Envlron. Sei. Technol. 1984, 18, 262-264
Photochemical Separation of Xenon and Krypton Chrlstlne E. Geosllng* and Terence Donohue"
Laser Physics Branch, Naval Research Laboratory, Washington, DC 20375
rn A technique is described for the separation of xenon from krypton using photolytically produced fluorine atoms as the reactive species. Stable XeF, is produced when a mixture of xenon, krypton, and fluorine gases is irradiated with near-ultraviolet light. KrF, is not stable at the near room temperature conditions of the experiment. Separation factors (enrichments) greater than lo00 for Xe/Kr are reported for yields of less than 98%. Even higher yields (resulting from longer photolyses) are possible, but then separation factors become slightly smaller.
Introduction There is increasing interest in new methods for noble gas separation for reasons of both cost and safety. The prevalent separation method for xenon and krypton is cryogenic distillation (1)generally as a byproduct from the manufacture of liquid air. This process, while well developed and reliable, is relatively energy intensive. Another source for these noble gases may become available in the future, as the products from nuclear fission, found as offgases from either nuclear reactors or reprocessing plants (2-5). A number of novel rare gas separation methods have been recently developed (6)to both take advantage of this new source and also reduce or eliminate the radioactive hazards found with these offgases. The separation costs arise primarily from the thermodynamically inefficient production of low temperatures required for separation, near liquid nitrogen temperatures (-196 "C). While the amount of rare gases separated constitute only a very small fraction of the volume processed by a commercial liquid air plant, noble gas purity requirements involve extensive amounts of distillation. Elimination of this feature, as found in the room temperature process described here, would reduce processing costs. Furthermore, a safety hazard is found when the cryogenic process is used with the gaseous products from spent nuclear fuel. Radiolysis of the air and water always present produces ozone (61, which becomes explosive when condensed as a liquid at temperatures below about -120 "C. The incentives for developing a noncryogenic separation technique is clear in this situation. Of the isotopes of Xe and Kr produced by nuclear fission, only one, 86Kr, has a significant half-life (of -10 years). Furthermore, krypton constitutes only 6.5% of the total rare gases, so that a simple elemental separation process is sufficient to produce xenon free of any radioactive isotopes (2). Further photochemical enrichment of this isotope is possible but not necessary for most applications. We have, in fact, investigated such a molecular laser isotope separation scheme (7). The discovery of the first noble gas compounds was soon followed by the first reports that some of these compounds could be synthesized photochemically (8). XeF,, XeF,, and XeFBcan be made thermally from Xe and Fz; however, temperatures of 250 "C are necessary for establishment of equilibrium on a reasonable time scale. It is relatively straightforward to make XeFz photochemically, but photochemical production of KrF, requires quite severe conditions such as in the cryogenic liquid or solid states (9, 10). This difference in stabilities between the two difluorides is the basis for the separation technique presented here. N
262
Environ. Sci. Technol., Vol. 18, No. 4, 1984
Xenon difluoride is produced when a mixture of xenon and fluorine is irradiated in the ultraviolet (UV), into the absorption band of the fluorine molecule. This band is relatively weak (e of -3.0 L-mol-lscm-') and broad, peaking at 285 nm, and photolysis in this region results in the production of fluorine atoms (11). Any photolytic source can be used as long as the wavelength is appropriate, from about 250 to 350 nm. It is not necessary that the mechanism resulting in XeF, production be totally understood. Indeed, some of the details have yet to be fully explained (12). The significant feature is that a photochemical process involving the rare gases can be made selective for xenon, where the products are molecular and can be readily separated from reagents.
Experimental Section The xenon and krypton used in these experiments were 99.995% pure from Air Products. The fluorine was 98% from Matheson and the nitric oxide 99.0% from Matheson. The gases were handled in an all-metal vacuum line passivated with fluorine. Removable traps were used to transfer samples to the mass spectrometer for analysis. Two photolysis cells were used. A quartz cell with fused quartz windows was employed in the initial experiments, but production of SiFl by action of fluorine on the cell walls was found to interfere with the mass spectral product analysis. For most of the experiments, a stainless steel cell with CaF, UV quality windows was used. This cell was also passivated with fluorine before use. In a typical experiment about 150 torr of krypton and 150 torr of xenon were admitted to a holding trap and an analysis trap. The analysis trap was then closed and taken to the mass spectrometer (Finnigan Model 3100) for a reference spectrum to allow exact determination of the Xe/Kr ratio. The remainder of the rare gas mixture was then condensed into the photolysis cell by lowering the temperature to -196 "C using liquid nitrogen. About 250 torr of fluorine was then added to the photolysis cell which was then sealed and removed for photolysis. The pressures of reagents were calculated to give a small excess of fluorine over that required for complete conversion to XeF,. The photolysis cell was maintained at room temperature with its cold finger immersed in a dry ice/2-propanol slush at -78 "C and irradiated with either a 150-W xenon lamp (Oriel), a 500-W medium-pressure mercury lamp (Oriel), or a XeF laser (351 nm; Lumonics Model 261, operated at an average power of 2 W). As irradiation proceeded, fine crystals of XeFz were observed to condense on the inside of the cell near the mouth of the cold finger. No observable or measureable amounts of KrF, were ever isolated in any experiment, not surprising considering the poor stability of this compound. In other experiments, XeF, was found to condense at temperatures as warm as 10 "C, but -78 OC was used to ensure maximum product yields. After photolysis, the cell was reattached to the vacuum line and the unreacted fluorine pumped off at -196 "C. A dry ice/2-propanol slush was then placed on the cell, and the unreacted Xe and Kr were condensed into another trap at -196 "C to allow mass spectral analysis of the product yield. Cryogenic temperatures were used here for ease of product analyses and complete material recovery in a single
Not subject to U.S. Copyright. Published 1984 by the American Chemical Society
stage “proof of principle” set of experiments. The product XeF, crystals were then subjected to three freeze/ pump/thaw cycles between -78 and 25 “C, then transferred to the analysis trap, and finally gently heated to decompose the compound into Xe, and Kr and F2for mass spectral analysis. This trap contained clean copper wool to remove the fluorine produced during this process.
Analysis The separation factor, p, was calculated from the mass spectral peak heights of several of the Xe and Kr isotopes according to the equation Xe/Kr(products) (1) = Xe/Kr(reactants) where Xe/Kr(products) is the ratio of xenon and krypton obtained from decomposed product crystals after all volatiles have been removed (see below). Xe/Kr(reactants) is simply the initial ratio of reactant gases. The &Kr peak was used exclusively in the calculations, since it was found to be least affected by impurity peaks in the mass spectrum. However, there was always enough background interferences to preclude exact measurement of the Kr peak intensities; thus, all measurements were lower bound estimates. However, there were no background interferences whatsoever in the region where Xe peaks were found, and the measured xenon isotope peak intensities always reflected the natural abundance isotope ratio. Calculations using any of the xenon peaks gave essentially similar separation factors, but the 129Xepeak, among the most intense, was generally used in the calculations. The yield was calculated from the amounts of unreacted xenon and krypton remaining after photolysis according to Xe / Kr(unreacted) 100% Xe / Kr (reactants)
(2)
where Xe/Kr(unreacted) is the ratio of xenon and krypton remaining volatile over product crystals after irradiation (see below). Xe/Kr(reactants) is as previously defined. It was possible to obtain yields near 100% with long irradiation times, but at the expense of small drops in p. This is most likely due to inclusion of Kr (or, less likely, KrF,) in the product XeF2 during crystallization.
Results and Discussion The separation factors and yields for photochemical separation of xenon and krypton are given in Table I. The yield rises linearly with photolysis time and then levels off as 100% reaction is approached. The separation factors, however, do not give such a monotonic trend, except possibly that larger extents of photolysis tend to give poorer separations. This could possibly be due to the relative difficulty of degasing the larger amount of material deposited in the longer photolysis experiments. In most cases, the magnitude of the separation factor obtained was dependent upon the ability to distinguish and measure very small product Kr peaks among mass spectral background and impurities near the limit of sensitivity of the mass spectrometer. These factors varied from experiment to experiment. Thus, actual separations are always larger than those listed in Table I. Runs 14 and 15 were carried out with added impurity gases, air and NO, respectively, since these are the most likely impurities encountered in nuclear reactor offgases. Added air had essentially no effect on the yield or p. Added NO, however, increased the separation factor
Table I. Separation Factors and Yields for Xe/Kr Separations irrad time, h
run no.
source
P
yield, %
1.o 1.5 1.5 2.0 4.0 6.0 1.25’ 3.0a 4.0b 4.0‘ 1.0
11 12 7 10 9 13 2 1 14 15 8
Hglamp Hglamp Hglamp Hglamp Hglamp Hg lamp Hglamp Xelamp Hg lamp Hglamp XeFlaserd
1100 1300 1700 350 100 500 300 7 1000 7000 300
34 50 e 74 98.7 99.8 e e
98 64 7
50 torr of room air added. a Quartz photolysis cell. 351 nm. e The yield was not 50 torr of NO added. measured in these experiments.
greatly while decreasing the yield. There were no new peaks in the product mass spectrum of the sample with air; however, the product spectra of the sample with NO contained significant peaks corresponding to both N20 and NOz. These results show that the photochemical separation method can be successfully employed without the need for preliminary separation steps to eliminate nonnoble gas constituents ( 3 , 4 ) . In conclusion, we have demonstrated a technique for separating xenon from krypton by a selective reaction with photolytically produced fluorine atoms near room temperature and atmospheric pressure conditions. Separation factors greater than 1000 have been achieved, and product yields greater than 99% are possible, with some reduction in product purities. The conditions used here, however, are not optimum for a full-scale separation process. Even though Kr is only 0.02 % by volume of a reprocessing offgas stream and Xe constitutes 2.1 % , recovery of F2, use of several cascaded stages and a continuously flowing system can improve efficiencies markedly (8,13). Larger mercury lamps can be made to operate more efficiently than those used in the present system. In fact, a single 10-kW Hg lamp would be an adequate photolytic source for processing the entire offgases from a typical nuclear reprocessing plant. The costs involved in cooling are small, since XeF2 can be condensed at near room temperature. The process can be operated at or near atmospheric pressure, which can give large throughputs while avoiding the risks of operating at higher pressures. Possibly the largest barrier to a practical application of this concept will be solving the problem of handling fluorine in these environments. Registry No. Krypton, 7439-90-9; xenon, 7440-63-3; fluorine, 7782-41-4.
Literature Cited (1) Handley, G. G. In “Noble Gases”; Stanley, R. E.; Moghissi, A. A,, Eds.; ERDA TIC CONF-730915 (avail. NTIS): Springfield, VA, 1973; p 90. (2) Rohrmann, C. A. h o t . Radiat. Technol. 1971, 8, 253. (3) Mastera, S. G.; Laser, M.; Merz, E. Enuiron. Sci. Technol. 1978, 12, 85. (4) Bohnenstingl, J.; Heidendael, M.; Laser, M.; Mastera, S.; Merz, E. In “Management of Radioactive Wastes from the Nuclear Fuel Cycle”; International Atomic Energy Agency: Vienna, 1976; p 129. (5) Stein, L. In “Noble Gases”; Stanley, R. E.; Moghissi, A. A., Eds.; ERDA TIC COW-730915 (avail. NTIS): Springfield, VA, 1973; p 376. (6) Bendixsen, C. L.; Buckham, J. A. In “Noble Gases”; Stanley, R. E.; Moghissi, A. A., Eds.; ERDA TIC CONF-730915 Environ. Sci. Technol., Vol. 18, No. 4, 1984
283
Environ. Sci. Technol. 1984, 18, 264-271
(avail. NTIS): Springfield, VA, 1973;p 290. Donohue, T., U.S. Naval Research Laboratory, unpublished data, 1981. Weeks, J. L.;Chernick, C. L.; Matheson, M. S. J. Am. Chem. SOC.1962,84,4612. Streng, L. V.;Streng, A. G.Inorg. Chem. 1966,5, 329. Turner, J. J.; Pimentel, G.C. Science (Washington,D.C.)
1963,140,974. Calvert, J. G.;Pitts, J. N. “Photochemistry”; Wiley: New York, 1966;p 184.
(12) Messing, I.; Smith, A. L. J. Phys. Chem. 1982,86, 927. (13) Weeks, J. L.; Matheson, M. S. In “Noble Gas Compounds”; Hyman,H. H., Ed.; University of Chicago Press: Chicago, IL, 1963;p 89. Received for review April 1,1983.Revised manuscript received August 23,1983.Accepted September 19,1983. This work was partially supported by the Office of Basic Energy Sciences of the Department of Energy.
Aggregation and Colloidal Stability of Fine-Particle Coal Suspensions Paul R. Schroeder” and Alan J. Rubln
Water Resources Center, The Ohio State University, Columbus, Ohio 43210 The aggregation and colloidal stability of colloidal coal suspensions in the presence of varying concentrations of hydrogen ions, neutral salts, and aluminum sulfate were investigated. Critical concentration and critical pH values for coagulation and stabilization were determined from turbidity changes during settling following aggregation. Two colloidal suspensions of a bituminous coal representing stability extremes due to oxidation were compared. In the absence of other coagulants, vigorous oxidation lowered the isoelectric point of the coal sol from pH 5.1 to pH 1.1and the pH for stabilization from 7.5 to 2.6. The coagulation of the suspensions followed the Schulze-Hardy rule as hydrophobic sols although the oxidized coal sol was slightly less sensitive to neutral salts. The entire log aluminum sulfate concentration-pH stability limit diagram for the oxidized coal sol was established. The boundaries of settling of the coal in the presence of aluminum sulfate were similar to other hydrophobic sols except for small differences in alkaline solutions. Regions of ionic coagulation, rapid coagulation due to enmeshment in aluminum hydroxide precipitate, and restabilization were also observed and delineated.
Introduction Billions of gallons of fine-par ticle coal suspensions are generated annually by wet cleaning and sizing methods in the coal industry (1). These black wash waters are usually treated by plain settling and then discharged to our nation’s stream carrying large quantities of fine coal particles (2,3). Chemical clarification should be employed to promote the settling of colloidal particles in order to meet the suspended solids’ effluent discharge limit of 0.07 g/L imposed on the industry (4). Therefore, the aggregation and colloidal stability of fine-particle coal suspensions are of practical importance as well as of theoretical interest. In ordinary treatment practice, clarification is usually accomplished upon aggregation of the colloidal particles using iron or aluminum salts. These metal salts hydrolyze to form complex species or precipitate depending upon solution pH. Generally, the salts are used as aggregating agents at or near neutrality where the insoluble hydroxide is the predominate form. Under these conditions, it is apparent that mechanisms of aggregation occur other than simple ionic coagulation as predicted by the Schulze*Address correspondence to this author at the U.S.Army Engineer Waterways Experiment Station, WESEE, Vicksburg, MS 39180. 264
Environ. Scl. Technol., Vol. 18,
No. 4, 1984
Hardy rule. With hydrophobic colloids these mechanisms are controlled primarily by solution pH and metal salt concentration and are less dependent on the properties of the sol itself ( 5 , 6 ) . The effects of a specific sol, however, must be determined experimentally. As with many other sols the colloidal stability of coal is dependent on particle size and the presence of surface ionizable groups. Oxidation of coal surfaces occurs rapidly in nature by weathering and exposure to the atmosphere (7). Marinov (8)and Czuchajowski (9) in studies with air and other oxidants have shown that the quantity of hydroxyl and ionizable carboxyl surface groups increases with the extent of oxidation. In this study aqueous colloidal suspensions of a bitmuinous coal representing extremes in stability were compared. One group of suspensions was freshly ground, with minimum oxidation resulting only from exposure to air, while others were vigorously oxidized with hydrogen peroxide. The fine-particle coal suspensions were characterized by experimental techniques used in past studies in this laboratory (e.g., 10-12). Reductions in turbidity during quiescent settling following addition of coagulants were used to indicate aggregation. The first series of experiments with these sols involved the determination of their ioselectric points (pHJ by microelectrophoresis and their critical pH for stabilization (pH,) in the absence of aggregants except hydrogen ion. Next, neutral salts were employed as coagulants at constant pH and the sols were compared by using TeZak’s formulation of the SchulzeHardy rule (13,14). The final study was run to establish the entire aluminum(II1) concentration-pH stability limit diagram (“domain”) for oxidized colloidal coal. These diagrams delineate the regions or zones of aggregation and stabilization of a sol as determined by the hydrolysis products of the aluminum(II1) salt (15, 16). The Al(II1) concentration range in this study was varied from lo-’ to 10-1M over the pH range 2-13.
Experimental Section Solutions and Suspensions. Carbonate-free distilled water was used in the preparation of the solutions and suspensions, All chemicals were reagent grade. Stock solutions of NaN03 and ErC13.6Hz0(Ventron Alfa Chemicals) were prepared directly from the dried salts and not standardized. The stock solutions of BaC12.2Hz0,Ca(N03)z.4Hz0,A12(S04)3.18H20,and A1(N03)s.9H20 were standardized by titration with ethylenediaminetetraacetic acid (EDTA) using Eriochrome Black T as the indicator (17).Solutions of NaOH and HN03, which were stand-
0013-936X/84/0918-0264$01.50/0
@ 1984 American Chemical Society
