Downloaded by UNIV OF NEW SOUTH WALES on August 26, 2015 | http://pubs.acs.org Publication Date: September 1, 1980 | doi: 10.1021/i360075a033
Ind. Eng. Chem. Prod. Res. Dev. 1980, 19, 459-467
almost linearly as a function of percent nitrogen removal when doctored raw anthracene oil is used as a feedstock. Catalysts made from Ketjen 007-1.5E untreated with steam and those made from the same support with steam treating are identical in every respect except for the physical properties of the catalysts. Steam treating shifted the most frequent pore radius from 33 to 38 A and reduced the surface area from 291 to 231 m2/g. When undoctored raw anthracene oil was hydrotreated over these two catalysts, the differences in hydrodesulfurization activity were negligible (Table IV). But the reduction in surface area did lead to a reduction of the denitrogenation activity (62 vs. 48% removal at 1 hour space time) (Sivasubramanian and Crynes, 1979). However, as can be seen from Figure 6, addition of quinoline to raw anthracene oil changes the relative activities of tbese two catalysts. Kirsch et al. (1959) found that there is a minimum concentration a t which basic nitrogen has an immediate effect on selectivity towards hydrodesulfurization of petroleum fractions. Hence it seems probable that the nitrogen concentration in raw anthracene oil may be below the minimum concent,ration required to have an effect on selectivity. But with the addition of quinoline the concentration of basic nitrogen in the feedstock increases, and, since the catalyst made from steam treated support is less active for nitrogen removal than the catalyst made from untreated support, that leaves large quantities of basic nitrogen which may completely block the first type of sites on the catalyst. This may explain why the steam-treated catalyst was affected more than the other two catalysts by increasing nitrogen concentrations. However, when only raw anthracene oil was used as a feedstock, the basic ni-
459
trogen concentration may not have been large enough to completely block the sites of the first type. It is known that denitrogenation is more difficult than desulfurization and that an active denitrogenation catalyst has always been a very active desulfurization catalyst. Results presented here show that one of the reasons for this may be due to the fact that an active denitrogenation catalyst leaves less basic nitrogen to block the active sites and hence the desulfurization activity remains high. The catalysts used in this study were Cc-Mo-alumina catalysts. Since nickel catalysts are known to be better denitrogenation catalysts, studies should be done on nickel catalysts, and other studies should be conducted with varying concentration and conversion levels of nitrogen in order to determine the minimum concentrations at which nitrogen begins to affect the selectivity of hydrodesulfurization. Acknowledgment
We acknowledge the generous support of the U S . Department of Energy, Fossil Energy, and Oklahoma State University that made this work possible. Literature Cited Ahuja, S . P., Derrien, M. L., LePage, J. F., Ind. Eng. Chem. Prod. Res. Dev., 9, 272 (1970). Doleman, J., WorM Pet. Congr., Proc., 6th, 3, 247 (1963). Kiovsky, J. R., Berg, G. H., paper presented at the 52nd National AIChE Meeting, Memphis, TN, 1964. Kirsch, F. W., Shalit, H., Heineman, H., Ind. Eng. Chem., 51, 1379 (1959). Satterfield, C. N., Mcdell, M., Mayer, J. F., AIChEJ.. 21, 1100 (1975). Sivasubramanian, R., Crynes, B. L., Ind. Eng. Chem. Prod. Res. Dev., 18, 179 (1979). Sooter, M. C., Ph.D. Thesis, Oklahoma State University, Stillwater, OK, 1974.
Received for review M a y 7, 1979 Accepted April 17, 1980
Chemical Flowsheet Conditions for Preparing Urania Spheres by Internal Gelation Paul A,. Haas," John M. Begovich, Allen D. Ryon, and John S. Vavruska Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830
Small, ceramic urania spheres can be prepared for use as nuclear fuel by internal chemical gelation of uranyl nitrate solution droplets. Acid-deficient uranyl nitrate solutions up to 3.4 M in uranium with N03-/U mole ratios of 1.5 to 1.7 are prepared by dissolution of U308or U03. Decomposition of hexamethylenetetraminedissolved in the uranyl nitrate solution releases ammonia to precipitate hydrated U 0 3 . Previously established flowsheet conditions have been irnproved and modified at ORNL and have been applied to prepare dense U02 spheres with average diameters of 1200, 300, and 30 p m . The 1200- and 300-pm U02 spheres were prepared by gelation in trichloroethylene at 50 to 65 OC; 2-ethyl-1-hexanol was used as the gelation medium to prepare 30-pm U 0 2 spheres.
Introduction
As part of a nuclear fuel refabrication program, 50100-kg quantities of both depleted and enriched UOz spheres were recently prepared for use in fabrication development and irradiation tests. Three sizes of spheres having diameter ratios of about 40:lO:l were necessary for a packed-particle (Sphere-Pac) fabrication concept. The intermediate, 300-pm diameter, spheres are easy to prepare; thus the process selection was determined by the need for 1200- and 30-pm products. 0 196-432 1/80/12 19-0459$01 .OO/O
The schematic flowsheet for the gel-sphere process is shown as Figure 1. In this process, decomposition of hexamethylenetetramine (HMTA) dissolved in uranyl nitrate solutions releases ammonia to precipitate hydrated UO,. The internal gelation process was first developed and described for the production of nuclear fuels at KEMA by van der Brugghen et al. (1970) and Kanij et al. (1974) (Keuring van Electrotechnische Materialen at Arnhem in the Netherlands). The conditions now being used at ORNL show important differences from and improve0 1980 American
Chemical Society
w-4 HNO~---+
ADUN PREPARATION
HMTA MAKEUP
-
"20
venient and stable for storage and metering a t ambient temperatures. The acid-deficient uranyl nitrate (ADUN) solution should have N03-/U mole ratios of 1.5 to 1.7 and uranium concentrations of 12.9 M. Since U02(N03)2is equivalent to a N03-/U mole ratio of 2.0, the lower ratios correspond to compositions represented by U02(OH),(N03)2-x,where x is 0.3 to 0.5. The most important reason for using acid-deficient solutions is to promote better gel-sphere properties, but the acid deficiency is also necessary for high uranium concentrations. The solubility of uranium at room temperature increases from 2.4 M at a NO,-/U mole ratio of 2.0, or x = 0 in the chemical formula, to about 3.6 M at x = 0.4. The ADUN solutions are three-component systems which may be represented as mixtures of U 0 3 H N 0 3 + H 2 0 or U 0 3 + U02(N03)2+ H20. From the Gibbs phase rule, an ADUN solution (no solid phase) has two degrees of freedom in addition to temperature and pressure. Therefore, the solution is completely specified if two other independent variables are measured or fixed. We commonly use molar concentration of uranium and N03-/U mole ratio as the two convenient variables to specify an ADUN solution. For practical operation, the density and the pH can be measured to completely determine or specify an ADUN solution composition; however, the solution must be at equilibrium and at a known temperature and pressure. The density of the ADUN solution is determined mainly by the uranium concentration. Botts et al. (1978) used the following equation to correlate experimental measurements for ADUN solutions p(ADUN) = p(H20) + 0.2659[U] + 0.0282[N03] (1) where p(H20) = the density of H 2 0 at the same temperature, g/cm3, [U] = the molarity of uranium, and [NO3] = the molarity of NO3-. For N03-/U mole ratios of 1.5 to 1.7, the effect of NO; can be combined with the uranium concentration to give p(ADUN) = p(Hz0) + 0.31[U] (2)
Downloaded by UNIV OF NEW SOUTH WALES on August 26, 2015 | http://pubs.acs.org Publication Date: September 1, 1980 | doi: 10.1021/i360075a033
+
0.5
AOUEOUS HMTA,
M
Hz
-
SINTERING TO 1600 OC
-
GASES
1
Preparation of ADUN and Feed Broth The feed to the gelation operation should have a long gelation time at a low temperature so premature gelation does not occur and a short gelation time at a higher temperature. The properties of the gel spheres generally improve as the uranium concentration increases. Therefore, the uranyl nitrate and hexamethylenetetramine (HMTA) solutions are prepared to the highest concentrations con-
While the experimental data were obtained for uranium concentrations below 1.4 M, the correlations adequately represent our measurements for higher concentrations and also give the density of U02(N03)2.6H20crystals to within 0.3% of the handbook value. The pH of an ADUN solution is dependent on both the NO72 h) or transfer of the partially dried gel to an oven to steam-dry for removal of the final 10% of the water. These fines showed intermediate densities. The preferred calcining and sintering conditions and typical results are briefly as follows. One product catcher full of dried spheres (gel weight, -800 to 900 g) is loaded into a stainless steel tray to a depth of about 1in. As many as ten of these trays are then heated under flowing argon-4% hydrogen at 100 "C/h to 450 "C for 1 h. This process eliminates most of the residual water, ammonium salts, and organic materials without any significant particle shrinkage or reduction. The particles are then removed from the protective atmosphere and loaded into molybdenum trays, again to a depth of about 1in., and sintered under pure hydrogen in a cold-wall-design, tungsten-element furnace. They are heated rapidly (3 h) to 1600 "C and held for 4 h; then they are cooled at the natural furnace cooldown rate (-3 h). This step completely eliminates the volatile species, reduces the U 0 3 to U02, and sinters the particles to a high-density product. Several kilogram-size batches of U 0 2 of the three standard sizes (i.e., large-coarse, small-coarse, and fines) were prepared by internal gelation and characterized in depth. Each batch of product spheres exhibits chemical compositions
Table IV uranium oxygen/uranium ratio density carbon metallic impurities chlorine, fluorine gas release a t 1600 ' C moisture content
88.2 wt % 1.98 t o 2.00 99+ % of theoretical 4 t o 7 ppm, by weight Q 100 ppm total; < 10 pprn individual < 5 ppm each < 0.085 cm3/g of uranium < 0.004 wt %
and impurity contents well within the ASTM/ANSI specification for sintered U02 pellets, except for gas release and moisture content. The latter parameters are expected to be higher for spheres than for pellets due to the greater surface area of the spheres. Before the spheres are loaded into rods, however, a simple low-temperature vacuum outgassing step will decrease the moisture and gas levels sufficiently to meet specifications. Typical results of the chemical characterization of the UOz batches are listed in Table IV. Summary a n d Conclusions Improvements and modifications to the previously established flowsheet conditions for internal gelation were developed and applied at ORNL to prepare dense UOz spheres with average diameters of 1200,300, and 30 pm. The gel-sphere processes generally require close duplication of process conditions or "recipes". Details of the chemical flowsheet conditions were selected, in part, to control three problems associated with the internal gelation process: (1)the feed solution or broth is temperature-sensitive and must be mixed at temperatures around 0 "C to prevent premature gelation; (2) partially denitrated or acid-deficient metal nitrate feed solutions of high metal concentration are required to produce good gelation behavior; and (3) gelation takes place in a warm, or hot, organic liquid which must be removed from the spheres and recycled. The ADUN solutions are three-component systems which may be represented as mixtures of U03, "OB, and HzO. In the absence of a solid phase, the ADUN solutions have two degrees of freedom in addition to temperature and pressure. We use the molar concentration of uranium and the N03-/U mole ratio as two convenient variables to specify the composition of an ADUN solution. In normal operations, monitoring the density and the pH can provide a convenient means of process control. Solutions with a uranium concentration of about 3.5 M and a N03-/U ratio of 1.6 were consistently prepared by use of excess uranium oxides a t the optimum temperature and nitrate concentration. (Uranium in ADUN has a maximum solubility at intermediate nitrate concentrations and intermediate temperatures.) The rate of gelation and the gel properties depend on the gelation conditions. For the large-coarse (0.4 cm diameter) droplets, the gelation times are approximately the same as the times calculated for heat transfer from the organic liquids to the droplets. In the case of fines, the times calculated for heat transfer are 0.1 s or less, and the gelation times are probably controlled by the reactions of HMTA. The best internal gelation results are for HMTA/N03- mole ratio of about 0.8, in which nitrate preneutralized by addition of NH40H is not included. For continuous mixing of solutions, the uranium and HMTA concentrations must be near the solubility limits to give the desired mixed concentrations. The effect of temperature on gelation is based on the droplet temperature during gelation. The gelation rates become inconveniently slow for a gelation column at tem-
Downloaded by UNIV OF NEW SOUTH WALES on August 26, 2015 | http://pubs.acs.org Publication Date: September 1, 1980 | doi: 10.1021/i360075a033
Ind. Eng. Chem. Prod. Res. Dev., Vol. 19, No. 3, 1980 467
peratures below 50 "C,but 50 "C will give good gel properties. At 70 "C or higher, the characteristics of the gel are less desirable. Therefore, the preferred organic liquid temperatures increase from 50 to 55 "C for fines, where heat transfer is rapid, to about 65 "C for large-coarse droplets. Dehydration of the gel surface by the organic liquid is important for the preparation of fines and also appears to be desirable for larger-size product; however, it can confuse the primary effect of gelation temperature. The preferred forming organic medium for fines is 2EH, because it provides more dehydration of the gel surface then does TCE. The high density and higher interfacial tension of TCE are necessary to keep the 0.4-cm droplets spherical during gelation. The 0.1-cm droplets can be gelled in either 2EH or TCE, but TCE allows easier separation and washing. The minimum wash times and volumes are determined by the requirements for leaching soluble constituents from the gel. Empirically determined wash procedures and measurements for removal of NH4N03from the gel agree with literature models and calculations for stripping ionexchange resins. Wet aging and drying effects are complex and only partly understood, but many variations in the dried gel do
not result in any differences in the sintered UOpspheres. L i t e r a t u r e Cited Botts, J. L., R a r i i n , R. J., Costanzo, D. A,, "Density, Acidity and Conductivity Measurements of Uranyl Nitrate-Nbic Acid Solutions", ORNL/TM-6491 (Oct 1978). b a s , P. A., Begovich, J. M., Ryon, A. D., Vavruska, J. S., Chemical Flowsheet Conditions for Preparing Uranla Spheres by Internal Gelation, ORNL/TM6850 (July 1979). Haas, P. A., Am. Ceram. SOC. Bull., 58(9), 873 (1979). Kanij, J. B. W.. Noothout, A. J., Votocek. O., "The KEMA U(V1) Process for the Production of UO, Microspheres," pp 185-95 in "SoCGeI Processes for Fuel Fabrication", IAEA-161, International Atomic Energy Agency, Vienna, 1974. Rosen, J. B., J . Chern. Phys., 20, 387 (1952). Rosen, J. B., Ind. Eng. Chem., 45(8), 1590 (1954). Schneider, P. J., "Temperature Response Charts", Wiley, New York, 1983. van der Brugghen, F. W., et al., "A U(V1) Process for Microsphere Production", pp 253-63 In "Symposium on Sol-Gel Processes and Reactor Fuel Cycles", CONF-700502 (May 1970).
Received for review January 28, 1980 Accepted May 1, 1980 Presented at 34th Northwest Regional Meeting of the American Chemical Society, Richland, Wash., June 13-15,1979. Research sponsored by the Nuclear Power Development Division, U S . Department of Energy, under Contract W-7405-eng-26 with the Union Carbide Corporation and Purchase Order 64357 AG from Battelle-Pacific Northwest Laboratories.
