I
M. J. RASMUSSEN and H. H. HOPKINS, Jr. Hanford Atomic Products Operation, General Electric Co., Richland, Wash.
Preparing Plutonium Metal
. . . viu the
Chloride Process
Using the chloride, rather than the fluoride, greatly reduces neutron exposure
*
r r H E BASIC CHEMISTRY of the chloride route from plutonium nitrate to plutonium metal was known in 1944 but apparently has not previously been developed for industrial use. In general, a fluoride has been preferred to chloride in plutonium metal preparation. An objectisn to a plutonium fluoride is the high neutron density resulting from intense alpha radiation of the fluorine atoms. By using a halogen of higher atomic number, the neutron exposure of personnel would be reduced more than by any reasonable shielding. The most promising substitute, PuC13, would lessen the neutron radiation by a factor of 100 in the dry salt. A second advantage of a chloride process is a matter of purity. The volatility of several metallic chlorides serves to remove or decrease these impurities in the halide powder and the metal. The process selected (70) €or further development was the continuous chlorination of PuOz powder by phosgene gas in a heated vibrating tube, followed by the conversion of PuC13 to plutonium metal by heating with calcium in a hermetically sealed bomb. This chloride route for plutonium metal preparation is one of many variations possible in a general method proceeding from nitrate solution to oxalate and/or oxide, to halide, and to the metal (3, 9, 77). This report emphasizes the chloride process selected. Excellent development work has been performed previously by the author's colleagues on steps before and after chlorination. The relatively-high purities of plutonium obtained, together with the demonstration of a workable process, show that preparation of plutonium metal via the chloride is a practical method. The exposure of personnel to neutrons may not decrease by a factor of 100; but on a single comparison of equal amounts of plutonium, the fluoride was shown to emit neutron radiation 64 times that of PuC13.
Literature Background Subject Early studies of chloride process for Pu metal Chlorinations Bomb reductions Other early work Variations possible in chloride route Pertinent Pu chemistry More recent PuCh preparations Other significant work leading to chlorination process selected Oxide preparation Chlorination by phosgene
Ref. (7) (1, $1
(13)
(9, 14) (4-6, 8, 11)
(3)
(17)
Experimental Oxalate Preparation. A stream of about 1M plutonium nitrate solution and one of 1M oxalic acid solution plus 0.25M HzO2 were pumped steadily into a 4-inch-diameter precipitator. The peroxide served to convert the plutonium to the quadrivalent state. An air-driven stirrer prevented settling of the oxalate precipitate and aided the overflow of slurry into a round-bottomed pan (flowsheet, below) in which a moving blade kept the slurry agitated. Vacuum (18 inches or more of water) applied to a rotating filter drum removed the filtrate from the pan and deposited
the precipitate on synthetic fiber filter cloth. The wet plutonium(1V) oxalate was scraped from the drum and dropped into a calciner. Losses of plutonium into the filtrate tank were about 2 to 5%. After destroying excess oxalate with permanganate, this solution was recycled into the process by which the plutonium nitrate was prepared. Oxide Preparation. The screw calciner received the wet oxalate, Pu(C~04)26H20, passed it through 30 inches of furnace, and converted it to free-flowing oxide in about 25 minutes. The 2-inch-diameter screw in this calciner was made of Hastelloy C. Within the Type 304 stainless steel furnace tube, the powder was pulverized in a device similar to a kitchen meat grinder. Usual calcining temperatures were 250' to 350' C. Chloride Preparation. The chlorinator was in a gloved hood adjoining the calciner hood, but there was n o + a t mospheric flow be,ween them. The chlorination hood was supplied with a dry atmosphere. The PuOz dropped from the calciner into the chlorinator through two ball valves which opened and closed on cycles timed to prevent gas flow between these units. Carbon dioxide entering the valves swept the phosgene downstream. A Teflon bellows isolated these valves from the vibration of the chlorinator.
Simplified flowsheet shows continuous flow process for PuC13 preparation VOL. 53, NO. 6
JUNE 1961
453
Table 1. Typical Operating Details o f Continuous Chlorinator to Yield 98% PuC13 Throughput rate Residence time Bed depth Tube inside diameter Furnace temp. Heated length Filter temp. Chlorinator slope Vibration cycle Gas pressure Phosgene flow
250 grams Pu/hr. 30-40 min. 0.25-0.4 in. Oval, 1.25 X 2.25in. 500' C. max. 40 in. (29 in. above 450' C.) 350'-400° C. 40 10 sec. on, 25 sec. off 1-20 in. of water vacuum 6 moles/mole PUOZ
Intermittent vibration was achieved by an automatic timer and an F-115 Syntron vibratory feeder rated at 6 tons of sand per hour. The intensity of vibration was adjustable. Gases left the reactor at a tee about 8 inches from the upper end. The filter housing and ceramic filter mounted at this point were subject to vibrations of the reactor. The vibrations served to clean the filter and empty this powder back into the chlorinator. As the powder vibrated down the Hastelloy B tube, it was heated during a 40-inch path to temperatures rising to 500' C. The temperature in the first 15 inches could become excessive and sinter the powder unless the electric heat was adjusted downward as exothermic heat incrtased. Phosgene gas entered at both ends of the tube after being preheated to about 200' C. Typical operating details of the chlorinator are given in Table I. These conditions yielded about 98y0plutonium trichloride, whereas powder flowing at 500 grams of plutonium per hour showed
about 80% conversion. The maximum throughput with acceptable product was not determined. Metal Preparation. The preparation of plutonium metal by calcium reduction of PuC13 was done in the same developmental equipment as was used for PuF4. The pressure bomb and lid with pressure gage are shown (below). Also shown are containers of iodine, calcium, and PuC18, a crucible in a can, and a typical plutonium metal button. -4 calcium and iodine reaction was used to boost the temperature of the principal trichloride and calcium reaction. These chemicals were mixed and placed in a magnesia crucible in the stainless steel can. This crucible had been previously sand-packed and fused into the can by impregnating the magnesia with molten CaC12. The can was used to prevent the molten chloride salts of a reduction from fusing the crucible and sand into the bottom of the bomb. The slag and metal were eventually routed or broken out of the crucible, and the crucible was reused many times. After a reduction bomb was charged and the lid bolted in place, it was leaktested and purged at least two times with argon. Firing was initiated by heating in a 10-kc. induction coil for about 15 minutes to a bomb skin temperature of about 700' C. Firing was normally indicated by an upward change on the temperature chart and by a sudden drop in pressure. Bomb pressures were usually 50 ==! 30 p.s.i.g. and rarely exceeded 100 p.s.i.g. About 200 p.s.i.g. was considered the maximum safe pressure. LVhile the reduction charge was molten, the plutonium metal collected in the bottom of the crucible, essentially free of any slag and excess calcium. The flowsheet (p. 455) shows the relative amounts of chloride, iodine, and calcium. Soine (76) recently discussed the variables in the calcium reduction of PuC13. Plutonium metal was prepared in batches ranging from 30 to 1400 grams with good success. Slag and broken crucibles from the metal reduction step contained plutonium which had to be recovered. Smith (75) developed techniques for dissolution of these wastes in HNOI and recovery of the plutonium as nitrate solution. Small amounts of scrap plutonium mctal, powders, and liquid were similarly recovered. Liquid and solid wastes from this recovery work were sent to underground disposal along with other contaminated wastes. Results and Discussion
Equipment
used
in
PuFs
reduction
was also used for PuCl3 reduction Top-gage on lid, pressure bomb, reusable crucible; center--jars of iodine, calcium, PuC13; bottom-typical metal button
454
Process Conditions. The optimum conditions of oxalate preparation required an 0.1M excess of oxalic acid in the supernatant stream. Nitric acid in the plutonium nitrate stream was 3.5 to 5M. Under these conditions, the plutonium losses to the filtrate were smallest. and the filter cake was easily
INDUSTRIAL A N D ENGINEERING CHEMISTRY
handled. If these conditions were not maintained, the precipitate would not filter readily and/or would cling to the trough rather than fall freely into the calciner. The oxide preparation was stopped short of conversion to the stoichiometric PuOZ. The powder could Iose about 10% of its weight on further heating to 1000' C. The objective was to produce a free-flowing powder which would be easily chlorinated. Calcining temperatures were varied from 150' to 400' C. with noticeable changes in powder color shade. All of the oxide powder was readily chlorinated, so optimum temperature was not defined. Prior to use of the chlorinator described here, several other chlorinating agents were tried. Hydrogen and hydrogen chloride: gaseous carbon tetrachloride, and C12-CO mixture were all compared with phosgene and found less reactive. Use of phosgene permitted a choice of a 100° C. lower temperature for an equivalent reaction rate or a higher chlorination rate at the same temperature as the other reagents. The vibrating tube Tvas selected as the reactor because of other local experience with vibrating tubes and also because it has no moving parts exposed to high temperature and corrosive atmosphere. The minor nickel impurity found in the chlorinated powder indicates the possibilities of using moving parts and higher temperature in a Hastelloy B chlorinator. Phosgene flow was 6 moles per mole of plutonium (Table I), or three times the stoichiometric requirement. This may have been more than was necessary. Under the specified conditions, the principal and minor chemical reactions were as follows : PUOZ 2COC12 + PuCl3 2C02 1/2C12 (principal) PUO? COCl,--t COa 1/2Cla (minor) PuOCl PUOCl COCl? PuCls CO2 (minor) COCla -+ CO C19 (minor) H 2 0 COCl? --t 2HC1 C O Z(minor) The phosgene hydrolysis indicated in the last equation occurred when moisture entered with the PuOa or when moist air leaked into the gas exhaust system. Chlorination did not occur when air leaked into the reactor tube. Small air leaks were the cause of incomplete chlorination on several occasions. The partly chlorinated powders from such runs were recycled through the chlorination step with good results. The conversion of PuC13 to the PUOZ by high temperature air was utilized in the handling of some impure powders such as hood sweepings. Some of these were freed of chloride by heating in air at 400' to 500' C.: then dissolved by acids in stainless steel equipment. These may also be recovered, as mentioned.
+
+ +
+
+
+
+
--f
+
+
+ +
PLUTONIUM METAL Most of the time the PuClr formed in lumps which were largely pulverized by the vibration in the tube and in the product receiver. The inside material of the lumps was chlorinated and offered no problem in reduction to metal. When the reaction tube was plugged by lumps, it was due to experimental conditions. Three things were done to eliminate any serious problem from plugging: the best operating conditions were defined, the chloride exit was made oversize to prevent interference with flow, and the tube end was altered to permit easy rodding if a plug developed. Powder flow control problems were greater than anticipated. The resonance of the vibrating tube was influenced by suspension points, powder load, filter housing weight, and vibrator mounting. Experimentation was necessary to eliminate standing waves which interfered with powder flow. The vibrating tube served its purpose in producing high quality PuC13 which was readily reduced to plutonium metal. A study of the variables in the metal preparation (76) showed that good mixing of the charge was not always necessary. The charge could be unmixed if the PuC13 was on top. The presence of moisture in the reduction could lead to hazardous bomb pressures; 0.5% water was permissible in the trichloride. Because the reduction work was performed in a gloved hood a t 40y0 relative humidity, the rate of hydration of the trichloride was of concern. Several powder samples of 0.75-inch depth were exposed to air at controlled humidities. The rates of weight increase are shown (p. 456). Less than 0.01% increase was observed in 1000 hours at dew point, -18’ C. There
was no evidence of hydration at dew points of -20’ and -28’ C. The permissible time of exposure a t various humidities may be approximated from the graphs. Hydration is accompanied by a powder volume increase and a change to bright blue. Relative exposed area is critical in hydration rate. Gases. Phosgene has been used for over two years without harmful incident. Procedures for safe handling of H F are adequate for phosgene. Some psychological resistance was experienced. I n cold weather, the phosgene cylinders mus; be allowed to warm to room temperature to provide the normal 30 p.s.i.g. pressure. The gas flow control valve was heated to eliminate liquifying the phosgene by the pressure drop from 30 p.s.i.g. to below atmospheric normal. Exhaust gases from the process were filtered to prevent losses of entrained plutonium powders and contamination of the environment. The gases from the calciner were kept hot through a cyclone which trapped some of the oxide dust. A filter of glass cloth and then a water vapor condenser further prepared the off-gases for release into the general laboratory vacuum system. Exhaust gases from the chlorinator were filtered through a heated ceramic filter. Filter cylinders cut into 5-inch lengths from products of two manufacturers were both used with such success that no efforts were made to use more expensive metal filters. At the flow rates used, pressure drop across a filter was nearly 0 to 10 inches of water. Filters are described in Table 11. The filters permitted the “bleeding” of some fine dust until a cake formed. I n 30 or 40 hours of test operation, less than 0.1 gram of plutonium passed either
, RE D UC TAN T
BOOSTER Ca
33.58
I2
2 13.0
Ca
CHARGE
3149
I I
L
PuCI~ I
I
14458
1
J
I
-
SAND AND CRUCIBLE
c
BOMB REDUCTION
I BUTTON PICKLING IH20 I LITER
Si 0 2
r
BUTTON Pu
973.59
PICKLE SOLUTION TO RECOVERY
S L A G TO RECOVERY ~~
Flowsheet for batch reduction o f PuC13 shows relative amounts o f chloride, iodine, and calcium required
Table II. Commercial Ceramic Filters W e r e Satisfactory for Chlorinator Exhaust Filter 5
B XF
0.2P
0.29
A
Grade Porosity Pore size Wall thickness Material a
65-M dia.a loo-& dia. 7/8 in. 3/16 in. Aloxite Unglazed porcelain
Approximate.
of these filters. When a filter was operated cold, it became plugged by volatile chlorides. When a filter was heated to 350’ to 400’ C., these chlorides passed through and condensed beyond the filter. Spectrographic analyses of these condensates showed significant amounts of the following elements (listed in approximate decreasing order of concentration) : iron, phosphorus, molybdenum, chromium, zinc, boron, bismuth, aluminum, nickel, silicon, potassium, titanium, manganese, and sodium. Most (but not all) of these are explained by the volatility of their chlorides and carbonyls. After a second filtration and cooling of the exhaust stream, the gases for disposal were found (72) to be phosgene, COz, chlorine, hydrogen chloride, and CO. The boiling points of these make it easy to separate the phosgene (boiling point = 8.3’ C.) in a cold trap. Phosgene thus recovered was reused with success. The economics of this step have not been evaluated. Direct release of the filtered exhaust gases to atmosphere via a large dilution with other gases in the exhaust stack appeared feasible, but was impractical for this development program because of the long run of poly(viny1 chloride) pipe needed to reach the stack. A secondbest method of gas disposal was adopted. The filtered gases were carried into the bottom of a NaOH scrubber tower by applying vacuum to the top of the tower. The caustic from a 25-gallon reservoir was recirculated continuously through the tower which contained graphite Raschig rings. The initial NaOH concentration was GM. More concentrated solution was impractical because of carbonate precipitation in the tower. The caustic was discarded when still 0.5 to 1.OM because of impending carbonate precipitation. This scrubber effectively removed corrosive gases and permitted release of the remaining gases into the hood exhaust system. Materials. Most of the problems in selecting materials for this process were due to the chemical reactivity of phosgene and hydrogen chloride. When dry and a t room temperature, phosgene VOL. 53,
NO. 6
JUNE 1961
455
Table 111. Five Materials Are Satisfactory in Phosgene. (78)
Table IV.
Corrosion R a t e , Mils Penetration/Month 90'400' 500' BOOo 100" C. C. C. C.
Xaterial HastelloyB HastelloyC Nickel, A Chlorimet-2 Ni-0-Ne1 304L stainless
7.0 9.0 11 11 19 68.5
0.02 0.02
0.9 0.6 0.6
... ... ...
0.8
0.9
Suitability of Materials Tested for Use in Phosgene Depends on Environment
Test Conditiona
Hastelloy C Hastelloy B Nickel Ni-0-Ne1 Chlorimet-2 96% silica Porcelain Quartz Borosilicate glass Asbestos
u p to 2000 C.b
For corrosion resistance at 100' C. Gaskets, tubing, miscellaneous parts
Teflon Kel-F Type 304L stainless (for short life)
25' C.
Gaskets
Neoprene Koroseal
Gum rubber Silicone rubber
25" C.
Tubing, plastic bags, fittings, pipe
Saran Polyethlene Nylon Poly(viny1 chloride), rigid
Poly(viny1 acetate), poly(viny1 chloride), plasticized a s in flexible tubing or sheeting
25' C.
Protective spray paint
25" C.
Phosgene supply lines and miscellaneous parts
7.2 6.1 4.4 5.6 12
Phosgene plus 50% water-saturated air, flowing a t 2 liters/hr. a
is safely contained in metals such as black iron. copper, Monel. and stainless steel. When hot, none of these is satisfactory. Table I11 shows five materials suitable for use in phosgene at 500' C., but much less suitable at 600' C The high corrosion rates at 90' to 100' C. show the effect of moisture and phosgene (essentially hot HC1) a t condensation temperatures. Several nonmetallic materials are satisfactory a t this temperature. At room temperature, the corrosion of Type 304L stainless steel was not as severe. Phosgene and PuCI:, were used for two years in one gloved hood where the air was usually 40 to 60% relative humidity, without serious damage to the Type 304L stainless steel. The attack of steel by phosgene and PuC13 powder at room conditions is through hydrolysis and the formation of hydrogen chloride. The effect of various humidities was investigated as a means of limiting corrosion. At a relative humidity of 40y0 or higher, thin films of powder deliquesced (formed puddles). This is a corrosive
Type 304 stainless Tantalum Titanium, A-55 Gold Platinum Alloy : 6070 gold25y0 palladium1574 platinum
For reactor, filters, and gaskets at highest temp.
... ...
67
Materials Unsuitable
Materials Useful
Use Considered
About 500' C.b
Methyl methacrylate resin in solution Type 304 stainless Monel Copper" BrassC IronC
Zinc Aluminum Aluminum-silicon alloy
'
Metals a t elevated temperaa Temperature shown not necessarily maximum for material. Severe corrosion i f moist air contures tested in phosgene plus 50% water-saturated air (18). tacts phosgene and metal.
situation. A group of test coupons showed that corrosion decreased as humidity decreased. At dew point, -20' C., there \vas negligible corrosion to test coupons of Type 304 stainless steel when in contact with PuC13. A dry atmosphere of -20' C. dew point or better is readily available with commercial supply units. Such an atmosphere
prevents corrosion and prevents hydration of powder from carrying moisture into the reduction step. A number of materials \vere tested in phosgene and are classified loosely in Table I\' as useful or unsuitable for use in this program. Classification is subject to change with change in environment. For example, although plasticized poly(viny1 chloride-acetate) bags are unsuitable in phosgene. these bags gave satisfactory service on the chlorination hood because phosgene was generally confined to the process lines. Product Quality. The quality of the final product is related to the qualities of the intermediates. Some impurity concentrations in the process are shown later. Significant impurity removal was achieved in the precipitation of plutonium oxalate from the nitrate solution. Impurities in the oxalate may be assumed to correspond to those in the oxide. The PuOz was of olive-green color, entirely free of lumps, and of particle size about 1 micron or less. Its bulk density was approximately 1.5 to 2.0 grams per cc. Normal plutonium content was 80 to 85%. The PuC13 color was various shades of greenish-blue when 90% or more chlorinated. When chlorination was incomplete, the oxide present was visiblr. as a brownish tinge. From this color, a
-
5
I O
il
05
Exposure of powder samples io air at controlled humidity
w 4
t w 3 0 3
co5
001
1
,
I
l l I 1 I
5.
IO
I
I
I 1 1 1 1 1 1
50
100
E X P O S U R E . HOURS
456
INDUSTRIAL AND ENGINEERING CHEMISTRY
,
I
I I I I I I
500 I O (IO
-
PLUTONIUM M E T A L V). This expression was insensitive to Table V. Analyses of PuCl8 Powders Show Almost Complete Chlorination Ratio Water Run No. Cl/Pu Insoluble, % ’ 16 previous runs Av. 2.99 17 18 19 20 21 22 23
2.96 2.62 2.94 2.90 2.98 2.97 2.98
24
...
... ... ... ... ...
...
8 13 12.4 3.7 4.7 7.4 0.03
25
Nil Nil 1.5 Nil 1.5
26
27 28
Nil
29
Nil 2.9
trained person may estimate the unchlorinated portion with accuracy of a factor of two. Laboratory analysis of the chlorinated powder was approached in the following two ways: An unweighed sample was analyzed for plutonium and for chloride content. The results were expressed as a ratio, Cl/Pu. This served to confirm that the material was primarily PuC13 (Table
trace amounts of unchlorinated powder. The analytical method failed when oxide in the sample failed to dissolve, as in Runs 22 and 23. I n these runs, a much higher chlorinator flow rate was attempted. 0 The second analytical approach was more satisfactory for detection of unchlorinated powder. A weighed sample was stirred into water and the percentage insoluble determined. The water-insoluble material included all of the PUOZand most, if not all, of any PuOC1. This method has not been thoroughly investigated, but the agreement between some duplicate samples is seen in Table V. The PuC13 from numerous other runs was not analyzed in the laboratory, but the water-insoluble content was visually determined to be less than 5y0by comparison with other samples. The bulk density of PuCla varied from 2 to 3 grams per cc. The average particle size was found to be 4.1 and 6.1 microns in two determinations by a Fisher sub-sieve sizer. The trichloride powder may be safely stored indefinitely at mom temperature in containers sealed against moisture. Glass jars with neoprene-lined caps were adequate, whereas paper-lined caps or polyethylene jars permitted water vapor to reach the powder. The powder is not stable against oxidation at room temperature if moisture is present. The greenish-blue PuCls turns to the bright-blue
Table VI.
Typical Impurities“ in Chloride Process Show Progressive Changes and Product Quality Impurity, Parts/Million Parts Pu Element Nitrate Oxide Chloride Metal Ag A1
As B Be Bi Ca
Cd Cr cu Fe
Ge
IC La Li Mg Mo Mn
Na Ni
P Pb Si Sn
T1
V Zn
Total
