— L 3 445k 035822 5 ORNI/TM-11955 ATION APPLICABLE TO THE REACTION OF URANIUM OXIDES WITH CHI.ORINE TO PREPARE URANIUM TETRACHLORIDE P. A Haas February 1992 L. 5, - iy ~ -J.-».nt' i i-\ Bodpn amanem s Thig rzpor has hasn rearkinsd ; ¢ oM ing noat avaush ooy aehie aad Techn Availtztz to DCL ana GOE ca! '“’r-‘f.'-ai;(m, i .O. Box 62, Oak Hicdos, TN 37831 pricas avaisi's 7 am (S15) auttc from the Nate~s! Yechrical - =f Comrnegioe, BIRS Dori foyal Rd e an 2oocunt of work gponsoral Mlaitter the Lnitd Siates ~ She - ] -;:-.".:.“.5 n0r s o Ny o ot OF represstit et i u3T WD not HyOneS g e Sarein 10 any SO0 COMMmSTTia n;oduct, trade nams, trsdomark, manufacturac. or other=izg, doca o - imply ite endoirsemant, rocommsndation, or favoning Foay thereot, Ing do not necoeaority state o i of any Acancy thoraod, — | = — ey any gy - 289 B ORNL/TM-11955 Dist. Category UC-501 (Chemistry) Chemical Technology Division LITERATURE INFORMATION APPLICABLE TO THE REACTION OF URANIUM OXIDES WITH CHLORINE TO PREPARE URANIUM TETRACHLORIDE Paul A. Haas Date Published: February 1992 Prepared by the OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37831 managed by MARTIN MARIETTA ENERGY SYSTEMS, INC. for the U. S. DEPARTMENT OF ENERGY under contract DE-AC05-840R21400 A TMAIE T SN 151 e R f I 4450 D35280a o CONTENTS ABSTRACT . . ettt e e e e e e e e e e e e e e e et et e 1 INTRODUCTION .............. B 1 LITERATURE INFORMATION ..... e PR 3 21 THERMOCHEMICAL DATA . ..ottt e e aeeeaaenns 3 2.2 PHASE DIAGRAMS ...... e 5 23 EXPERIMENTAL RESULTS FOR CHLORINATION OF URANIUM OXIDES . . e e et et e e e e e e e e e e e e e 14 DISCUSSIONS AND CONCLUSIONS . ...\ tutitiiiietiieararaaeannns. 17 3.1 THERMOCHEMICAL CONSIDERATIONS . .. ..vutinenearanannnnn. 17 32 SELECTION OF CHLORINATION CONDITIONS ..........c.cv.n.. 22 33 SALT PROPERTIES AND CONTROL CONSIDERATIONS ........... 24 3.4 PREPARATION OF TiCl,, ZtCl,, SiCl,, and ThCl, FROM 10). 1) X 27 ACKNOWLEDGEMENT . .\ oottt e e e e e e e 28 REFERENCES . . vttt e et e e e e e e e e e e e e 29 T LIST OF TABLES Physical properties of U-O-Clcompounds ........... ... ... .. ... ... ... 4 Thermochemical data . ... ... ... .. . i i i e 6 Vapor pressure CQUAtioNS ... itn oo s tnnnonecrtonanneeeenannnnns 8 Heats of formation for U-O-Cl compounds at 298 K .............. ... ... ... 18 Free energies of formation for U-O-Cl compounds at 900 K (627°C) .......... 19 LIST OF FIGURES Vapor pressures of uranium chlorides ............ .. ... i i, 7 Phase diagram for UCI,-UQ, ... ... . . i e 9 Phasc diagrams for the compounds of uranium and chlorine ................ 10 Liquidus temperatures (°C) for UCI-MgCl,-NaCl ............... ... ... ... 12 Diagram for log P vs log PC12 at400°C . ... 13 Conversion reactions for U-O-Cl compounds and free energies at 900 K ........ 20 Solubility of UQ, in UCl,and UF, ...... ... .. 25 iv LITERATURE INFORMATION APPLICABLE TO THE REACTION OF URANIUM OXIDES WITH CHLORINE TO PREPARE URANIUM TETRACHLORIDE Paul A. Haas ABSTRACT The reactions of uranium oxides and chlorine to prepare anhydrous uranium tetrachloride (UCl,) are important to more economical prepara- tion of uranium metal. The most practical reactions require carbon or carbon monoxide (CO) to give CO or carbon dioxide (CO,) as waste gases. The chemistry of U-O-Cl compounds is very complex with valances of 3, 4, 5, and 6 and with stable oxychlorides. Literature was reviewed to collect thermochemical data, phase equilibrium information, and results of experi- mental studies. Calculations using thermodynamic data can identify the probable reactions, but the results are uncertain. All the U-O-Cl compounds have large free energies of formation and the calculations give uncertain small differences of large numbers. The phase diagram for UCI,- UO, shows a reaction to form uranium oxychloride (UOCI,) that has a good solubility in molten UCl,. This appears more favorable to good rates of reaction than reaction of solids and gases. There is limited information on U-O-Cl salt properties. Information on the preparation of titanium, zirconium, silicon, and thorium tetrachlorides (TiCl,, ZrCl,, SiCl,, ThCl,) by reaction of oxides with chlorine (Cl,) and carbon has application to the preparation of UCI,. 1. INTRODUCTION "An anhydrous UCI, salt has the properties to be an important intermediate chemical for processing and applications of uranium compounds. This was recognized during the World War II program to prepare nuclear weapons; preparation of UCl, was studied at that time. These studies showed that uranium tetrafluoride (UF,) was much easier to prepare and handle than UCl,. Also, the uranium fluorides were better than chlorides for gaseous diffusion separation of isotopes and for batch, bomb reductions to uranium metal. Therefore, most of the uranium that has been mined, concentrated, and purified to give uranium ore concentrates (uranium oxides) has been converted to UF,. The use of uranium chlorides in place of uranium fluorides would have important economic advantages. The hydrogen fluoride (HF) and fluorine (F,) required to prepare the uranium fluorides are expensive chemicals. Processes that use fluorine compounds end up with toxic and troublesome wastes, such as magnesium and calcium fluoride (MgF, and CaF,) and isotopically depleted uranium hexafluoride (UF). If uranium chlorides were used, the recycle or reuse of the chlorides is more practical. The proposed installation of a new industry for enrichment of uranium isotopes could benefit from the economic advantages of using uranium chlorides. The feed to an Atomic Vapor Laser Isotope Separation (AVLIS) process will be uranium metal.! The principal production of uranium metal for nuclear fuel cycles has previously been by batch metallothermic reductions of UF, using magnesium or calcium metal. For a large enrich- ment plant (=10* ton Ufyear), the costs of the HF feed, the calcium (Ca) or magnesium (Mg) feed, and the disposal of MgF, or CaF, waste are major parts of the total uranium enrichment costs. Some alternate processes for preparation of uranium metal from UCl, allow recycle of Cl, from electrolytic cells. The application of these processes requires the reaction of uranium oxides with Cl, to prepare UCl,. The purpose and scope of this review is to collect, organize, and discuss the litera- ture information useful to the reactions of uranium oxides and chlorine to prepare anhydrous UCl,. The review is selective in that only one set of consistent and useful results is presented without reference to less useful or inconsistent information. An excellent comprehensive and critical review of the chemistry of uranium was prepared as an account of work and information from the U. S. Manhattan Project.? Such a review for the preparation of UCI, will not be repeated here. Well-organized and more complete presentations of thermochemical data for uranium compounds were published by Rand and Kubaschewski in 1963,> Fuger et al in 1983,* and Barin in 1989.> An assessment of thermochemical data for the system uranium-oxygen-chlorine by Cordfunke and Kubaschewski® illustrates the scatter of individual values, the limits of accuracies, and the dependence on estimated values. Selected values will be listed and discussed in the following sections without detailed reference to these limitations. 2. LITERATURE INFORMATION The processes of interest for the preparation of the anhydrous UCl, are to react the uranium oxide feeds with chlorine (ah oxidizing agent) and carbon or carbon monoxide (reducing agents). Oxidation and reduction reactions will take place and all possible uranium oxides, uranium chlorides, and uranium oxychlorides must be considered. Physical properties for these compounds are tabulated (Table 1). A consistent set of thermodynam- ic data for these uranium compounds is needed to allow calculations to identify the probable reactions. The data for C, CO, and CO, as reactants or products and for H,O and HCl as impurities are included for convenience. Phase diagrams are important as they present useful equilibrium results. Finally, results are reviewed for the reported experi- mental studies of the reactions of uranium oxides, chlorine, and a reducing agent. The principal component of a molten salt for a chlorination will probably be UCI,. Other physical information reported for UCI, includes:’ Heat of fusion at 863 K: 44.8 kJ/mol Free energy of vaporization at 863 K: 218 kJ/mol Entropy of vaporization at 1062 K: 133 J/mol K Densities of molten UCl, are: Temperature Density (°C) (g/om’) 590 3.57 600 3.55 650 3.45 700 3.36 750 3.26 The properties, preparation, and chemistry of the uranium chlorides and oxychlorides are comprehensively reviewed by Brown.” 2.1 THERMOCHEMICAL DATA The application for the thermodynamic data is to make calculations at the chlorina- tion reactor conditions. The most useful values are the free energies of formation at 700 to 1100 K. A temperature of 700 K (427°C) is about the lowest temperature of interest for both practical rates of reaction and the use of molten chloride salts. The 1100 K Table 1. Physical properties of U-O-Cl compounds Density Melting Boiling Molecular at 298 K, point point Compound weight (g/cm?) (K) (K) Color at 298 K U 238.03 19.05 1405 4091 Silver gray UCl, 344.39 5.44 1114 1930 Olive-green UuOoCdi 289.48 Dark red Ud(l, 379.84 4.87 863 1065 Dark green UQCl, 324.94 Green Uo, 270.03 10.96 3110 Brown-black (U0),Cl 685.33 U,0, 1096.12 10.9 decom.? Black UCl 415.30 3.8 ~ 600 decom. Red-brown UGCl, 360.39 Brown UO,(Cl 305.48 decom. Brown-violet U,0, 842.09 8.30 decom. Greenish-black (UO,),Cl, 646.42 Black-brown UCl 450.75 3.5 452 decom. Black or dark green UOd], 395.85 uo.Cl, 340.93 5.34 851 decom. Yellow U0, 286.03 7.29 decom. Orange-yellow *decom. = decomposes without phase change. (823°C) is above the boiling point of UCl, and is near the highest practical temperature. The enthalpies and entropies of formation at reference conditions are available for nearly all the uranium chlorides and oxychlorides, but the high-temperature data is much less complete. The enthalpies and free energies in recent (since 1975) assessments and collections of data are mostly 20 to 30 kJ/mol smaller (less negative) than those listed before 1970. 1t is probably inconsistent and misleading to use carly and recent data together in one calculation. A recommended set of data for calculations is tabulated (Table 2). This data is from the more recent publications.**#1% Data published for uranium chlorides and oxychlorides before 1975 is not consistent with this more recent data. The significance of and conclusions from the thermochemical data are discussed in Sect. 3.1. Krahe listed vapor pressure equations as shown in Table 3.' Calculated values from these equations are shown in Fig. 1. The decomposition of UCl; (or UCly) into Cl, and UCI, must be considered; the UCIs or UCl; are stable only when excess Cl, is present. 22 PHASE DIAGRAMS Phase diagrams present equilibrium information in several different ways. With only two components, a composition versus temperature type of phase diagram can show a complete representation of the solid, liquid, and gaseous phases present. The diagram for UQ,-UCl, gives information important to understanding the chlorination behavior. Most of the other two-component phase diagrams for uranium oxides, oxychlorides, and chlorides are not available in published literature. For three components, a triangular diagram can show one variable (usually the liquidus temperature) vs all compositions. A third type of phase diagram can be calculated from the thermodynamic data. The calculations give the equilibrium concentration or solid phases present vs two of the concentrations as variables. Table 2. Thermochemical data Free energy of formation (-A,G°) -AH at 298 K Sat 298K (kI/mol) Compound (kJ/mol U) (J/mol-K) 700 K 900 K 1100 K UCl, 862.1 159.0 712.2 6709 629.7 UOCI 9473 102.9 836.8 807.2 795(E) udci, 1018.8 197.23 814.6 762.3 720(E) UOCl, 1069.4 138.32 904.9 863(E) na® uo, 1084.9 77.03 963.5 930.8 897.7 (U0),Cls 2197.4 326.4 na na na U,0, 4510.8 335.93 3972.9 3827.2 3680.1 UCI, 1041.4 246.9 816.3 763(E) na Uod], 1140.1 169.9 946.8 888.3 835(E) UO,Cl 1169.4 112.5 na na na U,04 3574.8 282.59 3114.0 2994.3 2874.5 (U0,),Cl,y 2404.5 276.1 na na na UCI, 1068.2 285.8 812(E) 760(E) na UOC], na na na na na UO,Cl, 1145.8 150.6 1017.7 960(E) 900(E) U0, 1223.8 96.11 1043.0 992.9 942.3 CO 110.53 197.65 173.52 191.42 209.08 CO, 393.52 213.80 395.40 395.75 396.00 COCl, 220.08 283.80 187.05 177.84 168.66 Cd, 95.98 309.81 -1.74 -28.34 -54.56 HCI 92.31 186.90 98.75 100.15 101.43 H,0(g) 241.83 188.83 208.81 198.08 187.03 *(E) indicates estimated values. *The term "na" indicates that values are not available in any of the known references. ORNL DWG 91A-11 1000 800 [ 600 [ 400 [ 20 A mp: 590°C VAPOR PRESSURE (mm Hg) o o i A mp: 327°C . L UCls o | >800 °C 2 L | ] | L 100 200 300 400 500 600 700 800 TEMPERATURE ( °C) Fig. 1. Vapor pressures of uranium chlorides. Table 3. Vapor pressure equations® A B C Temperature Compound (X) UCl; (s) 19.224 15,760 3.02 298 - 1110 UCH (1) 24.044 14,340 5.03 1110 - 1950 UCI, (s) 20.329 11,350 3.02 298 - 863 udi, () 26.079 9,950 5.53 863 - 1062 UCl; (s) 21.810 7,450 4.03 298 - 600 UCL (1) 26.027 6,210 6.29 600 - 800 UCl (s) 22.317 4,765 5.03 298 - 453 UCl, () 26.120 4,060 7.04 453 - 650 logP., = A-BT-ClogT. The phase relationships between tetravalent uranium oxide and chloride are shown in Fig. 2.1 The phasc diagram shows that there are three stable compounds over the entire range of composition--UCI,, UOCI,, and UQ,. There is a eutectic reaction between UCl, and the intermediate compound, UOCI, The melting point of pure UCl, is 590°C. A minimum melting point of 545°C occurs at the eutectic composition of UCl,; + 6.9 mol % UQ,. A maximum solubility of about 13 mol % UO, in molten UCI, is reported at 810°C. At temperatures from 810 to 855°C, UCI, vapor is in equilibrium with solid UOCI, UQOCIL, decomposes at 855°C. At higher temperatures, vapor and solid UQ, are in equilibrium. This phase diagram suggests practical limitations on the useful chlorination conditions and will be discussed further in this respect (Sect. 3.2). A phase diagram with the Cl/U atom ratio as the concentration variable shows the melting points of the uranium chlorides and their eutectics (Fig. 3).!' This diagram is a scries of binary diagrams for U-UCI,-UCl,-UCIs-UClg as no more than two of these compounds can be present in equilibrium. Fig. 2. Phase diagram for UCI,-UO,.”? | wt % 99 97 95 90 85, T TTTT T T T T 900 ; Vopor + U0, — Vapor A___“ggs_:___ e 1 apor + UOCI, . 800 »={8I10° _| 700 — 600 - L 3 f (6.9%) N 500H — i I 400H- — | I00H UCly ss + UOCiz — { | | 200 - N 100 S - S oL 1 Lt 1 1 1 1t ., | UCI42 6 10 i4 I8 50 U, — mol % 11 For reasons discussed in Sect. 3.2, a ternary mixture of UCl,-MgCl,-NaCl might be the preferred melt for reaction of UO,, Cl,, and C. While this ternary diagram has not been determined, the three binary diagrams (UCl,-MgCl,, UCl,-NaCl, MgCl,-NaCl) have been published.”® These binary diagrams give the liquidus temperatures for the three sides of a ternary MgCl,-NaCl-UCl, phase diagram. The three binary diagrams are simple, and simple liquidus curves for the ternary are very probable. Estimated curves were drawn (Fig. 4) with shapes similar to those for other published ternary diagrams. These curves are derived from the binary data and should be considered interpolations between them instead of extrapolations. | An important use of thermochemical data is to calculate the equilibrium composi- tions at specified temperatures. Uranium has five major valences (0, 3, 4, 5, 6) and also has some stable compounds of apparent intermediate valances (4.5, 5.33, 5.5). Only two of the major valences can be in equilibrium at a specified condition. Uranium and chlorine give a series of compounds (U, UCl;, UC,, UCl;, UCly). Each composition from Cl/U=0 to Cl/U=6 will have an equilibrium overpressure of Cl, gas. These equilibrium concentra- tions can be calculated from thermodynamic data. They can be conveniently represented by equations with temperature as a variable and do not require phase diagrams. The melting points and eutectics were shown in Fig. 3. | Thermochemical data can also be used to calculate equilibrium compositions for the U-O-Cl system. Because of the multiple valances of uranium and the formation of oxychlorides, over twenty U-O-Cl compounds are possible. At least seventeen of these compounds have been reported experimentally.5 The results can be presented as diagrams showing the composition of the solid phase with a specified temperature as a constant and two gas activities (Cl, and O, or CO) as the variables. A result of interest for the chlorination of UQ, to UC}, is reported by Krahe (Fig. 5)."! Chlorine pressures near 1 atm give a melt that consists of UCl, and UCl;. UCI, is the stable composition for a wide range of both Cl, and CO (or O,) concentrations. Krahe's result looks very good with respect to utilization of Cl, and the formation of UCl, as the product. Cordfunke discusses the U-O-Cl phase calculations and the limitations from the precision of the data.® He shows that small differences in the data can cause phases to appear or disappear from the calculated results. He estimates a need for 0.25% precision for enthalpies and 1% for entropies. The data for uranium oxychlorides is probably not this good. The poor TEMPERATURE (°C) 10 ORNL DWG 91A-634 1200 7 1 1130° | 1000 ..o 8150 Ef&? goo F— — — |- 590° 600 - 550° 400 F N\ 3270 \ N\ 170° ‘\\ 200 | 1700 N U —~UCly _UCl, _UCls UClg_ 0 | 0 1 3 4 5 6 ATOM RATIO OF CI/U Fig. 3. Phase diagrams for the compounds of uranium and chlorine.’ 12 : —PHA IIIIIIIIIIIIIIIIIIIII RNEX AR PN 50 /\ "‘“»’0“;}"@0’" INaCI. ucl, DOPOK L PPE N - DPOLKASXE5D » X ‘:" ;.0:’1 f{‘:f"/':('l”"/ 700 0 0‘0 L /AA g !* One such process was used at Oak Ridge for producing calutron feed for isotope separations. The CCl, cannot be produced by a simple reaction of C with Cl,. Therefore, the reactions of CCl, with uranium oxides do not provide efficient overall reactions of Ci, to prepare UCI,. There are many literature descriptions concerning the reactions of chlorine with uranium oxides. Some of these were intended to produce uranium chlorides with some results for carbon as a reactant and CO, or CO as producis. Experiments using chlorine gas feed to graphite distributors immersed in molten salts showed good rates of reaction of Cl,. These graphite distributors provide a carbon source of relatively low-surface area and reactivity, and the rates of formation of CO and CO, are much better than might be expected. Canning demonstrated nearly complete utilization of Cl, for up to 90% chlorination of uranium oxides in NaCl-MgCl, at 700 and 800°C."> Analyses indicate 80 to 85% UC], and 15 to 20% UCI; at the end of chlorination. Gibson claims a similar result with all the uranium soluble in the NaCl-KC! melt at the end of chlorination.'® Gens studied the volatilization of uranium chlorides from nuclear fuels and appeared to find the formation of some non-volatile UO,Cl,.!” Lyon reported rapid reactions of uranium oxides in molten NaCI-KClI at 850°C with Cl, to give UO,CL,.'* The teactions of uranium oxides with CCl, have been more carefully studied than the reaction with Cl, and C or CO. Since the free energy of formation of CCl, is positive above 415°C, the use of CCl, above this temperature is somewhat thermodynamically 15 equivalent to use of C and Cl,. Budayev provides good thermodynamic analyses and experimental results of reactions with CCl,.'* Reaction products at 200 and 300°C were UQO, and UCl,. Reaction products at 400 to 700°C included UO,, UCl,, UOCl,, UOCI,, and U,0,Cl;. The experimental results showed stepwise reaction with many intermediate products, including CO, COCl,, Cl,, and all the uranium oxychlorides. Gens reported that the UyO;4 treated with CCl,-Cl, is first converted to UO,Cl, and is then further reacted to give UCl,, UCl;, and UCI."® Jangg found high conversions to volatile UCI, and UCI; at 700 to 900°C using CCl, while Cl, gave mostly UO,CL,.” Katz and Kabonowitch published an excellent review of the literature on the chemistry of uranium up through 1946.> The overall results for preparation of uranium chlorides indicate the following conclusions: 1. There were no complete and practical conversions of UO, to pure UCl, by reactions with chlorine and C or CO. 2. UO, was clearly the preferred uranium oxide feed. Uranyl compounds were much less reactive. Higher oxides, such as U,Oz and UQ,, gave larger amounts of UCl as compared to the UC yield from UO, at similar conditions. 3. The conversion of UO, to UOC!, appears to liberate more energy than the conversion of UQCI, to UCl,. Therefore, an incomplete conversion is likely to leave large amounts of UOCI, instead of unreacted UO, and UCl,. 4. The first reported preparation of UC], was by the reaction of a UO,-C mixture with Cl, gas. The major disadvantages were the major yield of UCI,, the high reaction tempera- ture, and the phosgene in the waste gas. 5. The reaction of UO, with CCl, proceeds at lower temperature and gives less UCl; and phosgene than UO, with C and Cl,, Common conditions were 350 to 450°C in CCl, vapor or 150 to 250°C in liquid CCl, under pressure. 6. Phosgene (COCl,) was an effective reagent at temperatures of 450°C or higher. 7. The higher chlorides (UCl;, UCL) are formed when the higher oxides (U;0,, UQ;) are converted to chlorides and are also formed by reaction of UCl, with Cl,. 17 3. DISCUSSIONS AND CONCLUSIONS The application of literature information to plan a development program for UCI, preparation is given here. The use of thermodynamic data is the logical first step, but has important limitations. The experimental results reported in the literature show reactions of chiorine and uranium oxides with little information on what reactions are occurring. The chemistry and preparation of TiCl,, ZrCl,, ThCl,, and SiCl, have important similarities to thase of UCl,. Therefore, references for preparation of these compounds are reviewed as sources of information applicable to UCl,. 3.1 THERMOCHEMICAL CONSIDERATIONS Uranium dioxide is one of the most stable metal oxides and has a larger free energy of formation than UCIl,. This means that many of the reactions that might convert UO, to UC], are thermodynamically unfavorable. The displacement of the oxygen in UO, by reaction with Cl, is not practical. While the rate of reaction of UCl, with air is low at room temperature, the thérmodynamic eqdilibrium is a high ratio of Cl,/O, in the gas. At 225°C or higher, UC], reacts with air to release Cl,. The practical conversion of UO, to UCI, by a chemical reaction requires a reducing agent which yields an oxide product that is more stable than the chloride (i.e., it does not react with UCl,). Carbon and carbon monoxide meet this criteria. Both have large free energies of formation while CCl, has a zero value at about 415°C and decomposes thermally at higher temperatures. Hydrogen oxide and hydrogen chloride are about equally stable, and the oxide (water) reacts with uranium chlorides to form HCl and UQCI, or UQO,. The equilibrium pressure of H,O over UQ, in HC! gas is very small, and conversion to UCl, by countercurrent treatment of UO, with HC] (as used to prepare UF, using HF) is completely impractical. For the same reasons, UCl, prepared in aqueous solutions cannot be dehydrated to anhydrous UCl,. Thermal or other treatments give removal of HCI leaving UO, or UOCI, as the product. Thermodynamic calculations are a logical first step to identifying the probable chemical reactions for conversion of UO, to UCl,. The use of thermochemical data does not identify the probable reactions with any certainty or degree of confidence. There are 18 several major causes of uncertainty. Uranium chemistry is complex with stable valences of 3, 4, 5, and 6, and with stable oxychlorides. UQ,Cl, and UOCI, are well known compounds and others are possible. All of the possible products must be considered. The uranium chlorides are more volatile with increasing valence, but UCl and UCl; are less stable with increasing temperature and decreasing Cl, partial pressure. All of the uranium compounds have large heats of formation, and calculating the free energy of these reactions usually results in a small difference from two large numbers. Small percentage uncertainties for the large numbers give large uncertainties for the differences. The relationships between the U-O-Cl compounds can be illustrated by a matrix listing. Data are available for heats and entropics of formation at 298 K for nearly all of these compounds (Tables 2 and 4). The free energies of formation at 900 K would allow more realistic and useful calculations. However, because the free energy of formation data at high temperatures is much less complete, this matrix listing contains less certain or estimated values (Table 5). Table 4. Hceats of formation for U-O-Cl compounds at 298 K - {Heats of formation), kJ/mol U Valence All dl Onec O Two O >Two O 0 u? U U U 3 UCl, UOoCl 862.1 9473 4 ucCl, Uodi, uo, 1018.8 1069.4 1084.9 4.5 (U0),Clg U,0, 1098.7 1127.7 5 UCl UOd, Uuo,Cl 1041.4 1140.1 1169.4 5.33 (UO,).Cl, U,;04 or 5.5 1202.3 1191.6 o UCI, uQod, UQO,Cl, U0, 1068.2 ~ 1140 1145.8 1223.8 aAfI’IO = {), 19 Table 5. Free energies of formation for U-O-Cl compounds at 900 K (627°C) : - (Free energy of formation), kJ/mol U Valence All Cl One O Two O >Two O 0 U U - U U ucl, Uocl 670.9 807.2 4 ucl, Uoc, U0, 762.3 863 930.8 4.5 - (U0),Clq U,0, 880 (Est.) 956.8 5 UCl uoC, UO,Cl 763 888.3 960 (Est.) 5.33 (U0,),Cl, U0, or 5.5 990 (Est.) 998.1 6 UCy, uoci, Uo,ClL, U0, 760 900 (Est.) 960 992.9 The complexity of the uranium conversion chemistry is partly shown by a diagram giving the free energies of the simple oxidation and reduction reactions (Fig. 6). This diagram was simplified by omitting the compounds of U(4.5) and U(5.5) vailences. It also does not show the reactions of two U-O-Cl compounds to give a third U-O-Cl compound. The reductions are shown for 4C to 4CO,. Similar conclusions would apply for C to CO or CO to CO, as the three free energics are 197.9, 191.4, and 204.3 kJ at 900 K. The three free energies are equal at 973 K (700°C), and the probable products from carbon are CO, below 700 K and CO at higher temperatures. Some general conclusions from examinations of Tables 4 and 5 and Fig. 6 are: ® At a given valence state, the uranium oxychlorides are more stable than the chlorides. ® At a given valence state, the uranium oxides are more stable than the chlorides. 20 ORNL DWG 91A-632 C 0/U, ATOM/ATOM URANIUM VALENCE 0 ‘ 2 >2 U(0) U U U U * : : : -224 ~538 | | , s : ' ' -136 —-465 1 I U(III) UCly UocCl 8 ; : = ' | -91 g —-56 g I ~374 : i i ¢ ! u(Iv) uc uocl, uo, | . - . - l : 25 § -29 } 90 | ~1 4 : : { i § i UsOsg u(v) uc UOCIy Uo,Cl (VALENCE—5.33) . » ° l : -12 2 0 3 : : s | 3 3 : $ +16 3 : $ ! ¥ { ¥ ' —-14 . -B0 —~33 U(VI) UCI6 Ww UQCI, seesrtsssmsssinsioe UO2Clo sesersrsrrerenssm UO3 LEGEND: esasesaessap CHLORINATIONS; FREE ENERGY IN kJ/ATOM CI ————s OXIDATIONS; FREE ENERGY IN kJ/(0 + Cl,) sessssssrenssn DISPLACEMENTS; FREE ENERGY IN kd/[0 — Clg] e REDUCTIONS; FREE ENERGY IN kJ/(C-CO) Fig. 6. Conversion reactions for U-O-Cl compounds and free energies at 900 K. 21 e All additions of chlorine to oxides or oxychlorides of lower valence {less than U(VI)] are favorable to yield oxychlorides of higher valence. ® The oxychlorides can be formed by both direct reaction of chlorine and by reaction of a uranium chloride with an oxide or oxychloride. Since oxychlorides are the intermediate compounds for conversion of UO, to UCI,, their stability can be of critical importance to complete conversions. A conversion may appear favorable overall, but one of the steps for the intermediate may be much less favorable. For example, consider the following overall reaction: U0, + 2C + 2Cl, - UC, + 2CO. At 900 K, AG = -214.3 k], but individual steps show: UO, + Cl, - UO,(Cl, AG = -29.2; UO(l, + C - UOCI, + CO, AG = -94.4; uod, + Cl, » U0, AG = -37.0; and UoCl, + C-»UCl, + CO, AG = -54. These numbers indicate that the reactions should take place, but older data indicate they could stop at UO,Cl, or UOCI,. If using CO to give CO, as the product is considered, then the two reduction reactions change to: UO,CL, + CO - UOCI, + CO,, AG = -107.3 and UOC], + CO - UC), + CO,, AG = -66.6. The uncertainties for A(G® values of UO,Cl, and UOCI, may be larger than the above AG values, so it is difficult to be certain that the reactions are thermodynamically favorable. The more favorable calculation for CO as compared to C may also be misleading. The C would be present as a solid with a thermodynamic activity of 1, while CO would be mixed with other gases and would have a lower activity for 1 atm total pressure. Many of the reactions to change between the U-O-Cl compounds are shown in Fig. 6. Large values for the negatives of the free energies of reaction (kJ/equiv at 900 K) show reactions that are thermodynamically favored. Positive values in Fig. 6 indicate that the 22 reactions are not thermodynamically favorable. The data for the oxychlorides are uncer- tain, and values ranging from -20 to +20 kJ do not justify predictions. The formation of UCI; and UCI appears to require excesses of Cl, and will not be complete. Otherwise, all additions of Cl, are favorable, and oxychlorides should add chlorine to give UO,CI, or. UOC],. The reductions of UO; or U;0,4 to UQO, are highly favorable. The reduction of UQCI,; to give UCI, is unfavorable. Some of the other reactions give free energies of rcaction that are too near zero to justify predictions. A practical preparation of UCI, probably requires that the chlorinations to U(VI) and the reductions of UO,Cl, and UOCI, be possible since the intermediates would otherwise accumulate as stable products. The equilibrium mixture among the uranium compounds for normal operation of a chlorination reactor will always be UCl, and uranium oxychlorides containing one oxygen. The feed of UQO, will be limited to prevent excessive solids; therefore, O/U ratios will be much less than one. Because of the lower stability of the chlorides as compared to the oxychlorides, any uranium compounds containing two oxygens will react with a uranium chloride to form two oxychlorides with one oxygen in each. Any U(VI) or U(V) will be much more stable as oxychlorides than as UClg or UCl;. Therefore, the amounts of UC]; and UCl, will be small unless the moles of U(VI) and U(V) exceed the moles of oxy- chlorides. Some of the reactions and their free energies at 900 K are: UCl, +UO, » 2UOCl,, AG = -32.9 kJ; UCl; + UO,Cl - 2UOCH, AG = -53.9 kJ; UCl, + UO,CL, - 2UOC],, AG = -80 kJ; UC!, + UO,CL, - 2U0C, AG = -54.3 kI; UCl, + UO, » UO,ClL, + UCl,, AG = -31.5 kJ; and UCI, + UOCL, - UOC, + UCl, AG = -39.3 k. The first five reactions indicate that the uranium chlorides will react to give oxychlorides. The first four show that an oxychloride containing two oxygens will react with uranium chlorides to give two oxychlorides. The last two show that UCl, will react with a U(IV) oxychloride to give UCl, and an U(VI) oxychloride. 3.2 SELECTION OF CHLORINATION CONDITIONS One feed material must be the uranium oxides from ore refineries. Recycling consider- ations to eliminate large amounts of waste require the use of Cl, gas from electrolytic cells. 23 Thermochemical data show larger free energies of formation for uranium oxides than for chlorides of the same valence. Therefore, the overall reaction must include a reducing agent that has a much more stable oxide than chloride. Carbon or carbon monoxide are the mostk practical reducing agents that meet this requirement. Hydrogen (as H, or hydrocarbons) does not meet this requirement since H,O will react with uranium chlorides to form HCl. Sulfur and phosphorus (and some of their compounds) can meet this thermochemical requirement, but they are expensive feeds and give troublesome waste oxides in‘comparison {0 carbon. _ The overall reaction from the above considerations shows one or two solids (uranium oxides, C) and gaseous feed (Cl,, perhaps CO) and a gaseous product (CO, or CO). Even with stepwise reactions, the reactions between two phases with one of them a solid tend to be slow or incomplete. The phase diagram for UO, and UCI4 (Fig. 2) shows several indications toward practical conditions foriconversion of U'O2 into UCl,. The use of a liquid melt with reaction of UQ, and UCl, to give UOC, in solution appezirs favorable to high rates of reaction. Then the Cl, and carbon can react with the UOCI, or other oxychlorides in solution. The Cl, can also react with UCl, to give UCl; or UCl6 as soluble chlorinating agents in the melt. | Both the temperature and the fraction of UQO, in the charge must be limited to maintain the desirable liquid melt condition. All of the UQ, reacts with UCl, to form UOCl, and any UOC], above the solubility is present as solids. An equimolar mixture of UG, and UCl, gives all UOCI, solids without any melt. Since C will also be a solid, the preferred concentration to ensure a fluid melt will be less UO, than the solubility limit (from 6.9 mol % at 545°C to about 13 mol % at 810°C). The preferred temperatures will be intermediate between the melting point and boiling point of UCI, (590 to 792°C). The vapor pressure of UC, is also an important consideration. The addition of a diluent salt to the melt would relax some of the composition and temperature limits indicated by the UCl,-UO, phase diagram. The diluent salt should be unreactive with the carbon, Cl,, UO,, and UCL,, should have a low volatility at 600 to 800°C, and should melt below 600°C. An equimolar mixture of MgCl,-NaCl meets these requirements better than any single salt. Mixtures of MgCl,-NaCl are commonly used for electrolytic production of Mg and Cl, and have good properties. There are no published phase diagrams for UC],-UO,-MgCl,-NaCl or for any ternary mixtures of these components. However, reasonable liquidus temperatures and UO, 24 solubilities can be estimated from four binary phase diagrams. Binary diagrams for MgCl,- NaCl, MgCl,-UCl,, and NaCl-UCI," can be used to give liquidus temperatures for the three sides of a ternary MgCl,-NaCl-UCl, phase diagram. The three known binary diagrams are all simple and simple liquidus curves for the ternary are very probable. Estimated curves were drawn (Fig. 4) with shapes similar to those for other published ternary diagrams. The UCI,-UO, binary phase diagram gives the data needed for a UQO, solubility versus temperature in units of moles UQ,/UCI, (Fig. 7). Experimental data for fluoride salt mixtures show that the UQ, solubilities are proportional to the UF, concentra- tion; that is, moles UO,/moles UF, is dependent on the temperature and the other salt compositions but independent of the UF, concentration. Figure 7 shows the data for the solubility of UO, in UF, also. The data for UCl, and UF, could be represented by a single curve. Considering the MgCl,-NaCl-UCJ, ternary together with the UO,-UCI, binary, it is likely that replacement of part of the UCl, by UQ, up to the solubility limits shown by Fig. 7 would have only small effects on the liquidus temperatures of Fig. 4. As an example of the results of these assumptions, a charge of 20 mol % MgCl,-20 mol % NaCl-60 mol % (UC1,+U0O,) would be expected to be all liquid above 500°C with UQ, solubilities from 4 mol % at 500°C to 8 mol % at 800°C. 3.3 SALT PROPERTIES AND PROCESS CONTROL CONSIDERATIONS The practical chlorination of uranium ore concentrates to UCI, on a large scale would require continuous processes with controlled inventories of the process materials. The literature data indicate that a molten salt reaction with feeds of UQ,, Cl,, and C or CO might be most practical. The UCI, product must have low concentrations of unreacted carbon or oxygen. Large amounts of unteacted chlorine in the product gas are also very undesirable. Countercurrent tlows with high conversion to UCI, in a melt (low oxychlo- rides, UCI and UCIy) would be extremely difficult to accomplish. A more practical concept is to remove UCI, as a vapor from a melt with compositions favorable to high utilizations of the Cl, feed. The chlorination reactor will have three reactant feed streams and two products leaving the reactor. The most likely overall reaction will be one of: U0, + C + 2CL, -» UC], + CO, U0, + 2CO + 2Cl, - UC], + 2C0, . U0, SOLUBILITY (mol UG, /mol UX,) 0.40 0.30 G.20 ORNL DWG 91A4-631 1 r 1 ! &’ /' a "”’, - UOCI, DECOMPOSES ’/,/' AT »855° C . — ... — 0‘. - ) . ® FROM UCIl, —UO, PHASE DIAGRAM;'? ",,‘ DISSOLVED U0, IS UoCl, e ® DATA FOR UF, (GREENFIELD ~ PR AND HYDE IN AERE-R 6463) B e | ! 1 | 1 i 400 600 800 1000 1200 1400 1600 TEMPERATURE (°C) Fig. 7. Solubility of UQ, in UCl, and UF,. ¢c 26 Reactions with either CO or carbon and with other uranium oxides as feed will also have five primary flows of feeds and products. Small amounts of unreacted Cl, or of CO in addition to CO, and large amounts of a nonreactive diluent gas (N, or CO,) do not change the need to control five process inventories or process flows. All gases including the gaseous products, unreacted Cl, or CO, and nonreactive diluent gas leave by displacement without need for a control measurement. This leaves four other flows requiring control. One flow can be set to cstablish the system capacity. This is most logically the Cl, feed rate. Since there is little or no chlorine inventory in the melt, a change in the chlorine feed rate can give an immediate change in the rate of Cl, reactions. It would normally be desirable to have a good inventory of UO,, C, or CO and the molten chloride salt to allow high utilizations of Cl,. Depleting one of these charge components to provide control of the reaction rate contributes to high, undesitable losses of unreacted Cl,. The condensed phase in the chlorination reactor is likely to contain more UCI, than any other component. Therefore, it is logical to measure the condensed phase level or volume and usc this measurement to control the UC], exit rate. The gas leaving the charge should be at a vapor pressure equilibrium; that is, saturated with UCI, vapor. The UCI, vapor rate could be changed by changing the charge temperature (to increase the vapor pressure of UCI,) or by changing a diluent gas rate. Liquid or solid UCl, could be recycled to the charge to control the net outflow of UCI,. Use of nonvolatile diluent salts such as CaCl,, MgCl,, or NaCl would reduce the vapor pressure of UCI, by reducing the UCI, concentration. The diluent salts could provide some automatic changes in the rate of UCI], vapor. When the UCI, is depleted, the lower concentration of UCl, would result in a lower rate of vaporization. Excess UCl, would result in a higher rate of UCI, vaporization. For purc UCI,, concentration in the melt and the vapor pressure do not change as the amount in the charge varies. After the rates of gas exit flow, Cl, feed, and UC], vapor are determined as described, the feed rates of UO, and C or CO remain to be controlled. While the rate of Cl, losses to the exit gas can depend on the charge inventories, measurement of Cl, losses is a poor control criteria. A high utilization of Cl, is desirable and conditions that allow increascs in Cl, losses are undesirable. Also, increases in Cl, losses could result from deficiencies in either UQ,, or C-CO and could also result from excessive UOCI, or C solids 27 with poor gas-charge contact. Since a high Cl, loss might have several causes, it would not be a dependable control measurement for one feed. The UO, and C feed rates should be controlled on a long-term basis by some direct measurement on the reactor charge—either measurements of salt properties or analyses of samples for C and oxygen. A review of the limited literature on uranium salt properties did not indicate any promising possibilities for use of measurements of the salt properties at 600 to 800°C. Based on the lack of information for measurement and use of salt properties for control, the UQ, and C feed rates will probably be sct to agree with the Cl, feed rate with periodic adjustments from analyses of charge samples. Good inventories of UO, and C appear desirable both to ensure good utilization of Cl, and to make control from periodic samples more practical. 3.4 PREPARATION OF TiCl,, ZrCl,, SiCl,, AND ThCl, FROM OXIDES The chemistry and preparation of these metal chlorides have important similarities to those of UCI,. For all of them, the tetrachlorides have been produced from oxides using Cl, and carbon as follows: MO, + C + 2C], - COy,y + MCly,, or MO, + 2C + 2Cl, » 2C0O, + MCl,,,. The ThCl, preparations were small scale. The other three conversions have been produc- tion processes and should provide practical information that applies to production of UCl,. For silicon, the free energy of reaction is positive, but the reactions can be completed by removal of the SiCl, gas. This conversion of TiO, to TiCl, is used on a large scale (10° tonsfyear) to prepare titanium dioxide pigments. The TiCl, is a more volatile product (mp 248 K and bp 409 K) than UCl,, and can be handled as a liquid at room temperature. While the process concepts for preparation of TiCl, are old and well known, the details of plant design and operation are proprietary with little publication as technical literature. The descriptions published in the 1950s*! and 1980s* are very similar. 28 One description of a plant operation is as follows: A mixture of 20 to 30 wt % calcined coke and 70 to 80 wt % TiO, ore is fluidized at 900 to 1000°C using Cl, gas. The reactor is lined with SiO, brick and some oxygen feed may be used to preheat the reactor or replace heat losses (the chlorination reactions are exothermic). The exit gases contain the TiCl, product as vapor, some excess Cl,, CO,, and CO (ratios near 2 mol/mol), and volatile chlorides of impurity metals such as Fe, Mn, Cr, and Al. The chlorides are separated using the differences in volatility and melting points to give a purified TiCl, liquid as the product. An alternate to the fluidized bed is to flow the chlorine up through a fixed bed of TiO,-carbon briquettes. Electrical resistance heating may be used to generate supplementary heat in the fixed bed. The use of SiO, brick linings without excessive attack is possible as a result of the low surface area of the brick as compared to the mixture of fine SiO, and carbon that is used for preparation of SiCl,. A similar chlorination of ZrG,-C briquettes at 600-800°C was developed for prepara- tion of ZrCl,” Fixed beds of 66-cm diam were operated using ZrO,-C briquettes to produce 100 mol ZrCl/h. The ZrCl, leaves the reactor as vapor and is collected as con- densed solids. ThCl, has been prepared using by reacting crushed ThO,-C compacts with Cl, in a KCIi-NaCl melt at 900°C.?* Careful operation was required to give complete conversions to ThCl, without residues of ThOCI,. The containers were silica (quariz). 4. ACKNOWLEDGEMENT This review resulted from a development program for preparation of uranium metal feed for the AVLIS (Atomic Vapor Laser Isotope Separation) program. Discussions with and suggestions from H. W. Hayden, J. C. Mailen, and M. J. Stephenson contributed to the evolution of the review. Some references were supplied by H. W. Hayden and W. D. Bond. Any oversights or errors in the selection of data and the discussion and conclusions arc the responsibility of the author. The information in this review was used to plan and correlate results for a laboratory study. The experimental results show high rates of chemical reaction, good utilizations of chlorine, and collection of condensed UCI, product.® 10Q. 11. 12. 13. 14. 15. 29 5. REFERENCES Phillip G. Sewell and Norman Haberman, "AVLIS Program Powers Ahead in the United States," Nucl. Eng. Int., pp. 17-19 (October 1988). Joseph J. Katz and Eugene Rabinowitch, The Chemistry of Uranium, Part 1. The Element, Iis Binary and Related Compounds, McGraw Hill (1951). M. H. Rand and O. Kubaschewski, The Thermochemical Properties of Uranium Compounds, Interscience Publishers (1963). J. Fuger et al,, The Chemical Thermodynamics of Actinide Elements and Compounds- Part 8: The Actinide Halides, International Atomic Energy Agency, Vienna (1983). Thsan Barin, Thermochemical Data of Pure Substances, Cambridge, New York (1989). E. H. P. Cordfunke, "Basic Thermodynamic Data in Nuclear Technoldgy: New Developments, Old Problems," J. of Nucl Mater., 130, 82-93 (1985). David Brown, "Compounds of Uranium with Chlorine, Bromine, Iodiné," Gmelin Handbuch der Anorganischen Chemie, Vol. C9 (1979). E. H. P. Cordfunke et al., Thermochemical Data for Reactor Materials an Fission Products, Eur.-Contract no. ETSN-0005-NL (1988). E. H. P. Cordfunke and O. Kubaschewski, "The Thermochemical Properties of the System Uranium-Oxygen-Chlorine," Thermochim. Acta, 74, 235-245 (1984). M. W. Chase et al., JANAF Thermochemical Tables, Third Edition, Amer. Chem. Soc., Am. Inst. Physics, and National Bureau of Standards (1986). Josef Krahe and Franz Miller, "Zur Thyermochemie der Stoffsysteme U, Th, Pa, C, O,, Cl,," Institute fir Chemisehe Technologie, Jul 565-CT (December 1968). Y. M. Sterlin and V. V. Artamonov, cited by E. M. Levin and H. ¥. McMurdie, Phase Diagrams for Ceramists: 1975 Supplement, The American Chemical Society, Colum- bus, OH, 1975, p. 396. H. F. McMurdie, Editor, Phase Diagrams for Ceramists, Vol I to VII, American Ceramic Society (1964 to 1989). I. V. Budayev and A. N. Volsky, "The Chiorination of Uranium Dioxide and Plutoni- um Dioxide by Carbon Tetrachloride," Proc. of Second United Nations International Conf. on the Peaceful Uses of Atomic Energy, p.2195, pp. 316-330, {1959). R. G. Canning, "The Production of Uranium Metal Powders by Electrolysis in Molten Chlorides,” Australian Atomic Energy Symposium-1958, pp. 115-122 (June 1958). 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 30 A. R. Gibson et al., "Processes for the Production of Uranium Oxide," U. S. Patent 3,117,386 (January 14, 1964). T. A. Gens, Laboratory Development of Chloride Volatility Processes for the Recovery of Uranmium Directly from Spent Rover Fuel or from its Combustion Ash, ORNL-3376 (June 1963). W. L. Lyon and E. E. Voiland, The Preparation of Uranium Dioxide from a Molten Salt Solution of Uranyl Chloride, HW-62431 (October 1959). T. A. Gens, Chloride Volatility Experimental Studies: The Reaction of U;04 with Carbon Tetrachloride and Mixtures of Carbon Tetrachlonde and Chlorine, ORNL/TM- 1258 (August 1965). G. Jangg et al., "Chlorination of Uranium Oxides," Afompraxis, 7, 332-336 (1961). A. D. Mcquillan and M. K. McQuillan, Titanium, Academic Press (1956). Minoru Qgawa et al., "A Study of Titanium Resources and Its Chlorination Process," Titanium *80 Science and Technology, pp. 1936-1945, Kyoto, Japan (May 19-22, 1980). S. M. Skelton et al., "Zirconium Metal Production,” Proc. of the International Conf. on the Peaceful Uses of Atomic Energy, 8, 505 (1956). A. R. Gibson et al., "Thorium Metal Production by a Chlorination Process," Proc. of Second United Nations International Conf. on the Peaceful Uses of Atomic Energy, 4, 237-242 (1959). P. A. Haas et al., Reaction of Uranium Oxides with Chlorine and Carbon or Carbon Monoxide to Prepare Uranium Chlorides, ORNL/TM-11792, Oak Ridge National Laboratory (November 1991). ORNL/TM-11955 Dist. Category UC-501 | (Chemistry) INTERNAL DISTRIBUTION 1. D. H. Andrews 16. J. C. Mailen 2. J.T. Bell 17. J. R. Merriman 3. W. Fulkerson 18. J. A. Pashley 4. R. K Genung 19. A. S. Quist 5-9. P. A Haas 20. M. H. Randolph 10. H. W. Hayden 21. K-25 Records Depart. 11. W. H. Hermes 22. Enrichment Technol. 12. J. R. Hightower 23. Cen. Research Library 13. K H. King-Jones 24. Document Ref. Section 14. F. E. Kosinski 25. Lab Records, ORNL RC 15. L. A Lundberg 26. ORNL Patent Section EXTERNAL DISTRIBUTION 27-36. Office of Scientific and Technical Information, Department of Energy, P. O. Box 2001, Oak Ridge, TN 37831 37. Office of Assistant Manager, Energy Research and Development, Oak Ridge Operations, P. O. Box 2001, Oak Ridge, TN 37831 38. N. Habermann, U. S. Department of Energy, Washington, D.C. 20585 39. J. A Horton, Lawrence Livermore National Laboratory, P. O. Box 808, Livermore, CA 94550 40-136. Given distribution as UC-501, Nuclear Energy (Chemistry)