I
J. R. FLANARY and J. H. GOODE O a k Ridge National Laboratory, Oak Ridge, Tenn.
Recovery of Neptunium-237 from Process Residues by Solvent Extraction I
This process extends the use of solvent extraction with tributyl phosphate to recovery and separation of neptunium isotopes from uranium process residues
EARLY
in 1955 several thousand pounds of process residues became available, which contained 0.05 to 75 grams of neptunium per ton. This material is a source of the alpha emitter, neptunium-237 ( 6 ) . The general chemistry of neptunium and the recovery of trace amounts of its 237 isotope have been investigated by a number of workers (7-5, 7-27). T h e cost of thenoyltrifluoroacetone and tributylamine rules them out for largescale operations. The main disadvantage of hexone is that neptunium must be oxidized to neptunium(V1) before extraction; this also increases the extraction of ruthenium and plutonium and contaminates the product stream. The similarity between the 20% tributyl phosphate (77) and the O R N L 25-TBP process (5) suggested that the 25-TBP process might be modified for recovering neptunium and uranium from the residues. Neptunium(V1) is the form most readily extracted by tributyl phosphate (76), but is difficult to hold in the sexivalent state. It can be extracted when excess nitrite ion is present (27). Because some of the oxidants required, such as 0.1M bromate, may also be extracted by tributyl phosphate, 0.01 to 0.02M ferrous sulfamate, was added to the aqueous feeds and scrubs to keep the neptunium in the quadrivalent state.
Experimental Feed Preparation. Dissolution of the residue (Table I) in various reagents, including nitric acid, aluminum nitrate, and dibasic aluminum nitrate, was studied. Boiling aluminum nitrate-nitric acid mixtures appeared most satisfactory from the standpoint of solubility, chemical costs, and ease of handling. T h e simplest procedure consisted in adding 1 kg. of residue to 13 liters o f a mixture containing 1.8M aluminum nitrate and 1 M nitric acid, mixing, and digesting at the boiling point under full reflux for 2 to 4 hours. The digestion coagulated and dehydrated silica in the feed, so that emulsions and column operational difficulties were minimized or eliminated. Lower concentrations of aluminum nitrate and nitric acid could be used, and the final concentrations, 1.8M aluminum nitrate and 1M nitric acid for a uranium concentration of 50 grams per liter, obtained by an evaporation-digestion step. I
IA (Extraction-Scrubbing) Column Conditions. CHOICE OF SOLVENT CONBecause the dissolution procedure is expected to yield a feed containing approximately 50 grams of uranium per liter, 15 volume % of tributyl phosphate in Amsco 125-82 was chosen as an optimum solvent concentration CENTRATION.
based on the complexing power of tributyl phosphate for uranium and satisfactory organic-aqueous flow ratios for smooth column operation.
Flowsheet for Recovery of Neptunium. Uranium and neptunium are recovered in a battery of three pulsed columns arranged in cascade. I n the first (A) column (shown a t left in flowsheet) they are coextracted (four theoretical stages) from the feed with 15Yo tributyl phosphate and scrubbed with aluminum nitrate-nitric acid solution to remove extracted impurities. The organic phase then flows to a second (B) column (center), where neptunium is preferentially
Table 1.
Chemical Analyses of Initial Samples of Residues Sample Wt. %
A B 29.50 46.70 Fe 8.55 0.09 F 35.54 24.61 Ni 0.87 0.79 Ca 2.52 1.91 Mg 3.28 2.37 Na 2.06 2.57 82 .3a 79.0a PO4, p.p.m. 330 225 SiOz, p.p.m. 524 240 Ng3', g./ton 12.7 34.5 a Remainder probably oxygen and water.
Component
U
-
RESIDUE
u
650~
Literature Background Subject General chemistry of Np and recovery of trace amounts of 237 isotope 30 to 40 grams isolated from radioactive wastes and byproduct streams 20% tributyl phosphate solvent in kerosine-type diluent extracts Np and U from Pu in nitrate solutions Extracting agents Ether Thenoyltrifluoroacetone dissolved in chloroform or other inert solvent Tributylamine in hexone Hexone Ketones and glycol ethers
U
Ref.
STEP
21"Ol. -
OOlM
HNO,
(1-5, 7-21)
(14-18)
(11)
(9,18, 20) ( 1 , 9,12)
(4) (7,9) (9,10)
Flowsheet for recovery of neptunium A stable nanemuisifying feed solution is produced of high nitrate salting strength and maximum
uranium and neptunium concentrations \
VOL.61, NO. 1
JANUARY 1959
55
contain 40 grams per liter. As percentage saturation of the solvent with uranium increases, the uranium extraction coefficient slo\vly decreases. More Figure 1. A pulsed important, the neptunium(1V) extraccolumn 3 / 4 inch in tion coefficient decreases very rapidly. diameter was arThis is expected because, as the perranged for solvent centage uranium saturation increases, extraction studies less uncomplexed tributyl phosphate remains for neptunium extraction. A, 8 , C. Column designaThe equilibration data were obtained tion 1-6. Process head with synthetic feed spiked with neptutanks nium-237 tracer, containing 50 grams of 11-13. P r o d u c t reuranium per liter, 1.8M aluminum niceivers trate, 1M nitric acid, and 0.01.M ferrous 14-16. Remote heads of pulse gensulfamate. Aliquots (2.5 ml.) were erators batched-extracted Ivith 20 to 50 ml. of 22-25. Pulse generasolvent. The uranium and neptuniumtors. Interface (KV)distribution coefficients were calcontrollers a r e on wall ( r i g h f ) culated from analyses of the organic and control and aqueous phases after equilibration. valves a r e near IB (Partitioning) Column Condibottom of coltions. A series of batch equilibrations umns. Process pumps also a p (Table 111) indicated that an aqueous p e a r ( l e f t and phase concentration of 1,VI nitrate right) should be near optimum for separation of neptunium and uranium. At this concentration the neptunium distribution coefficient is low, 0.03, promoting transfer to the aqueous phase, while that of uranium is high, 1.5, which is favorable for extraction of uranium by the to approach a steady state in solventextraction systems involving little reflux. solvent phase. The aqueous strip volor of the EFFECT OF URANIUMSATURATION. ume can be as low as organic phase volume, permitting about To define A column extraction condithreefold concentration of neptunium. tions more closely, it was necessary to Five volume changes were sufficient for determine the effect on neptunium(1V) close approach to steady state. After extraction of percentage saturation of the solbent by uranium. Batch equilifour stripping stages the neptunium loss was 1.170, and after four back-extracbration data (Table 11) showed that tion stages the aqueous neptunium prodneptunium(I\') extraction is favorable up to, hut not appreciably above, 6070 uct contained 0.01 gram of uranium uranium saturation. At 6070 uranium per liter, showing a separation factor of 2.1 X lo4. Under these conditions. s2turation the neptunium(I\') distribuwith residue containing 12 grams of tion coefficient is 1.7. From experience neptunium per ton, the product LT,'Np \vith 15% tributyl phosphate systems, 69% uranium saturation of the solvent weight ratio would be about 311, and further processing-e.g., by solvent exis known to be satisfactory for uranium traction or ion exchange-would be extraction. needed to complete the separation. A 15 volume yo solution of tributyl phosphate containing 62 5 grams of d4queous feed containing 50 grams of uranium per liter, 1.8M aluminum niuranium per liter is saturated with uranium; a 60% saturated solution would
stripped (four theoretical stages) from the solvent with 0.541 nitric acid. Finally, the organic phase flows to a third (C) column (right) where uranium is stripped with 0.01M nitric acid (five theoretical stages), These conditions were tested in laboratory-scale batch countercurrent experiments and found highly satisfactory. Only two extraction stages were needed to reduce the uranium extraction loss to 0.01%. For neptunium(IV), the extraction loss after four stages was 0.77,. Uranium and neptunium reflux values in the scrub section Tvere about 1 and 2yG',? respectively. The feed was prepared from the .residue and spiked with neptunium-239 tracer. Four extraction and four scrub stages were used, and the operation was carried through five volume changes. These conditions are normally sufficient
Table
II.
Neptunium(1V) Extraction Is Favorable up to 6070 Uranium Saturation
Aqueous feed. 50 grams Ulliter, 1.8.U Al(KOa)3, 1 X HXOa, 0.0151 Fe(NH?SOz)?,1.8 X lo7 Npe3gyc./m./ml. Solvent. 15YGTBP in Ainsco 125-82 25-ml. aliquots of feed equilibrated 3 minute- with 20-50 ml. of bolvent
L-
Solvent-Aqueous Volunie Ratio 0.8 1.0 1.2 1.4 1.6 1.8 2.0 a
56
Satn. of Solvent,
% 86.5 72 61 54 50 45.5 36
U
Conci?. in Solvent, G./Liter 54 45 38 34 31.3 28.5 22.5
Distribution Coeff., Eaaa U NP 5.7 0.15 22.5 0.72 54 1.93 3.81 103 130 5.64 158 6.55 187 10.5
-
Ratio of solute concentration in solvent to that in aqueous phase, at equilibrium
IMDUSTRIAL AND ENGINEERING CHEMISTRY
Table 111. An Aqueous Phase Concentration of 1M Nitrate Is Optimum for Separation of Neptunium and Uranium B column organic phase.
150/; TRP in Amsco 125 82, 28 grams U/liter, 0.08M "03, Np239tracer Organic-aqueous volunie ratio. 6:1 Equilibration time. 3 minute.5 "03 in Strip S o h , M
0.1 0.5 1
3
5
Distribution Coeff. U Sp 0.8 1.1 1.5 4.9 7.5
0.02 0.03 0.05 0.2 0.4
Sepn. Factor (U D.C.1 Np D.C.) 40 37 30 24 19
__
. __”
-
NUCLEAR T E C H N O L O G Y trate, and 1 M nitric acid, was spiked with neptunium-239, reduced with ferrous sulfamate, and extracted with 1.5 volumes of 15% tributyl phosphate. Neptunium seems to remain in the quadrivalent state in the B column, and no holding reductant is used. The organic phase, containing 35 grams of uranium per liter, 0.01M nitric acid, and neptunium(IV), was blended with 0.3 volume of additional solvent to simulate the organic phase . a t the feed stage. Aliquots of the organic phase were equilibrated for 3 minutes with i/6 volume of aqueous neptunium strip solution, 0.1 to 5M in nitric acid. Organic and aqueous phases were analyzed for uranium, neptunium, and nitric acid. IC (Stripping) Column Conditions. After the neptunium is stripped from the solvent stream, the solvent is cascaded into the C column, where the uranium is stripped into 0.01M nitric acid. The stripped solvent is reconditioned by washing with dilute sodium carbonate and nitric acid and returned to process.
Pulsed-Column Flowsheet Demonstrations
A. number of test runs were made with feed prepared from the residues inch diameter pulsed column in ”4 equipment to demonstrate flowsheets developed in the laboratory in batch countercurrent experiments. In initial test runs nitric acid carryover from the extraction column by the solvent to the partitioning column failed to provide sufficient nitrate ion salting strength for adequate uranium re-extraction, resulting in excessive uranium contamination of the neptunium product. The acidity of the A column scrub was increased from 0.5M to 2 M nitric acid to achieve optimum concentration of 0.18M nitric acid in the solvent from the A column and 1M nitric acid in the neptunium product. Under these conditions the uranium concentration in the neptunium product averaged about 0.040 mg. per ml. with 6 feet of backextraction height. Batch countercurrent data showed. that the uranium concentration of this stream could be
-
Table IV.
-
reduced to 0.01 mg. per ml. with four theoretical stages. I n an extended run, 6 to 12 hours was required to bring neptunium to a steady state; uranium reached equilibrium in about 4 hours. After 12 hours of operation, the neptunium material balance (feed to product) was >99%. The uranium-neptunium separation factor was 2.4 X lo3 under the following conditions. Feed, 100 volume. 50 g. U/liter, 1.8M Al(N03)8, 1 M “ 0 3 , 0.01M Fe(NH2SOs)r Scrub, 20 vol me. 0.’5M Al(NOI)?, 2M “ 0 3 , 0.0”M Fe(NHpS03)Z Extractant, 150 volume. 15% TBP Back-extractant, 60 volume. 15% TBP Np strip, 40 volume. 0.5M “ 0 3 U strip, 210 volume. 0.01M HNOl Uranium losses were low, totaling 0.023%. Table IV shows typical analytical data from flowing stream samples taken after 12 and 24 hours of operation. Column operation also proved that a feed digestion step was necessary to prevent intolerable interfacial emulsions in all three columns. Undigested feed often resulted in column shutdowns, owing to failure of interface controls after only a few hours’ operation. Digestion for 2 to 4 hours coagulated and dehydrated siliceous materials so that only the usual small amounts of dense solids formed a t the interfaces. With digested feed runs as long as 30 hours were made. Approximately 85 mg. of neptunium237 was recovered and purified in the process demonstration runs along with 4.55 kg. of uranium. The average flowing stream loss of uranium was 0.02%; neptunium loss was below the limit of analytical detection.
Discussion and Conclusions A solvent extraction process, utilizing tributyl phosphate as the extractant, was developed and proved on a semiworks scale for the separation of neptunium from processing residues. This extends the usefulness of tributyl phosphate as a solvent for the recovery of the heavy elements, adding neptunium to thorium, uranium, and plutonium. Standard laboratory techniques of batch equilibrations and batch countercurrent
Extended Column Operation for Recovery of Neptunium from Process Residues Shows Process Is Efficient Composition -12 Hours 24 Hours U, “01, NP, U, “08, NP , Flowing Stream mg./ml. M c./m./ml. mg./ml. M c./m./ml. Organic from A column 29.4 0.16 532 29.7 0.18 ... Aqueousfrom A column 0.0015 0,99 0 0.0009 0.99 0 Np product from B column 0.024 0.95 1482 0.060 0.89 1497 U product from C column 13.7 0.08 11 16.6 0.07 0 0.001
