UNCLASSIFIED ORNL-2116 Chemistry Contract No. W-T405-eng-26 SOLID STATE DIVISION MEASUREMENT AND ANALYSIS OF THE HOLDUP OF GAS MIXTURES BY CHARCOAL ADSORPTION TRAPS W. E. Browmning C. C. Bolta DATE I8SUED OAK RIDGE NATIONAL LABCORATORY Operated by UNION CARBIDE NUCLEAR COMPANY A Division of Union Carbide and Carbon Corporation : Post Office Box P Oak Ridge, Tennessee UNCLASSIFIED UNCLASSIFIED . Co H. ~111- INTERNAL DISTRIBUTION Center . Blology Library 1 2 3. Health Physics Library 5 6 . Central Research Library 48, . Reactor Expserimental Engineering Library . Laboratory Records Department 27. Laboratory Records, ORNL R.C. 28, A. M. 29. L. B. 30. J. 31. J. 32. E. 3k, D, ® ¥ Ll N = CREPTURECODEE Y [ ha. P, L. 24'31: AQ H:fl . T, H. . Z. A. S. 8. E. W. M. J. _ . T. 56. D. D. » = SO b@@pf}?kflfifigb - WJ1 g a n o G6 . 97. 98. 99-103., Rd FQ G, L. Stienl, PU HO Weinberg Bmlet (K-25) . Murray (¥-12) Swartout Taylor Shipley . Biilington Welson . Jordan . Keim Lane Frye, Jr. Livingston Tind Culler Snell » Hollasender Kelley Clewett Morgen Lincoln Househonlder Harrill Winters Cardwell King Miller Howe Cowen 72, 73, 7l 5. 76. 7T 78. 79, 80, 81. 82, 83. 84, 859 86. 87. 88. 8¢9, 90, 91. 92, 93, ok, 95. EXTERNAL DISTRIBUTION Bacher, Brmmett, R. H. ORNL-2116 Chemistry ?Z 52?, . Charpie E. Boyd J. Skinner R. Dickison J. McHargue D. Ackley E. Beall H. Blewitt ¢ r R E A B + . Bolta ", Boudreau . Brosi . Browning . Conlin . Cottrell H. Crawford E. Guss J. Cromer L. Hemphill H. Jenkins W. Keilholtz . V. Klaus 0. Kolb F. Osborne E. Robertson . T. Robinson W, Savage P. Bhields . Sisman B. Trauger . Van Winkle . U. Wehster ¢, Wilson . G. MacPherson {consultant) L. Sproull {consultant) Brooks (consultant) MIT Practice School ORNL -~ Y-12 Technical Library, Document Reference Section Californis Institute of Technology Convair Division of General Dynamics Corporation, San Diego Johns Hopkins Pratt & Whitney Aircraft Division {Fox Project) (1 copy ea. to H. Hersghenson, J. ¥. Krause, V. Scotti, R. Strough, and J. J. Wesgbecher) 10k, 105-788, {300 copies - OTS) Division of Research and Development, AEC, (RO Given distribution as shown in TID-4500 (11lth ed.) under Chemistry ¢ategory DISTRIBUTION PAGE TO BE REMOVED IF REPORT IS GIVEN PUBLIC DISTRIBUTION UNCLASS IFIED MEASUREMENT AND ANAIYSIS OF THE HOLDUP OF GAS MIXTURES BY CHARCOAL ADSORPTION TRAPS by W. E. Browning and C. C. Bolia SUMNATY The purpose of this study is to determine holdup behavior of gaseous fission | products in charcoal traps, The holdup behavior for a given ges is a function of trap geomeltry, type and amount of charcoal, trap temperature, flow rate, and type of inert gas used to purge the trap. Radiokrypton, used to typify gaseous fission products, was swept through.the trap with either nitrogen or helium purge gas. Trap temperatures studied were from + 16°C to ~110°C. It was determined that holdup times of radickrypton were greater for helium purge gas than for nitrogen purge gas. In all cases holdup times increased with decreasing trap temperature. The maximum concentration of krypton in effluent gas was lower for helium than for nitrogen purge gas, and was lower at:lewer temperatures. For a given trap diameter, doubling the length of trap doubleas the time to peak and increases break through time and time to reach the end of the peak. Doubling trap length reduces the maximum effluent krypton concentration by half, | An analytical expression is presented from which It appears possible to predict holdup.curves for gases when trap geometry and the linear adsorption ithherm for the gas mixture on charcoal are knowma Optimized iraps may be designed for any purpose with this information. introduction Gaseous fission products, from regions of nuclear processes which may be con- taminated by such gases, usually cannot ; vented directly to the atmosphere. Charcosl traps have been used to reduce to a safe lewel the rate at which the activity is vented to the atmosphere. o 2 o The traps reduce the amount of activity by holding it up while it decays and in the case of a sudden burst of activity they reduce the rate of activity release by releasing it slowly over a long period of time. To determine the amount of activity leaving the traps as a function of time, the time the radiocactive gases are delayed (holdup time) in the traps must be known, as well as the length of time over which they are released. Both factors for a given gas are a function of trap geometry, type and amount of charcoal, trap temperature and type and flow rate of inert purging gas. The function of these parameters must be known to design and determine the effectiveness of charcoal traps for gaseous fission products. An investigation of the literature on adsorption of gases yielded two pertinent reports. The paper by E. Wicke’ published in 1940 presents a comprehensive study of the "microscopic! mechanism of adscrption in long, small diameter tubes. High con- centrations of COp in nitrogen as carrier gas were studied at 0°C and at 100°C. A mathematical analysis of gas mixing and sorption velocities is presented. As stated in the report, the results obtained are not directly applicable to many industrial processes,; but nevertheless, they afford an excellent foundation for experiments to develop industrial uses of adsorption. D. Gussl reported a study of the adsorption of radickrypton on charcoal in a Solid State Division Semiannual Report. Break through times of radickrypton at room temperature were reported for several traps of different geometry. The objective of the experiment was to determine the length of time before the adsorbate first appeared in the effluent gas from the trap, rather than concentration changes with time after break through. The need for information on thils type of work was increased and the present experiment was a natural outgrowth of Guss' worke Experimental Procedure An experiment was set up (see Fig, 1) using radickrypton to typify fission gases, Radickrypton was chosen because it is the first fission gas to emerge from a -3 - éharcoal trap. Ten year krypton was used in this work so decay is negligible, Two different purging pases, nitrogen and helium,were used to sweep the radiokryptbn through the trap. The trap was made_from 2% schedule 40 stainless steel pipe and was 13" long. It contained 3/4 1b of 8 to 14 mesh Columbia ACA Charcoal. Trap temperatures were kep® constant using freezing sclutions of Calls brine for the wfiOOGQ and freezing sclution of Freon 11 and 22 for the -110°C temperature, Idquid nitrogen was used tc keep the material at its freezing point. For each run the trap was filled with purge gas, the manifold was evacuated and radiokrypton allowed te fiow into the krypton chamber. The chamber was sealed and the rest of the manifoid evacuated, Carrier gas was then allowed to push the radickrypten inte the traps at constant flow rete of 5.0 £ per hr. The concentra~ tion of radiokrypton in the effliuent gas and thersfore its partial pressure can be determined.by:msasuring the activity of the gas. The relative activity of the effivent gas was measured ag it passed through a small cell, one wall of whieh‘was the end windc& of a Geiger tube. The gés activity was registered.on a log~coufiting rate reccrder; Experimental Resulis Figures 2, 3, 4, % show the resulis obtained for nitrogen purge gas using one trap, and usiflg two identical traps io aéries, at ¢ léOG, §5OC, ~51°C and alloéc respectively. In sach case.the activity injected wag determined By integration of the activity vs, éimfi curves, All were narmalifiéd to the same area under the curve. It can be seen that for a given trap geometry the holdup times, i.e., time for activity to break through, time for activiiy to peak, and time for last appearance of activity, all increase with.fieereafiing temperature., This can be predicted since the energy of a:gas molecule‘is iess at low %amperaturg and it eannct as readlily desord from the charcoal surface into the moving gas siream, The maximum concentration of radio. krypton in effluent gas i3 lower for lower temperatures, -l - The effect of using twe identical traps in series with nitrogen purge gas is also shown on figures 2, 3, 4 and 5., These curves show that the time to peak is approximately twice as long as for one trap and that the break-through time is from twice as long to three times as long as for one trap. The time to the end of the effluent activity is also increased and the maximmunm radiokrypton concentration is about half as great. Figure 6 shows two curves for helium purge gas at ~5°C and ~51°C. The shape of these curves is somewhat uncertain since their accuracy was limited by the sensitivity of the flowmeter in its lower range when used with He. Several runs were made at each temperature and the holdup times are significent even if the exact shape of the curve may be in doubit. These curves show that radiokrypton is held up for longer periods of time in the presence of flowing helium than in the presence of flowing nitrogen. Since nitrogen is itselif adsorbed in large amounts on charcoal, the charcoal surface is not as free to adsord krypton. Helium will pass over char- coal with very little adsorption, Hence, krypton is retained on the charcoal for longer periocds of time in the presence of flowing helium. Meximum radiokrypton concentrations are significantly lower for helium than for nitrogen. Figures 7 and 8 show the same data as figures 2, 4, and 5 plotted together for comparison. The time to reach maximum activity and the height and duration of the peaks are clearly indicated as a function of temperature. Analvtical Theory & theoretical analysis of the holdup process in charcoal traps is postulated, making use of a number of theoretical chercoal-filled chambers, N, cormected in series, occupying the total volume of the trap, This analysis is similar to the continuous dilution tank problem in chemical engineering, It is assumed thalt as gas enters each chamber it is instantly spread and brought to adsorpiion equilibrium throughout the entire theoretical chamber, These conditions apply only to systems whers dllution - applicable t¢ liquid chramatcgraphy and ion exchange where these procssses are slower, The rate of removal of radiokrypton in each chamber is first order with respect to its partial pressure, and therefare to its concentration in that chamber, iaefis\ N differential equations of this type for the N chambers are solved simultaneously. The gemeral equation for the BB chamber iss o e = 1) LN -1) Rt km where Py = Partial pressure of radickrypton {atm) it Amount of radickrypion injected into first theoretical chamber (cc-atm) ig Number of theoretical chambers = Flow rate (cc/min) it Time after injection of pulse (min) " o+ = Slope of the linear isatherm.; = kP for radickrypton in the mixture of i inert purge gas and radlokryptan o (LZrgim m = Amount of adsorbent {charcoal) in trap - {gms) x = Amount of gas (iadiekrypton) adsorbed isothermally (cé et | . km is e messure of the adsarptive capacity of the charcoal trap for radlokrypton in the presence of purge gas at the given iemperaturefi In order te fit the expression te the experimental curves of activity vs. time, experimental data were used from thres points on each curve: time to reach maximum partial pressure and the $w§ times to reach a definité fraction of the maximum.partial pressure (1/10 or 1/2), The time to, to reach meximum partial pressure P, is -6 - obtained from equation (1) by setting ‘%%“ =0, (N =~ 1) (km) 2 % e s it (2) —— == Substituting equation (2) into equation (1) gives: N -1 L N(N - 1) A ew(N - 1) (3) P w1 ) Equation (3) will give the maximum concentration of radiokrypton in the effluent gas, and F P is the maximum rate of release of activity, Dividing equation (3) vy 10 (or 2) and using experimental values for tp.y, t at 1/10 (or 1/2), P .. and F, equations (2) and (3) are solved simultéfieously for N and for km., These values of N and km are used in equation (1). Analvtical Resultis Figures 2 through 5 show how closely the curves from the analytical expression fit the experimental curves. The deviation is within experimental error, most of which comes from counting statisties, Figure 9 is a plot of parameters N and ln (km) as a function of tempsrature as determined by fitting equation {1) to the data in figures 2 to 5. N has the same value at all temperatures, Figure 9 indicates that km and therefore k increases exponentially as 1/T increases (decreasing temperature) since the curves are for constant charcoal mass m. Analtyical Interpretation In the analytical expression, equation (1), km and N determine the time %o maximum krypton concentration, and the narrowness of the peaks respectively. The quantity (km) will determine the time to reach maximum partial pressure and therefore concentration, regardless of trap geometry for a-given trap volume. km is directly proportional to trap volume (amount of charecoal) since the slope k of the linear - T e isotherm g = kPl/n is constant for a given temperature (where n = 1 for low partial pressure and x < amount of gas adscrbed:isothermally-by m grams of adscrbent with equilibrium pressure P).,2 One would expect that for krypton on charcoal saturated with helium the slope ky will be greater than for-kryfiton on charcoal saturated with nitrogen, because of interference b}iédsorbed nitrogen. Therefore the holdup times for krypton with helium purge gas should be greater than those for nitrogen as observed. If km, the parameter in equation (1) which specifies holdup time, is the slope of the linear isotherm, multiplied by the amount of charcoal, then the logarithm of km plotted ageinst % should yield a straight line as shown in Fig. 9. The acti@tion energy caleulated from this curve is 3.0 k cal., for adsorption of krypton on nitrogfinnsaturated charcoal, This straight line relationship indicates that the para- meter k is proportional to the siope of the linear isotherm. Whether the proportion- ality factor is unity has not yet been determined. At a given temperaiure, km,:for two traps in series, is twice as large as for ome trap. This should be true since k is constant for a given temperature and m is doubled. In Fig. 9 it appears that the mumber of theoretical chambers N is a character- istic of the geometry of the charcoal trap since N does not vary for the different temperatures sfudied@ Thisgshould be true if the model used to derive the expression for the curve has physical Significancea: For two traps in series,fl.is twice as large as for one trap, indicating.that ¥ may be a first order function of the length of the trap divided bj the diametefia, | N is a function of trap geometry and does not effect the time to reach maximum concentrationQ: It does effect the shape of the concentrstion vs. time curve, and therefore, the maximum concentration for a given amount of adsorbate. For small N, the w 8 curve will be flat and broad; for large N, the curve will be more sharply peaked. For example: Effluent Krypton Concentration = Time At a given trap temperature (or value of km), the breakthrough time and time to end of pulse is influenced by N. This effect is shown clearly by comparing the varlous | curves for one and for two traps. The time 1o reach maximum concentration is.twice ag great for two traps in series as for one trap since km is twice as large. The breakthrough time for two traps is about four times as long as for one trap and the tlme the last actlv1fy appeared is somewhat greater than twice as longe. If N is a function of trap geomeiry only, it appears that the function is the length of trap divided by trap diameter. For nitrogen, N is constant. Since the data for helium is somewhat in doubt because the flow rate was erratic, the value of N obtained from these curves varles. Values of N for the hellum purge gas runs varied from 6 to 14, This is a feirly good indication that the N obtained is not greatly different from those obtained using nitrogen purge gas. A more complete study of N with He must be made, -0 - Equation {3) is of particular interest since Ppax indicates the maximum concen- tration of fission gases in the effluent gas and F o Pyay is the maximm rate of release of fission gases tb the atmosphete, Both quantities are inversely propor- tional to trap size, m, and %o k. Since k is very sénsitive to temperature, as discussed above, it is more efficient to reduce F o Ppoy by chilling the trap than by using a larger trap, especially if trap size 1s important as when radiation shield- ing is required. F # Ppay is smaller for small N as in short large diameter traps. The flow rate, F, should be minimized.tokeep F Ppax 2t & minimm. For application to gas chromatography, both equations {2) and (3) must be considered. To obtain maximum separation of different adsorbate gases the difference between tpay for the differefit adsorbates should be a8 large as possible., However, to gain sensitivity of detection Py, should be large. These two conditions place opposite requirements on k, the only parameter which is expected to depend on adscrbate gas. Improved resolution cen be obtained by using large N so that concentration peaks are high and narrow, since %, iz relatively insensitive to N while Pp,y depends strongly on N. The long thin traps conventiomally used in gas chromatography provide the large N required for optimum design. Conclusiong The expression of equation (1},derived in a previous section can be used to determine holdup times, maximum concentration, and the shape of %hé concentration curve of a gas carried by an inert purge gas when valunes for N and k are known. A close estimate of N may be obtainéd for the narrow range tested from the trap geometry using N equal to length of trap divided by trap diameter. The slope of the linear isotherm k mist be obtained for mixtures sither from the literature or by experiment. If k at two temperatures is known; then k at any cther temperature msy be obtained by plotting In k ve. temperature as a straight line., Witk this information, an opiimized trap can be designed for any purpase; If a low maxdimum efflivent concentration of adsorbate is - 10 = needed, then N should be small and the trap temperature as low as possible. For a given volume, & short wide diameter trap would be best, If a trap is needed which will holdup the adsorbate gas for a relatively long time and then release the gas in a short, high concentration peak as is done in gas chromatography, then a long, small diameter trap is best, This geometry will meke N large. For a given volume of charcoal, the time delay that is nseded before the gas leaves the trap will determine the trap temperature and therefore k., The cholce of purge gas will also determine the time delay of adsorbate in the trap. For long time delays, a purge gas that is not itself adsorbed on charcoal to a great extent is better than a purge gas that has appreciable adsorption, This analysis makes it possible to prediet the concentration as a function of time of mixtures of gas flowing through charcoal traps. Future Work A study is in progress to determine if N is always a linear function of trap length divided by trap diameter by studying traps of wvarying length to diameter ratio with the same volume of charcoal used in the work presented in this report, These traps will be purged with a variety of gases some of which are nitrogen, helium, argon and alr, Adsorbates used will be radlokrypton and radioxenon. Exact amounts of adsor- bate injected will be measured, Saturation effects of adsorbate on charcoal will be studied. This will be done with a radioactlve tracer of Xr mixed with inactive Kr to increase the XKr partial pressure many fold, Studies will be mede with varylng trap geometry, temperature and gases and with adsorbate continucusly injected into the traps instead of a single short duration injection. Saturation valuss and equilibrium concentration will be correlated. Pre- diction of the results will be attempted using an extenslon of the theory presented in this report. - 11 = Experimental measurements of isotherms will be made for gases in mixtures of interest to determine whether ths k determined in this study is indeed identical with the adsorption isotherm, UNCLASSIFIED ORNL-LR-DWS 10822 il v — KRYPTON TRAP / / GM COUNTER Vs VACUUM PUMP / o CHARCOAL TRAP — / KRYPTON CHAMBER > ~ T — FLOWMETER ™ a " — | FREEZING SOLUTION UNCLASSIFIED 1(:)5 ORNL LR-DWG 13046 A EXPER!MENTAL ONE-TRAP SYSTEM AT16° — A ANALYTICAL; ONE ~-TRAP SYSTEM ff.‘..;.:'_f;fj;__________ff; S5 | N=T83 k,m=1487 x 10% cm3 e - EXPERIMENTAL TWO-TRAP SYSTEM AT 18°C ~1 7 0 ANALYTICAL: TWO TRAP SYSTEM 5 | VN=15.00, am = 3.29 x 10% cm® RELATIVE EFFLUENT ACTIVITY (counts/min) O 10 20 30 40 50 60 70 TIME (min}) Fig. 2. Holdup of Radickrypton in Nitrogen-Purged Charcoal Traps as a Function of Temperature. One trap held at 16°C; two traps in series held at 15°C. - 4 UNCLASSIFIED ORNL—LR~DWG 13047 10° e - | e — ] — | A EXPERIMENTAL ONE TRAP SYSTEM AT 5C o 5 | & ANALYTICAL; ONE - TRAP SYSTEM -~ — N=T727, /rm = 2 .05 x ‘10 'Ci'!‘\3 — 'fi—*fi @ EXPERIMENTAL: TWO-TRAP SYSTEM AT OC 2 -0 ANALYTICAL; TWO TRAP SYSTEM N =144.2, /rm = 4.82 x 1O cm?® S LY wn N o W w N RELATIVE EFFLUENT ACTIVITY {counts/min) o Mo TIME (min) Fig. 3. Holdup of Radiokrypton in Nitrogen-Purged Charcoal Traps as a Function of Temperature. One trap held at +5°C; two traps in series held at 0°C, _]5... UNGLASSIFIED o5 ORNL-LR-DWG 13048 - & EXPERIMENTAL: ONE-TRAP SYSTEM AT —51°C 5 A ANALYTICAL: ONE-TRAP SYSTEM N=7.43 km=1.09 x 10° ¢cm> L ® EXPERIMENTAL . TWO-TRAP SYSTEM AT —50°C o ANALYTICAL;, TWO-TRAP SYSTEM N=15.0. km =1.90 x10° cm> 10 _~ONE TRAP TWO TRAPS 10 e Ao e 10° |4 \\ o RELATIVE EFFLUENT ACTIVITY (counts /min) 10 O 40 80 120 160 200 240 TIME (min) Fig. 4. Holdup of Radiokrypton in Nitrogen-Purged Charcoal Traps as a Function of Temperature, One trap held at =51°C; two traps in series held at -50°C, RELATIVE EFFLUENT ACTIVITY {counts/min) e 10 {0 ..,]6... UNGCLASSIFIED ORNL-LR-DWG 43049 A EXPERIMENTAL . ONE-TRAP SYSTEM AT —#10°C —— A ANALYTICAL. ONE-TRAP SYSTEM _ N=8.03, km=6.74 x 10° cm® e EXPERIMENTAL: TWO-TRAP SYSTEM AT —#{0°C | o ANALYTICAL: TWO-TRAP SYSTEM N=151. xm=152 x 10® cm? 500 TIME (min) {000 Fig. 5. Holdup of Radiockrypton in Nitrogen-Purged Charcoal Traps as a Function of Temperature. Traps held at ~110°C, 1500 UNCLASSIFIED ORNL-LR-DWG 43050 104 & EXPERIMENTAL; ONE - TRAP | SYSTEM AT —-5°C A ANALYTICAL; ONE~TRAP - SYSTEM | N o= 44,21, - km = 2.0 x 104 cm? RELATIVE EFFLUENT ACTIVITY {counts/min) 0 20 40 60 80 100 120 TIME (min) ‘103 [ T f T Yo o] - r— A EXPERIMENTAL; ~ ONE-TRAP SYSTEM ~ | Mo S— 5 |t AT 50 s N=14.85 .. . 4 ANALYTICAL: RELATIVE EFFLUENT ACTIVITY (counis/min) ZOQ TIME {(min) R N o ) AV Y SN i N 10 ‘ o 100 300 400 500 600 Fig. 6. Holdup of Radiokrypton in Helium-Purged Charcoal Traps at -5 and -50°C. ..,'{8.,. UNCLASSIFIED 5 ORNL-LR~-DWG 13051 RELATIVE EFFLUENT ACTIVITY (counts/min) 0 100 200 300 400 500 600 700 TIME (min) Fig. 7. Comparison of Experimental Data on Holdup of Radiokrypton in Nitrogen- Purged One-Trap Systems at Varicus Temperatures., RELATIVE EFFLUENT ACTIVITY (counts/min) UNCLASSIFIED ' ORNL~LR-DWG 13052 105 . T : T : : S NS SRS - : e e i q 5 O ",¥,,7ffi4_a____“_..q..._ . e — — - — — I N SR S S S S 2 * | —] ot : e e e — =) T T e . | 4 L3 S l I i | | | T i ; g e e ks e -s—-—-»--\-] | ! | I & e 1 { | 1 3 o -4-.-.....-—-—-{-.._.}..-._-_ 1 > " -“"'b i Mo R R o e e e L e ] ! i P —ee B | | o \ v . \ 1 | i { . 1 — f j N F L o S | fi el wlohe s T i s e i B S W S gt Y A o | o * T ! ~ _— X) S e | — ey < R oh A é — | e \‘ 3 £/ . : : - .__—‘—f-,.___(‘i - - — e — J‘ — 100 200 300 400 500 600 700 800 S00 1000 1100 {200 1300 1400 TiME (min) Fig. 8. Compdrison of Experimental Data on Holdup of Radickrypton in Nitrogen~ Purged Two-Trap Systems at Various Temperatures. -&1= -20- UNCLASSIFIED ORNL~LR-DWG 13053 TEMPERATURE (°C) O -50 —{00 3X106 3X10 10 5 N 2 1 5 z 10! 0.0030 ' 0.0040 Q.0050 0.0060 Vr (°K ™ {) Fig. 9. Comparison of Analytical Constants N and km for One~ and Two-Trap Systems as a Function of Temperature. - 2D - Bibliogra Guss, D., Solid State Division Semiannual Progress Report August 31, 1955 RNI~1944 lewis, Squires and Broughton -~ ®Industrial Chemistry of Collodial and Amorphous Materials® MeMillan, 1952 Wicke, E., Koll. Z., 93, 129 (1940)