09
R-8
4485
4337
5340
2nd
R-9
3ra
R-10
4th
4300 :
R-11
5th
3283
.
0.09
system.
The maximum loss in any rin was less than 2 tsnths, and the
total loss for the series was less then a tenth of a .
LIGATS, PLEASE.
The molten-salt fluoride-volatility process offers several advantaees
over conventional equeous methods for fuel processing. The most obvious
one is the high degree of decontenination from fission products. Another
advantage is the reduced process and waste volumes inherent in the
volatility process; you may recall from the flow-sheet that all the
high-activity-level waste from a run is in the 65 liters of solid salt
which is buried. An additional advantage is the increased nuclear safety
which results from the absence of an aqueous moderator.
rccre are the principal disadvantages of the process. Corrosion
in the dissolver and in the fluorirator is high, but is predictable and
.
.
R
can be taken into account in vessel design.
The latest corrosion measureronta
28
4
14---
T
.
.
-
.
-
-
.
...
HALLY
7
.
:
in the dissolver show that after 40 runs the average rate 18 one half
rll per run; the deepest ineasured penetration was equivalent to
seven-tenths mil per run. The corrosion rate in the fluorinator 1s
olmilar. The other disadvantage is that at present the process is
economically limited to high enrichment fuels, since plutonium recovery
v 15
has not been demonstrated an a pilot plant scale.
The necessary basic information is available for the design of a
comercial rolton-salt fluoride-volatility plant for processing irračiated
uraniwn-zirconium alloy fuels. With one exception, such a plant oesign
would be primarily a scale-up of the present pilot plant. The exception
is the economic reed "o provide for either recycling part of the waste
salt 2fter fluorination back to the dissolver is part of the barren
salt or retaining part of the salt in the dissolver after dissolution es
a bigh-zirconiun heel. Such a provision would permit the use of a
bine y salt of lithium and sodium fluorides for barren salt feed,
avoiding the addition of expensive zirconium fluoride. A recycle flow
sheei es outlined has been proven on a somewhɛt smaller scale. From
a chemical composition standpoint, the pilot plant runs simulated such
a flow sheet.
-
To briežly summarize the results of the work reported, operability
oi toe molton-salt fluoride-volatility process as applied to uranium-
zirconium alloy fuels has been proven. Capability for sustained operat:0::
with Ima 2-; shori-cooled fuel has been decoustzated via 10: losses
arrá 0x0028: Cacoatacination.
Hir. Cuimd, is there time for questions?
...*
.
lii
you
can see

.
.
22
DATE FILMED
2 / 0 765
then
.
RE
r
"
LEGAL NOTICE -
1.
!
0
1
This roport was proparod u an account of Govornment sponsored work, Noithor the United
Statos, por tho Commission, nor any person acting on behalf of the Commission:
A. Makos any warranty or roprosentation, oxprossod or impliod, with rospect to the accu-
racy, complotoneos, or usefulnous of the inlormation containod in this roport, or that the wo
of any Information, apparatus, molhod, or procon disclosod in wis roport may not Infringe
privately owned rights; or
B. Assumos any liabilities with respect to the use of, or for damagos rosulung from tho
10 of any information, apparatus, molhod, or process disclosed in this roport.
As wond in the abovo, "por son acting on boball of the Commission" Includes any om-
ployoo or contractor of the Commission, or omployer of such contractor, to the oxtont that
such omployoo or contractor of the Commission, or omployee of such contractor proparoi,
disseminatos, or provides accos. ro, any information pursuant to his omploymont or contract
with the Commission, or his omployment with such contractor.
St
.
1
.
1
.
17.
LO
..
'MA
"
FL
I
END
V