I
ROBERT LEE MOORE, C. A. GOODALL,
J. L.
HEPWORTH, and R. A. WATTS, Jr.
Hanford Atomic Products Operation, General Electric Co., Richland, Wash.
Nitric Acid Dissolution of Thorium Kinetics of FI u o ride - Catalyzed Reaction In processing spent reactor fuel, reaction rates for dissolving thorium can be correlated with variables by an empirical equation
THE
f
dissolution of thorium metal, or of thorium oxide, in nitric acid is the first step in the recovery of metallurgical scrap or in the chemical reprocessing of blanket material from breeder-type power reactors (7). Unlike uranium, however, thorium does not react rapidly even with boiling concentrated nitric acid. Thus, the dissolving rate in 60% acid is approximately 1 mg. per hour per sq. cm., over a thousandfold less than that of uranium under the same circumstances. However, the reaction is strongly catalyzed by small concentrations of fluoride ion, a n effect discovered by Schuler, Steahly, and Stoughton ( 2 , 4 ) ,and advantage has been taken of this fact in various processing schemes (7). The present investigation was undertaken to obtain quantitative information on the rate of the reaction, to determine the effect of the variables involved, and to develop correlations which would aid general application. Experimental Two types of experiments were performed : measuring the instantaneous penetration rates and integral dissolvings, the latter simulating a batch-type dissolving operation. Penetration rates were measured by the procedure used in earlier studies of uranium dissolution. Thorium wafers, slightly over 1 inch in diameter and approximately 1 / ~ inch thick, were immersed for periods of 2 to 4 minutes in a 250-ml. boiling solution of the desired composition in a stainless steel beaker. After removal from the
dissolving medium the wafers were washed immediately in water and then in acetone and the weight loss and final dimensions determined. The specific penetration rate, in terms of weight loss per unit time per unit area (mg. per hour per sq. cm.), was calculated from these data. The pitting effects of the wafer and change in composition of the solution were ignored and the error from these sources was believed to be negligible. A new wafer was used for each measurement. Metal fabricated by both extrusion and rolling was used ; no difference was found in the dissolving rate The integral dissolvings were carried out in a pot-type dissolver equipped with an updraft condenser. The dissolver consisted of a 1-liter stainless steel beaker fitted with a Teflon-gasketed top and heated with a hot plate. The condenser was a water-jacketed stainless steel tube, 30 inches in length and 0.5 inch in diameter. The charge to the dissolver was, in most experiments, a 2-inch length of thorium rod and weighed approximately 500 grams. The initial specific surface was 0.022 square inch per gram. The quantity of nitric acid was based on the empirical equation: Th
+ 6HNOa
P
HYDROFLUORIC ACID (h4)
Figure 1. Penetration rate of thorium and thoria in boiling 13M nitric acid
Th(N0a)r
+
Discussion arid Results The rate of penetration was determined as a function of concentration of fluoride, nitric acid, and dissolved thorium. All experiments were run at the boiling point of the solutions. A few experiments were performed to determine the effect of additives, such as aluminum, and to compare the effective-
-5: v)
hI Br
o/"-O,
Y
W
+ NO2
z 9
4- N O
f-----;/
ci
and was computed so that a 200% heel of unreacted metal would remain on completion of the reaction-i.e., one third of the charge was dissolved. The course of the dissolution was followed by analysis for thorium and nitric acid of 0.5-ml. samples withdrawn as a function of time. The unreacted metal was also periodically removed and weighed as a cross check on the thorium analyses.
3Hz0
10,000
0
Ga
Table 1.
YI
+
All dissolutions were carried out a t the boiling point, approximately 110' C. Because of the extreme vigor and exothermic character of the reaction, external heating was stopped during early stages to prevent boilover. Dilution water was added at the end of the dissolving to make up for loss due to evaporation and to prevent solidification on cooling.
I-
a a I-
W
z
W
a N I T R I C A C I D (h4)
Figure 2. Effect of nitric acid on penetration rates of thorium and uranium
Effect of Fluoride on Penetration Rate of Thorium by Nitric Acid
HF, M
Pen. Rate, Mg./Hour/Sq. Cm.
HF M
0 0.0001 0.0002 0.0005 0.001 0.002 0.0025 0.005
1.3 23 83 230 470 810 1710 3410 3020
0.01 0.05
I
Pen. Rate Mg./Hour/Sq. Cm.
0.1 0.2
0.5 1.0
VOL. 49, NO. 5
5200 8910 7060 10440 11120 9130 6950
MAY 1957
885
leads to a residue of ”blue thorium” in neutron-irradiated material (7). The c thoria used (produced a t Oak Ridge) contained about 1% calcium oxide, had been pressed a t 20,000 pounds per square inch, and sintered at 1800° C. to 96.97’ of the theoretical (crystallographic) density. The results of pene3 tration rate measurements in 13M nitric acid are compared with those of thorium metal in Figure 1 . A white coating, believed to be thorium fluoride or calcium fluoride, was observed on the Figure 3. Typical integral dissolving surface of the thoria pieces for hydroInitially 13M nitric acid and 0.04M fluoride at fluoric acid concentrations of 0.1M about 110’ C. or greater. These results illustrate the reason for the slow dissolution of the thorium oxide impurity and the resultant ness of fluoride salts, such as fluoborate difficulty in dissolving “blue thorium.” and fluosilicate, to hydrofluoric acid. NITRICACID. The effect of nitric acid Effect of Variables on Penetration Rate. FLUORIDECONCENTRATION. concentration. at constant fluoride concentrations of 0.005and O.O5M, is shown The effect of fluoride concentration on the initial rate of attack of thoruim by boiling 0% nitric acid ( 1 3 M ) is shown in Table I and in Figure 1. The disTable 11. Effect of Dissolved Thorium solving rate is nearly linear in fluoride on Thorium Penetration Rates concentration at low catalyst concentraAll solution. 0.05-TI hydrofluoric acid tions. It reaches a maximum at about Rate, 0.1M hydrofluoric acid and decreases at Mg./Hour/ higher concentrations, apparently due to Th, M “Os, *VI S q . Cm. exceeding the solubility product of tho0 13 8000 rium fluoride. The maximum rate of the 0.2 13 5690 catalyzed reaction was approximately 0.5 13 4010 10,0@0times the rate in the absence of 0.72 13 2950 1.0 13 1790 fluoride (1.3 mg. per hour per sq. cm.). Also, for fluoride concentrations greater 0.5 1 110 4 580 than about 0.002M the rate equals or 1390 exceeds the (uncatalyzed) dissolving 3550 10 rate of uranium, which is 1300 mg. per 13 4000 hour per sq. cm. a t this acidity. 14.3 4010 A few measurements also were made with thoria, because of reported difficulties in dissolving the oxide impurity in Figure 2. Dissolution rates of urapresent in most thorium metal. This nium without fluoride catalysis are included for comparison. Initial rates are greater for thorium than for uranium, 100t both for 0.05 and 0.005M fluoride, \I except at the highest acidity. The disTO2MTh solution rate of thorium passes through a tmaximum at about 13M nitric acid. THORIUM AND ALUMINUM. Dissolved thorium markedly decreases the penetration rate. undoubtedly through complexing of the fluoride (Table 11). The experiments were not extended to thorium concentrations greater than 1M because the thorium nitrate solubility is limited at high acidities. Several experiments were also run to determine the effect of aluminum (a potential cladding material) on thorium penetration. Aluminum ion, like thorium, complexes fluoride strongly at room temperature and low ionic strengths. , ’ /,I Aluminum nitrate, 0.5M, in 13M nitric 0.1 ,001 .01 0.I acid, 0.05M hydrofluoric acid solutions gave penetration rates of about 250 mg. HYDROFLUORIC ACID (M) per hour per sq. cm. (mean of four deterFigure 4. Effect of fluoride concentraminations ranging from 91 to 410 mg.) tion of dissolving times us. 8000 mg. in the absence of aluminum. In another experiment, a rate of 890 mg. All initially 13M nitric acid at about 110’ C. 250
7
k
1
886
:\
,
‘
I
INDUSTRIAL AND ENGINEERING CHEMISTRY
was observed in a solution of 7 M nitric acid, O . O j M hydrofluoric acid, 0.5M thorium. and 0.5M aluminum. This may be compared with 1390 mg. for the same composition, less the aluminum. Therefore the reduction of dissolving rate is most dramatic at the start of a dissolving operation, less marked at lower acidity and in the presence of dissolved thorium, which competes with aluminum for available fluoride. Other Fluoride Compounds. Fluoborate and fluosilicate were suggested as substitutes for hydrofluoric acid on the theory that they might be less corrosive to materials of construction. Fluosilicate yielded a precipitate (presumably silica) in the presence of excess dissolved thorium? and the penetration rates suggested that it was largely hydrolyzed to free fluoride. Fluoboric acid also appeared to be completely hydrolyzed or dissociated in boiling nitric acid solutions. Thus, the catalytic effect of 0.01M fluoboric acid was identical to that of 0.04M hydrofluoric for nitric acid concentrations from 2 to 15.7M. Accordingly, the corrosion problem was assumed to be the same as with fluoride. Integral Dissolvings. Because of the complexities of the thorium-hydrofluoric acid-nitric acid system and the interrelated effects of the various variables, it is difficult to estimate the time cycles of a large scale dissolving from instantaneous penetration rate measurements. In most pot-type “integral” dissolvings, a 200% heel was employed, all of the nitric acid was added at the beginning, and all of the dilution water at the end of the dissolution. In a hypothetical plant operation, however, the acid would probably be added in several stages (for better control during the vigorous first phase of the reaction) and the dilution water prior to termination of the dissolving to prevent solidification on cooling. The “equilibrium heel” accumulated in a dissolver would also probably have a larger specific surface than in the laboratory dissolvings, where the pieces were used only once, and this would result in somewhat more favorable rates than predicted from the laboratory results. Also, the boundary conditions may have a marked effect on dissolving rate and time cycles, particularly in the adverse effect of dissolved thorium on the fluoride catalysis. Two series of experiments were carried out. In one. initial nitric acid concentration was held constant at 13M and fluoride concentration was varied from 0.005 to 0.075M. In the other, catalyst concentration was constant at 0.04M and initial nitric acid varied from 8 to 16-44. Results of a typical dissolving are shown in Figure 3. The initial rate is very rapid, but it decreases markedly after the solution has become about 1M in dissolved thorium. This takes only about 30 minutes, whereas over 6 hours
THORIUM D I S S O L U T I O N computed using in each case the experimental values of the latter quantity. Inferior correlations were obtained when volume concentrations were substituted for weight losses. A logarithmic plot of the slopes (or pseudo rate constants) is shown in Figure 6. The rate constant is first order in catalyst concentration and appears to be about second order in initial nitric acid concentration. The k values can thus be calculated by
0.04
0.03
-?
0.02
rs\
-0.01
Oil ' ' ' I
' '
' '
5
Figure 5.
' ' IO
I
' ' '
'
'
' ' ' '
'
25
' ' ' ' I 30
Correlation of integral dissolving data
are required to reach the crossover point, at which the molar concentrations of dissolved thorium and nitric acid are equal. The weight loss curve, rather than thorium molarity, actually indicates more accurately the extent of dissolution, since the solution volume decreases with time as a consequence of nitric acid consumption and the loss of some water and acid around the condenser. I n the variable fluoride series, the time intervals required to dissolve to thorium concentrations of l M , 2M, or to a 200% heel are shown by the parallel lines in Figure 4,which also shows the effects of fluoride concentration and the time penalty for dissolving to high final thorium concentrations. The effect of initial nitric acid concentration was less dramatic than that of fluoride, because the initial acidity was varied by only a factor of 2 (8 to 1 6 M ) and because of dilution by extra water present in low acid solutions. Lower thorium concentrations (for the same weight dissolved) would result and thus minimize the complexing of fluoride.
INITIAL NITRIC ACID (M)
8 Dm
100
'
15 20 TIME IN HOURS
I
I
!
Mathematical Correlation of Batch Dissolving Data. Thorium dissolution is complex and would not lend itself readily to rigorous mathematical treatment. Thus hydrofluoric acid is relatively weak ( K = 7.2 X l o 4 a t room temperature) (3) and therefore is present largely as undissociated hydrofluoric acid in the presence of excess nitric acid. Thorium forms a series of very stable fluoride complexes and inhibits fluoride catalysis. The dissolver solutions are also concentrated, and the laws of dilute solutions probably do not hold, nor are the activity coefficients available. For these reasons, the data for the various dissolvings (Figure 3) failed to fit any simple idealized reaction rate equation. However, the weight loss curves could be reproduced by assuming that the reaction behaved as if it were second order in the number of moles of nitric acid present (contrasted to molar concentration). This is equivalent to d- W = dt
k(W,
-
W)Z
where W a is the total number of grams of thorium dissolved by acid charged to the dissolver-i.e., weight loss a t infinite time-and W is weight loss a t time t. Constant k includes the surface area, which varied little in these experiments. I t is also a function of catalyst concentration and the initial concentration of acid. Integration of Equation 1 yields
m
0 X
X
,001
C)I
0.1
HYDROFLUORIC ACID (MI
Figure 6. Correlation of rate constants 0. Constant initial nitric acid, varying fluoride
0.
Constant fluoride, varying initial nitric acid
where a plot of l/(W,,, - W ) vs. time should yield a straight line of slope k. Figure 5 shows a correlation for the series of dissolvings with variable fluoride. Similar results were obtained with constant fluoride and variable initial nitric acid. The fit is rather sensitive to the value chosen for W ----Le., the value assumed for the number of moles of nitric acid consumed per mole of thorium dissolved-and the points plotted were
where HF is the molar concentration of fluoride catalyst, HNOs is the initial molar concentration of nitric acid, A is the area of metal exposed to the solution, and NHNO, is the number of moles of nitric acid charged to the dissolver. The factor, A/PHNO,, corrects for the effect of specific area and for the gross weight losses, rather than concentrations, used in Equations 1 and 2. The constant, K , has the approximate value of 2.2 X lo4 when A is measured in square centimeters and time in hours. Alternatively, Equation 3 may be written k =
K'A (HF) ("Os)' Wm2
(4)
in which case K' has the value of about 0.36 if W is in grams. T h e instantaneous penetration rates at different stages of a dissolution can be obtained from Equations 1 and 3 or 1 and 4 by solving for dW/Adt. Thus, effective penetration rates can be obtained for solution compositions which would be difficult to duplicate synthetically because room temperature solubility is limited. literature Cited
(1) Gresky, A. T., "Solvent Extraction Separation of U-233 and Thorium from Fission Products by Means of Tributyl Phosphate," Geneva Atomic Conference, Paper 540, July 20, 1955. ( 2 ) Schuler, F. W., Steahly, F. L., Stoughton, R. W., "Production and Separation of U-233-Collected Papers," NNES-IV-l7B, pp. 346-56 (1952). (3) Sidgwick, N. V., "The Chemical Elements and Their Compounds," p. 1106, Oxford Univ. Press, hTew York, 1950. (4) Steahly, F. L., Stoughton, R . W. (to Atomic Energy Commission), U. S. Patent 2,546,933 (March 27, 1951). RECEIVED for review June 23, 1956 ACCEPTED November 13, 1956
Northwest Regional Meeting, ACS, Seattle, Wash., June 12, 1956. VOL. 49, NO. 5
M A Y 1957
887
