/255 COCUMENT CONFIRMED A3 THIS DO L ASSIFIED DW\S!/G)&OF CLASSIFICATION ~— N DATE | 2/(C |2 e en CRINL-4473 UC-10 — Chemical Sepa ations Processes for Pistonium ani Uranium MASTER e~ . e~ ENCAPSULATION OF NOSBLE FISSTON PRODUCT GASES IN SOLID MEDIA PRIOR TO TRANSPORTATION AND STORAGE W. E. Clark R. ¢. Blanco OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the V.S, ATOMIC EMERGY COMMISSION B k] L S S S et < e e g T T T e ety e o ot s it et dbiiliokine iy 3 watadh. BLANK PAGE 1 r . - E E 1 ; 3 E Printed ir. the United Stotes of America. Available from Llsaringhouse for Tederal Scientific and Technical information, Nationa! Bureau =f Stendards, U.S. Departaent of Commerce, !pringfield, Virqinia 22151 Price: Printed Copy $3.00; Micr-Z'che $0.65 b - LEGAL MOTICE -- This recort was prepared 03 an occount c{Govounnni sponsoraéd work. Neither the United Stotes, nor the Commission nor eny persoy acting an behaif of ths Commirsion: A. Mokes cny waronty or regresentation, ~xpressed x implied. with respest o vhe accurocy, completennss, o usefviness of the information contained in “his report, or 1nat the use of ony infonation, opporatus, muthyd, or precess disclosed in this repert moy not infrings privatsly cwned ~lgm's; or , B. Assvenes eny linbilitie . with riuspact to the use of, or for domuages re.uiting from the uso of eny informotion, opporotus, swthod, or process disclozed in this report. As usad in the cbcvs, "persom ccting on behalf of the Coamission'’ inclwi:s any employes or contractor of tive Coam ission, or wmployee =f such controctor, tc the extent thet such emplon. ot comiractor of the Commission, or employes of such confractor preparus, disseminates, of provides access to. cay aformotion pursvont to his swploymer? or contract with the Commission, i b s smployment with tech zontractor. - ey mee et o Rdan B Y e TR T TY O TR TR ORNL- 4473 Contract No. W-7405-eng-26 CHEMICAL TECHFNOLOGY DIVISION Chemica! Development Section B ENCAPSULATION OF NOBLE FISSION PRODUCT GASES IN SOLID MEDIA PRIOR TO TRANSPORTATION AND STORAGE W. E. Clark R. E. Blanco FEBRUARY 1970 OAK RIDGE NATIONAL LABORATCORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the | U. 5. ATOMIC ENERGY COMMISSION T PP PR~ S — Abstract 1. 2. . Alternative Methods for Secondary Containment . Discussion, Conclusicns, and Recommendations . Acknowledgment LR N ] i CONTENTS L ] L ] * ® a & @& » # P o &8 &5 a & & S & & P s © & 8 & & 5 & 5 & 85 @9 & B & & 2 & o S & a & » Inrtroduction ...................................... Assumptions for Survey and Safety Criteria .................. Estimated Volumes of Fission Product Gases ............... ................ ................................... References ....................................... s i s ok e mm il i i s s £© 4 dad o ENCAPSULATION OF NCELE FISSION PRODUCT GASES IN SOLID MEDIA PRIOR TO TRANSPORTATION AND STORAGE W. E. Ciarx R. E. Blanco ABRSTRACT The encapsviation of fission product gases in various solid media is being considered ot ORNL as a gassible method for immobilizing these goses during interim storug:2, transportation, and ultimate storoge. This type of in-mobilization would decrease the fossibiiity of the un- controlled release of such materials. In the study reported here, three media — glass, plastics, and metals — were investigated for use in en- caosulation. The combination of known techniques and extrapolated te.+ results showed that gas loadings of up to 50% of those currently obtained in cylinder storoge are obtainable by using either pressurized steel bulbs or molecular sieves in a matrix of epoxy resin. Loadings of up to at least 7.5% should be obtainable by direct dispersion of the gases in glass. Other possible encapsulation methods were also con- sidered, and the odvantages and limitations of each are discussed. The volumes of fission product gases produced in reccicrs fuslea with 27%y (LVFBR), 233G (M2 BR), and 235U (PWR-1) wer: estimated. It was assumed that tritium and ivdine will be separated from the noble gases and will be converted to stable solid compounds for permanent storage. The combined volumes of krypton and xenon are considered to be 25.0, 27.6, and 30.9 liters (STP) per 1000 Mwd for the LMFBR, MSBR, and PWR-1, respectively. The volumes of xenon and krypton generated daily in a 5-metric-ton-per—doy reprocessing plant for LMFBR fue! represent about 81% of the capacity of a standard gas cylinder. If the relatively short-lived xenon were separated from krypton, the daily volume of krypton would occupy less than one-tenth the volume of c standard gas cylinder. Bosic techrology is already available for encepsulating radio- active gas in solid matrices to yield a final product containing 25 to 50%, by volume, of the gaseous component. Engineering and economic evaluations are needed to determine whether the added scfety factor obtained by immobilizing suc™ a gas wamrants the additional expense. T B AN SY. Spen Iy W e - eyl b e, i gl 1. INTRODUCTION The growth of the nuclear power industry has resulted in an increasing awareness of the possible cumulative effects on the environment of the release of even very low levels of long-lived radicisotopes. Whereas high-level radioactive wastes have always been carefully stored under surveillance, it has been custorary tc either discharge low-level streams directly to the environment or to give them the minimum treatment necessary o decresse the oclivity below specified levels. With an increasing number of nuclear power plants and the resultir.g fue! reprocessing facilities, more efficient reroval of redioactive components from off-gas streams will be necessary to meet Federal and USAEC regulations cnd to prevent a buildup of the long-lived radionuclides in the environment. Necrmally, industrial gases are stored and trarsported in steel cylinders under pressures up to about 2600 psig at 70°F. The shipment and the handling of such cylinders require precautions because of the potential for rupture and sudden release of pressure. Storage and transporiation of highly radioactive gases in cylinders require secondary containment that is rugged enough to prevent the escape of these gases during the following sequentiai t&sts:‘ (1) a -ft fall onto an unyielding surface, (2) c puncture test consisting of a 40-in. fall onto the end of an unyielding vertical steel bar, (3) exposure for 30 min to a temperature of 1375°F, ond (4) immersion in 3 ft or more of water for ot least 24 nr. The purpose of our studies was to investigate the rechnical feasibility of either solidifying the fission product gases or of dispersing them in stable solid edia as a means of minimizing the hazards in case of an accident and/or reducing the size, complexity, and cost of the secondary shipping container. Another advantage of solidification or dispersion would be increased safety against acciderital release during interim and final storoge. This report summarizes the results of our scoping porrn b PR ESIAAS DA SIS & DN B i A i, S Wl ARSI Mol it it ot R R e [E!:},u;?:,“..;. Tt . LRI Sy tests, compares the various proposed methods for secondarv containment, and contains estimotes of the amounts of radioactive gases expected to be generoted by the re- processing of reactor fuels. 2. ASSUMPTIONS FOR SURVEY AND SAFETY CRITERIA The long-lived radioactive isctopes found in gaseous waste streams from reactors ]29| I . | 1n7 \ 3: I 4 e ] and nuclear fuel r;; cessing plants — AUy = 1.&x W yearsj, H ‘il/Z - 12.26 yearsy, and — Kr (r]/2 10.76 years) ~ are of primary interest. Since sepcration of stable and radioactive isotcpes may not be economically atractive with present technology, the statle isotopes must be included in the volumes of gases to be treated and stored. All fission product isotopes of xenon are relatively short-lived (ti/2 £ 12 days). Xenon can, therefore, be safely released after a holdup pe-iod of a few months (Fig. 1). In order to obtain minimum storage volumes, iodine ana tritium must be separated from the nsble gases and converted to stable solid compounds for storage. Xenc:h shouic be separoted from krypton?-4 and then eventually be released to the atmospherz after decay to an acceptably low level. This lecves only krypton to be stored as a gas for long periods. Our scoping studies were, therefore, primarily de- signed to test *he feasibility of encapsulating the noble gases, specifically krypton. Kryptcn is now separated from xenon and other off-gases at the Idaho Nuclear . 4 Corporation Plant, where it :s stored in conventional gas cylinders at pressures up to 2000 psig at 70°F (21. 1°C) The major fraction of the stored gas consists of stable isotopes. After a 1-day decay period, essentially all of the remaining activity is 85 from K. We have assumed (1) that the separation of stable and radicactive isotopes of the same element will not be economically attractive in the near future and (2) that rKr constitutes 7.76% of the total amount of krypton (LMFBR core). For purposes of compariscn, we have defined a "standard gas cylinder"* (ICC-3A2000) os a *ICC reguiations do not specify cylinder volumes. These vary appreciably even fer the same general class of‘lcylilnderu In this report, we have assumed a cylinder having a nominai iength of 51 in. (exclusive of valve and valve shield) and a nominai OD of 9-1/8 in. Thls cylincer corresponds in size to thot of a cylinder somehmes. called a "K" cylmder - ey s ] B e st B B M A [pRr g P TR T T T T T W W S——— o ———— e . | e T www TN v wy oy e we s =y ww g i Bk ORNL DYWG._ 49-10606 ] 1D 40° fil“l — T =3 10~ /' KRYTON ’. H ?(r THIH[ {0 r ITTHH! ACTIVITY 'curies) [ IHIIHI n TT_!TTTWEI?'_I_T—I FT.IH‘ S o~ I ] — XENON < | | 100 DAYS AFTER SHUTDOWN 200 Fig. !. Decay of Xe and Kv After Shutdown (LMFBR Core). : _',____,__‘,M,_,‘,,,,u__“‘,‘ o TRl cylinder having an absolute volume of 40 liters at 70°F (21.1°C). At the ratzd pres- sure of 2000 psig, this cylinder will contain 5089 liters (179.7 ft3) or 227 g-moles of gas at STP. This approximates current storage conditions at the Idaho plonf.s The internal heat generated by 85Kr on removal from the reacztor will be 2150.2 Btu,/hr'ffs;* ond the total absorbable radiation energy during complete decay will amount to 4.35 X 10]8 ergs. Of the radiation energy, all excepi 3.41% i. attributable io the 0.67-Mev beta emission.7 The small gamma component has an energy of 0.52 Mev. More than 993 of the radiation wiil, therefore, he obsorbed inside the container. 3. ESTIMATED VOLUMES OF FISSION PRODUCT GASES Estimates ar2 made for gaseous wastes produced by the reprocessing of fuels from three tyoes of reactors: the liquid metal-cooled fast breeder reactor (LMFBR), the molten salt breecer reactor (MSBR), and the pressurized water reactor (PWR=-1). These estimates are based upon recent computer calculations and are subject to change os more exact information becomes avaiiable. In the cases of the LMFBR and the MSBR, minimum amounts of jaseous impurities derived from the fuel are alsc estimated. How- ever, these estimatec inpurities may be negligible as compared with the impurities that will be nresent as a resu!t of in-leakoge. 3.1 Cose 1: The Liquid Metal-Cocled Fas* Breeder Reactor (LMFBR) The amcunts of xenon, krypton, and icdine expected ro be present during the reprocessing of an LMFBR core after cooling times of O to 30 days (Fig. 2) were cal- culated using the RIBDOR code, based on the operation of the Atomics Internatiznol reference LMF3R at an average specific power of 150 Mw(thermai)/metric ton for 540 dc:ys.8 The amount of tritium produced was calculafed from the fission yields . 9 recommended by Dudey. The largest uncertainty among the individual constituents is in the value for tritium, which does not include any allowance for (n, p) reactions *Provision must, therefore, be made for heat dissipation. **Tritium is generated by the thermal and tast fission of 239Pu (yield about 0.00024%) and 235y (yield about 0.00012%). L e e e Eaam x ORNL DWG. 09-10610 95 - — XENON 90 J W q 44— O Lt O : KRYPTON o ? IODINE = < [+ 4 O — X 9 10 -— 0.0030 - 1 TRITIUM 0.0075 - 0] 10 20 30 DAYS AFTER REACTOR SHUTDOWN Fig. 2. Change in Amounts of Fission Product Gases After Reactor Shutdown. LMFBR core and blankets, 33,000 Mwd in 540 days. T TR e b g in the clodding and, in this respeci, reprasents a minimum value. However, the contribution of tritium to the total amcunt of gases is negligible either with or with- out the product of the (n, p) reactions. The amounts of ‘vater, carbon dios:ide, carbon, hydrocarbons, and nitrogen (see Table 1) estimated to be present during processing are based on the snecifications 1 suggested by Olsen for impurities in the fuel. 0 It is assumed that the hydrocarbon L —® e s e e e ! | f“rrr "LGf would give 1S MGX i i g®m I 3 AN WU iu Jyivye [} sl fraction is in the form of methaone (i e, the volume on oxidation) and that this is converted entirely to CO;: and HZO' It is assumed that all the carbon is converted to C02. The volumes of xenon and krypton estimated to be generated per 10CO0 Mwd are 22.4 liters and 2.59 liters respectively; the estimated quantities of iodine and tritium P T T S T U are 5.72 x 10-2 g-moles and 1.08 x 10'-3 g-moles, respectively. If both xenon and krypton were stored or encapsulated, the output of a 5-metric-ton-per—day plant would represent about 81% of the volume of a standard gas cylinder (see Sect. 2) at 2000 psig and 70°C. If the xenon were separated from krypton and were released to the atmosphere after holdup and only the krypton were stored, only about 9% of a standard gas cvlinder would be required per day. 3.2 Case 2: The Molten Salt Breeder Reactor (MSBR}) The molten salt reactor considered here is a single-fluid breeder containing 1461 ft3 of fuel salt of the nominal composition LiF-Ber-ThF4-233LF4 (7V.7-16~12- 0.3 mole %)'.] ]The assumed power level is 2250 Mw(thermal), and the volume of volatile preducts is caiculated both for one day's operation at the planned power level of 2250 Mw and for 1000 Mwd (Tab!e 2).'2 Noble goses anda tritium* are continuously removed from tha rezctor by a helium purge. Llarge quantities of other fission products (tellurium, noble metals, etc.) are - *Triticm isgprofiuced by the 7Li(g_, (o} 2)3H reaction and in smaller quantities by the Li(7, @)“H reaction. - Table 1. Amounts of Gases and Potentially Volatile Materials Estimated to be Present During the Processing of LMFRR? Core Daily Quontity Quantity par From 5-Metric-Ton Metric Ton Quantity per Constituent Assumed Form (U+ Pu)/day P'=nt (U + Pu) 1000 Mwd Fission Product Goses ot 30 Days Cooiing Tritium, g-moles SHHO 0.028 0.179 0.00105 Krypton, lite: (STP) Kr 85.4 427.1 2,59 lodine, g~moles > 1.89 9.43 0.0572 " Xenon, liters (STP) Xe 739.2 3496.1 22.40 Helium (bonding),? liters (STP) He 86.0 429.8 2.62 Estimated Impurifies,c liters (STP) Water (50 ppm) H,O 70.7 353.3 2.4 Carbon (100 ppm) C52 212.0 1060.0 6.42 Hydrocarbons (50 ppm) CH, = COy + 2H0O 238.5 1192.5 7.23 Nitrogen (50 ppm) Ny 45.4 227.1 1.38 “Atomics International Reference Oxide LMEBR. Burnup: core - 80,000 Mwd; axial blanket - 2500 Mwd; radial blanket (undifferentiated) - 8100 Mwd. Fuel weight (in metric tons): core - 12.027; axial blanket - 7.318; radial blonket (undifferentiated) - 26.564. b!=or core and axial blanket, the volume of helium is estimaied to be equal to the volume of the oxide fuel; for the radial blanket, it is estimated to be equal to 0.05 of the volume of the oxide fuel (A. R. Irvine, ORNL, personal communication, Nov. 20, 1948). cFuq\sl specifications {ref. 10) state that the volume of gas evolved from the fuel on heating to 1800°C is less than 0.03 cm*/g. This would include water, Ny, hydrocarbons, and all adsorbed or entrapped atmospheric gases. 1r does not allow for oxic ation of carbon or of hyurocarvons. Total impurities listed here are roughly ten times the specified maximum "gas content,” which would be 56.8 !iters/metric ton (U + Pu). Table 2. Amounts of Gases Estimated to be Present During the Continuous Processing of MSBR Fuel Salt® Power - 2250 Mw{thermal) Salt Discard Cycle - 800 doye Continuous helium sparge, 20 scfm Amounts per Day from the Reactor Amounts After a Holdup Time ai 30 days) Eiement Assumed Form of 30 doys per 1000 Mwd Fission Gases Tritium, liters (STP) “HMH or SHF 0.132 0.059 Krypton, liters (STP) Kr 15.0 6.7 Xenon, liters (STP) Xe 47.1 20.9 lodine, g-moles I2 0.134 0.05%4 Estimated lmpurifies,b liters (STP) ; Water (100 ppm) HF 48.2 21.4 % Sulfur (10 ppm) 1 atom/molecule 1.35 0.60 i [ < Helium Sparge Gas, liters (STP) 8.16 x 107 3.63 x 105 e e —m— = S S ®Fuel salt has the nominal composition: LiF-BeF ,=ThF ;~233UF, (71.7-16-12-0.3 mole %). bJ. H. Shaffer, ORNL, private communication, Oct. 15, 1968. The amount: of impurities listed here fali within the limits specified (i.e., water is 10% of the specified limit, and sulfur is 40% of thz specified limit) for fuel components in ref. 13. These amounts of impurities may decrease with recycle of the fuel. 10 also removed by the purge gas, but are believed tc be in the form of entrained solids rather than volatile compounds. Tellurium, in tum, decays to iodine. Most of the iodine is retained in the fused fuel sa't and is released in the processing plant during the fluorination step to recover uren.um. A large percentoge of the total volume of noble gas is produced by the decay of these entrained elemeits. Minute amounts of the fission product gases can be expected to appear at two or three poirts in the various reprocessing side strears; however, their contribution tc the volume of the gas to be processed will be negligible. 3.3 Case 3: Pressurized Water Reactor (PWR-1) Fission product gas dah8’9 for o PWR are listed in Table 3. No estimate was mode of gases that would be produced from impurities ir this fuel. However, the amounts of gases estimated to be produced by fuel impurities for the LMFBR (see Table 1) are indicative of the order of mognitude of the amounts which can be ex~ pected for a PWR. A total of 30.9 liters of noble gases (Xe and Kr) is generated per 1000 Mwd of burnup. 4, ALTERNATIVE METHODS FOR SECONDARY CONTAINMENT 4.1 General Considerations Altemnative secondary containment methods include: 1. Double containment by simple mechanicol means. The containment cylinder can be enclosed in-a secondary pressure vessel fabricated of metal or other material. [f warranted, a layer of shock—absorbing material can be placed between the containers. 2. Enclosure »f gos in small contoiners (capsules) of metal, glass, etc., which are then incorporated in a suitable matrix of glass, metal, plastic, concrete, etc. Fracture of the exterior container would resulr in breakage of none, or - only a few, of these primary containers. The storage of gos in zeolite . - — ———— - - e ——— e T Table 3. Amounts of Gaies Estimated to be Present During the Reprocesting of PWR-1 Fuel Basis = | metric ton uranium Burmnup - 20334.0 Mwd in 605 days Amounts at Amounts (at 30 days) Element Assumed Form 0 doys 30 days ____1 year ‘pear 1000 Mwd Tritium, g=mole 3HH or SHHC 0.00138 0.00138 0.00132 0.00069 Krypton, liters (STP) Kr 65.0 65.0 64.7 3.20 Xenon, liters (STP) Xe 563.1 543.6 563.4 27.72 lodine, g-mole | (). 607 0.58 0.59 0.029 2 R, e et i, s it Wt Wt bt ot i Bi 12 structur&sé' 14 or in ciathrates is considered to be a veriation of this method. 3. Dispersions of gas in a solid matrix (i.e. as bubblss in glass, metal, plastic, etc.). In a very viscous mairix (e.g., molten glass), the release of activity would be slow even ot high temperatures. oresently be corsidered @ a mears of sforing nobie gases. For excmpie, none of the known compounds of krypton are stable at temperatures higher than about S0°C. 15 Moreover, adsorbates must always be in equilibrium with the free gas and the re- quired pressure of the free gas rises rapidly with increasing temperature. On the other hand, there is o possibility that either compounds or adsorbates could prove useful as intermediates in the formation of dispersions although, of present, other methods of generation appear to be more satisfactory in this respect. All methads of containment must make provision for the contimious removal of decay heat. An excessive risc in temperature will, of course, result in a rapid rise in the pressure in the containe.r. in estimating the amounts of gas that can be shipped or stored in various forms (Ta!e 4), we have assumed that the encapsulated gas within the matrix is stored in a container of the sc me shape and size or that of Gui standard gas cylinder; that is, encapsulation is used only to in:rease the sofety factor, not as a substitute for the gas cylinder. Thus, the cylinder becomes the secondory containment barrier, while the encapsulating medium becomes the primary barrier. In the evert that the secondary bharrier is breached, a neglijible amount of gas would be released or low temperotures and only a slow release would occur at temperatures where the inedium is molten. 4.2 Factors Affecting Choice of Matrix Glass, plastics, and metals have been suggested as riatrices for encapsulating I [ ] ‘6 * L] ® [ ] L] radioactive gases. ~ The ideal mcterial would be mechanically strong, exhibit stability when exposed to heat and to radiation, and have a sufficiently high themal 13 Toble 4. Conditiom for Storing or Shipping Encopsuloted Noble Gases Encopsulation Encepsulotion Sk ndard Liters Pressure Tempercatyre of Gos per liter Type of Storage (psig) (°O) of Storoge Volume Remarks A. Cylinder Storage 1. Maximum 2640 21.1 167.7 ICC-3AA 2400 (or "T" cylinder with 10% overlood pressure. 2. Actval 2000 21.1 127.2 Approximotes conditions used by Idahc Nuclear Corp. Plont. B. Dispersions (“Fooms™) 1. Inglos 25% by voiume 4.7 350 0.085 Experimeniai; up 0 Z5% .n polybutene; 23% in glass. 1469.6 550 8.3 Calculated. 50% by volume '4569.6 550 16.6 Caolculated; very speculative. 2. In epoxy resins 67% by volume 14.7 100 0.732 Experimental; top of resin only. Limited by imadiation - - 2.92 Colculoted; assuming a rodiction dose of 2 x 10'0 roos. C. Entrapment 1. In moleciiar sieves 62,500 350 168 Linde potent (vores need seol- ing). 2. In clothrates 294 5 £7.7 Experimentsl (iiterature); probably limited by heot ard irradiation levels. 3. Insteel buls encap- suloted in resin, glass, or metal 1500 21.1 53.8 Cadlculated using commerciolly avoilable gas bulbs of 29.5— moderate terperatures for encapsulation. 4.2.3 Metals Metals are the most satisfactory materials available with respect to high therma! corductivity and mechanical strength. The meiting temperature and other physical propetias can be varied widely by alloying. On the other hand, most metals melt T P ] 17 sharply ot well-defined temperatures, and molten metals are generally much less viscous than glasses. Metals are, therefore, not particularly attractive as media for direct containment of gos bubbles, but would ke very attractive as matrices for secondary containment of capsules, molecular sieves, etc. 4.3 Cor:parison of Specific Methods Conditions foi the storage of krypton in cylinders have already been discussed (see Sect. 2). The current practice at the ldaho Nuclear Corporation Plant amounts “o storage of 127.2 standard liters of gas per liter of storage volume.5 By using the higher-pressure "T" cylinder (ICC-3AA2400) with a 10% overload pressure, this could be increased to 167.7 standard liters per liter of sioraje volume6 (Toble 4). 4.3.2 Incorporation of Loaded Capsules, Loaded Zeolites and Clatirates into Stable Media Capsules. = An carly suggestion for mechanical ancarsulation wos to pressurize gas in smali-bore glass tubes, which would ke sealed off into sousage-like sections. These sactions would then be incorp:.ated into a glass matrix. Using an internal pressure of 2500 psig for a tube with on inside diometer of C.04 in. and 0).033-in. walls, ard assumirg on irside cy!inder length equct to the inside diometer of the tube, the calculated value for storage wouid equal 6.94 standard liters per liter of storage volume. The use of glass contdiners at such high pressures is doubtless un- realistic. Commercial glass ampules (volume =2 cc each) were encapsulated in giass (Fig. 3) and in epoxy resin (Fig. 4) in our scoping studies. These ampules, which were very thin-walled, were filled with air at room temperature (~ 26°C) and atmospheric pressure. They contained between 32 and 35%, by volume, of the total storage volume, or betveen 0.288 and 0.315 standard liter of air per liter of storage volume. et it e s o PPN § Y o BT T —— TN, W Al ¢ 7 18 PHOTO 963%4 Fig. 3. Commerciol Glass Ampules Encopsulated in Glass. Cross section obtained by sawing through stoinless steel contuiner. The encapsulating gloss was not melted to sufficient fluidity to remove all voids. [ PN, 19 gl PHOTO 96393 ~ e e e ——————— e i) ’ L] i A . ) ' } T . . - TR Ea y - . P Y e wE L. ; Y- . Y B i ‘: e . R . T L gel. 1 f" 1§ l OAK RIDGE NATIONAL LABORATORY Fig. 4. Commercial Glass Ampules Encapsulated in Epoxy Resin. 20 A few of the ampules cracked during encapsulation in clear epoxy. It is probable that a larger number cracked during encapsulation in the glass at about 600°C. Steel ampules would be much more practical than glass. particularly for encapsulction in resin or in metel. Commercially available industrial gases are sompressed tc 1500 psig (at 21°C) in steel pressure bulbs that may have void volumes os small as 5.5 cc. Using a similar commercial bulb,* about 54 standard liters of gas per liter of storage space could be readily obtained. ke of higher encapsulation nrocciirme n /N lnm6rhl"‘1( rhnlrl Mllu 1™ A/ chH C'l\‘ ur\l wno A fize nar b me n& r' W W Wew 4 a , ST wwywrw $IFw Tawiwiifyw wi u .Iv vv U storage space to 50% of that obtained in current practice. These storage volumes appear to be obtainable under much sy odmitting gas molecule: larger than those which are normaiiy aliowed to pass. Cooling of the system, while maintaining the pressure, causes the gas to be physically entrapped within the sieve. Heating of the entrapped gas to encapsulation temperature under zero partial pressure will result in the eventual release of all of the gas. As much argon or krypton con be encapsulated in these materials as is normally compressed into corresponding standard gas cylinder:‘.,é'l4 provided a sufficiently high encapsulation pressure is employed. The data in Table 4 assume a pressure of 62,500 sig at 350°C. Cata for argon indicate that coout 50% as much gas is encapsulated at S000 psig. sing Type A Linde molecular sieve, some leakage of the encopsulated argon always occurred; nowever, zero leakage of krypton is claimed during 30 doys' storage of the krypton encapsulated in a special sieve material hoving a K/Na atom ratis of 40/6C. 14 This entrapment technique must not be confused with the more common use of molecuiar sieves as adsorbents. Aciually, it corresporas more nearly to the behavior of a clathrate; that is, once entrapment has raken place, the gas in the structure is no longer in equilibrium with extemal gos but is essentially in o micro-container. Although the *Part No. 237 manufuctured by Knapp~Monarch Co.; volume = 20 cc. e T T ya entrapped gos is released upon heating to 350-400°C, the sieve structure is stable to obove 700°C. Therefore, we need to find some way of sealing the pores at temperatures up to those at which the sieve structure is destroyed. In our scoping studies, we entroppec betwzen 5.7 and 6.1 cc of argon per gram of sieve by heating the sieve to 350°C under 1000 psig of argon for 1-1/2 to 2 hr and ailowing it to cool overnight under the pressurized argon. MNolten Pemco 41G glass (~ 600°C) was poured onto the unheated sieve in a stainless steel tube. Analysis of the resulting ccnglomerate showed thet cpproximstely 40% of the entrapped argon had been retcined in the mixture. Cther experiments showed that the loaded sieve is difficult to coat with glass because the sieve releases gas rapidly at these *emperctures. Encapsulation of the loaded sieve in Araldite epoxy resin produced almost no bubbling, and the few bubbles that were produced were held by the matrix. The epoxy penetrates the sieve material readily, and most of the gas remains effectively entrapped during des!riction of the resin with solvents and subsequent treatment of the sieve with hot water. Products containing about equal volumes of resin, aluminum powder, and sieve were easily prepared. Loaded sieve encapsulated in molten Wood's metal retained about 1.8% of the entrapped gos (Table 4). A similor technique would involve the preparation of a clothrate of krypton, which could then be encapsulated in o suitable maiiix. For example, a recurring suggestion is to encapsulate the well-characterized hydroquinone clathrote (up to 57 standard liters of gos per liter of clc:i'hrute)22 in plastic. Although clathrates are thermodynamically unstable at all temperatures,23 the hydroquinone clathrate structure decomposes only slowly below 172.5°C, the melting point of hydroquinone. Gas is lost rather rapidly at temperatures of 130°C and higher. The hydroquinone clathrate of krypton decomposes more slowly than its argon counterpart. The hydro- quinore clathrate is surprisingly stable - irradiation; argon was lost more slowly from samples irradiated up to total doses of IO9 rads than from unirradiated somples.23 This radiation phenomenon has not been adequately investigated. a—y T 22 We prepared the hydroquinone clathrate of krypton, a random sompie of which, on analysis, was found to contain 6.7%, by weight, of krypton. When we encapsulated selected crystals of the clathrate in clear Araldite epexy r=sin, a very slight reaction occurred between the resin and rthe hydroarinone. This soon cecsed. Only ¢ negligible amount of gas was released into the resin, and no appreciable changes were noticex: in the shape or the structure of the clathrate crystals. i il depeind upon ihe deveiop- ment of clothrates exhibiting stability to a highly radioactive environment, to the temperatures necessary for encapsulation, and toward the matrix used. The zeolitic moleculaor sieves proposed for entropment approximate such a nioterial. It seems possible that true inorganic clathrates with the necessary stability can be developed if sufficient need exists. It also seems pcssible to use organic clathrctes as inter- mediates in the formation of gas dispersions (see Sect. 4.3.3 below). 4.3.3 Direct Dispersions® of Gas in Stable Matrices Dispersions of gas in liquid can be generated in a variety of ways; the following methods are being considered in this program: (1) Agitation of the interface between liquid and gas in such a monner that the gas is dispersed into the liquid in smoll bubbles (e.g., "blending™). (2) Introduction of the gas into the liquid in the form of bubbles that are sufficiently small to form a stable dispersion. The two-fluid nozzle is ote method of forming these small bubbles. (3) Incorporation of gas-hearing materials (e.g., compounds, clathrates, adsorhates, etc.) into a matrix, followea by treatment to produce gas evolution in place. Commercial foam glass is made :n this manner. *The word "dispersion’ is considered more svitable here than "foam". Foam is defined (Meriam-Webster, 1966) as a "|ighl’, frothy mass of fine bubbles...." and connotes mechanical weakness. - —— 23 (4) Incorporaticn of very small capsules of gas into a matrix in which the copsule walls dissolve or otherwise lose their identity. The use of very small glass capsules in a glass matrix would be an example of this type of generation. We have investigated method (1), in some detail, using Newtonian fluids (polybutenes)* as matrices instead of materials (2.g., glass) that weuld suosequently solidify. Our experimental arrangements inciuded ordinary laboratory stirre.s, commercial blenders (Waring Blendor), and specially designed equipment comisting of impeliers that were carefully machined to give accurote angles of blode pitch and are driven from below like those of the blender. Dispersions containing a meximum of about 25% of gas, by volume, were obtained in both flowing and static systems. These maximum loadings were achieved only at impeller speeds equal to, or greater than, about 5000 rpm. The maximum loading of about 25% seems to represent the point at which the rate of bubble coalescence and escupe becomes equal to the rate of bubbie formation in t!e systems studied {(Fig. 5). At loodings below the maximum, the looding rate increased as tha rate of power input into the system increased (Fig. 6). Gas could be convenientiy fed into the blender from abave (i.e., through the vortex}; bubbiing the gas through the liquid increased the rate of loading only s'ightly and aid not offect moximum loading. Formation of an open vortex reaching from the gas-liquid interface to the impeller was a critical process. The loading rate increased instantaneously when such a vortex was formed. The pitch of the impeller blades was important primarily as it affected vortex formation and power input; that is, maximum loading was independent of impeller blade pite! (Fig. 7), but ¢ high angle of pitct: resulted in a more rapid transfer of powzr to the liquid and, consequently, in a more rapid attainment of maxirium loading. *Obtained from the Oronite Division, Chevrr~ Chemical Co. The gradzs that were studied had absolute viscosities ranging from about 1,000 to about 24,000 centipoises at 100°F (37.8°C). However, only the leost-viscous grade wos studied at length. More ctudies are needed to define adequately the effect of changing viscosity. ORNL DVG, 69-10607 30 B ® /’.....—3 ® '2 - O 20 / p— o Z o O 7/ O L / a ./ Wl >3 2 O 10— > ® 0 | l | 1 | I I O 4 2 3 4q 5 6 7 TIME (min) Fig. 5. Gas Loading os a Function of Blending Time. Batch study in a blender. (From ORNL-MIT-81) v NRNL DWG_ £9.10608 30 n-73 @ - 95 O A/ >~ ya A-170 _e 2 0~ @-550 o 9 FLOW RATES (ml/min) v 2§ O h /‘ Q | £ ol L O ! > ® A S/ 0 50 100 150 200 250 POWER INPUT (w) Fig. 6. Percentage Loading as a Function of Power Input and Liquid Flow Rates in a Flowing System Using Polybutene No. 16. (From ORNL~-MIT-81) —— e 26 ORNL DWG. 69-10609 30 O Z 9( 20+ < O A A O\O - {0+ O > 0 | | | B | 9] 20 40 60 80 IMPELLER BLADE PITCH (°) Fig. 7. Effect of Impeller Blade Pitch on Gas Loading in Poly>utene 16 at Constant Power Input (155 w at a flow rcte of 74 ml/min). (From ORNL-MIT-81) 27 There were indications that higher loadings might be possible with the more viscous grades of polybutene, but the effeci, if real, tended to be cbscured by the decrease in viscosity ccused by the temperature rise accompanying the rapid transfer of energy to the fluid. Additional work needs to be done to define the effect of viscosity on gas loading. The disoersiors produced were quite stable for periods of time sufficient to allow pouring, measurement of viscosity, etc. The dispersions were also Newtoniar in behavior. Further detaiis of the polybutene work are reported elsewhere.24'25 Similar experiments in which air was dispersed in lead borate glass,* using a |laboratory-type stirre~impeller attached to a high-speed,varicble-speed (max = 20,000 rpm) motor at temperatures between 500 and 700°C, resuited in products that contained a maximum of 23% gas, by volume, after cooling. Much more sophis- ti cated equipment must be used if closely controlled experiments are to be done with glass. Attempts at producing a dispersion in Wood's metal were unsuccessful; o maximum of only about 3%, by volume, of gas was obtained. [n Araldite epoxy resin, uniform dispersions were inifially obtained using a blender; Lowever, the bubbles portially coalesced and rose to the top of the samples, resulting in sharply defined bubble layers representing dbout % of the total resin-gas volume. Within these layers, the gas accounted for about 67% of the volume; on the cther hand, the lower paits of the somplgs were almost free frcm gos. Attempts were mode to circumvent this phenomenon and produce a uniform dispersion by postponing the blénding step until immediateiy before the resin underwent its initial "set." Under these circum- stances, however, nearly all of the gas was rejected before the resin hardened. The generation of stable dispersions of gas in glass is certainly possible by the use of blending techniques. The application of this method to the encapsuiaiion of radioactive gas on a practical scale will involve the solution of difficult engineering problems resulting from the simultaneous use of high temperatures, high impeller speeds, and moderately high pressures. The glass used in cur experiments was selected *Pemco 41G manufactured by the Glidden Co. PRSPy v e o Ty ST e —rR—WTrS < nm L because of its relatively low softening temperature and rapid decrease in viscosity with incieasing temperatures. For routine use in a practical process, other churacter istics would alsc need to be considered (see Sect. 4.2.1). As noted cbove, gases can be dispersed in piastics and resins. However, the dis- persions that we prepared did not have the necessary stobility in the particular plastic em; loyed. It is almost certain that plastics with the required physical properties and hordening characterisiic; are commercially avoiloble. The use of unreinforced piastic matrices will be limited by their stability to heat and to radiatio.: as discussed in Sect. 4.2.2. The experiments with Wood's metal demonstrcied the difficulty of obtaining high gos loadings in relorively nonviscous molien metals. Although better metal candidates than Woed 's metal can be found, the operational problems listed for glass apply generally to metuls. N :merous attempts were made to produce dispersions through the use of the two- fluid nozzle [ method (2) listed in Sect. 4.3.3; see also Fig. 8] in which the gas was introduced via a hypodemic needle into a small tube through which the matrix fluid flowed. Using glycerin, polybutenes, and freshly mixed Araldite epoxy resin as matrix fluids, we were never able to obtain a dizpersion that approached uniformity. By using the nozzie in a downflow position, it was possible to deposit bubbles on a cooled sizrface in such a manner that foom which must locally have centained up to 50% gas, by volume, was formed. However, bubble coalescence and escape occurred at temperatures ranging from ambient to those obtained when dry ize mixtures were used as coolants. Significant volumes of the foam wire never obtaiired. Upflow exgeriinents were equally unsatisfactory. The operction of the two~fluid nozzle is strongly dependent upon the flow ratio ot gas to fivid. Maintaining the proper fiow ratio becomes increasingly difficult as the viscosity of the fluid increases. In principla, it should be possibie to operate a iarg2 number of such devices in a manifold which could be used to fill a steel cv!inder by beginning at t} 2 bottom and slowiy withdrawing 1he manifold as the cylinder is filled. However, the operational problems appear to he vary sevare for a high-temperature process i 1 which gloss is the matrix maerict. e kel i e i i L st At akotn B N b 4 I T S T - .. \\% . D, . W, D, 3 i~ . PER. L. ORNL. DWG. 69-12611 _— GAS INLET e / -&— V[SCOUS FLUID INLET ' SEPARATED BUBBLE 1 / O Fig. 8. Two-Fluid Nozzle as ¢ Bubble Generator (Schematic Diagram). L N § S Mot o e ot 1 st e st K1 We have rt attempted to produce dispersions by interml generation of gases from the dispertion of solid compounds, adsorbates, or encapsulates of gas in matrix materials. However, our experiments aimed ot the encapsulation of gas-loaded molecular sieve: (see above) indicate that the encopsulated gus-bearing materiol must be very finely civided if the bubbles are to be small ond evenly distributed. In glass, for example, much of the antrapped argon escoped in the form of relatively large bubbles. In epoxy resin, where the rate of gas evolution was much slower, bubble cize wos corresnondingly smallar. Encamsulation of the ariginal gas~beoring material ir. the matrix appears 1o be generally more desirable than the generation of a dispession. We briefly tested atomization as ¢ method for producing dispersions. Quaiitatively. the product that was fcrmed f-om glycero! anc air apnecred to be as satisfactory as that produced in the blender. The technique seems impractical for scale-up, parti- culorly with gloss o- metal matricas. 5. DISCUSSION, CONCLUSIONS, AND RECOMMENDATIONS These initial studies serve ro outline the possibilities and to define the probleins thot wili be involved in the development of practical processes for encopsulating fission product gases in solid media. A large number of processes were screened; none were studied intensively. No particular olternative process is clearly superior to the others. Tf the processes considereu (Table 4), the encapsulation cf steel bulbs containing pressurized krypton in metal-filled plastic (e.g., epoxy) oppears to be the most immediately amenable to practical use. Development of new technology will not be necessary. The temperature required for encapsulation would be relatively low and ecsily controlied. The radiation resistance of commercial plcstics is odequate in instances where beta is the chief type of emitted radiation; this type of radiation will be obsorbed by capsule containers. However, the radiation resistance of such plasiics is probably not adequate to permi? direct dispersion of gas into them. En- trapment in molecular sieves is ar intriguing and potentially useful method of storage, e n e e ey cn A e o bt e it il s b+ o 31 provided that an effective method of sealing the sieve pore. can be developed and demonstrated on a practical scale. Relatively high pressures a-e required, although lower loadings ot lower pressures may be acceptable. ‘the encapsulotion of the looded sieve in the matrix would not require high pressure. The slow rate ot which the process proceeds may be a disodvantage; for example, the pressurized gas requires 1 to Z hr (minimum) at 350 to 400°C to fil! the interior of the sieve, and the pressure must be maintained whiie the loaded sieve is cooled. Ti& produciion of a dispersion containing practical amounts of gas also requires operation at fairly high pressures. Whenr o high-rmelting matrix (e.g., glass) is used, relatively less gas will enter the dispersion than would be true ot lower temperatures. Regardless of the method used for preparation of the dispersion, the mechenical ond materials probiems v+iil muiripiy as the operationcl temperature and pressure increase. Dispersions in highly viscous glasses should pussess very desirable safety properties, but their preparation is difficult. Motrices of metal (e.g., aluminum) approach the ideal mediur for encapsulation in terms of mechanicai strangth, the>mal conductivity, and resistance to radiation domage. Our few experiments with low-meiring alloys have shown less promise than those with either glass or epoxy resins. Stable inorgonic clathrates would also approach the ideal vehicle for long-term storage in every respect except possibly that of thermal conductivity. If ihey were stable at high remperatures, they might not require secondary containment. Un- fortunately, such highly stable inorgonic clothrotes are unkicwn ot presont. Future wo~ should be directed toword the solution of specific probiems that would allow early testing of the encapsulation concept on an engineering scale. Suggested studies are listed below in order of nrobable importonce. 1. For plastic~resin systems, determine which specific resins have the most desirable characteristics of compatibility with metals and with mclecular sieves, stability to irradiation at ievels greater ihon 2 x .IOlo rods ot temperatures up to 150°C, and operational properties that wo:.'d e e e ————— - . 32 make encapsulation of either looded molecular sieve or pressurized steel copsules in a plastic-metal matrix feasible. The thermal con- ductivity of the encapsulate should be measured, and the efficiency with which the matrix seals the pores of the molecular sieve should be detemined. 2. For glass systems, cietermine (on a loboratory scale) the feasibility of preparing dispersions containing 20% or more, by volume, of gas in gloss, of pouring these dispersions into st them to give small=scale prototypes of the proposed shipping-storage contairers. The studies would involve selection of the most suitable type of glass, based on the exparsion coefficients of glass and con- persion, and the nature of the viscosity-vs-temperature curve for the glass. They should clso yield data for use in designing scaled-up equipment for generating dispersions and in predicting the lifetime of such equipment. This would involve corrosion/erosion and metallo- grophic studies of the materials of construction of the generator equipment as a function of operating time. 3. For metal systems, determine the feasibility of encopsulating pressurized steel copsules in aluminum or in other suitable metal matrices. 6. ACKNOWLEDGMENT A major port of the polybutene work was carried out by M. H. Jorris, W. A. Heath, and L. S. Bowers of the MI" School of Chemical Engineering Practice. We are indebted to R. H. Mayer and P. J. Wood of the MIT staff for helpful suggestions in designing the polybutenc experiments and in interpreting the results, and to W. F. Schaffer of ORMNL for the engineering design of the equipment. We are especially indebted to D. E. Spangler for many practical ideas as well as for assistance in all phases of the laboratory work. 10. 7. REFERENCES "Safety Standards for the Packaging of Radioactive and Fissile Materials,” USAEC Manual, Chap. OR 0529, TN-0500-33, 19¢6. R. H. Rainey, W. L. Carter, S. Blumkin, and D. E. Fain, "Separation of Radio~ active Xenon and Krypton from Other Gases by Use of Permselective Membranes," paper SM 110/27, pp. 323-42 in the Proceedings of the Symoosium on Operating and Deveiooments! Exserience in the Tr eatment of Aircome Radioactive Wastes, New York, August 26-30, 1968. J. R. Merrimun, J. H. Pashley, K. E. Habiyer, M. J. Stephenson, ard L. W. Anderson, "Concentration and Collection of Krypton and Xenon by Selective Absorption in Fluorocaibon Solvents,” paper SM 110/25 presented at the Symposium on Operating and Developmental Experience in the Treatment of Airborne Radioactive Wastes, New York, August 26-30, 1968. C. L. Bendixsen and G. F. Offutt, Rare Gas Recovery Facility at the Idaho Chemical Processing Plant, IN-1221 (April 1969); cf. C. E. Stevenson, IDO- 14453 (1958). G. F. Offutt, Idaho Nuclear Comp., personal communication. G. A. Cook, Argon, Helium, and the Rore Gases. Interscience, New York, 1961, p. 229. C. M. Lederer, J. M. Hollander, and |. Perlman, Table of Isotopes, 6th ed., Wiley, New York, 1967. RIBDOR code; A. R. Irvine, ORNL, personai communication. N. ©. Dudey, Review of Low-Mass Atom Production in Fast Reactors, ANL-7434 (April 1968). A. R. Olsen, ORNL, personal communication, May 20, 1968. i A e e ime 11. 12. 13. 14. 15. 16. 17. 18. 23. 24, 34 MSR Program Semiann. Progr. Rept. Aug. 31. 1968, ORNL-4344, Sect. 5. M. L Bell ond L. E. Mchieese, Unit Operations Section Quarterly Progress Report, October-December 1968, OR'NL-4448 (to be published). MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL~3708, p. 292. L. H. Shaffer ond W. J. Sesny, U.S. Patent 3,316,691 {May 2, 1967). J. H. Holloway, Noble Gas Chemistry, Methuen and Co., London, 1968. R. E. Blanco, "Ultimate Storoge of Volatile Radioactive Wastes in Soiid Foams," letter to D. E. Ferguson, ORNL, dated Mar. 4, 1968. W. W. Parkinson, "Radiation-Resistant Polymers," in Encyzlopedia of Polymer Science and Technology, Interscience, New York (in oress). R. Sheldon and G. B. Stapleton, The Effect of High Energy Rodiation on the Mechanical Properties of Epoxy Resin Systems Used for Particle Accelerator Construction, RHEL,R 152 (1968). Chem. Eng. News, p. 58 (Apr. 13, 1964). Chem. Ena. News, p. 38 (May 17, 1965). H. Lee and K. Neviile, Handbook of Epoxy Resins, McGraw=-Hill, New York, 1967. H. M. Powell, J. Chem. Soc. (London) 1950, 300. K. O. Lindquist and W. S. Diethcrn, Irtem. J. Appl. Radiation and Isotopes 19, 333-44 (1968). M. H. Jorris and L. S. Bowers, A Study of the Dispersion of Gases in Viscous Liquids with Application to Rodioactive Krypton and Xenon Disposal, Part i, ORNL-MIT-81 (May 8, 1969). | W. A. Heath and M. H. Jorris, A Study of the Dispersion of Geses in Viscous Liquids with Agglicofion to Rodioactive Krypton and Xenon Disposal, Part Il, ORNL-M!T-84 (June 4, 1969). D e ——— 1 e At