vv International Atomic Energy Agency IWGGCR/2 i INTERNATIONAL WORKING GROUP ON GAS-COOLED REACTORS

Specialists Meeting on Coolant Chemistry, Plate-out and Decontamination in Gas-cooled Reactors

Juelich, Federal Republic of Germany 2-4 December 1980

SUMMARY REPORT INTERNATIONAL ATOMIC ENERGY AGENCY

Specialists' Meeting on Coolant Chemistry, Plate-out and Decontamination in Gas-cooled Reactors

Juelich, Federal Republic of Germany 2-4 December 1980

Chairman: C.-B. von der Decken Kernforschungsanlage Juelich GmbH. Institut fuer Reaktorbauelemente Juelich, Federal Republic of Germany

Scientific Secretary: J. Kupitz International Atomic Energy Agency Vienna, Austria

SUMMARY REPORT Printed by the IAEA in Austria May 1981 CONTENTS

1. INTRODUCTION 5

2. SUMMARY AND CONCLUSIONS

2A. Fission Product Plnte-out 1 & 2 5 2П. Decontamination of Activity 6 2C. Coolant Chemistry 6 21). Discussions, Conclusions and Recommendations 7

3. CONTRIBUTIONS

In-pile helium loop "Comédie" 8 J J. Abnssin, R.J. Blanchard, J. Gentil Out-ot'-pile helium loop for liftoff experiments 17 R.J. Blanchard, A. Bros, J. Gentil Experimental facilities for plate-out investigations and future work 26 K. Muenchow, II. Dederichs, N. Iniotakis, B. Sackmann Results from plate-out investigations 35 N. Iniotakis, J. Malinowaki, H. Gottaut, K. Muenchow F.P. plate-out study using in-pile loop OGL-1 44 0. Baba Fission product behavior in the peach bottom and Fort St. Vrain HTGRs 49 D.L. Hanson, N.L. Baldwin, D.E. Strong Iodine sorption and desorption from low-alloy steel. and graphite 55 R.P. Wichner, M.F. Osborne, R.A. Lorenz, R.B. Briggs Remarks on possibilities and limitations of theoretical approach to plate-out problems 64 E. Obryk Fission product behaviour in the primary circuit of an HTR 68 C.-B. von der Decken, N. Iniotakis Derivation of criteria for primary circuit activity in an HTGR 75 S.D. Su, A.W. Barseil The infLucnce of dust on the hazard potential of a depressurization accident of a high temperature reactor 81 N. Iniotakis, C.-B. von der Decken Modelling of plate-out under gas-cooled reactor (GCR), accident conditions 87 A.R. Ta ig Safety research on iodine plate-out during postulated HTGR coie heatup events 94 A.W. Barsell, O.P. Chawla, CG. Hoot Plate-out measurements and decontamination of a component of the AVR reactor at Juelich 99 J. Hanulik, H. Schmied, J. Wahl Evaluation of n decontamination model 107 O.W.T. Rippin, Л. Hanul.ik,^. Schenker, G. Ullrich Decontamination and high temperature materials 114 E. Schenker, G. Ullrich, .1. Ilanulik, W.B. Waeber, K.M. Wiedemann Maintenance concept of the gas turbine in a 1640 MW direct cycle HTR 11H II. Schmied, H. Karaus, К. Rocllig, E. Schenker Coolant Chemistry of the advanced carbon dioxide cooled reactor .... 125 R.L. Faircloth, K.S. Norwood, H.A. Prior Primary coolant chemistry of the Peach Bottom and Fort St.Vrain high-temperature gas-cooled reactors 132 11.D. Burnette, N.L. Baldwin Chemical reactions during nuclear drying of the AVR primary circuit following a water ingress 137 U. Nieder, К. Vey Predictions on an HTR coolant composition after operational experience with experimental reactors 144 R. Nieder Tritium behaviour in an HTR-system based on AVR-experience 153 W. Steinwarz, II.D. Roehrig, R. Nieder

4. AGENDA OF THE MEETING 161

5. LIST OF PARTICIPANTS 163 1. INTRODUCTION 2. SUMMARY AND CONCLUSIONS 2A. FISSION PRODUCT PLATE-OUT 1 & 2

The Specialists' Meeting on "Coolant Chemistry, Plate-out and Session chairmen: Decontamination in Gas-cooled Reactors" was held at the Fission product plate out 1: Mr. Blanchard Kernforschungsanlage Juelich GmbH., Juelich, Federal Republic of Germany, 2-U December 1980. The meeting was sponsored by the International Atomic Fission product plate out 2г.. Кг. ВаЪа Energy Agency (IAEA) on the recommendation of the International Working Group on Gas-cooled Reactors (IWGGCR) and was attended by 38 participants and observers from France, Federal Republic of Germany, Italy, Japan, Prediction of distribution of fission products which are deposited Poland, Switzerland, the United Kingdom of Great Britain and Northern on surfaces in primary circuits of gas-cooled reactors depends on îmow— Ireland and the United States of America. ledge of the value of a large number of parameters. These paraceters are being estimated by experiments in various test facilities. The purpose of the meeting was to provide a forum for exchange of information on experimental and theoretical results of fission product Two categories of test—facilities are used: behaviour in primary loops of gas-cooled reeactors in order to provide comprehensive review of the present status and of directions for future — in-pile loops applications and development. — out-of-pile loops

The meeting was divided into five sessions: Specimens of irradiated fuel from gas-cooled reactors are used as the source of fission products in the in-pile loops, ue fuel is irradiated A. Fission Product Plate-out I further during the test by neutrons produced in test reactors. In order to

B. Fission Product Plate-out II get an appropriate fission product source burnt-up fuel orf in case of C. Decontamination of Activity coated particles, fuel with a certain amount of defect particles is used. D. Coolant Chemistry Fission products are then carried by the heated coolant gas through, the E. Round Table Discussion plate-out section where some deposit on various material samples to be examined later. Subsequently, the gas is filtered, purified ana flows back to the fuel section. During the meeting papers were presented by the participants on behalf of their countries or organizations. Each presentation was In-pile experiments have been performed or are planned in the following followed by an open discussion in the general area covered by the paper. loops:

- Saphir (Pégase reactor) CSA, Cadarache, Prance - Comédie (Silo'é reactor) CSA, Grenoble, France - Vampyr I (AVR reactor) KFA, Jülich, Federal Republic of Germany - Vampyr II (AVR reactor) KFA, Jülich, Federal Rerublic of Germany - OGL 1 (JMTR reactor) JAERI, Tokai Mura, Japan.

In addition to these in-pile loops, samples of materials that have been loaded with fission products in main circuits of gas-cooled reactors such as in the Dragon Reactor, Peach-Bottom and FSV have been examined.

Concerning the out-of—pile loops, single ion sources such as Cesium, Iodine and Silver are used instead of reactor fuel. These loops allow the study of interaction of selected isotopes with various reactor materials.

Out-of-pile experiments are performed in, or planned for the followir^ loops: б - Scafex (КРА, Jülich, Federal Republic of Germany) 2B. DECONTAMINATION OF ACTIVITY - Smoc (KFA, Federal Republic of Germany) - Lift-off experiments (CEA, Grenoble, France) - Iodine sorption/desorption test facility (ORNL, Oak Ridge, USA) Session chairman; E. Obryk - Helium loop (EIR, WUrenlingen, Switzerland)

Radiation levels caused by fission and activation products deposited in primary circuits of nuclear power plants require decontamination of Experiments are performed under various combinations of several reactor components in order to conduct inspection and repair. The basic parameters such as flow rate, time, temperature, fission product concen­ concept of decontamination by wet chemical treatment is that the activity tration and composition of coolant gas. near the surface boundary is removed as the surface itself is dissolved away by the decontaminating fluid. Comparison of experimental and theo­ Results demonstrated that fission product deposition does not only retical results have proved that the extent of penetration of radioactive depend on adsorption and desorption mechanisms at the surface but also on material into the wall material of reactor components can be predicted. irreversible mechanisms such as diffusion into the wall-raaterial. Adsorption Thus, the depth of the surface material to be dissolved can be determined. and desorption are of main importance at lower temperatures than },Q0oC, but Chemical solutions have been developed in order to remove the desired at temperature above 500° diffusion is dominant more and more with increasing surface layer. These substances have been tested successfully on samples temperature and operation time. loaded with fission products and on components of "the Arbeitsgemeinschaft Versuchsreaktor (AVR) in Jülich. Experimental results together with power plant design parameters are the input for computer codes that have been developed for the prediction of The maintenance concepts for the gas "turbine for a 1640 KW direct fission product behaviour. Codes are available in the Federal Republic of cycle HTR were reported and extensively discussed. The maintenance concepts Germany (KFA/IKB), Japan (JAERI), UK (UKAEA) and USA (CAC). foresee minor inspections at two years intervals and zajor inspectiez every six years. The major inspections require disassembling and dismantling of In addition to Iodine the main fission products to be considered are the turbo-machine and the inspection of the individual parts. The "turbo- Ag 110 m, Cs 134, Cs I37 and Ba/La 140. The influence of dust as a transport machine will be transferred by a traction a?chine into a dismantling cell. mechanism for fission products must be taken into account for the prediction The machine wiT.1 be dismantled by remote—handling. For the disassembling of fisBion product plate-out. of the machine blades two variants were reported:

— remote controlled disassembly of machine blades — manual disassembly of machine blades with pre—decontamination.

After having disassembled the blades they will be decontaminated separately and examined using a special inspection procedure.

After having inspected all components of the turbo-machine it will be re—assembled, mounted and connected into the primary circuit. The power plant shut down tine for the major inspection is expected to be ^0-60 days. 2C. COOLANT CHEMISTRY 2D. DISCUSSIONS, CONCLUSIONS AND RECOMMENDATIONS

Session chairman: R. Wichner Session chairman: С. B. von der Decken

The operation of the Peach Bottom and Port St. Vrain (FSV) HTRs has Safe prediction of fission product behaviour in primary circuits of provided valuable data and experience on the coolant chemistry of primary gas-cooled reactors is an important aspect in the lay—out of reactor ciruits. In Peach Bottom oil contamination was the main source of chemical plants. Deposition of fission products in components of primary circuits imparities, while in FSV moisture ingress and subsequent long term component will influence the plant concepts of maintenance and repair and could be outgassing occurred. In addition, a hydrogen injection experiment was per­ a potential danger to the environment in case of severe accidents. formed in Peach Bottom. Irapuritier of the Peach Bottom reactor mainly consist of hydrogen and methane due to the oil ingress, which led to a coating of Fission product plate-out strongly depends on several parameters, virtually all metallic surfaces with carbon deposits. The deposits caused such as choice of material, coolant gas composition, flow rate, gas and no untoward effects on reactor operation nor on the metallurgy of the wall temperature and time. underlying materials. Influence of these parameters needs to be studied in various experi­ A water ingress from the heat exchanger of the AVR experimental power ments. In order to limit future effort the recommendation was made to station into the primary circuit led to a shut down period of about one make as much use as possible of the already available results in partici­ zear. During this time the defective heat exchanger was repaired and the pating countries. This mainly concerns the experience gained "by the entire plant was inspected. 30 tons of water entered througji a small leak operation of the British Hagnox and Advanced Gas-cooled Reactors. The in the super—heater of the steam generator into the primary circuit while Representatives from the United Kingdom offered design parameters of a GCR-power plant in order to compare their own fission product predictions theQreactor was shut down. Because of the low temperature of about 500 C— 600 С only an extremely small part of the water reacted with the graphite with those made by using the German computer code PATRAS. of the fuel elements. The water was removed through the valves by evacua­ tion and by stepwise nuclear heating of the core and the reactor is now A general comparison of all assumptions and analytical aethods cade again in operation. in computer codes for plate-out prediction was suggested. The Federal Republic of Germany will take the lead and invite a group of experts in In the British Advanced Gas-cooled Reactors (ACH) carbon monoxide and this area and report its meeting results to the IAEA. methane are added to the coolant gas (carbon dioxide) in order to reduce moderator corrosion. Since additions of methane increase the possibility Participants recommended to emphasize in future programmes the follow­ of producing carbonaceous deposits on hot surfaces, such as fuel pins, ing activities: a special coolant strategy has been developed and successfully supplied. - the influence of dust on the deposition The primary circuits of HTRs will continuously be contaminated by a - of fission products in reactor components large number of minor sources, such as hydrogen diffusion processes from - fission product behaviour in the reactor containment the steam circuit, methane production by a radiolytic reaction with hydrogen - accident analysis with specific regard to fission products and carbon monoxide from minor air ingresses. The composition of the coolant in order to safely predict fission product behaviour. gas also influences fission product behaviour in addition to the influence of materials properties and must also be taken into account in the design The participants recommended to publish a summary of this meeting in of experiments. an international journal in order to disseminate information to a wider range of interested readers. The behaviour of Tritium in HTRs is of major importance because it is the only radioactive isotope which diffuses from primary into secondary circuits. In order to reduce the amount of Tritium in the produced gas of process heat plants, oxide layers for the steam side of the heat exchanger and duplex tubes are being developed. 8 3. CONTRIBUTIONS - the fission products balance trapped in the different filter components. - the cumulated released fraction of solid fission products. IN PILE HELIUM LOOP "COMEDIE" DESCRIPTION OF THE EXPERIMENTAL FACILITIES. J.J. ABASSIN, R.J. BLANCHARD, J. GENTIL Centre d'Etudes Nucléaires de Grenoble The loop general assembly is shown on figure 1.

Grenoble The principal parts of the loop are : France 2.1 An in-pile section which includes (see figure 2) ABSTRACT - the irradiation section with the fuel loading, source of fission pro­ ducts and heat production. Above the fuel is placed a graphite diffusion

The SRI test in the COMEDIE loop has permitted to desmonstrate block designed to study the cesium diffusion. particularly the device operation reliability with a fuel loading. - the plateout section where the fission products are deposited. It works

The post-irradiation examinations have pointed out the good as a gas counterflow heat exchanger. filter efficiency and have enable to determine the deposition profiles either for the activation products (e.g : 5lCr, 60Co) or for the fission - the filter which traps iodine, solid fission products and the neutron- products (e.g. : Ag, I, Cs, Cs) activated products.

- the electrical heater (power 50 K'tl) which provides and controls the gas 1. INTRODUCTION. temperature before the irradiation inlet.

The loop is designed to study the release, migration and deposition of fission products formed in high temperature gas cooled reactor. The experiments 2.2. An out of pile section wich includes (fig.1) : are carried out in the SILOE reactor at "CENTRE D'ETUDES NUCLEAIRES DE GRENOBLE" - the filters - the cooler - the blower - the out of pile test section - the electrical heater (38 KW). A first experiment SRO /1/, /2/, enabled us to determine the loop characte­ - the facilities for analysis, gas purification and injection of different ristics and possibilities related to thermal, thermodynamic, chemical and neu- impurities. tronic properties. - the measurement and control of the loop parameters, mainly : teraperatu—

, flow-rate, pressure, permanent gases, rare-fission gases. A second experiment SRI /3/, /4/, has been carried out in high temperature gas cooled reactor operating conditions. It enables us to determine particulary The main loop characteristics are as follows : - the deposition axial profile of activation and fission products in the operating pressure : 20 to 70 bars plateout section constituting the heat exchanger. useful diameter for fuel loading : 70 mm 9 hélium flow rate 16 to 45 g.s 6. POST IRRADIATION EXAMINATION.

fuel surface temperature 800 to 1100°C After the in-pile section dismantling, the examinations have been carried plate out section inlet gas temperature 600 to 835°C out on : out of pile test section temperature 600 to 850°C - the diffusion block - the plateout section - the filter.

The sepctrometry y measurements are performed by means of 2 counting faci­ lities : 3. SR1 EXPERIMENT. LOOP OPERATION. - one for counting the solutions, graphite and molecular sieve - the other A three runs experiment has been carried out with a HTR fuel element. The placed under an automatic conveyor enabling to carry out the test tubes SR1 temperatures in the in-pile section are given on figure 2 with the follo­ Y scanning. In the 2 cases, the counting facility is composed of a

1 wing conditions : pressure 60 bars, flow-rate 40 g.s . Ge(Li) detector connected to a data processing system using a MITRA 15 computer and treated by the NATHALIE code /5/. 4. SR1 EXPERIMENT. FUEL LOADING.

6.1. Filter. The graphite block housing the fuel is a 1160 MWe type, manufactured in The filter schematic diagram is shown on figure 5. At the plateout section PECHINEY, P3JHAN graphite. It has 3 cooling channels of 21.03 mm diameter and 4 outlet, the gas crosses the filtering cartridge which includes the wire gauze fuel channels of 16.20 mm diameter. in stainless steel, the DEXTER paper X 2080 and the sintered stainless steel cartridge (P0RAL). After that, the gas flows through 3 baskets filled with sil- The fuel loading is composed of 32 fuel rods made of BISO fissiles parti­ o vered molecular sieve type 5 A, followed by 2 baskets of normal molecular sieve

cles Th 02, U02, enrichment 92.2% and fertiles particles Th (fig.3). о type Mi 0 4 A.

The source of fission products necessary for the plateout is obtained from After dismantling, each element is recovered and the aliquot portions of the components are analysed : 2282 bare kernels Th 02> U02 arranged in 4 particle holders and placed at the

core mid-plane in the fuel channels (fig.4). The total heavy metal loading - the wire gauze is complety dissolved by cnloronitric acid.

(fissile and fertile) is 4.683 g of U and 187.705 g of Th. - the DEXTER paper is Complety dissolved by fluorhydric acid. - the sintered cartridge is washed during 1 hour with chloronitric acid. 5. FISSION GAS ANALYSIS. - the metallic baskets housing the molecular sieve are washed with chlo­

The loop is designed to sample the gas in 4 sampling points along the gas ronitric acid. circuit. It enables to analyse the fission gases and to determine the R/B by - the molecular sieves are counted in their state. means of NATHALIE and JEROME computer codes /5/. The Y measurement results of the different filter components are given in the table 1 and, the activity distribution are represented : The S5mKr R/8 release rate is about 10 3 for the bare kernels and 10 5 for 235 51 54 59 53 60 the external fuel rod contamination in U /3/. - figure 6 for Cr, Mn, Fe, Co, Co.

126T 131T 129mT 132T - figure 7 for I, I, Te, Te. - figure 8 for 134Cs, 136Cs, 137Cs, 14°3a. 10 It stems from this results .that : the filtering cartridge has an efficien­ 6.3 Diffusion block.

cy maximum for the activation products (100% of efficiency considering the trap- It is an annulus cylinder made in PECHINEY graphite P3JRA2N, 50x17 mm dia­ 51 meter, 80 mm length, placed downstream from the fuel loading. A part of the gas ped quantity). Moreover the Cr has an activity 100 times greater than the flow is forced to cross through the cylinder in order to study the fission other activation products. products permeation through the graphite. After recovery the diffusion block • f • ^ и 110m„ 124ei. 132т 126т 134o 136^ is peeled longitudinally into slices of 0.25 mm thickness. The у spectrometry Concerning the fission products Ag, Sb, Те, I, Cs, Cs 13^Cs, 141Ce, the filtering cartridge efficiency is near 100% of the trapped of the collected powder enables to determine the fission products concentration 131 129m profile as a fonction of the distance from the outer surface. quantity. On the other hand, the molecular sieve mainly traps I, Те, 140„ Ba.

In conclusion, the filter efficiency is near 100% for activation and fis­ 7. NUCLIDES BALANCE. sion products except for iodine and baruym. The table 2 gives the cumulated atom number, released and deposited during the 3 irradiation runs on the following elements : diffusion block, plateout 6.2 Plateout section. section and filter. The heat exchanger is composed of 3 identical bundles including each 7 test tubes of 2800 mm length, 8x10 mm diameter. The choosen materials for the 8. CUMULATED RELEASED FRACTION (F). SR1 bundles are Incoloy 800, Hastelloy X and Stainless Steel AISI 347 (see fi­ gure 9). After the bundle dismantling each test tube is cut into 3 parts of 880 The atom number created in the fuel at the end of the irradiation is calcu­ mm length and placed into a Stainless Steel container with thin cladding ^0,1 lated by means of the CREATION code from : - the fertile and fissile atom num­ mm thickness). The у scanning is performed by displacing the container closely ber contained in the fuel - the irradiation fluxes - the capture, fission and to the Ge(Li) detector shielding. The test tube length seen by the detector is absorption cross sections related to gas cooled reactors. For example, the

40 mm which corresponds to 22 countings per third of tube (see figure 10). created atom number for 131I, 13^Te, 13^Cs, 14°Ba is given for the bare kernels (table 3). Considering the SR1 experimental conditions we can admit that the The activity axial profile for activation and fission products is carried fission products collected are stemming only from the bare kernels. This enable out completely on 3 test tubes of each materials and partly on the others, the us to calculate the corresponding F values from the table 2. whole representing 702 spectra determinations. The activation products which are measured on the tubes are : Cr, Mn, Fe, Co, Co. It has been possi- _ amount released out of the fuel at the time t 89 *1 *1 Om ~ amount createa at the same time t. ble to determine the auantity of the following fission products Sr, Ag,

124_ 125_' 129m,. 132,. 126T 131 _ 134„ 137. 140„ Sb, Sb, Те, Те, I, I, Cs, Cs, Ba. t = elapsed time at the end of the SRI irradiation.

The total balance for the whole plateout section is given in table 1. 9. CONCLUSION. The figures 11, 12, 13, 14, 15, 16 represent an example of the plateout profiles obtained on the total length of the three materials studied, after The SR1 test has permitted to demontraste particularly : smoothing of the values. - the Comédie loop operation reliability, either for thermal or chemical purposes TABUE I TOTAL dALANCE С? Ti-E *i-C'_S ?'-ДТ= CUT SE<"T'»x, - the possibility to determine in the plateout section the deposition pro­

files for the different nuclides. i P.F. INCOtO 3CO 1 -из: 5S 3-17 1 X T3T4I. a Ci -J "i. 1 -j Ci 1

\ 51 Cr 2~ti Î55 J 137*.Г* П.* j 1506 . 147 5345.7 . 517 3.23 14 • 3.~Э £ •л - the satisfying filter efficiency for the activation and fission products. «л [ 5* «in 36.6 'S.« t 20,2 7 Q.7 i H«4 Г о.а Г3.2 7" 5.3 1.13 £ va - J.TO £ * i 53 F* 12.3 7 1.3 1 £.3 7 0,7 ' 1 .1 - 0,1 го.э 7: 1 .7 4.3Т 12 — с. * 74,3 . 3 4.4 5.35 £ - 13 - 3.03 Е • These results should permit to determine the adsorption isotherms and the I :a сэ 31 7 3,1 1 , 24,4 - 3.5 , •зэ.з - 13 i 17.5 7 i .* ! -о.з 7 0.7 ! 23.2 ~ 0.3 31,5 Т 1 .5 4.53 - - а.*4 £ *4 i i CO Co 1 -375 3.53 I аз Sr *223 i есзз, 7 31Э se*3 •?25Э i;a-:3 т: = 13 t З.АЗ = -5 i fission products penetration parameters for the project of high temperature j ПОЛд 32.5 7 "•3 1 24,7 7 *.T 33,3 - 3.5 1СС.Э * 2.53 1.-5 - •А - 2.23 = - i [ 12Л 5a •3.1 z У,3 i J6.6 ~ 5.î . 15.3 - - 1.2 55.2 » 13.э 1,31 £ - 1j Г 2.4л £ 13 1 22 1.Э 5.3 57 27.5 1* - cooled reactors. z 27 1 T.3 7 :7SJ Г Г z.ez - 7 1,^ £ • I 125 1 SO -.53 7 :*.з I 2В.Э ~ 3.1 94,1 - ifl.3 275,Э С =9 4,г* - *3 - 3.53 = "1 1 132 Те 2619 Г a:a 1 Ю35 * 2*05,4 Г 331 ».23 - 13 7 т.23 = - i3 : 1 5239.4 Г - • 1.25 £ 2*9 Г 125 1 iSO.3 7 2C3„* Г 5Э2.Т Г 179 3.4* £ -з тЗ i i 123 I 1 131 I 53.3 7 -л.± ï ла.э 7 7.A 37,3 » 1.1 ; Iii » 17.1 5).50 Е - •г 2 * 2.5* 5. - 12 J O-î Cs 7 i 1 2 7 3.1 1,7 - 3,2 1 5.5 7 1 .4 i г.гз Е 13 - з.*а = - -з ; ! 137 c* 20.9 7 1-1.4 ; э.з £ 1.3 , 25.7 - 11.3 55.3 Г 13.4 г.зэ с * t; 7 £ - 15 ' 1 1« 3*. 31.a з.э 1 2*2.5 » 3.3 33,1 - 2.3 «.7 Г 5.5 * - 2.33 S - а .за REFERENCES.

/1/ J.J. ABASSIN, R.J. BLANCHARO, J. GENTIL, F.J. VEYRAT. TAS1-S Z - LOC? TOTAL BALANCE Uto™ nuMltri Programme de l'expérience SRO dans la boucle d'irradiation à hélium COMEDIE

CR DMG № DR 42/77 diffusion ЫосС 1 Pliïe-cuc lectian | fil;:r Total А^гл

_ - к ; 51 Cr l 1.3T E - 13 1 s.aa E «- Ii 1 6.55 = l-i -I.S5 £

Si Mn 2 E » 12 1 1.13 E » Ii 1 1.29 с 1-i 2.-Ü 121 J.J. ABASSIN, R.J. SLANCHARO, J. GENTIL, J.F. VEYRAT. I .55 - E • 1- I 53 Fe I 3.9Б E * 10 1 -i-31 E - 1Z 1 3.13 £ 12 T-.Î3 £ - 12 { Expérience SRO dans la boucle d'irradiation à hélium COMEDIE 53 Сэ I 1.05 E » 12 1 5.96 с * 13 I 1.21 - 14 1.32 - - i CR DMG № 46/79 SO Сэ I 1.9T E * 13 1 i.SS E * Ii 1 з.-г E 1— ' г.-ss £ 89 Sr I 1 3.53 £ » 15 i 3.Î3 £ - 15 j

ПОГА5 1 l.iZ E - 11 1 i.CS E * Ii [ г.33 E - 13 1-34 £ - Ii 1 /3/ J.J. ABASSIN, R.J. BLANCHARD, J. GENTIL, J.F. VEYRAT. 12J S3 1 1 1.91 E - 13 1 1.СЗ E - 12 -».31 £ • i3 ; Boucle COMEDIE. Dossier de 1' ixpérience SR1. [ 125 SB 1 l г.ег £ - Ii 1 e - 14 \

1 129С Те 1 E - 13 I 1.5i E ** £ CR OMG № 33/80. - • " i i 132 T; 1 1 3.23 £ * 13 1 1.1S £ i* 1 2-11 - i

1 125 I 1 1 3.« E - 13 I з.;г S 13 5-M £ - 13 î /4/ J.J. ABASSIN, R.J. BLANCHARD. 1 131 I 1 Î 5.53 £ * 12 ï 5.55 E - 14 s.ez £ - 14 l 1 13i Cs 1 1 Z.25, £ — 13 1 3.5i E 14 E - -~ t Etude post-irradiatoire de l'expérience SR1 9.23 E - П 3.33 £ - 1 137 Cs 1 Z.ii £ - 13 1 г.зэ Г- 15 I - 15 5-75 £ - 15 ! CR to be published. 1 -го Si 1 s.га £ - 1Z \ i -5î* s 1 1 * *г " -12 1

/5/ J. COUSTOLS, J.P. HAIRION, R.WARLCP. Description et utilisation du programme NATHALIE (Mitra 15).

CR DMG № 55/SO t I ! î I !1 ! Г i out-of pile section ;

in-pile section I

лГ Ht«- -i fr.-« filter |i h,

plate-out Ii section _*-_(> j p ... Й1 •я Iii irradiation section ИЗ \ fr Siloé core 9 Z; <-} displacem FIG:1 FIG:4 PARTICLES HOLDER 14 (131 I)

INCOLOY 800 1 T

i l rr th.

-t -„-4 1 I л iV 1 1 _ ^~ 1 1 =1 " „ 1 1 1

^ lube 14 j 1 1

-1 i ! i "E 1 i.i i i

i. t . t 1 , i , Л ( » i » i 1 » i '! i J • t • i 1 » J

1 1 1 SS AISI 347 .J u 10 1 ~I 3 i j i ï i L ! -J- ш л-"" i I , J*M I

M i к^н" " i j A J*^ 1 i =P M - i - i

_j, Iube24 1 J 1 "+ I ' J. 1 1 1 1 1 1 1 1. , , , i , . . ,h , . . , i . • .i. i . , . . t . J

SO 100 ISO 200 250 —». length (err

FIG.-10 FIG:9 COMEDIE PLATE-OUT SECTION EXAMPLE OF AN AXIAL PROFILE OBTAINED SRI EXPERIMENT BY MEASUREMENT WITH A 40 mm COLLIM ATATION

COMEDIE FILTER FIG:6 UCTIViTY DISTRIBUTION OK THE FILTER COMPONENTS

FIG:15 PLATE-OUT FIG:16 PLATE-OUT

Cs134 [es 13jJ

FIG:14 PLATE-OUT 10 'л INCOLOY 800 КГ' X л * "\ 4 INCOLûY 800 - / X a * a

INCOLOY 800 ~ О N 10"'

Il J — х U 4 ti Л 1iA II 0 и a • M I I I i i i 1 i i i i [ i t t M м м 1 t i Iff- I ' ' j ' I ' 1 ' 1 I 1 ' ' 1 I ' ' ' ' I ' 1

a - HASTELLOY X

— -3J

& 31 * 3J + 34 Д

31 Л 3t О 3t П i i i i j i i m I i i i i 1 m i i j m i i I i i IJ 1,1 1 1 t i i t i 1 i i i i 1 ) i t i ! , i

SS AISI 347 SS A1SI 347

77 О

I ' I I M I j I i i i 1 ч и 1 i i i i I I ' 10' • •• t I • t i i I t i i i I i i i i I t i t

50 100 150 200 250 50 100 150 200 250 S - 50 100 150 200 250 length (cm) gos length (cm) gas . length [en 19 2. TEST DEFINITION. OUT-OF-PILE HELIUM LOOP FOR LIFTOFF EXPERIMENTS In the case of a gas flowing in the tube axis direction the shear ratio R.J. BLANCHARD, A. BROS, J. GENTIL SR (ratio- of the wall shear stress under blowdown conditions to that under Centre d'Etudes Nucléaires de Grenoble normal conditions) is givetv by the equation : Grenoble Trance p 0.75 v 1 .75 j -0.58

SR= (^) (^) (/) N N N

ABSTRACT. where :

P = absolute pressure A helium loop has been developed to study the fission products re-en- V = bulk velocity trainment. Its principal characteristics are described, as also the experi­ T = absolute temperature mental method which has been used.

The subscripts В and N refer respectively to blowdown and normal condi­ The studies have shown that the principal parameter affecting the lift­ tions. off of fission products and dust particles is the shear ratio. In The case of a turbulent flow perpendicular to a row of tubes (PEACH

A preliminary study has been started concerning the dust particles, par­ BOTTOM) the shear ratio expression is more complicated an can be written for ticularly about their nature and their contribution to the fission products a given temperature : transportation.

SR = KV1'75

1. INTRODUCTION.

3. DESCRIPTION OF THE EXPERIMENTAL FACILITIES. The study of fission products release from high temperature reactor fuel elements, their entrainment in the primary circuitry and possibly their re­ The experiments are carried out in a loop (fig 1) which is composed of: lease into the atmosphere, are very important for the operation and safety of - a gas bearing blower

reactors. - a heater

- a test section Some accident possibilities can involve a rapid depressurization of the - a by pass with a laminar valve primary cooling circuit and one can suppose that a small fraction of the pla­ - a measuring equipment for : flow rate - pressure and temperature teout activity may be re-entrained by a modification of the wall shear forces. - a purification unit. The main loop characteristics are as follows : deposition on its internal or external surface- The quantity of fission pro­ ducts and dust particles re-entrained during the test is collected on the - Pressure 10 to 60 bars

1 1 filter located downstream of the specimen. - Flow rate 140 g.s" (at 40°C) to 25 g.s (at 600°C) - Temperature from room temperature to 600°C The study parameters are : temperature — velocity — specimen nature — - useful diameter 23 mm. test duration. The wall shear forces depend on the first parameters : tempera­ ture and velocity. The specimens issued from the PEACH BOTTOM reactor have the following characteristics : length 100 mm, variable inner diameter and outer diameter A blowdown experiment includes the following operation : with a maximum of 19 mm, materials : Incoloy 800 or carbon steel. The pla­ teout is located at the external surface of the tubes. The specimen is cente­ - The test section being isolated with the by-pass valves, placing or red by means of two centering cylinders enabling to define also the hydraulic the specimen in the locating device and cf the filters. diameter with a h of 40. H - Starting of the loop, rising of pressure and temperature. flo« adjust­ ment in the by-pass by the controlling valve, and purification of the The dust particles and fission products re-entrained are trapped by two gas. filters located downstream of the experimental zone. The first filter compo­ sition is : - Transition of the by-pass gas flow on the test section. The flow rate does not change (same opening of the by-pass and test section).

Elements Cut Size Thickness - Isolation and emptying of the test section Fiberglass 0.3 pm 0.1 mm - Dismantling and recuperation of filters and specimen. Sintered bronze 75 urn 3 mm Silvered molecular sieve 5 A 24 mm Sintered bronze 75 Urn 3 mm 5. TEST RESULTS.

The second filter has the same composition but the molecular sieve is Only three nuclides y emitters are detected on the specimens ; t*o fis— 134,. 137,. replaced by activated charcoal. sion products, Cs, Cs and an activation product Co. Besides, the if S emitter, is also detected.

These filters have been calculated for a gas velocity lower than Im.s 1. According to some preliminary tests, it has been established that after

4. EXPERIMENTAL METHOD. one minute of the test, the fission products quantity extracted »as not va­ rying at the some kinetic as during the test beginning. Consequently the A gas flow, dry or wet, circulates during a determined duration, at a tests lasted two minutes. given temperature, in or outside a specimen which has a fission products 21 5.1. Influence of the shear forces ratio. tron microscope. The following questions arise :

5.1.1. у emitters. - nature of different dust types - dust contribution to the fission products transportation For SR varying from 1 to 16 there is an increase in the liftoff nuclides - dust formation mechanism quantity, either for cesium or cobalt (fig.2 and 3).

5.1.2. ß emitters. Only the first question has been solved, using a fluorescence X spectro­ The results are the same (fig.4). meter connected to an electron microscope. The figures ô and 7 show an ехг-з- ple of the observed pictures; 5.2. Influence of the materials.

The superheater tubes are Incoloy 800, evaporator and economizer tubes Two types of dust can be observed : are carbon steel. Owing to the scattering of results it is difficult to show - metallic dust composed essentially of Cr and re, constitutive elements of the matrix but which appear with concentrations different froa an influence of the material. On the other hand one can notice a rather good those in the matrix (particles richer in Cr than in re whereas the nomogeneity for the Incoloy 800 results when they are widely scattered for matrix is richer in Fe than in Cr). carbon steel (fig.5). This difference could be explained on account of the Incoloy 800 oxydation resistance which is higher and consequently entails a - dust particles whose main elements are the silicium and aluminium, more homogeneous surface state. о both elements present, in the molecular sieve 4 A, in the filters and thermal insulation. Concerning the ß emitters it is still more difficult to draw an inferen­ Concerning the PEACH BOTTOM specimens a dust recuperation system has ce because most of the measurements have been performed on the same materials (Incoloy 800). been developed (fig.8) in order to answer to the question of dust contribu­ tion to the fission products transportation. This system located up-stream of

5.3 Influence of the temperature. the total filter was designed to separate according to five granulations, 200, 100, 40, 10, 5 pm, by means of copper or bronr.e sieve. 3ut it was not The temperature has no influence from 30 to 400°C, either for S or у possible to use the sieve 5 and 10 vim. because of pressure drop problems. emitters. Although in low quantity the particles of diameter 40 jm represent 3X of the transported Cs activity. As the Co is trapped completely in the total filter, 5.4. Influence of the water concentration. it can be admitted that it is transported by particles of diameter lower than From 10 to 300 vpm the water concentration has no influence on the lift­ 40 Pm. The nature of these particles is different from those of PEGASE, the off, even at 400°C, at least for the test duration (2 minutes). main components are Fe anr carbon and also few particles containing alkali metals and alkaline earth elements (K, Ca).

6. DUST STUDIES. 7. CONCLUSION. In the case of experiments carried out in the PEGASE reactor at CADARA- The studies have enabled to develop methods and technics for liftoff CHE the dust particles collected have been examinated with a scanning elec­ studies, dust counting - recuperation and characterization. The quantity of fission products re—entrained is proportional to the shear ratio SR, for example in the PEACH BOTTOM Incoloy tubes case, the re-entrained quantity in % is for the Co = 0.67 x SR and for the Sr = 0.98 x SR + 0.2.

The studies related to the nature of the dust particles and their contri­ bution in the fission products transportation have been approached and some partial answers have been given.

Yl^y2^».'^'ectr'ca' valves j.Vg. Flow control valve T. Temperature measurement P. Pressure м AR Flow rate n N.Circulator rotation speed

OUT-OF-PILE HELIUM LOOP FOR LIFTOFF EXPERIMENTS СЕ5ШП ЪЕ-ЕНТЯЯШПЕНТ FROH lliCOLOY

F/3:7 . SUO 15.1

• SUO 36.1 1?<2..enfro

ejuanhty j SU| 52 ,

о SUI 52.2

45J

30.

15.

3

0 1 5 10 SR 1 5 10 S R

COBALT RE-EtlTRAIttFIEriT FROH IfiCOLOY STROHTlim •RE-EffTRfl/NnENT FIG Л cesium nt-tmmmENT FIG 5 25

Image en fluo. X Image en fluo. X

du Silicium , des de I ' Aluminium;

poussières de poussières de la

la photo n 14. photo n 14.

Fig.7

Û.T?ing . "R-.гЯцяс

CRîng. Tcrbun&n

CRAYLOC." connection Tillênoy cJùccfcg» sp.5 -Sînf'greo? brenne

'„Ш77ят

(a) Jiei/c Ян Gi *VECÛ*

(b) jtet/e flOp Bravât "SflÜLAS J / rs • / / ' / (с).^Ге«/« 4 ЮО р * N • A (d) itevcZDOp brass "

FigzS гаг 26 in an in-pile position and an auxiliary circuit with carçressor, purifi­ EXPERIMENTAL FACILITIES FOR PLATE-OUT INVESTIGATIONS cation plant, He- supply and analysis equipment which wis connected in a AND FUTURE WORK bypass to the main circuit.

K. MUNCHOW, H. DEDERICHS, N. INIOTAKIS, B. SACKMANN Kernforschungsanlage Jülich СтЗН. Jülich Federal Republic of Germany

The safety of HTR under normal operation and accident conditions, the possibility of inspection, maintenance and repair or decontamination of single primary components as well as the safety of maintenance personnel are essentially determined by.the transport- and deposition behaviour of the non gaseous fission - and activation products in the primary loop of the reactor. A compreshensive program has been started in 1969 in KFA in collaboration with various industrial firms and foreign institutions to investigate these problems. The program includes in-pile and out-pile experiments, simulating reactor conditions and also different laboratory experiments and extensive theoretical investigations. The aim of these efforts is to test experimentally the models and computercodes, which are 1: Fig. 2: used for prediction of transport and deposition behaviour of fission pro­ ducts for HTR's as well under normal as under accident conditions. Further more a verified dataset is to be established.

The test section (fig. 2) consisted of the fuel unit with К HTR spheres In this paper a survey is given of the experimental facilities carried out mounted inline in an isolated graphite sleeve, the exchangeable test tube by KFA or in cooperation with KFA. of 3o mm inner diameter and 2 m length respectively 2o mn jS- and 1,2 n length, the connection tube of 1 n;, in the last runs of 3 ш length and 1. Terminated experiments 16 mm ф. as well as the absolute and control filter.

A He flow of 12 respectively lo g/s later on, transported the fission pro­ ducts released from the fuel through test section and filter, in which they

The first plate-out experiment was started in 197o in the loop Saphir in were retained quantitatively. In the auxiliary cicuit the gaseous impuri­ the reactor Pégase in Cadarache in close cooperation with CEA (fig. 1). ties of the He and the gaseous fission products were analysed. In a bypass This loop consisted of the main circuit with blower, heater and testsection the required He purification in the ppm range was realized by a filter plant of three different filters connected in series. At last the purified 2o mm (>j through an absolute filter, installed inside of the thermocouple

He was pumped back into the main loop by a compressor. tube, and a second outside control filter into the purification system and back to the primary loop (fig. 4). After operation times of 4 respectively

The gas inlet temperature into the testsection (table 1) was about 82o 12 months test tube and filter were dismounted and quantitatively analysed respectively 97o °C, the wall temperatures were nearly isotherm 53o°C in the first four runs. In the two last runs it decreased from 72o to 48o °C respectively from 81o to 61o °C. In total 6 runs were completed success­ fully. OPERATING CONDITIONS

After each run tostsection, connection tube and filter were removed, SAPHIR VAMPYR I TEST-ISOL. disassembled and analysed quantitatively for ß- and -activity in the PEGASE AVR DRAGON laboratories of CEA and KFA. So the axial deposition profiles as well as EXPER. TIME (HEEKS) 6.8,1оЛ2Л9 4. 12 7o radial diffusion profiles were measured on 6 different materials. In this FLOH ( experiment we could indicate the penetration of the fission products into G/S) 12/10 0.7 I600 TURBULEhT the material for the first time. TURBULENT LAMINAR PRESSURE (в) 2o 43 11 1*2 d9LSU5_§^EliG9_tybe_VamgYr_I/AyR GASINLET TEMP.(°C) 755-900 7oo 82o/97o At the end of 1972 the hot gas sampling tube Vampyr I was started in close HALL TEMP (°C) 87o-55o/ 7oo ISOTH cooperation with AVR. This experiment was installed in one of the three .53o ISOTH/ 550-150 72o-48o tubes for interchangeable thermocouples in the- third layer of the coal 8I0-6I0 bridge, immediately above the core of AVR (fig. 3). By a hole bored in this MATERIAL 4541. 15Mo3 Tl. 15Mo3 DIN 4981 «41. 4961 INCOLOY 800 tube in a distance of 225 mm from the reflector wall, a He stream of about HASTELLOY. AS 2-500 ST 35.8 AISI 347 15 Nm /h was led by way of a replaceable test tube of 2,2 m length and INCONEL 625 1OCRMO91O AISI 316 NIMOCAST 713LC INCONEL Б25 (RESTRICTED)

GEOMETRY (мм) 3o/2o 0, 2o 0. 193.7 0, 2ooo/12oo L 23oo L 600 U BOBBIN

RUNS (Ho) 6 17 (23)

PARTIAL PRESSURE CONST. CONST. CONST.

FISSION PROD. TOT.SPECTRUM TOT.SPECTRUM TOT.SPECTRUM

Fig. 3: Cross-section through AVR Fl9- ^- Vampyr I scheme in the laboratories of IRB. The mean He coolant temperature was varied The gas leading tube consisted of 5 so called Bobbins of different material; between 77o to 95o °C and with this also the inlet temperature of the test DIN 4981, Incoloy 8oo, AISI 347, AISI 316 and Inconel 625. The inner dia­ tube between 755 to 99o °C. The temperature at the end of the tube was meter was 193,7 mm, the wall temperature about 68o°C, the gas-flow 1,6 kg/s about loo to 16o°C (table 1). So far 22 runs were carried out, 17 of them and the experimental time 49o d (table 1). were successful.

The 8obbins were cut by EIR in parts of lo cm length which were -spektro- The aims of Vampyr I were the measurement of deposition - and as far as metrically analysed in IRB to obtain the deposition profiles along the possible of diffusion-profiles at different materials under laminar flow whole tube. Subsequently each tube was divided in 6 segments. At one part conditions and the determination of the concentrations of the non gaseous decontamination investigations were performed by EIR, at the other parts fission and activation products in the hot coolant of the AVR direct above KFA carried out diffusion profile measurements. Before starting the experi­ the core. mental investigations, Iniotakis had performed a precalculation of the expected deposition, diffusion and dekontamination factors.

Essential results of these three experiments are:

To investigate the behaviour of hotgas isolations a metallic foil isolation 1. At temperatures above about 68o to 73o°C the deposited nuclides was installed in 1974 in the duct F of the Dragon reactor (fig. 5). After penetrate predominantly into the material, whereas in the colder the unexpected shutdown of the reactor this equipment was removed for parts the adsorption dominates. deposition and decontamination investigations.

2. Kind and property of the material surface are of great influence on decontamination and deposition.

- IRB

IrHl 77 - K/01 3. According to the respective conditions the flow influences decisively the deposition.

-11

IMreO'COUpl» FOIl- 4. In AVR Ag-llom was detected for the first time, which has to be TESTIS01ATI0N considered at all safety aspects besides Cs-, J- and Sr-isotopes. IN DRAGON - REACTOR 5. The evaluation of the experiments under very different test condi­ J tions with special evaluation methods based on the model of Inio­ takis and with the code PATRAS-S and the good agreement between experimental and theoretical results confirm the validity of the

Fig. 5: model of Iniotakis. A first set of deposition data was set up. Because of the limited variation range of the test conditions and the In the auxiliary bypass He is purified by three different kinds of filters. difficult measuring technique and experimental evaluation conditions in The gaseous impurities in the He are measured continously. The operation the experiments terminated so far, the following out-of-pile and in-pile conditions (table 2) are variable in a wide range: gastemperatures from experiments were planed, which at present are in the state of cold testing 2oo to 9oo °C, wall temperatures from 15o to 9oo °C, flow till 24o g/s at or constructing, in order to confirm and complete the set of plate-out 9oo °C, 8o bar and 83o g/s at loo°C, 8o bar, pressure from 5-8o bar. data.

In this experiment the time slope of the axial deposition profiles of 2. Experiments under construction and testing particular nuclides is measured under turbulent flow conditions on all interested materials with special regard to the surface conditions. By 2-1 QytrofrBilË.çir.çyiLS'ESE disconnecting the source the desorptionphenomena are investigated. The main aim of these investigations are the determination of all plate-out parameters in a wide and very exactly measurable range of test conditions In the control section an out-of-pile circuit, Smoc, is mounted which con­ and the verification of the validity of the heat-mass analogy. sists of a primary loop (fig. 6), a purification system, 5 coolant circuits and the He supply, as well as extensive safety and regulation equipments. 2.2 Laboratory_circuit_Scafex Driven by a gas bearing blower, He flows through an 1 MW heater, a test section of 6 m length and 3o mm t>., a cooler and 2 filters working at loo°C and utilizing the adsorption on metallic foils. In front of the Because of the great difficulties in the realisation of sources for out of test section in the first runs Cs-134 is injected by an evaporating source, -12 pile experiments in the range

Fig. 6: Fig. 7: Hon-Crcurt Smoc SCAFEX-CIRCUIT 30 2.3 Degosition_looD_Yampy£_II_in_AVR

LAY OUT CONDITIONS Due to the high activity level by neutron activation in the forward part of the test section of Vampyr I the choice of material was severely liaited in VAMPYR II SMC SCAFEX/FILTER this interest temperature range above 55o°C. More over it was not possible AVR IRB IRB to measure the temperature during the runs. The temperature profiles bad to EXPER.TIME (WEEKS) 16.24 -2 -1 be determined in premilinary runs. So in 1973 KFA and AVR started the design FLOW (G/S) -7 -83o(loo°C) -3.6 cf a second experiment Vampyr II in AVR only for plate out examinations. At TURBULENT -24o(9oo°C) LAM.+TURB. present the manufacturing of the main components is terminated (fig. S). In 1UBULENT this experiment hot cooling gas is led through an insulated tube installed 5-8o 2-6 PRESSURE (в) 11 in the coal stone bridge 12o° shifted against Vaapyr I into the test equipment in the protection room of AVR. Here activation of test naterial GASINLET 4oo-85o 2oo-9oo loo TEMP (°C) is excluded so that all interesting materials can be installed. Gas- and wall temperatures can be measured accurately during the runs. By means of WALL ТБЧР (°C) 25o-85o 15o-9oo a hot gas distributor in front of the hot equipment the gas inlet tempera­ MATERIAL UNRESTRICTED UNRESTRICTED UNRESTRICTED tures can be changed between about 85o and 4oo °C. Behind the hot section

GEOMETRY (MM) 3o 0. 2o-8o 0, 5. lo 0, the gas flows through a cooler, two absolute filters and a conpresscr, which 54oo L ISOTH:2OOOL -3oooL allows to adjust turbulent flow till to 15o MM3/h or 7 g/s. The test REKUP:4OOOL equipment (fig. 9) has a length of 5,4 m, the inner dianeter of the test

PARTIAL PRESSURE CONST VARIABLE VARIABLE tube is 3o mm. Through an annular clearance, cooled He can be led counter currently along the outside of the test tube effecting axial temperature FISSION PROD. TOT. SPECTR. SINGLE SINGLE gradient conditions. Between this tube and the outer pressure pipe a heater

The He purified in a bypass (fig. 7) is pumped by a compressor about the source, through the control tube with a length of 15oo mm and an inner diameter of lo mm and the testfilter section, in which 1-8 tubes of diffe­ rent lengths and diameters can be investigated. In an absolute filter be­ hind this section the radioactive nuclides are adsorbed quantitatively. The 3 main test conditions are shown in table 2: flow 2-12 m/h, gastemperature loo °C, pressure 2-6 bar. The time slope of the deposition of the investi­ Fig. 8: FIG. S; gated nuclide is measured at 3 points of the control tube as well as before « sect** avir I and behind the filter section. and a fiber insolation is mounted. This allows to provide experiments also under isothermal conditions. The wall temperature can be varied between LAY OUT CONDITIONS OF HOT GAS FILTER EXPERIMENT 25o and 85o °C. IN ШИJMTR-JAERI

Because of the AVR accident this experiment can start not before the end of 1981. So the number of possible runs is stroughly restricted against the EXPERIMENTAL TIME 5 MONTHS origional planning. It is intended to perform runs of б months to measure deposition - and diffusion profiles in some of the most interesting HTR- FLOW 5o G/S materials. Because it is not possible to determine the concentration of PRESSURE Зо в the fission products in the AVR coolant with Vampyr II, furthermore Vampyr I GASINLET TEMP. 73o °C will run parallel to Vampyr II restricted to this task.

MATERIAL/GEOMETRY (мм)

HGF1 INCONEL Боо

81 0A. lo 0,. 25o L An important result of the knowledge obtained so far from the above descri­ SHEET THICKNESS 2 bed experiments was the development of a hot gas filter for the primary circuit of HTR /1/. In this filter the diffusion of fission - and activation PLATE-OUT TUBES products into the material above about ca. 45o °C is utilized, to exclude JAERI HASTELLOY x; HASTELLOY XR. the non noble gaseous nuclides out of the He coolant so reducing the contami­ STAINLESS STEEL TYPE 315 nation of the primary components remarkably. To investigate the efficiency HASTELLOY X. INCOLOY 8oo. KFA of such kind of filter a hot gas filter experiment was layed out by KFA and INCOLOY 6oo lo 0,. loo L fabricated and mounted in the mean time by JAERI in the loop 0GL1 of JMTR

(fig. lo). This consists of the hot gas filter HGF1, outer i> 81 mm, kernel INCONEL 6oo HGF2 t 4o mm, length 125 mm, a plate-out section with б tubes circularly arranged 81 0A. lo 0,. 5oo L around the axial kernel with a length of 4oo ran and a second hot gas filter SHEET-THICKNESS 2

similar to HGF1 but 5oo mm long, serving as an absolute filter. The operation conditions (table 3) are: flow 5o g/s, pressure 3o bar, gas inlet temperature at HGF1 73o °C, operation time 5 months. The filter foil material is com­ posed of Inconel 6oo with a wall thickness of 2 ran, the plate out tubes of Hastelloy X, Hastelloy XR, Inconel 6oo, Incoloy 8oo and stainless steel type 316 with an inner 0 of lo mm. The radiation will start ad the end of 1-4: measuring points hot gas filter experiment 198o and will be stopped in the beginning of 1982. In the post irradiation in OGL1 / JMIR R9- l0: examinations the activity of the nuclides deposited on the measuring points 1 to 4 is analysed quantitatively to determine the efficiency of the fil­ bundles and 3 connection tubes. Each test bundle is composed by 2 tubes with ters. Besides the activity distribution along the filter is measured by 14 mm (>. length of about 26oo mm and a concentrically adjusted rod of lo mm £ -scan and the deposition profiles of the plate-out tubes and if possible diameter with a length of about 27oo mm. Each bundle is connected with the also diffussion profiles are determined ^ -spectrometricly. filter plant by a tube of 18 mm and about 6oo mm length.

3. Planned experiments Test bundles and connection tubes are cooled by He which flows counter- currently along the outside of the test tubes. The filter plant consists of

3.1 DeDosition_looD_Comedie_in^ a hot gas filter, a cooled prefilter designed similar to the hot gas filter and a filter with loose material.

Because of the strougly restricted test program of Vampyr II and in order In the out-of-pile section the He passes a so called comparison loop with to guarante the whole plate-out program and to clear up open questions not the same test conditions as fixed in the test section and with samples of investigated till now by experiments, CEA, EIR, HRB and KFA agreed to per­ the same material inserted in the test section. In this zone are also form additional and complementary plate-out-, decontamination-, material- placed the purification plant, the He supply and the control - safety - and also fuel specific investigations in the Comedie-loop in Siloe/Grenoble and analyse equipments. He

The operation conditions (table 4) are: pressure 2o bar, flow 16 g/s, gas temperatures along the test section 835 - 5oo °C, wall temperatures 78o - 45o °C. Test tube materials are composed of Incoloy 8oo H, Hastelloy X, test rods of IN 617, IN 519, Incoloy 8oo H, Nimonic 86, TZM and Nimocast 713 LC, connection tubes of 15МоЗ, 1оСОНо91о and 4541.

During the runs the gaseous impurities before and behind the test section are measured continously just.as the gaseous fission products to estimate the R/B values. During each normal shut down y-spectrometric measurements are performed at the entrance of test bundles, connection tubes and hot gas filter to control the time slope of the source strength. In the post irradia­ tion examinations deposition and diffusion profiles are measured and exten­ sive material investigations are performed. scheme Comédie p-jg, ц. Loop/ Siloe The following aims are predominant: Performing 3 runs with 12, 2 and 4 months under always equal operation In the fuel section (fig. 11) 8 HTR-spheres with 58 mm й , consisting of conditions, which simulates the HTR conditions as realistically as a possible, the time dependence shall be determined so that the extra­ 6 driver and 2 source elements, are arranged in line in a graphite sleeve polation of the data evaluated till now to realistic reactor operation of 64 mm fi.. A He-flow transports the released fission products through the times can be confirmed with more assurance. test section into the filter plant. The test section consist of 3 test LAY OUT CONDITIONS OF COMEDIE-LOOP The installation of a comparison loop in the out-of-pile section

operating under the same conditions as the testsection enables the

performance of comparative material investigations about the influence EXPERIMENTAL 12.2. 4 MONTHS of the He-atmosphere and the fission products to the material proper­ FLOH 16 G/S ties.

PRESSURE 2o в The first run shall start in beginning of 1982. PARTIAL-PRESSURE: Cs-137 1.5 la'}} в AG-11OM 4 lo­ ll 3.2 QustexgerimentinAVR TEST SECTION: In the primary circuit of HTR dust quantities can result of some tons GASTEMPERATURE 835-500 °C after 3o years operation time by abrasion of graphitic fuel elements and

HALL TEMPERATURE 78o-45o °C by corrosion of graphitic core components. Therefore it is necessary to obtain more evidence about the behaviour of dust in the primary circuit, MATERIAL/GEOMETRY (мм) INCOLOY 8oo H it's interaction with the fission - and activation products, the portion of

TEST TUBES HASTELLOY X free and dust bound activity, its quantity and grain size distribution.

16 0Д. 14 0[. 26oo L

TEST RODS INCOLOY 8oo H. NIMONIC 86 To get more knowledge of these problems in the cold part of AVR a dust IN 617. IN 519, TZM experiment shall be installed at the bifurcation to the purification system NIMOCAST 713 LC lo 0, 27oo L (fig. 12). The design has been started in the middle of 198o.

CONNECTION TUBES 15Mo3, 1OCRMO91O. 4541

21 0A, 18 0,. 6oo L

Of dusllillef GAS IMPURITIES CO = 2.5 VPM reactor vessel , ctntpnmenl absolute filter I GC 505 cooler С\\ц - 0,25 VPM ~-' TBF tube bundle filter N2 = 0.25 VPM

H20 * 0.5 VPM

H2/H20 = lo + 5

The test time of 12 mounths and the requested Cs and Ag partial

pressure shall enable the examination of long time effects as

evaporation respectively solubility. Fig. 12: The arrangement of the test bundles enables to investigate the dust experiment in AVR

influence of the deposited and penetrated nuclides as well as the

influence of the decontamination on the properties of the different

HTR specific materials. More over deposition and decontamination are Directly at the bifurcation valve cold He is led through a sampling tube of

investigated on replaced and already decontaminated samples. 2,5 m length, 2 shutoffvalves and a gas cooler. Behind the cooler the He is divided into a parallel flow. In one tube there is mounted a dust and a The knowledge about the influence of the solubility respectiveely cooled absolute filter, in which all activity and dust is trapped. In the evaporation, rising with increasing operation time, on the deposi­ other one an addional tube bundle filter is connected in series, which tion respectively contamination and also on the possibility of filters the free activity respectively the activity not dust bound out of decontamination and maintenance of primary components in the higher the He cool gas. temperature range of the HTR could not be investigated in experiments till now. At present we are checking the possibility of realizing such

Behind the parallel arms the He flows through a compressor laied out for 7o investigations at different institutes.

Nm3/h. After different operation times between 8 days and 8 weeks the Specific experiments for transport and deposition of dust will be filters are removed and analysed relative to dust and activity. One obtains aspired in the furture. the total dust quantitiy, the grain size distribution and the free and dust

bounded activity. By means of the cooler the gas temperature can be varied Iniotakis, N. between 27o and 5o °C enabling the investigation of the interaction between Filterkonzept Tür Spaltprodukte im Hauptkühlkreislauf von HTR's. dust and fission products, of reversible and irreversible dust bound acti­ Jül-1353 (1976) vity and of aerosol fraction in AVR-primery cooling gas. This experiment is performed in close cooperation with AVR. The design is performed by KFA-ZAT. The runs shall start end of 1982.

4. Future planning

The total activity deposition program provides for following further expe­ riments:

In THTR-Зоо sampling tubes shall be installed in the hot and cold part of the primary circuit, by means of which the concentration of non gaseous fission - and activation products in both parts and if possible also the dust quantity shall be measured. The CEA hot gas isolation mounted in duct D of the Dragon reactor as well as the main cooler belonging to the same duct by EIR and KFA with regard to decontamination and plate-out. The operation time of the cooler amounted to 12oo days that of the isolation to 49o days.

To the present there are no experimental investigations about the fission product behaviour under accident conditions. Therefore lateron definite accident conditions shall be simulated if possible in Smoc and Scafex, or in new test circuits. of FP and dust with the surfaces of the components and the interaction of

FP with the dust particles, a physical and mathematical Eodel has been RESULTS -FROM PLATE-OUT INVESTIGATIONS developed by Iniotakis in the Institute for Reactor Components of the KFA Jülich /5,8,9,1о/. N. INIOTAKIS, J. MALINOWSKI, H. GOTTAUT, K. MÜNCHOW Kernforschungsanlage Jülich GmbH. Jülich This model is shown schematically in Fig. 1.

Federal Repolie of Germany

Introduction

The fission- and activation products in the primary circuits of HTR's belong to the essential factors for the security of the plant during the normal work and in the case of accident. The deposition of these products on the components of primary circuits, specially their diffusion in the wall material, can cause a more or less intense contamination of the components and play a decisive role in the solving of such problems as inspection, servicing and repair.

The knowledge of fission product [FP] behaviour in the primary circuit of a

HTR is an essential requirement for the estimations of the availability of Schemctc riustration of transport and deposition mechcnisms tor fission products and dust pcrticles the reactor plant in normal operation, of the hazards to personnel during

inspection and repair and of the potential danger to the environment from Fig. 1: severe accidents. The transport and plate-out mechanisms identified are: In the Insitute for Reactor Components of the KFA Jülich, much theoretical 1. Macroscopic transport of FP and dust particles in the bulk of the and experimental work was done /1 bis lo/, with the aim to clarify this flowing stream. complex problem. In this paper we would like to report some essential results 2. Microscopic transport of FP and dust particles perpendicular to the of these investigations with special attention to the possibility of extra­ wall. The adhesive forces which have a macroscopic range are taken into polation of these results for long periods of time. account.

3. Reflection of FP and dust particles on the wall. Theoretical principles and model

4. Adsorption of FP and dust particles on the wall/reversible process/. Starting from a systematic analysis of the elementary processes which can 5. Diffusion of FP into the wall material /irrevesible process/. influence the transport of FP and dust particles in helium, the interaction 6. Desorption of FP and dust particles from the wall; the influence of in the conventional way, Iniotakis developed further the method of Papoulis condensation effects on the desorption energy is considered. and successfully applied it for the discussed problem.

7. Evaporation of FP from the wall as a result of local condensation in On this basis the calculation code "PATRAS" (without considering the presence the bulk material immediately at the surface, as a result of evaporation of dust) was developed; now the calculation code "PATRAS-S" (with considera­ of the wall material itself or as a result of erosion. tion of dust) ist being developed. 8. Collision of dust particles with FP. Adsorption, diffusion, desorption or evaporation of FP on or from the dust can occur. The calculation code PATRAS enables the calculation of the transport and

9. Strip off of the dust particles by shear stresses. Furthermore, the deposition of FP with, or without mother activities, for open or closed reac­

growth rate of oxide layers and the influence of this on the diffusion tor circuits, with arbitrary sources, for normal run and for some cases of

of FP are taken into account. accident.

As no significant amounts of dust were found in our loop experiments until Many deposition experiments (Vampyr, Saphir, Dragon) were evaluated using now, its influence will not be considered in our particular case. special procedures based on the model and the code PATRAS mentioned above. The theory was confirmed through the evaluation and pre-calculation of many

The transport of FP to the wall is characterised by the mass-transfer number KFA experiments. "h" which includes such factors as flow conditions, the temperature of the gas and the binary diffusion constant in the gaseous phase. In order to clarify the open questions and for completing the existing experimental data, more investigations were planned and carried out, basing

The interaction of the FP with the wall is described by several parameters: on the existing theory. The experimental procedures were developed to fit the sticking probability "oi.", the penetration coefficient "1-ß", the desorp­ the requirements connected with the solution of this task. tion constant "iJ" " and the diffusion constant "D" of considered FP on, or in the wall material. Experimental procedures

Considering diffusion; it must be differentiated between the diffusion in The tubes from Saphir and Vampyr experiments were cut into lo cm long pieces. the oxide layer (if present) and the diffusion in the bulk material. Moreover, Their diameter ranged from 2o to 35 mm. The hot gas duct from the Dragon the surface properties of the material like the roughness, represented by the reactor was cut into rings, lo cm high. These rings were sub-divided in roughness factor "X ", or general properties, like the solubility of the FP such way, that slightly curved samples, lo x lo cm were obtained. in the material represented by the solubility "0^, ", must be taken into The samples were analyzed by -spectrometry for active FP, using self-made account. standards of the same geometry and material.

The whole problem is extremely complex and can be described only with more After the 5f-spectrometric analysis was done, the samples were surface than 2o non-linear partial differential equations. For the case of low dust leached; for some samples 2M nitric acid was used for this purpose, other concentrations it was possible to establish the Laplace-transform of the were surface leached using water at different temperatures, with, or without analytical solution. For the inverse transformation, which is not possible ultrasonic treatment. Next to the surface leaching, the oxide layer was removed, in most cases in Pégase in Cadarache. The presented results originate from the б and 11 two steps: in the first step a very weak attack was ^one, in the second step runs of the series, in which ferritic steel 15 Mo 3, stainless steels the total removal of the oxide layer was accomplished. 4541 and 4961, Inconel 625 and Nimocast 713 LC were examinated. The

0 deposition took place in the temperature range from 8oo to 9o °C(from

Finally, after the removal of the oxide layer, layers of the bulk material were turbulent gas flow. taken off electrolytically. The thickness of the removed layers was determined 3. The third experiment was performed in the facility Vampyr of the AVR by chemical analysis of resulting solutions or by weighing the samples before reactor in Jülich. The results reported here originate from the 12, 19 and after treatment. and 2o runs of the series, in which Titanium and ferritic steels 15 Mo 3 and St 35.8 were examinated. The deposition took place in the temperature Before applying the electrolytical removing (in the anodical polishing mode) 0 0 range from 9oo to loo C(from turbulent gas flow. to the radioactive samples, a comprehensive study of the process was under­ taken for each alloy and each shape of sample on non-radioactive objects. It Some essential working data of the above mentioned experiments are given in was found that, under defined conditions, layers sufficiently parallel to the the Table 1. original surface were removed. The resolving power of the equipment used for measuring the thickness was 1 urn for 1 scale division. The total depth removed in many steps amounted to loo ^m or more. DEPOSITION GAS FLOW GAS Cs-137 TIME TEMPERATURE EXPERIMENT All resulting solutions were again analyzed by -spectrometry; after each PARTIAL PRESSURE ACTIVITY DEPOSITED DAYS G/S °C BAR fi Ci/см- chemical or electrochemical treatment, the rest-activities of the samples were determined and the corresponding balances verified. DRAGON HOT DUCT j 49o 16oo 7oo 7.7 lo"15 0.3 - o.l PÉGASE об 7o 12.1 82o-39o o.oo6-o.oo2 PÉGASE 11 155 lo 32o-39o 2.2 1о"ь 0.7-0.2 Results VAMPYR 12 34.2 0.7 870-75 2.1 lo-" o.ol5-o.oo3 VAMPYR 19 11.3 0.7 370-85 6.7 10--L5 0.6-0.05

13 In this paper results from following experiments will be considered: VAMPYR 2O 27.3 0.7 o7o-8o 4.3 lo" 0.3-0.o3

1. The facility for investigation of metal foil isolations was built in Table 1 the hot gas duct of the Dragon reactor. After the shut-down of this reactor the possibility was given to investigate the plate-out behaviour of FP on the duct materials: Inconel 625, AISI 316, AISI 347, Incoloy 8oo IRB-80-11/K OPERATION DATA OF THE EXPERIMENTS and DIN 4981. The effective deposition time was approximately 5oo days, the temperature about 7oo °C. The deposition took place at approximately

isothermal conditions and turbulent gas flow. This experiment will be Table 1: Operation data of the experiments. dealt with here under the name Dragon hot duct.

2. The second experiment considered here, from which some selected results The first column gives the deposition time, the second the gas flow, the will be presented, was performed in the loop Saphir of the reactor third the range of gas temperatures along the test tube, the fourth the partial pressure of Cs-137 at the beginning of the test tube, and the last The band A illustrates the expected deposition for a source of 3,3 x lo 3 4 3 the range of deposited specific activities of Cs-137. The differences in Atoms/cm , the band В for a source of 5,7 x lo Atoms/cm . The shaded surfaces the parameters of various experiments are remarkable. demonstrate the uncertainty of the predicted values resulting from the uncertainty of material data. The strength of the real source is not known,

In the case of Dragon hot duct, the theoretical precalculation of the expected but a greater probability was expected for the weaker one. The results deposition and diffusion of F.P. on and in the materials present was of obtained much later, confirmed this assumption. special interest because of the long running time at about 7oo °C and because the plate-out behaviour of these materials was not yet examined. The opportu­ Beside the deposition profiles, the predetermination of the diffusion profiles nity was also given to compare the theoretical predictions with the results and decontamination factors for Cs in and for different construction materials of later investigations; this in turn gives the impression about the degree was done. of uncertainty for the pre-calculation of reactor circuits done on the basis

of uncomplete material data sets. In Fig. 3 the comparison is done between a theoretically predetermined and the experimentally, in the Institute for Reactor Components, found diffusion

The plate-out specific data of the materials used were estimated on the •profile of Cs-137 in Incoloy 800 from Dragon hot duct. basis of their composition, lattice structure and the knowledge of deposi­

tion mechanisms. The behaviour of these materials in the primary gas atmos­ (m)-.RB

phere of the Dragon reactor was also considered. From our own previous work 77 -14/02 and from foreign publications, we posessed the informations about diffusion constant in similar materials and about the decrease of concentration of Cs-137 in Incoloy 800 at 800 °C.

Diffusions profiles of Cs-f37 in In Fig. 2 the expected course of deposition of Cs-137 for two different tncoloy BOO at T . 680 *C sources is given. I . 1.37 a

t •^Vttttttt?, '»^

Rsnge A source 1.3 Ю Atoms/cm

Rarqe В source5.7 Id'Atoms/cm1 Fig. 3: t eipenmenfal values

Icon*! ЕЯ , A3 316 , 15117 60 120 ZiO 300 The experiment confirmed well the theoretical calculations. Positon n [cml

Pre- calculated deposition profile of Cs-137 on the IRB78-07/11 Both experiments: Dragon hot duct and Pégase-11 included Inconel 625; hot gas duct of the Dragon experiment Fig. 2: moreover the steels AISI 347 and AISI 316 from Dragon hot duct are similar 39 to the steel 4541 from Pégase-11. It was self-suggesting to check the theo­ The picture illustrates the influence of the material on the deposition of

retical predictions relating the fidelity of extrapolation over long periods Cs in the high temperature zone. This influence is due principally to the

of time and broad ranges of partial pressure through comparison of both different solubilities of Cs in the materials. In addition, for Inconel 625

experiments. For this reason, the same data sets were used for Inconel 625 and stainless steel 4541 the influence of temperature is noticeable. It and steel 4541 in the evaluation of both Pégase-11 and Dragon hot duct expe­ results from the temperature dependence of the solubility. The adsorptively

riments. bound component of Cs activity is also temperature dependent. In the low temperature zone the flew conditions are dominant fcr the course of depo­

The set of data for Nimocast 713 LC was obtained by application of special sition; for this reason the positions of the bends on the tube can be seen evaluation procedures and experimental values obtained from two of total on the deposition curves.

thirteen segments present. As the result of comparison, an excellent agreement between theory and expe­

riment has been found. In Fig. 4 a comparison is given between the deposition profiles of Cs-137 and

Cs-134 along the deposition tube of the experiment Pégase-11. The continuous In our investigations special attention was paid to the behaviour of FP in line represents the course of deposition as calculated on the basis of theory, the case of diffusion into the wall material. The clearing of this question the points given the experimentally found values. The dotted curve represents is an essential condition for evaluation of the decontamination possibility the wall temperature along the tube. At the bottom of the figure, a schematic of HTR construction materials. view of the deposition tube is given. In positions marked ".K" the tube was bend four times under 9o 0 or 18o 0 angles. uCi/cm3

Cs-137 Theory.*Exp.; • Exp lube 22

Iw=69t°C

id2! t = 133 d

.) 0n=a4 Id" exp -W/sa

•)ст2Лес

0^*2.110"exp I--^P- )cm2/sH

3 £lt)477-10 1/Ti;expt-^-)[cml

t -1П Irfl

1 3579 11 13 61719212325 29 31 35 37 394 13 «1719 5153 5557 Exper POS* —

Theor Kiwi «190°) nnaa°j Saphirîl: Deposition of Cs-137 and Cs-134 along the tube IRB 80-04/01 Fig. 5: ol Experiment Pégase 11

[Saphir 11 ditfuson profile of {йЗаО-К/02 j Fig. 4: um ' Cs-137 in Kmaccst 7T3LC 1 t 0 20 50 100 60 In Fig. 5 the diffusion profile of Cs-137 in Nimocast 713 LC is given. On the Finally, in Fig. 6 the comparison is given between the precalculated and the abscissa, the penetration depth in um, and on the ordinate the Cs-137 activity experimentally determined diffusion profiles of Cs-137 in Inconel 625 froa 3 ' in ^jCi/cm are given. The stepped curve represents the experimental results. the experiment Pégase-11. The agreement between experiment and theory is very The curves 1/ and 2/ are precalculated on the theoretical basis. From the good; further, it can be seen from the shape of the curves, that the oxide evaluation of this one and other similar experiments a data set resulted for layer will influence the decontamination of UTR materials, at least regarding the diffusion constants in the oxide layer and in the bulk material, as well the decontamination from Cs. The thickness and the rate of build-up of this as for the growth of the oxide layer. layer depend from the composition of the primary gas atmosphere. It follows, that the composition of the cooling gas in the primary circuit and his oxida­ The influence of the oxide layer, and consequently, of the oxidation rate, is tive properties are important for the future decontamination from Cs. of primary importance in the case of Cs diffusion. A compact oxide layer creates a considerable diffusion barrier for Cs. The thicker and more compact During investigation of some Vampyr experiments, Cs-137 deposited on ferritic the oxide layer, the less Cs can penetrate into the bulk material. The steel and on titanium was tested for its behaviour against the action of water. destruction of the oxide layer e.g. through erosion, reduces its protection effect against Cs diffusion. In Fig. 7 the dependence of the decontamination factor f for Cs-137 deposited on ferritic steel at 555 °C from the time of water treatment at room temperature, is given. The last point, on the end of the dotted curve, gives the value of the decontamination factor after complete removal 10° Tub! 9 (Fi 10° of the oxide layer by chemical treatment. The shape of the curve allows the ф|р1 7во«: assumption, that the water treatment takes off only this part of Cs-137 "ОТ Dj.1.5 -10"'° «« 1-15500/«!). which is deposited on the surface by adsorption. If this assumption is correct, the amount of Cs-137 Teachable by water must depend from the kind of material and from the deposition temperature.

AH,O f = —r-MOO о A Exp. V-19-E 018 H О at roomtemperature

80 (555°C1 2 Cs-137 60 . after removal of the К oxide layer ДО

20 0 20 «1 0 20 10 60 • drplh luml 30 60 90 120 150 t [mini

Saphir ll.diffusion profiles ИВ7В-07Л2 Decontamination factor:dependence of Cs-137in Inconel 625 Fig. 6: from leaching time IRB80 12/02 Fig. 7: The Fig. 8 gives the proof for this assumption. code PATRAS, using the data from the Vampyr experiment (1-5 = o,o4 «, Q = 5o

kcal/Mol). Below 5ao °C about 80 % of the Ag-llo m acitivity was leached out; The diagram gives the dependence of the surface decontamination factor for at higher temperatures the decontamination factor goes strongly down. The Cs-137 on ferritic steel and titanium from the deposition temperature. The agreement between theory and experiment is good. decontamination was done in this case by a long term treatment of each sample with water at 80 °C A distinct jump on the curve on the joining point between steel and titanium can be observed. It proofs the differences in the adsorp­ tion capacity of steel and titanium for Cs-137.

AH,O

A steel 35.8 Exp. V-20 HO ot 80° с ~*—- 2 80 * 7 hours

— Theory 60 • о surface Lescft UR81 Cs-137 iO WC MC ^\Ti The surface decontamination of V-12 ÜRB75 1VÜ7 20 îi- tube from Ag-llOm deposition Temp. »•» Fig. 9:

0 00 200 300 400 500 600 700 8C0 SCO T°C The Fig. lo gives the course of deposition of Ag-llom along the titanium tube Oecontamination factor»dependence from from the experiment Vampyr-12. material and deposition Temperature IRB8012/01 Fig. 8: 10' MlpCi/orr] TI°C1 900 1-ß =0.04 %o IdTi] Q =50 kcol/Mol •800

If the deposited activity is not water soluble, the choice of a corresponding •700 10s- reagent guides to equal results. This case is considered in the Fig. 9, where — Temperature •600 the surface decontamination factor for Ag-llom on titanium is given, depending Theory --.o Experiment 500 from the depos tion temperature. •400

The diagram represents the results from the investigations on the Vampyr-12 •300 experiment. The points are experimental results obtained by surface leaching, •200 1-P =0,A0/« IST|1 done by washing the samples with 2M nitric acid at room temperature in an •100 Ian] ultrasonic bath, during 1 hour. The leached activities were analyzed by 10"' _ffi. 2 ' 3' 4 ' 5 ' Б 7 1Г9 1Û 11 12'lTÏ4 15 16 17 TS 13 20 21 22 23 24 >Ç -spectrometry. The contiuous line represents the course of the decontami­

Deposition of Ag-110m along the Titanium tube Va."npyr; V12 nation factor as calculated in advance. The calculation was done with the Fig, lo At temperatures higher than 85o °C ß-titanium is present, with the lattice The values in the table indicate the layer thickness to be removed during structure different from the one of cl -titanium, stable below this temperature. decontamination. They are normalized for deposition of Cs on different con­ This change in the material property is reflected in the course of the depo­ struction materials at different temperatures as to yield a decontamination sition profile of Ag-llom in the hot zone of the tube where the diffusion factor of 2oo. As basis for calculation, a non-oxidizing atmosphere in the process dominates. The penetration coefficient is in ß-titanium considerably primary circuit and a working time of 7 years were taken. For Nimocast 713 LC higher than in d -titanium, which causes the jump on the curve. With decrea­ the lower values, for Incoloy 800H and TZM the higher values seem to be more sing temperature the surface adsorption starts to dominate and the temperature probable. influence is more remarkable. Below 55o °C the flow conditions start to dominate, because the desorption can be neglected here, as compared with The similar calculation for Ag-llom delivers much higher values. The whole the adsorption. The Ag is bound in this low-temperature zone on the surface, problem of decontamination depends from many factors. If the gas atmosphere by adsorption; at higher temperatures, starting from 55o °C, the diffusion in the primary circuit was oxidizing, and dense oxide layers are present, of Ag in the bulk material begins to be remarkable. relatively thin layers of material must be removed, at least concerning the decontamination from Cs. In the opposite case, considerably thicker layers

After the theoretical considerations have been confirmed by the experiment, must be removed. the efforts were made to apply these coonsiderations to predict the behaviour of different construction materials during deposition of radioactive products It is to note, that the oxide layer has a much less preventing power against and the expected consequences for the future decontamination. An example of diffusion of such nuclides as Ag-llom, Eu-154 or Sb-125 which were investi­ such consideration is given in Table 2. gated in this Insitute too.

Summary

As conclusion, it can be stated, that on the basis of Iniotakis model it was possible to calculate successfully different experiments in a wide range of

T„ Pel INCONEL 625 ТТЛ NIMOCAST-713 LC IHCOLOY-8OOH source strength, temperature and working time.

95o 22o 23o - 33o 49o - 8oo 38o - 67o The possibility of reasonable extrapolation was proven, and many values 85o 16° 17o - 25o 36o - 59o 23o - 47o could be precalculated; these values were confirmed by the experiment. 75o llo 12o - 175 25o - 41o 21o - 34o 65o 7o 75 - llo 16o - 27o 13o - 23o These results justify the anticipation, that this model can be successfully 55o 5 15 - 25 6o - So 4o - 65 applied for postulations concerning the behaviour of the FP in primary circuits of HTR's. Table 2:

The necessary data sets must be enlarged and actualized by further experimen­ tal work. THE MATERIAL DEPTH С UJH] NECESSARY TO REMOVE,FOR REACHING IRB80-04/05 A DECONTAMINATION FACTOR BF " 2oo (Cs-137. REACTOR WORKING TIME 7 YEARS, NON OXIDIZING ATM05PHERE) References /6/ C.B. von der Decken, N. Iniotakis, К. Münchow Behaviour of Fission Products in Core of a Depressurisation Accident /1/ v.d. Decken, Gottaut, Malinowski, Münchow, Eßler CSNI Specialist meeting on HTGR Safety; Das Bestrahlungsexperiment Saphir in Reaktor Pégase in Cadarache Reactor Centrum Nederland, Petten, Mai 1975 KTG-Reaktortagung, Karlsruhe, 1973 /7/ Gottaut, Iniotakis, Malinowski, Münchow, Sackmann /2/ Engelhard, Gilli, Mehrens, v.d. Decken, Gottaut, Malinowski, Münchow Das Programm Spaltproduktablagerung im IRB der KFA Jülich Das Reaktorexperiment Vampyr im AVR KTG-Fachtagung "Spaltproduktfreisetzung bei Reaktorstörfällen" KTG-Reaktortagung, Karlsruhe, 1973 Karlsruhe, ol.-o2.o6.1976

/3/ Iniotakis, Gottaut, Münchow /8/ N. Iniotakis, C.B. von der Decken Theoretische Interpretation der Ablagerungsuntersuchungen in Reaktor­ The influence of dust on the behaviour of fission products in high experiment Vampyr temperature reactors, ENS/ANS Topical Meeting on Nuclear Power Reactor KTG-Reaktortagung, Berlin,1974 Safety, Brüssel 16.-19.lo.1978

/4/ Iniotakis, Münchow /9/ C.B. von der Decken, N. Iniotakis, F.P.O. Ashworth: A model for the Theoretische Auswertung und Interpretation der Ablagerungsuntersuchungen description of fission product behaviour in the primary circuit of a in den Reaktorexperimenten Vampyr/AVR und Saphir/Pégase high temperature reactor. In: Gas chemistry in nucl. reactors and large KTG-Reaktortagung, Nürnberg, 1975 industrial plant; Proc. of the Conf., Univ. Salford, UK, April 198o, p. 210-219. /5/ Iniotakis, Malinowski, Münchow Initial results of investigations into fission product deposition in /1о/ N. Iniotakis, H. Gottaut, J. Malinowski, K. Münchow in-pile experiments Ergebnisse zum Ablagerungsverhalten von Cäsium auf HTR-Werkstoffen Nuci engineering and design 34 (1975) 169-18o KTG-Reaktortagung, Berlin, 198o 1. Introduction The knowledge of fission product behavior in the primary circuit of a high temperature reactor is very important to estimate the hazard P. P. PLATE-OUT STUDY USING potential for operators during maintenance and inspection and the IN-PILE LOOP OGL-1 environment in case of severe accidents. The aims of fission product plate-out studies in Japan Atomic Energy Research Institute (JAERI) are 0. BABA (a) The knowledge of fission product in primary circuit of HTGS in Japan Atomic Energy Research Institute normal operation condition. Oarai Ibaraki (b) Evaluation of a plate-out model and plate-out analysis progra=e for normal operation condition. Japan (c) Evaluation of a computer programme for the calculation of fission Abstract product release to the environment in case of accidents. (d) Development of special techniques to decrease the amount of fission Fission product plate-rout measurements are carried out using the product plate-out In primary components and piping of ETG3. in-pile loop OGL-1 installed in the JMTR in Japan. The primary circuit (e) Development of decontamination technique. of OGL-1 is about 200 meters long including two regenerative heat ex­ At present our work is concentrated on (a) and (b) using the in— changers, primary cooler, dust filter and gas circulator. pile high temperature gas loop OGL-11'. In this paper the measuring F.P. measurement is done from the outside of the primary pipe by a method of fission product plate-out distribution applied in 0GL-J and pure Ge detector in lead shield with a collimator. The F.P. nuclide is recent results are described. identified by gamma energy and its amount calculated by count rate. -15 The calculated result by the plate-out analysis programme "PLAIN" fixed measuring points are distributed along about 10Q meters between is compared with measurements and a plate-out model is discussed. the fuel irradiation specimen and gas circulator. The primary gas tem­ perature during OGL-1 operation at each measuring point ranges from 700 to 30°C. The F.P. plate-out measuring is done after the end of each 2. Fission product plate-out measurement in OGL-1 operation of 500hours. The sensitivity of this system is lO"1* to 10-5 2.1 Description of OGL-1 pCi/cm2 for 20 hours counting, time at one point. The OGL-1 primary circuit which is shown in Fig-1 comprise an We have gathered plate-out data of 1-131, Cs-137, AgllOm etc. of 6 in—pile tube, No.l and No.2 regenerative hest exchangers, a primary operation cycles. These data are used for the verification of the plate­ cooler, a dust filter, gas circulators and an electric beater. out analysis programme "PLAIN", which was made by JAERI based on the The material of the components and piping used in the high tem­ theory set up by Iniotakis. perature region from the electric heater to No.l regenerative heat First fuel specimen was irradiated for 2 cycles and the second one exchanger is Hastelloy-X. The material used in other parts is was irradiated for 4 cycles. OGL-1 operation condition, such as gas austenitic stainless steel type 316. The total length of OGL-1 flow rate, temperature etc., was changed in order to meet the irradiation primary circuit is about 200 meters and the gas circulators are placed requirement of each fuel. about 100 meters from the fuel specimen exit. _ z In the case of 1-131, the plate-out activity was about the same in OGL-1 is normally operated at a pressure of 30 ^S f/ca (gauge) the first 2 cycles and decreased during following 4 cycles, suggesting and the flow ranges between 50 to 85 E/sec In order to meet the re­ that the F.P. release rate of the second fuel was lower than first one. quirement of the fuel irradiation test. The temperature of gas de­ 1-131 did not plate-out much in the high temperature region of the loop, creases from 1000°C at the fuel exit to 30°C at the cooler exit. The but in the low temperature region in the down stream a lot of 1-131 chemical impurity of the primary gas is kept below 10 ppn (volume) by was detected. The calculation for 1-131 was in good agreement with the the purification system. experiment for all 6 cycles. The type of fuel specimen irradiated in OGL-1 is TRISO-II which will be used in VHTR and the total heat generation of the fuel assembly In the case of Cs-137, the plate-out activity increased every cycle is around 70kw. The maximum fuel temperature is about 1400"C. The for the first fuel. A high quantity plated out in the high temperature fission product release rate ranges from 10-1* to 10~Б in R/B. The region and decreased with increasing distance from the fuel. During the fission product released from the fuel circulates in the primary irradiation of the second fuel, the activity of Cs-137 decreased in the system and plates out in it. high temperature region, and increased in the low temperature region. These results suggest that a part of Cs-137 plated out in the high tem­ perature region was carried to the down stream because of low partial 2.2 Selection of measuring method pressure in the primary gas. In order to perform plate-out studies efficiently, the choice of The calculation for Cs-137 was also in good agreement for first 2 measuring method of fission product Is very important. From the aims cycles but there was no good agreement between calculation and experiment of our experiment the measuring method should meet following require­ for the last 4 cycles. ments. (a) Fission product plate-out distribution along the primary gas flow each measurement point. In the first measurement, the collimator is to be measured in short time. window is closed with a lead plug during the counting tine, and in (b) Periodic measurement at the same points should be possible to the second it is opened. The net ga=ma-ray peak counting rate is observe the changes of plate-out activity after each OGL-1 obtained by subtracting the peak counting rate in the first measure- operation cycle, is • t from the corresponding counting rate in the second. The plate­ (c) Separate measurement of the fission product adsorbed on the metal out density is determined from the net gamma—ray peak counting rate itself. using a conversion coefficient. The conversion coefficient is cal­ There are two typical methods to measure the fission product culated by the detector calibration data obtained by experiment and plate-out. One of them is a destructive sampling method and the other the geometry data for each measurement point. one is a non-destructive method. As the above requirements from (a) The measurement points were selected and fixed at 15 points on to (c) cannot be met by single measuring method, we applied a non­ the outside of the primary duct from the fuel irradiation specimen to destructive method for Ca) and (b), and a destructive sampling method the circulator outlet. The gas temperature at these points ranges for (c). from 700 to 30°C. The non-destructive measuring method we applied is an external duct scanning method using in-situ gamma-ray spectrometers. This

2 method had previously been applied in the Teach Bottom Reactor ^ to 2.4 Results 2 measure the plate-out activity of about 1 uCi/cm , and it was improved The third fuel specimen of OGL-1 in-pile experiment was irradiated -5 2 to have a sensitivity up to ID uCi/cm by Terada et al. In JAERI. for two cycles from the 46th JMTR operation cycle in March in 1979. Measurements by this method have successfully been carried out during Tne fuel assembly had one fuel pin in a graphite holder. The fuel 6 OGL-1 operation cycles. pin consisted of 20 fuel compacts and a graphite sleeve. The average The destructive sampling method is used for small test pieces heat generation of the fuel and maximum fuel temperature during the installed in the primary* gas flow of OGL-1 to measure the adsorbed irradiation were calculated from the measured gas and fuel temperature and diffused fission products separately. The first measurement by tc be 39kw and 1350°C respectively. this method is now in progress at JAERI. Therefore results from these The fourth fuel specimen was irradiated for four cycles from the measurements are not available right now. 48th cycle in Oct. in 1S79. The fuel assembly had three fuel pins in a graphite holder. Each pin contained 20 fuel compacts. Total heat 2.3 Measurement system and points generation of three fuel pins was 76kw average during the whole irra­ As shown in Fig.2 for in-situ measurement of the fission product diation time. Maximum fuel temperature was 1340"C. These values plate-out, a portable pure Ge gamma-ray detector with a small liquid- were also calculated as the third fuel. nitrogen-dewar and a lead collimator was arranged at the outside of The plate—out measurement was carried out after each operation the primary duct in the OGL-1. The Ge detector system was designed cycle. in order to fit into narrow spaces in the primary duct room of the The following isotopes were found to be present ; Cs-134, Cs-137, OGL-1. 1-131, Ag-110m, Sb-124, Zn-65, Co-60, and other activation products. The pure Ge detector has a size of about 44mm diameter and 56inm long. Its full width at half maximum (FWHM) energy resolution and the peak detection efficiency for 1.33 MeV gamma-ray were 1.9 keV and 13Z (relative efficiency to 3"x3" Nal(Tl) detector) respectively. The lead collimator is of a cylindrical well type, having a rectangular window of 4cmxlcm on the side wall. The thickness of lead Is 100mm. The collimator is always arranged that way that the longer side of the window is in parallel to the direction of the axis 3. Verification of the plate-out analysis programme of the pipe at the measuring point, in order to eliminate measurement errors caused by the lack of accuracy of the distance between the pipe 3.1 Analysis programme and detector center due to different thermal expansions of inner liner The analysis programme "PLAIN" was made by JAERI to calculate and pressure pipe. Gamma-ray spectra acquired by the detector and the fission product plate-out in the primary duct of HTGR in normal accumulated In a pulse height analyzer are analyzed by a data process­ operation condition. The plate-out model of this programme is based ing system using a mini-computer. The peak energies and peak counting 4 on the theory developed by Inoitakis" ), which considers reversible rates are determined by spectral analysis of the gamma-rays. The and Irreversible processes. The reversible process is adsorption- nuclides are identified by the peak energies. desorption and the irreversible processes are diffusion and chemical In order to obtain a net counting rate of gamma-ray through the reactions. collimator window, gamma-ray spectrum measurement is made twice for The basic equations used in the plate-out model of PLAIN programme 3.2 Calculation results and discussion are as follows. Plate-out calculation with PLAIN programme was carried out for each 6 operation-cycles of OGL-1. For the calculation by the PLAIN

Щ 4hA 3N, programme, the value of fission product concentration in primary gas

+ (N N + x K + v x N = 0 at some point should be known and entered. But it was too low to be at ~ i- vi> i i IT - i-i i-i measured at any of six cycles. Therefore the concentration of the fission product at the fuel specimen outlet was selected for every

- h^Ni-N^) - 9lHol + ^У^Н^ - HhM « 0 calculation to give the same results of calculation and experiment at the nearest measuring point to the fuel. In Fig.3 to Fig.7 some of the calculation results compared with experiments are shown for the typical fission products 1-131 and Cs-137. ЭМО1 i 3 In the case of 1-131, the distribution of the plate-out activity after each operation cycle was constant for the first two cycles during the irradiation of the third OGL-1 fuel specimen. In the following 2 ЭФ1 Э Ф Эф i 1 1 four cycles when the fourth fuel specimen vas irradiated, the plate­ - D< { ^ +ï -âp-}+Vi-xi-i*i-i = 0 out activity decreased to about one hundredth and also was nearly constant for each cycle. This fact suggests that the fission product release rate of the fourth fuel was lower than that of third one by a 3MDi 3«i Ro 3*i factor one hundred approximately. + ^iMDi " *i-lMDi-l =-Di { Эр •ЦТ Ri Ri Эр Ro The plate-out calculation was made only for a specified operation cycle neglecting the effect of the previous cycle, because the life of 1-131 is short compared with the operation and shut down time of each cycle. Fig.3 and Fig.4 show the experimental and calculation n -37- + ЧМп! - Xi-iHDi-i = (I - h^is/J^l ^rt - i*iP=Ri results of the 47th and 51st cycle respectively. These figures show that the calculations are in good agreement with experiment in both, where the third and fourth, fuel irradiation tests. N : F.P. concentration in the main flow region In the case of Cs-137, the plate-out activity increased during Ни : F.P. concentration immediately above the wall the first two operation cycles proportionately to the operation time.

M0 : Number of F.P. atom adsorbed on the unit wall surface This fact suggests that the release rate was constant and there was ф : Concentration of diffused F.P. in the wall no equilibrium between either adsorption and desorption or penetration Mn : Total number of F.P. atom per unit wall surface which and sublimation during the whole operation. The calculation was done diffused into the wall for each cycle considering all the previous operation cycles. The i,i-l : Fission product and its precursor concentration of Cs-137 at the fuel exit was assumed to be constant and it was obtained from the experimental results of the 46th cycle Equation System in PLAIN Code by the same process mentioned above. In Fig.5 the calculation result for the 47th cycle shows a good agreement with the experiment. These equations are solved analytically using the Laplace transforms In the measurement of the 48th cycle, the first cycle of the and calculated numerically in the programme. fourth fuel, the plate-out activity of Cs-137 decreased a lot at the A diffusion constant used to calculate the mass transfer coef­ high temperature region where the penetration and diffusion processes ficient h in the equations above is obtained by the formula of Wilke are dominant and increased at the low temperature region where the and Lee5). A desorption energy used to calculate the desorption adsorption process is dominant. This suggests that the sublimation constant 9 is obtained by the theory of Levine and Gyftopoulos^*1. process was more dominant than penetration in the high temperature The sublimation constant ) is calculated assuming that the phenomenon region and the activity once plated out there transferred down stream can be interpreted as a combination of diffusion just beneath the to the low temperature, because of the low partial pressure of Cs-137 surface of substrate and desorption. The penetration coefficient in the primary coolant. In the following three cycles, the plate-out 1 - ß is not theoretically obtained yet but determined by parametric activity in the high temperature region increased a little at an almost survey and comparison between calculation and experimental results of constant rate. OGL-1. Gamma-ray Calculation was carried out for these last four cycles assuming spectrum that the concentration of Cs-137 at the fuel exit was one hundredth Primary of the previous two cycles. In the Fig.6 and Fig.7, the calculation duct results of the 48th and 51st cycles are compared with corresponding experimental results. There is no good agreement between them. The reason can be thought that the sublimation velocity of Cs-137 was not estimated correctly enough for the calculation of a circuit once contaminated. Portable Ge (Pure) — 4. Summary and Conclusions detector Plateout The plate-out measurement in OGL-1 gives us a useful information activity about fission product plate-out in normal operation condition of HTGR. The PLAIN programme is useful for the plate-out calculation of a circuit clean at initial condition, but it should be modified to give correct sublimation velocity in order to give better calculation results Lead also in a circuit already contaminated. cylinder If a more accurate theory for sublimation velocity is found, it will be possible to evaluate the fission product release to environment in case of HTGR accidents. Collimator window

Fig.2 In-situ Measurement of the Fission Product Plate-out in OGL-1

10 о XIO*

4 ln-pile Tube „ I0< о 10 E \ ° О v ° " Ю 8 о o Г J O" U 0 4. z2 о e Primary Cooling ^ ю 62 > v^_^ i- TEMPERATURE о E System & 10"'

a>

4 Gas Circulator I io No 2 HX ~11 >.._.. i vi 1 25 50 75 100 j=Cx3=j DISTONCE FROM FUEL EXIT ( M ) ПП If I о I b I С IDLLEL" No.1 HX FILTER Ф0 О : IN-PILE TUBE D:NO.L HX G : FILTER B : HIGH TEMP. DUCT- I E : NO.2 HX H : GAS CIRCULATOR Fig.l OGL-1 Flow Diagram С : HIGH TEMP. DUD-2 t COOLER

Fig.3 Plate-out Activity of 1-131 in OGL-1 (47th cycle) Distonce from Fuel Exit (m) Distance from Fuel Exit (m) I a I b I с 1 d I I el I f I 11 j I a I b I с I a I I el If! II j a : In-pile Tube d:No.lHX g : Filter g n a : In-pile Tube d-No. I HX g : filter g h b : High Temp. Duct-1 e:No.2HX h : Gas Circulator b : High Temp. Duct-1 e:No.2HX h : Gas Circulator с : High Temp. Duct- 2 f : Cooler с : High Temp. Duct-2 f : Cooler

Fig.4 Plate-out Activity of 1-131 in OGL-1 (51st cycle) Fig.6 Plate-out Activity of 1-131 in OGL-1 («8th cycle)

Distance from Fuel Exit (m) Distonce from Fuel Exit ( m ) I о I b I с I d I lei I f I II j • о I b I с I d I I el I f I II j o: In-pile Tube d : No. I HX g : Filter g h о : In-pile Tube d:Nol HX g : Filter g }, b: High Temp. Duct-1 e : No. 2 HX h : Gas Circulator b: High Temp. Duct- I e : No 2 HX h : Gos Circulator c: High Temp. Duct-2 f : Cooler c: High Tempt Ducf-2 f : Cooler Fig.5 Plate-out Activity of Cs-137 in ОС'. 1 (47th cycle) Fig.7 Plate-out Activity of Cs-137 In OGL-1 (51st cycle) References FISSION PRODUCT BEHAVIOR IN THE PEACH BOTTOM AND FORT ST. VRAIN HTGRs 1) S. Ouchi, Y. Okamoto, Reprint from "Gas Cooled Reactors with Emphasis on Advanced Systems" vol.II IAEA-SM-20.Q/15 (1976) 219-236. D..L, HANSON, N.L. BALDWIN, D.E. STRONG 2) F.F. Dyer, et al-, Distribution of Radionuclides in the Peach General Atomic Company Bottom HTGR Primary Circuit During Core 2 Operation, ORNL-5188 San Diego, California (1977). USA 3) H. Terada, et al., Non-destructive Measurement of Fission Product Plate-out using In-situ Ge(Li) Gamma-ray Spectrometer. Second U.S.-Japan Seminar on HTGR Safety Technology Vol.1 (1978) 151-161. ABSTRACT 4) N. Inoitakis, et al., Initial Results of Investigations into Fission Product Deposition in In-pile Experiments. Nucl. Eng. and Actual operating data from the Peach Bottom (FB) and Fort St. Vraie Design, 34 (1975) 169-180. (FSV) High-Temperature Gas-Cooled Reactors (HTGRs) have been compared with 5) Wilke and Lee, Ind. Eng..Chem., 47, 1253 (1955). code predictions to assess the validity of the methods used to predict the 6) J.D. Levine and E.P. Gyftopoulos, Adsorption Physics of Metals behavior of fission products in the primary coolant circuit. For both Practically Covered by Mettalic Particles II Desorption Rates of reactors the measured circuit activities were significantly below design Atoms and Ions, Surface Science 1 (1964) 225-241. values, and the observations generally verify the codes used for large HTGR design.

The PB primary circuit after seven years of operation was exceptionally clean. A fuel element purge system virtually eliminated the release of fis­ sion gases into the primary coolant circuit. Extensive examinations at end- of-life revealed that only Cs and trace amounts of Sr had placed out in the circuit. Their plateout distributions were in excellent agreement with PAD code predictions. Most of the deposited activity was associated with car­ bonaceous surface films which resulted from occasional small inleakages of lubricating oil.

Primary circuit activities in FSV during the first cycle were also very low. Noble gas activity was about 12 of the design limit; and the circulat­ ing iodines were at least one order of magnitude below the limit, although the measurement uncertainties are significant- The plateout per pass of the iodine isotopes increased with decreasing half-life (the value for 1-131 is about 12 per pass) as predicted with the PADLOC code. Gamma scanning of two helium circulators indicated very low plateout activities. Iodine-131 was the principal fission product observed, along with small amounts of Cs-134, Cs-137, зг-d Ba/La-140.

1. INTRODUCTION

The quantity and distribution of radionuclides in the primary coolant circuit are of fundamental importance to reactor design. They affect, for example, not only shielding requirements and maintenance procedures but safety analyses as well. In the case of the High Temperature Gas-Cooled Re­ actor (HTGR), actual operating experience is limited, and the circuit activ­ ities must be established largely by theoretical predictions. Typically, Surface Mass Balance 5fl these predictions are made with complex computer codes which must be verified to establish the adequacy of the design. The most rigorous test of 3S/3t b - XS + k(C - Cs) , (2) such transport codes is their ability to reproduce actual reactor perfor­ Source Decay Mass Transfer mance. Therefore, to assess the validity of the design methods used at Gen­ eral Atomic (GA), actual operating data from the Peach Bottom (PB) and Fort where S = surface concentration pg/cm2, St. Vrain (FSV) HTGRs have been compared with code predictions. For both b = surface source rate, Ug/cm2—sec. reactors the measured circuit activity levels were significantly below the design limits, and the observations generally verify the codes used for Equilibrium Adsorption Coupling large HTGR design.

in Cs = (a + ß/T) + (T + 6/T) in S , (3) Before the results are presented, the design methods used at GA to prfiict primary circuit activities will be briefly reviewed. where T = absolute surface temperature, °K, a, ß, i, and 6 are constants determined experimentally.

The code has three options for treating deposition: (1) no sorption 2. FISSION PRODUCT DESIGN METHODS (e.g., a nonadsorbing noble gas); (2) no desorption (the surface is a per­

fect sink or, more precisely, the Cg is zero for all surface concentra­ Prediction of primary circuit activities is a two-step process. First, tions); and (3) desorption (an adsorption isotherm is employed such that, at the release rates of fission products from the reactor core must be calcu­ a given surface temperature and coolant concentration, an equilibrium sur­ lated, and then their distribution in the primary circuit must be deter­ face concentration exists beyond which no further accumulation occurs, i.e., mined. Given the charter of this conference, the latter will be emphasized Cs = C). The material property data that serve as input to the present here. The distribution of coqdensible nuclides between the coolant and the plateout codes are gaseous diffusivities and sorption isotherms. Gaseous surfaces of the primary circuit and the distribution of deposited, or diffusitivities along with fluid properties are employed in empirical corre­ "plated out," activity must both be considered. lations to calculate convective mass transfer coefficients. The reference material property data used for the reactor design are reported in Ref. 3. At GA, plateout calculations are made with the PADLOC code (Ref. 1) or with its predecessor, the PAD code (Ref. 2). While the physical models are 3. DESIGN METHODS VERIFICATION analogous, PADLOC is more flexible and orders of magnitude more efficient than PAD. Both codes are transient, one-dimensional mass transfer codes. 3.1. Peach Bottom Experience Numerical solutions are obtained for the coupled, nonlinear differential equations describing the conservation of mass between the coolant and the 3.1.1. Observations. Considerable work was done to characterize the surface with a convective boundary condition. The coolant and surface behavior of fission products in the primary circuit of the Peach Bottom concentrations are coupled by an equilibrium adsorption process which can be HTGR. Oak Ridge National Laboratory (ORNL) conducted a surveillance program described by either linear or nonlinear sorption isotherms. A simplified throughout the course of Core 2 operation (Ref. 4), and an extensive end-of- set of equations is given below. life (EOL) program (Ref. 5) was performed after the reactor was shut down for decommissioning because of its uneconomically small size. Coolant Mass Balance The PB primary circuit after seven years of plant operation was

3C/3t + V3C/3X = В - AC - kP/A (C - Cs) , (1) exceptionally clean. A fuel element purge system virtually eliminated the Convection S" irce Decay Mass Transfer release of fission gases into the primary coolant; the total circulating activity never exceeded one Ci throughout Core 2 operation (Ref. 6). The where С «= coolant concentration, ug/cm^, plateout distributions of gamma-emitting nuclides in the primary circuit at t = time, sec, EOL were determined by extensive ijj situ gamma scanning. The dominant gamma X « distance, emitters were Cs-137 and Cs-I34; their relative distributions were similar. В = coolant source rate, um/cm^-sec, These cesium results were presented previously (Ref. 7). X = decay constant, sec-*, к = mass transfer coefficient, cm/sec, Following the gamma scanning, a large number of samples were destruc­ Cs " coolant concentraton in the boundary layer, iig/cm^ tively removed from the coolant ducts and steam generator tube bundle and P « wetted perimeter, cm, examined at CA. A.number of relevant observations were made. First, all A = cross-sectional area, cm2, the specimens were coated with a thin carbonaceous film which resulted from .V = velocity, cm/sec. the cracking of lubricating oil periodically leaked into the reactor. Gamma scanning of the specimens confirmed that only Cs-137 and Cs-134 were present above background, and the specific activities were reasonably consistent with the in situ scans. Radiochemical examination revealed Sr-90 to be the only other fission product present in measurable concentrations, but its specific activity was about three orders of magnitude less than that of cesium. Several specimens were also analyzed for Sr-90 content by 0RN1. and by the Commissariat a l'Energie Atomique (CEA). Neutron activation analysis failed to detect any 1-129. Significantly, >S0% of the cesium and strontium was associated with carbon film.

3.1.2. Predictions. The experimental and predicted plateout distributions for cesium and strontium are compared in Figs, la and lb, the format of which is the PAD code representation of the PB primary circuit. The in situ cesium data are shown, including the collapsed steam generator data (axially and radially averaged), and all the Sr-90 data; the scatter in Sr-90 data is immediately evident. The specific activity is plotted as a function of fractional cumulative surface area. (Note that the abscissa is drawn to scale within a given section but the scale differs from one section to an­ other.) Two PAD calculations for cesium are shown: (1) mass transfer con­ trol (i.e., the surfaces are perfect sinks for cesium), and (2) sorptivity control; only a perfect sink calculation is shown for strontium. In all cases, the time—average core release rate of cesium was adjusted so that the Fig. la. Plateout distribution of Cs-137 and Cs-134 in predicted specific activity at the evaporator inlet (shell side) was approx­ Peach Bottom HTGR imately equal to the measured value. Since the decay of Xe-137 produced negligible amounts of Cs-137 compared to the directly released component, the relative distributions shown in Fig. la apply equally well to Cs-134, which has no gaseous precursor.

Inspection of Fig. 1 indicates that the mass transfer control, or per­ fect sink, case (solid lines) resulted in good agreement everywhere except for cesium in the hot duct leading from the reactor vessel to the steam gen­ erator. Here the specific cesium activity is overpredlcted by an order of magnitude. Since the flow geometry is simple (a circular duct), prediction of the mass transfer coefficient should be reasonably accurate. Thus, the logical conclusion is that the cesium deposition in the hot duct is not lim­ ited by mass transfer effects but rather by the high surface temperature. Strontium deposition in the hot duct appears to be perfect sink, but the scatter in the data weakens the conclusion.

The major difficulty is choice of appropriate sorption isotherms to de­ scribe the cesium sorptive capacity of the surface. The hot duct cladding was constructed of SS304, for which sorption data are available (Ref. 8). Another complication was that the carbon film could be a significant sink for cesium. Given these uncertainties, the surface sorptivity was investi­ gated parametrically. In summary, the experimentally observed cesium and strontium plateout distributions in PB can be predicted almost exactly with the PAD code, providing appropriate sorption isotherms are employed. How­ ever, the observed sorption behavior is consistent with either assuming that the primary cesium sink is a relatively oxide-free SS304 surface or assuming Fig. lb. Plateout distribution of Sr-90 in that the carbon deposit has a cesium sorptivity intermediate to that of Peach Bottom Reactor TABLE 1 graphite and matrix. Since most of the radioactivity was associated with PLATEOUT ACTIVITY ON FSV CIRCULATOR carbon film, the latter appears more likely. Cesium and strontium deposi­ tion throughout the circuit wae apparently mass transfer controlled with the

exception of cesium In the hot duct. The profiles suggest that both cesium 2(a) Specific Activity (pCi/cm and strontium were transported primarily in atomic form despite the presence of carbonaceous dust. Details are given In Ref. 7. Nuclide Measured Predicted

3.2. FSV Experience 1-131 1.4 20 Cs-137 0.030, 0.024(b> 0-032 Since the FSV HTCR only began commercial operation in 1979, the fission Cs-134 0.005 0.001 product data base for FSV is more limited than for PB. Primary coolant Ba/La-140 0.024 0-33 activity levels have been monitored throughout the operation of the plant. Zn-65 0.039 The release rates of noble gases from the core are measured regularly, and — Co-60 0.018 periodic measurements of the circulating iodine levels are made with the — Cr-51 0.017 iodine monitor (Ref. 9). This device relies upon the measurement of the — Mn-54 0-007 xenon daughters of the collected iodines and, as such, provides data for — Fe-59 0.007 1-133 and 1-135; radiologically important 1-131 must be estimated by extrap­ — olation. Some Information on the plateout levels was obtained by examina­

tion of two circulators which have been removed from the prestressed con­ a t 7f scan results decayed back to time crete reactor vessel (PCRV) -.one for a safety inspection (a technical of removal. specification requirement) and the other to repair a failed shutdown seal. a ^ 't analysis of leach solution. More definitive information on the plateout levels will be obtained when the first plateout probe (Ref. 9) is removed.

3.2.1. Observations. Primary circuit activities during the first cycle observed in PB. Their specific activities did not increase appreciably with were very low. Typical coolant inventories, measured at —70% power, are operating time, and they were insignificant compared with the fission compared with the final safety analysis report (FSAR) design limits below; product activity. Hence, no detailed analysis of these data has been as expected, the activities are well below the design limits: attempted.

Circulating Activity (Ci) 3.2.2. Predictions. The FSV Cycle 1 operating history has been modeled Nuclide Measured Design with the reference large HTGR core design codes, and the predictions from the physics to the fission products have generally been confirmed. In the 1-135 0.8 53.8 fission product transport area, the emphasis has been on predicting the 1-133 1.2 34.4 coolant activity levels, because those measurements are the most extensive Total Kr and Xe 300 32,300 and reliable. The measured and predicted R/Bs (release rate Into coolant divided by birth rate in the fuel) for the reference nuclides Kr-85a and The plateout activities on the "C" circulator were estimated by gamma Xe-138 are compared in Fig. 2. The comparison for other isotopes is simi­ scanning a region of the impeller prior to disassembly of the machine; lar. Considering just the data taken at a reactor power of 70Z, the mea­ results are summarized in Table 1. Since the geometry was not ideal, there sured R/Bs varied as the half-life to the 0.4 power rather than the expected Is some uncertainty in the calibration; however, subsequent analysis of a 0.5 power. Since the subject of fuel performance is peripheral to this decontamination solution for Cs-137 gave reasonably consistent results. conference, this matter will not be discussed further here. Attempts were also made to analyze the decontamination solution for stron­ tium by radiochemical analysis and for 1-129 by (n,-y) analysis; however, The circulating iodine levels were also predicted. Here the comparison interference by unidentified elements in the proprietary decontamination is more complex because the steady-state circulating activity is established solution was so extensive that reliable results were not obtained. by the relative rates of core release, plateout, and helium purification. At GA it was decided to characterize the plateout rate by a so-called Small amounts of metallic, neutron activation products, such as Cr-51 "plateoüt per pass (P/P)," which is defined as the fractional depletion of and Co-60, were also detected on the FSV circulator. These activation the coolant inventory due to plateout during a single transit of the helium products may have been produced from small amounts of construction debris through the reactor circuit. In order to calculate the plateout per pass that became entrained in the helium and then activated when passing through from the measured 1-133 and 1-135 inventories, the 1-133 and 1-135 release the core. Similar low levels of these activation products were also rates from the core must be assumed. It has been standard practice to assume that iodine and xenon Isotopes of the same half-life have the same release characteristics. Making this assumption and using the measured O.A power Xe half-life dependence, the apparent P/P rates for 1-133 and 1-135 were calculated and are plotted versus half-life In Fig. 3. Also shown is the predicted relationship between P/P and ^1/2 as calculated with the PAD1.0C code using sorption isotherms measured by CEA (Ref. 10). Since the CEA isotherms are linear, the predicted P/P is independent of the core release rate and dependent only upon the half-life. Although the predicted slope appears consistent with the measurements, the absolute value is too low. When the absolute sorptivity is increased by a factor of five, the agreement is much improved. (There is considerable evidence from the CPL 2 test program that these isotherms have the correct temperature dependence but underpredict the absolute sorptivity.)

10"

10"

r~L_n |—•—^* i r» ~~EPöif

J 1~H" 00-0 0o o° Г-Lry О О The levels of fission products on the circulator were also predicted 10,- 6 С and compared with measurements in Table 1. Again, the problem is that these г predictions depend upon both the release rate from the core and the plateout distribution. The results are nixed. The Cs-137 activity was well pre­ dicted, but the Cs-134 was underpredicted by a factor of five; the latter • Kr-85m MEASURED discrepancy could result from errors in predicting the amount of Cs-134 pro­ О Xe-133 MEASURED duced by activation of Cs-133, the release rate, the plateout distribution, Kr-85m CALCULATED or a combination othereof. The specific Ba/La-140 activity was overpre— Xe—138 CALCULATED dieted by an order of magnitude. It is speculated that the release of the 14-sec Xe-140 precursor was overpredicted because the reference design methods do not account for the attenuation of short-lived gases by the graphite web of the fuel element. The amount of 1-131 was overpredicted by 10" I I I I I I I I I I I I I I I I I more than an order of magnitude. Some of the discrepancy likely resulted 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 IB0 170 from an overestimation of the core release rate. Nevertheless, the measured OPERATING TIME (EFPD) 1-131 plateout activity on the circulator and the 1-133 and 1-135 circulating activities do seem inconsistent. Data from the plateout probe Fig. 2. Comparison of measured and calculated fission gas release for FSV cycl should resolve this apparent discrepancy. 4. DISCUSSION REFERENCES

The levels of fission products in the primary circuits of both PB and 1. Hudritsch, W. W., and P. D. Smith, "PADLOC, A One-Dicensional Computer FSV were quite low, «eil below the design limit6. However, the "mix," or Program for Calculating Coolant and Plateout Fissioa Product composition, of fission products in the two HTGRs is quite different: the Concentrations," DOE Report GA-A1440L, General Atonic Cocpany, November PB circuit at EOL contained only the long-lived volatile fission metals and 1977. negligible fission gases; the FSV circuit during the first cycle contained primarily fission gases and their solid daughter products. Differences in 2- Vanslager, F. E., and L. D. Hears, "PAD: A Computer Cede for core design and in operating history are responsible. The purge system in Calculating the Plateout Activity Distribution in a Reactor Circuit," PB virtually eliminated the release of fission gases; even though the qual­ Gulf General Atomic Report GA-10460, January 1971. ity of fuel in FSV is far better than that used in PB, the nonpurged core design releases more fission gases into the primary circuit. PB Core 2 operated for 897 effective full-power days (EFPD); but FSV Cycle l «as 175 3. "HTGR Accident Initiation and Progression Analysis Status Report, Vol. EFPD, and the power level never exceeded 707.. As the burnup of the FSV core 5, AIPA Fission Product Source Teres," DOE Report GA-AÎ3617, General increases and the power level rises, the mix of fission products may «eil Atomic Company, February 1976. change. 4. Dyer, F. F., et al-, "Distribution of Radionuclides in the Peach Botto= Overall, the PB and FSV operational data verify the design methods used HTGR Primary Circuit During Core 2 Operation," ERDA Report 0RSI.-51S3, to predict circuit activities for the large HTGR. The PAD and PADLOC codes Oak Ridge National Laboratory, March 1977. appear to be adequate empirical tools; for prediction of plateout distribu­ tions provided that appropriate material data, the most important of which 5. Steward, K. P., "Final Sunnary Report on the Peach Bottoa End-of-Life are the sorption isotherms, are available. While these results are encour­ Program," DOE Report GA-A14404, General Atonic Company, July 197S. aging, additional design verification is needed, and the ongoing FSV sur­ veillance program should provide much of the needed data. 6. Hanson, D. L., and N. L. Baldwin, "Fission Gas Release froa Core 2 of the Peach Bottom HTGR," ASS Transactions TAKSO 28, 6SS (197S).

7. Hanson, D. L., N. L. Baldwin, and V. E. Selph, "Gacna Scanning the Primary Circuit of the Peach Bottoa HTGR," USERDA Report GA-A1&161, ACKNOWLEDGMENT General Atonic Conpany, October 31, 1976. Also AXS Transactions TANSO This work was supported by the U.S. Department of Energy under Contract 2±, 403 (1976). DE-AT03-76ET35300. 8. Milstead, С. E., and L..R. Zumwalt, "Cesiun Deposition on Stainless Steel," Sucl. Appl. J3« (1S67).

9. Hanson, D. L-, N. L. Baldwin, and D. Albertstein, "Fort St. Vrain Plateout Trobes," General Atonic Report GA-A14402, October, 1977.

10. Blanchard R., CEA, unpublished data. variability of from one to two orders of magnitude in the loading, which was clearly related to changes in the surface condition. The lowest IODINE SORPTION AND DESORPTION FROM loadings were obtained with specimens that appeared to be corroded; the LOW-ALLOY STEEL AND GRAPHITE highest loadings, which in this range was constant at 5 Vg/cn-, evidently represent monolayer coverage on the bare steel surface. A steep decline in loading was observed below 10~10 bar, with a value of M).05 ug/cri2 R.P. WICHNER, M.F. OSBORNE, R.A. LORENZ, R.B. BRIGGS typical at I0-11 bar. As with graphite, iodine desorption rates were Oak Ridge National Laboratory determined in tests involving either a drop in the iodine partial pres­ Oak Ridge, Tennessee sure or an increase in the specimen temperature. USA I. INTRODUCTION

ABSTRACT This paper describes work performed at Oak Ridge National Laboratory on the adsorption of iodine on low-chromium, steel (alloy T-22) and grade Tests have been performed at Oak Ridge National Laboratory on the H-451 graphite. The steel is of a variety that will form the boundary sorptive behavior of iodine on two prominent HTGR primary circuit of the cooler portions of an HTGR primary circuit which is expected to materials — type H-451 graphite, which is the selected core graphite, be an important zone for iodine deposition. Grade H-451 graphite is the and a low-alloy steel (T-22), which is employed in the lower temperature chosen reference core material. regions of the primary system. The determination of equilibrium sorp­ tive capacities at very low iodine partial pressures is emphasized in The principal purpose of this study is to improve the capability this work. Toward this goal, our test procedures allow on-line deter­ for predicting the locations of iodine deposition in the primary circuit mination of the sorptive capacity down to iodine partial pressures of by providing improved sorptivity data. The major problem to date regard­ 11 "VIO™ bar in helium at a total pressure of 1 bar. The apparatus ing sorptivities on steel has been that measurements have not been employed is of the helium flow-through type. Detectors monitor the carried out at sufficiently low iodine partial pressures. Most data are degree of sorption on specimens and traps, which, together with the based on partial pressures of I0-8 bar and above-, although primary-circuit helium flow rate and source vapor pressure, yields a continuous deter­ partial pressure are expected to be on the order of 10—bar. In our mination of iodine mass transport. experiments, therefore, we have sought to achieve as low iodine pressures as possible; the lowest we have thus far achieved is 10~bar, which For the graphite tests, temperatures ranged from 250 to 1000°C. At is about a factor of 100 below previously published data for sorption on each temperature, we find a linear relationship between the logarithm of steel. iodine loading (Ug/g graphite) and the logarithm of the iodine partial pressure. At 1000°C, the iodine loading ranged from 10*"5 Ug/g at an There have been only a few sorptivity studies on graphite that iodine partial pressure of 5 x 10-11 bar up to 2 x 10_ц Ug/g at a par­ relate to HTGR conditions because of the commonly held belief that holdup tial pressure of 8 x 10-3 bar. Equilibrium iodine loadings increased in graphite is not important due to its inherently low sorptivity. How­ with decreasing temperature, though not at a uniform rate. At the lowest ever, using published values for graphite sorptivity of iodine, such as test temperature (250°C), iodine loadings ranged from 2 x 10-2 Ug/g up to those of Osborne,1 one can readily show that the large internal surface 1.2 Ug/g for the same range of iodine concentrations in the helium. In available within the graphite more than compensates for its low sorptivity. addition to equilibrium tests, the rate of iodine desorption was measured No measurements have been made on the reference-core graphite material by two other types of tests: (1) desorption due to a rapid drop in iodine (grade H-451). In addition, earlier studies in the low partial pressure partial pressure and (2) desorption due to a temperature rise. range1 were performed on fragmentized material (<250-jm particles) that could yield uncharacteristically high sorptivities. Sorption tests on low-alloy steel concentrated initially on the 400°C isotherm. At iodine pressures in the range of 10~6 to 10-7 bar and a A second objective of this study is to measure rates of iodine temperature of 400"C, loadings increase with partial pressure evidently desorption from steel and graphite under conditions that may aid in is due to Fel2 formation. Between 10~7 and 10-10 bar, we observed a accident-consequence determination. Toward this goal, desorption rates have been determined by (1) diminution of iodine partial pressure at a constant temperature and (2) stepwise increase in temperature at a con­ stant iodine pressure. The first method approximates depressurization Research sponsored by the Gas-Cooled Reactor Programs Division, Office of conditions (wherein iodine partial pressures are reduced as the total Advanced Nuclear Systems and Projects, U.S. Department of Energy under pressure falls) as closely as possible in a 1-atm apparatus. The second contract W-7405-eng-26 with the Union Carbide Corporation. desorption procedure approximates core heatup conditions and is also use­ Consultant. ful in determining the range of desorption activation energies. At the time of this writing, the major portion of the experimental measuring 50 cm long. The inlet end contained three baffles to mix the work involved in this study hud been completed. However, the analysis incoming gas. Two independently controlled furnaces surrounding the and interpretation of the data are still in an early stage. Therefore, furnace tube were capable of heating the specimens to 1000°C at rapid all conclusions presented here that involve data analysis and interpreta­ heating rates ("Vl00°C/min) with stable temperature control ("Ы^'С). A tion must be regarded as preliminary. cartridge of activated charcoal capable of adsorbing all of the iodine source was mounted in the downstream end of the furnace tube. The count­ ing system was composed of three similar channels feeding data to a 2. DESCRIPTION OF THE EXPERIMENTAL SYSTEM printout control. At predetermined time intervals, the number of counts collected in each channel, the length of the time interval, and the 2.1 Atmospheric-Pressure, Helium-Flow Apparatus accumulated time from the beginning of the experiment were printed by a teletype and recorded on paper tape. The general features of the apparatus for iodine adsorption on steel

and graphite are shown in Fig. 1. Although this figure refers to the Tne 131I source material (typically 5 mCi) was obtained as carrier- experiments conducted with steel, a separate and almost identical system free Nal in aqueous solution. After the addition of a small amount of was employed for the graphite tests. Helium flow to the iodine saturator natural iodine (as Nal) and the extraction of standard samples, elerynral was provided at a rate of 2 to 10 std cm3/min from a purified helium iodine was prepared by the decomposition of Pdl2 in a vacuum. In order source with typically <0.02 ppra by volume oxygen. A second cylinder sup­ to obtain high specific activities, the amount of carrier iodine was

plied helium at faster flow rates — 200 to 400 std cm3/mir.-for the test limited to *vl mg. During preparation, the elemental iodine was deposited

conducted with steel and 10 to 80 std cm3/min for the tests conducted in an ampoule equipped with a magnetically operated break seal. This with graphite. In both streams, moisture was removed by means of a cold ampoule was attached to the experimental apparatus at point .4, Fig. 1.

trap held at liquid nitrogen temperatures, and oxygen was removed by hot- At a helium flow of 1-5 std cm3/min through the cold (-80°C) saturator metal getter. For most of the tests with steel, hydrogen was added to (low-temperature bath, Fig. 1) and purge line, the break seal was broken,

the main flow stream to a level of 5 x 10~L| bar. and the iodine was warmed to 40 to 50°C, permitting it to diffuse slowly into the helium stream and be carried to the saturator where it condensed on the inlet leg. After a few hours, most of the iodine had been trans­ ferred to the saturator, as verified by radioactivity measurements with

|MO»OC»*TROt. VAtVt^ î a survey meter, and the source ampoule was fused closed and removed. ВТ! TtbC .FUOMACE шсякочтягл. t Depending somewhat on sorption test conditions, the effective useful ыс/"'тод TwO-MhC Fun MCE lifetime of a typical l31I source (y2 to 5 uCi/ug I) was ^6 weeks, at ( Vi LL VACUA* / \ vWWWvWvJ I I which point the count rates for most loadings were only marginally useful. L-*.i. 1 • — MA 2.2 Low-Alloy Steel Specimens

The steel test specimens were formed from a low-chromium-content alloy with a nominal composition of 2-1/4% chromium and 1Z molybdenum.

MO&LC S"'ELD£D Dt TELTOWS Much of the lower temperature portions of an HTGR circuit may be fabri­ t£w-TCHPER*TURt cated from material of this type. Material obtained as 3/16-in. plate was rolled to a thickness of 0.25 mm and formed into specimen arrays as Fig. 1. Experimental apparatus for iodine adsorption-desorption studies. shown in Fig. 2.

The iodine saturator was a U-tube with a "v2.0-cm-diam bulb at the Attempts were made to determine the character of the surface by bend immersed in a propanol bath, which was cooled in a refrigerator to means of a mechanical roughness measurement, gas adsorption, and by the desired temperature (-80 to -50°C). A ball check-valve was located metallographic examination. Mechanical roughness measurements made by in the outlet from the saturator to impede the diffusion of iodine into means of a stylus indicated a surface roughness ratio* of 6.8. Gas the primary stream when no helium was flowing through the saturator prior adsorption measurements by means of the BET method indicated the true to the start of the run. Immediately downstream from the junction of the surface area to be M.40 times greater than the superficial area. How­ primary and saturator lines, a small side line was attached upstream from ever, the reliability of this value is low because the total sorption the furnace tube that allowed a small fraction (typically 3 to 6%) of the surface was muiginally low for the apparatus employed. (It will be flow to bypass the sample. This flow passed through a monitor charcoal shown later that iodine sorption data seem to point to an effective cartridge, mounted near a third detector, thus permitting a continuous roughness ratio of ^58.) and independent determination of the iodine partial pressure in the inlet * flow. The furnace consisted of a 30-mm-diam x 1.5-mm-wall quartz tube Defined as the ratio of the true surface area to the superficial surface area. ORHL-DWG 78-ЗЭЗЯ2 0.33 and 1.08 mz/g, with an average value of тЛ.8 mz/g. The average helium density of the specimens tested was 2.05 g/cm3. A list of impuri­ ties found in four specimens in concentrations >1 ppm is given in Table 1.

Table 1. Impurities found in grade H-451 graphite

in concentrations <1 ppa by neutron

activation analysis 14 mm

Range found in Inpurity four saoples (ppa)

Na 1.0 to 1.6

Hg 3.0 to 12

Al 0.7 to 2.4

CI 3.0 to 14 SPECIMEN i SPECIMEN 2 К 0.5 to 1.5 2 mm LONG 51 mm LONG Ti 1.6 to 2.7 V 6.0 to 6.5 Fig. 2. Low-alloy steel sorption specimens. Cr 2.4 to 2.S

Fe 3.4 to 33 Metallographic examinations were conducted on ten specimen plates; Си 0.2 to 3.0 two were as-fabricated controls, three had been exposed to a slightly Zn 1.0 to 7.S oxidizing helium atmosphere at three different temperatures (200, 500, Mo 0.9 to 1.0 and 800°C) for several days, and the remaining five plates were selected from five different adsorption tests. Examination at 1000X magnification revealed numerous small cracks, especially in the sheared edges. It appears likely that these cracks, which were produced during the rolling 3. EQUILIBRIUM SORPTION OF IODINE ON LOW-ALLOY STEEL and/or shearing operations when the plate was being fabricated, could significantly increase the surface area for gas adsorption. 3.1 Range of Test Data

2.3 Graphite Specimens Table 2 presents a partial list of the data taken during the course of the experiments with low-alloy steel. Specimen temperatures ranged The graphite specimens were formed from grade H-451, which at present from 200 to 800°C. In this series of tests, emphasis was placed on the is the reference core graphite that has been selected by General Atomic 400°C isotherm; experinental difficulties were analyzed at this tempera­ Company. This material is made from a near isotropic mix of calcined coke ture before proceeding to other temperatures. Iodine partial pressures filler particles with diameters <500 vm to which are added graphite fines. during the equilibrium experiments ranged from 5 x 10~b to 2 x 10-11 bar. The filler plus pitch binder are extruded into 50.8-cm-diam cylinders, (The partial pressure used here signifies the sum of the partial pressures baked, impregnated once, and then graphitized at 2800°C. The product is of all iodine-bearing species.) At temperatures >400°C and pressures 7 then purified using a halogen gas to diminish the content of iron and below M0" bar, thermodynamic equilibria calculations indicate that the other metals that form volatile halides. The specimens used in these major share of the iodine is dissociated, existing either as atomic iodine 2 tests were annular cylinders closed at one end (cup shaped). Iodine- or as HI. The equilibrium loadings ranged from тЮ.0012 vg I/cm , which 2 laden helium flowed into the open end and passed through the walls of the is approximately the limit of detection for the system, up to ^5 ug I/cm . specimen. The specimen was 2.54 cm long and had a diameter of 1.7 cm and a wall thickness of 0.60 cm. Vigorous cleansing methods were applied to In general, the iodine sorption experiments on low—alloy steel were remove dust and adsorbed moisture after the specimen was machined. Addi­ marked by experinental difficulties, many of which, we felt, were related tional treatment prior to the start of a sorption run included degassing to variations in surface character between similar specimens and also dur­ for ^24 h in a ^7 x 10~6-bar vacuum and heating at 1000°C in helium. ing runs. Initially, difficulties were encountered with specimen oxidation. Despite moisture removal at the liquid nitrogen temperatures and oxygen The measured surface BET area varies somewhat depending on the speci­ removal in a hot-metal trap, some type of oxidant was present that caused men and the measurement technique. Generally, measurements ranged between specimen corrosion in runs 4 to 6. An interesting effect observed during Table 2. Abstract of iodine sorption data on low-alloy steel

Run Temperature Iodine pressure0 Iodine loading Hydrogen pressure

No. (°C) (bar) (ug/cnJ) (bar) Remarks

4-1, 2 400 2-3 x ю-9 2-3 1.. 2 x 10-2 Reducing hydrogen pressure caused specimen oxidation 4-3, 4 400 2 x 10-' 0.1-0.2 3 X ю-5 5-1 400 5-6 x ю-' 0.4 •40".9 Specimen oxidation 200 5-6 x ю-' 0.4-2.0 •40"- 9 5-4 600 0 0.1 M0' -6 Desorption run

10 -a 6-1 400 0.8-1.. 7 x 10" 0.013 •40" Specimen oxidation; poor reproducibility 200 0.8-1.. 7 x КГ10 >0.034 •40" 10 -в 6-11,, 12, 13 400 0.4-14 x lO" 0.003 •40"- в 7-1 400 0.4-2.. 0 x 10-10 0.04-0.05 5 X ю-' Equilibrium plus desorptioa runs 7-2 400 1.0-1.9 x 10"' 0.4 5 X ю-' 7-3 400 1.4 x io- ' >2.1 5 X ю-' 8-2, 3 600-800 0 0.01-0.001 5 X ю-' Desorption runs 8-5 400 9.0-17 x 10"' 1.4 5 X ю-' Equilibrium data 8-6 600 1.3-1.. 4 x 10"' 0.3 5 ю-' - X 8-7 400 9.0-19 x Ю " 1.0 5 X ю-'

9-2 400 2-4 x ю-11 0.087 5 X ю-" 9-3 600 2-4 x io-11 0.003 5 X ю-' 9-5 400 4-6 x ю-11 0.064 5 X ю-'

12-1 400 0.7-1., 2 x ИГ10 <3.2 5 X ю-' Loadings were erratic for unknown reasons 400 2.2-3.. 0 x lO"10 >3.5 5 X ю-' 600 2.2-3.. 0 x Ю-10 0.1-0.2 5 X 10-' 13-1 400 2.6-3., 0 x io-10 5 X ю-' 13-2 500 3.0 x io-10 0.3-0.5 5 X ю-' 13-3 600 3.0 x ю-" 0.091 5 X 10"' 13-4 700 3.0 x 10"10 0.0062 5 X 10"' 13-5 800 3.0 x lO"1 • 0.0012 5 X ю-' Lowest loadings were difficult to distinguish from background 13-6 800 2.0-3.. 0 x 10"' 0.0014 5 X 10"' 13-7 700 2.0-3.0 x 10"» 0.010 5 X 10-' 13-8 600 2.0-3.0 x 10"' 0.046 5 X ю-"

Sum of the partial pressures of iodine species determined from thermodynamic equilibria estimates.

'Based on superficial surface area. these early runs was that the Iodine sorbed on the specimen during or The solid points in Fig. 3 represent data acquired at 400°C during prior to exposure to oxygen contaminant did not easily desorb. It was this experiment. The solid triangles refer to early experiments employing necessary to add hydrogen to the gas flow to remove iodine sorbed on the "purified" helium but with no added hydrogen, whereas the solid circles steel specimen in this fashion. refer to data acquired with 5 x 10-tt-bar hydrogen partial pressure in the helium. As noted earlier, employing "purified" helium resulted in obvious Prior to run 7, the apparatus was tightened with respect to oxygen corrosion of the metal specimen, which is reflected in the lower observed in-diffusion, and it became a regular practice to add hydrogen to the loadings for the solid triangular data points. The dashed line, labeled ц helium flow to a level of 5 x 10~ bar. This tended to control specimen segment D, is drawn through these data. For comparison, the 400°C isotherm corrosion; however, unexplained alterations in the loadings observed dur­ obtained for granular РезО^ (Osborne et al.1) is also shown. In comparing ing runs 12 and 13 were indicative of some change in the surface character the loadings for the oxidized metal specimens (segment D) with that obtained during the run. Despite these measures, some varying degree of specimen for Fe30t,, we note that the former are based on the superficial specimen corrosion was evident In subsequent runs. The source of the oxidant is not areas, whereas the ГезОц data are based on measured BET surface areas. The known. Unquestionably, the ability to accurately predict the degree of loadings indicated for segment D are "^30 to 100 times greater than those iodine sorption on steel, as well as the rate of desorption under depres- shown for the ГезОц data. Since this factor is within the range of esti­ surization or other accident conditions, is hindered by a lack of knowledge mated roughness ratios for the fabricated steel specimens, we conclude that regarding the surface character under these conditions. the apparent difference between the cases is due entirely to the roughness factor of the manufactured metallic specimen. 3.2 Adsorption Isotherms The solid circles in Fig. 3 represent runs in which the hydrogen The equilibrium loading data taken at 400°C are plotted in Fig. 3 as content of the helium flow was set at 5 x 10-tl bar. As noted earlier, the darkened circles and triangles. For comparison, data reported by despite the measures taken (hydrogen addition, liquid-nitrogen freeze 2 3 Milstead et al. and Neill are plotted in the same figure. As noted on traps, hot-metal oxygen gettering), some variable degree of surface cor­ the figure, the major portion of Milstead"s data (the original source is rosion was evident in these runs. Therefore, we feel that the range of ц 5 Burnette et а1. and Zumwalt et al. ) refers to Iodine partial pressures scatter shown for these data is due principally to varying degrees of sur­ 7 greater than 10~ bar, the point above which solid iron iodide is pre­ face corrosion. The most highly corroded specimens approach the values dicted to be stable. Therefore, it is reasonable to attribute the generally "observed for the runs performed with no hydrogen addition (segment D), upward trend of the open circles (labeled segment A) to be due to sublimation which were obviously corroded. Segment 3 of Fig. 3 is drawn across the of iron iodide. It should be mentioned that no hydrogen was added to the upper bound of the solid circular data points and hence represents loadings helium for Milstead's data, and that the oxygen contamination level was on largely noncorroded specimens. We note that this segment extends through quite high ("400 ppm). Therefore, the surface layer for this data was un­ a range of iodine pressures of ^2 x 10~10 to 5 x 10-8 bar at a constant doubtedly Iron oxide. Judging from the general level of the loading — that loading level of 5 ug 1/superficial cm2. Since this probably represents is, the fact that an upward trend is observed from segment B, which is monolayer coverage on bare steel, the true surface area of the metal speci­ believed to be monolayer coverage of unoxidized iron (see below) — it appears mens may be estimated by comparison with theoretically estimated monolayer that surface condition is not a sensitive factor in determining the degree loadings. A surface roughness ratio of 5/0.036 or 58 is obtained for the of loading under conditions where iron Iodides are stable solids. metallic specimens employed in this study using Compere's6 value of 0.086 ug/cm2 for monolayer iodine chemisorption on iron. This value falls I T" 1 г between the mechanically determined value of 6.8 and the BET value of "440. MIL5TEA0 ,

3 В. MONOLAYER. e Neill's data are shown as the three open triangular points in Fig. 3. BARE STEEL Since the gas compositions for these data were not clearly given, we assume that stern measures to avoid surface corrosion were not employed. In addi­ tion, no indication is given regarding the true surface roughness ratio. "»o Hence, all that can be said is that the two data points at 10~9-bar pressure Д NEILL agree roughly with the corroded sample data (segment D), and that the data О MIL5TEA0 point at 10~^-bar pressure, for which equilibrium was not attained, is con­

4 • 5 ж 10" Ьсг Н2 sistent with the high pressure segment of Fig. 3. Below partial pressures A CORROOEO 5URFACE of 'ью-10 bar at 400°C, we note a general reduction of loading, as shown by ,seoR»E segment £7. The data are sparse at these low pressures; hence, the precise manner of loading reduction below 10-10 bar is uncertain.

0.001 10' КГ» 10"" 10"' Summarizing then the observations regarding the 400°C data, the region IO0INE PRESSURE loot) of iron iodide deposition exists above 10~7 to 10~6 bar (depending on the Fig. 3. Measurements of iodine sorption on low-alloy steel at 400°C. hydrogen pressure). This is not chemisorption, and the degree of deposit Table 3. Range of conditions for iodine sorption tests depends solely on mass transfer effects (i.e., the ability with which \ on grade Н-Д51 graphite iodine may contact and react with the substrate iron). Between "^2 x 10~10- 8 and 5 x 10~ -bar iodine burden at 400°C, there occurs what is evidently a Graphite temperature Range of iodine Range of iodine zone of monolayer iodine coverage, which for our specimens yielded a load­ Run range pressure loadings ing of Vi pg/cm2 superficial surface. Below alO-1" bar, the loading No. CO (bar) (ug/g) diminished as the burden was reduced. 3 428 2 x I0"'-5 x JO"7 0.023-0.4 Equilibrium loadings obtained thus far at temperatures above 400°C, 4 600 2 x lO-'-Z x 10-7 8 x 10---0.045 specifically at 500, 600, 700, and 800°C, are shown in Fig. 4. We note, 5 250-1000 1.5 x ИТ7 2 x 10"*-1.2 as expected, the general reduction in equilibrium loading with increasing 6 250-1000 1.5 x 10-' 6.5 x 10~5-0.09

temperature. However, the data are too few and scattered to allow further 7 s 7 250-1000 9 x 10-"-3 x ИГ 3.3 x 10~ -0.6 discussion for these higher temperatures at this time.

Figure 5 illustrates the course of a typical test (run 5), the dura­ tion of which was 42 days. The ordinate signifies the measured iodine ORNL DWG 80-157Z loadings (jlg/g graphite) during the course of the run as determined from the on-line gamma detector reading. The initial segment (A) illustrates the approach to an equilibrium loading of 2.0 x 10-C pg/g at 1000°C in 'ьЗ days. Subsequently, the specimen temperature was reduced to 800°C, and segment 3 illustrates the approach to equilibrium, again in days. The following segment (C) was a time of test disruption due to a power failure beginning 6.0 x 10s s after the start of the run. Upon recovery, equilib­ 0 1 rium loadings were obtained at 600°C (D), 400°C (E), and 250°C (F). An E 5 isothermal desorption test was performed at ^21.0 x 10 s (segment G), a. TEST TEMPERATURE CO where the iodine partial pressure was reduced to zero for ^3 days, during which time the loading diminished from ^1.1 pg/g to 0.4 pg/g. Subsequently, ^•C О 600 I THIS EXPERIMENT the reproducibility of loadings obtained earlier in the run was determined р 4 ^ a TOO l нг ' « » "О" sar in segments D' (600°C), F' (250°C), and E' (400°C). As may be seen, the o.ooi • 800J - loadings obtained with repeated conditions deviated from those initially A 482~» OTHER DATA obtained by approximately -50, -10, and +25% for these three cases. This • 538J represents a far more consistent behavior than is normally observed for o.ooot J L_ the steel specimens. 10" 10 -в 10- I0D1NE PRESSURE (bur) 4.2 Adsorption Isotherms for Grade H-451 Graphite Fig. 4. Iodine sorption on low-alloy steel at temperatures 500 to 800°C. The equilibrium sorption data for all the test runs thus far con­ 4. EQUILIBRIUM SORPTION OF IODINE ON GRADE H-451 GRAPHITE ducted on graphite are plotted in Fig. 6. We note that the uncertainties in the loading are quite large at 1000°C, due mainly to poor counting 4.1 Range of Test Data statistics at the low levels involved. The limit of resolution for the present test arrangement is ^10-5 Pg/g. Except for the 428°C isotherm, The range of test conditions for the five test runs performed on all of the equilibrium adsorption isotherms shown were acquired using data grade H-451 graphite are summarized in Table 3. Initially, in runs 3 and from three or four different graphite compacts. The consistency of the 4, the specimen temperature was held constant while the iodine partial data (except for a single point at 250°C) indicates that the equilibrium pressure was increased in a stepwise manner after equilibrium loading was loadings are not strongly affected by the individual sample characteristics, achieved. The procedure was altered for the subsequent runs in order to although our samples were all cut from the same block. minimize the number of changes required in setting the iodine saturator conditions. Although the performance of the iodine saturator was satis­ The equilibrium loadings shown in Fig. 6 are substantially lower than factory overall, it was occasionally erratic for unknown reasons, especially those previously observed by Osborne et al.1 for grade H-327 graphite. following changes in refrigerator temperatures. Therefore, runs 5 and 7 Direct comparisons are given in Table 4 for two temperatures, each at two were performed with minimal changes in iodine saturator conditions. Equilib­ iodine pressures. We note that presently measured iodine loadings in rium loadings were altered as much as possible by progressively lowering the graphite are 47 to 640 tines less than predicted by the previous study. specimen temperature at a constant iodine pressure. Table A. Coaparison of observed equilibrium) loadings In grade H-451 graphite vith

previous âtudy^ using pulverised grade H-327 graphite

Data of 0sbomea Pulverized H-327 H-451 Iodine for pulverized loading assualng loading. Tenperature pressure grade H-327 0.8-mVg surface this study Pulverlied H-3272 Ce) (bar) (atons/cn*) (ug/g) (Ug/g) H-451

600 10-' 10" 0.17 5 x 10-- 340 io-13 3.5 x 101D 0.053 9 x Ю-5 640

400 ю-' 4.2 x 1011 0.70 0.015 47 io-•" 1.8 x 101* 0.30 3.2 x 10-' 94

^^M. F. Osbome, E. L. Cospere, and H. J. de Nardwall, Studies of lodirx Adocrption and Desorption or. STCF. Coolant Circuit Materials, ОКЛ./1Я-5094 (April 1976).

Undoubtedly, the fact that the earlier study employed pulverized graphite specimens contributed to the higher observed loadings. In addition, the experimental technique differed; the previous tests were performed in a static system at high vacuum, whereas the present tests were conducted in flowing helium at a total pressure of I bar. We are planning at this time to repeat our measurements employing the helium flow-through technique with grade H-327 in order to clarify the reason for this significant dif­ ference in observed loading.

5. IODINE DESORPTION RATES

5.1 Rate of Desorption from Low-Alloy Steel

The experiments to measure the equilibrium loadings of iodine on metal specimens included desorption tests. In some desorption tests, the iodine partial pressures was lowered while the specimen was held at the adsorption temperature. In others, the temperature was raised while the iodine pressure was held constant to obtain equilibrium loadings at a higher temperature. In a few, the iodine burden was reduced, and the temperature was raised.

Data from some of the desorptions are compiled in Table 5. As expected, the desorption rate increased with increasing temperature. However, the increase with temperature was less than would be expected for a constant acti­ vation energy of desorption. We therefore conclude that the iodine was adsorbed on sites having a range of activation energies for desorption. The average activation energy for the iodine that was desorbed increased when a desorption was carried out by raising the temperature in steps. We esti­ mate that 1-80% of the iodine was present on sites having an activation energy for desorption in the range of 2.4 x I05 to 2.8 x I05 J/mol. These conclu­ sions are based on the following analysis of the desorption curves.

We found that when the logarithm of the amount of iodine on a specimen is plotted vs time, the curves could be resolved, with a good fit to the data, to between one and four component curves, each of the form Table 5. Desorption rates of Iodine from low-alloy steel (run 13) The fact that more than one component curve was required to obtain a good Initial Time for 50Î fit to the data from most of the desorptions suggests that the iodine was Initial conditions Final conditions desorptlon change In Temperature Iodine pressure Loading Temperature Iodine pressure rate loading present in populations having different activation energies for desorption. CO (10-14 bar) (Ug/cm5) Cc) (10-14 bar) (fractlon/10* s) (10s s) In such circumstances, the desorption can be described by the equation

400 2.8 0.39 500 3.0 0.27 1.3 500 3.0 0.63 600 3.0 5.8 0.11 -е'/ЯГ

CL) M T—. 2 _ n 600 3.0 8.7 x № 700 3.0 20.0 0.035 (2) ж = CP;* - v£ vn 1 700 3.0 5.9 X 10- 800 3.0 29.0 0.022 H= 1 600 1-25.0 4.6 X 10- 2 600 14.5 6.6 0.069 In Eq. (2), the rate of change of the loading per unit area, âj/dt, is 400 1.39 0 >1.4 x 10"ja 50.0 3.1 400 expressed as the difference between the adsorption and desorption rates. 400 0 1.34 600 1.2.5 8.2 0.097 For the sorption rate, as given by the first term on the right side of the equation, it is assumed that the iodine partial pressure, Pr, is uniforn Desorptlon rate was limited by gas flow rate. along the surface, and the value of the sticking coefficient, s, for the sites of each activation energy are equal. The total desorption rate is given as the sum from all .7 sites of desorption activation energies, lnfe) = *(é"V' (D ej,...e^, where 17^ represents the loading per unit area of activation where energy, e^j. The term V is the vibrational frequency perpendicular to the surface, which is usually given a value of 101*/s. m, = amount of iodine on specimen at times t and t^, yg;

t = time, s; and Equation (2) may be used to calculate the value of each desorption

-1 rate constant, к , for site of desorption activation energy, ЕД, where for к = desorption rate constant, s . each site, One of the desorption curves and its component curves are shown in Fig. 7. The steady-state component is represented by an equation in which к = 0. к = ve " . (3) n ORNL DWG BO-137I If an appropriate value for the Pj is ascertained for the desorption run, then the time variation of the loading during the desorption nay be used to numerically calculate each desorption rate constant, and consequently, each desorption activation energy, via Eq. (3). Figure 7 illustrates how this may be done for a desorption test in run 13-3, which was carried out to an equilibrium loading at 600°C. The total curve, u(i), is seen to be given as the sum of three lines, each of the form given by Eq. (1). The corresponding desorption rate constants, and k^ (the equilibrium value yields k^ = 0), are shown.

Because of uncertainties in the data and in the assumptions nade in the analysis, we assume the results indicate only that the iodine was adsorbed on sites having a range of activation energies of desorption. For the values assigned to s (taken as 0.1) and v, the range was

5.2 Desorption Rates from Graphite

As for the tests with steel, two types of desorption runs were per­ formed: (1) isothermal desorptions, wherein the desorption was achieved О 0.4 0.8 1.2 1.6 2.0 by reducing the iodine pressure to zero, and (2) temperature-induced TIME ПО5») desorption, obtained by stepwise elevations of sample temperatures at a Fig. 7. Desorption curve for experiment 13-3 at 600°C in helium having an iodine pressure of 3 x 10"10 bar. Table 6. Summary of isothermal desorption rate tests on graphite constant iodine pressure. Figure 5, illustrating the course of run 5, shows both types of desorption tests. Segment G pertains to an iso­ thermal desorption from an equilibrium loading at 250°C, while segments D' Initial Time to desorb and E' represent temperature-induced desorptions. Initial Helium fractional 501 of initial Run Tenperature loading flow rate release rate loading No. CO (Ug/g) (cn'/n) (105 s)-1 (10s s) Table 6 presents a summary of the isothermal desorption test results. It may be seen that the initial desorption rate depended primarily on the 3 428 0.41 19 0.84 1.7

graphite temperature and also on the value *bf the initial loading and the 3 429 0.39 15 0.53 2.9 helium flow rate. The highest desorption rate observed occurred at 800°C, 5 250 0.21 71 0.43 where an initial fractional rate of 0.45%/s was measured. Table 7 lists 1 the results of the temperature-induced desorptions performed during run 7 5 600 0.020 52 127.0 4.6 x 10- with the iodine partial pressures reduced to zero. The initial observed 6 400 0.015 68 5.7 0.55 rates are comparable to the values obtained at the corresponding tempera­ 6 250 0.080 73 0.31 •^9.6

tures by the isothermal method. These desorptions were performed with 7 800 7.5 x 10-3 71 450.0 1.5 x 10"' the helium flow being forced through the graphite specimen, a standard rate being 'WO cm3/min at STP. Therefore, purely thermally induced desorp­ Table 7. Iodine desorption fron graphite tion rates, as for example due to accidental core heatup with no pressure by loss, would be substantially less. Hence, tests approximate more closely increasing temperature (run 7) conditions during graphite outgassing following a depressurization event. In any case, iodine desorption from a large graphite mass is a complex Initial fractional process involving (1) the intrinsic rate of desorption of the surface, Temperatures (eC)a Initial loading release rate (2) the rate of iodine convection caused by bulk helium flow through the Initial Final (ug/g) (10s s)"1 pores, and finally (3) diffusion from blind pores and small capillaries which are not significantly reached by the helium flow. 250 250* 0.48 0.34 250 300 0.47 3.9

300 400 0.44 6. CONCLUSIONS 16.0 400 500 0.33 64.0 Tests were performed to determine the equilibrium loadings of iodine 500 600 0.13 170.0 on a low-alloy steel and grade H-451 graphite in the temperature range 600 700 6.2 x 10-' 260.0 250 to 1000°C and iodine partial pressure range of VIO-7 to 10-11 bar. In addition, rates of iodine desorption were determined under a variety temperatures held for 50 min at each level. of conditions of two different types, namely, under isothermal conditions ^Iodine pressure reduced to zero at start of test with a drop in iodine partial pressure and desorption due to step increases in specimen temperature. All tests were performed at a total pressure of The measurements of equilibrium loadings on grade H-451 graphite 1 bar in a flowing helium system. were far more consistent and reproducible than those found for steel. As seen in Fig. 6, the logarithm of the loading varied linearly with the The tests on low-alloy steel were characterized by a large scatter logarithm of the iodine partial pressure throughout the experimental in the measured equilibrium loadings, which we feel was due primarily to range of temperatures and partial pressures. Where direct comparison variations in the condition of the surface. For this reason, wc examined is possible, these iodine loadings fall significantly below previous 1 one isotherm (400°C) most extensively in order to get a complete picture determinations by Osborne. This was most likely due to the fact that for at least one case. The tests at 400"C plus comparisons with earlier a crushed graphite sample was employed in the earlier study in which reports combine to present the following conclusions regarding the 400°C an abnormally high concentration of active sites would be expected. isotherm: (1) at iodine partial pressures above 10~6 to 10-7 bar, load­ The extent to which the graphite effected this difference (grade H-327 in the earlier tests) will be determined in future tests. ings increase with iodine pressures due to formation of Fel2; (2) between iodine pressures of 10-7 to 10-10 bar, we find a scatter in the data of one to two orders of magnitude. The lower bound was obtained on specimens Iodine desorption rates were determined for both the steel and that were clearly corroded, while the upper bound was constant loading of graphite specimens in two types of tests: (1) desorption due to reduc­ •v5 jjg I/cm2, which we feel is representative of monolayer coverage on the tion in iodine partial pressure at constant temperature and (2) desorp­ unoxidized steel surface; and (3) at iodine pressures less than MO"10 bar, tion due to step elevations in temperature. These data are summarized we begin to see a reduction in loading as the iodine pressure is decreased. in Tables 4 to 6. The latter method is useful in determining the range As yet, the data are very sparse in this regime, but the loadings below of available chemisorption sites. For steel, the estimated range of ^lO-10 bar (and down to ^lO"11 bar) vary approximately directly with activation energies of desorption was 230 to 330 kJ/mol, with *^80% of iodine pressure. the sites in the 240- to 280-kJ/mol range. REFERENCES REMARKS ON POSSIBILITIES AND LIMITATIONS 1. M. F. Osborne, E. L. Compere, and H. J. de Nordwall, Studies of OP THEORETICAL APPROACH TO PLATE-OUT PROBLEMS Iodine Adsorption and Desorption on HTGR Coolant Circuit Materials, ORNL/TM-5094 (April 1976). E. OBRYK 2. C. E. Milstead et al., "Deposition of Iodine on Low Chromium Alloy Institute of Nuclear Physics Steel," Nual. Appl. Teahnol. 7, 361-66 (October 1966). Krakow

3. F. H. Neill, Adsorption and Desorption of Iodine on Mild Steel, Poland ORNL/TM-2763 (April 1970). Introduction A. R. D. Burnette, D. R. Lofing, and C. L. Allen, Public Service Company of Colorado ЗЗО-Ш(е) High-Temperature Gas-Cooled Reactor Research and Development Program Quarterly Progress Report for the Tlio knowledge of the interaction of the fission /and activa­ Period Ending September 30, 1967, USAEC Report GA-8270, p. 143 (Oct. 30, 1962). tion/ products with primary curcuit materials of HTR is of essen­

5. L. R. Zumwalt, D. D. Busch, and B. D. Snow, Experimental Beryllium- tial importance to normal operational condit±ons as well as to Oxide Reactor Program Quarterly Progress Report for the Period Ending March 31, 1962, USAEC Report GA-2053 (Apr. 30, 1962). assessment of nnviromental risk connoctod with severe accidents.

6. E. L. Compere, M. F. Osborne, and H. J. de Nordwall, Iodine Behavior The Il'PR fuel development /coated particles in graphite matrix/ in an HTGR, ORNL/TM-4744 (1975). assure several effective barriers against fission product migra­

tion. Nevertheless through these barriers or due to their damage

a trace amount of the fission products reach the helium coolant.

The most important of thorn from point of view of the plate-out

phenomona are: Cs-137, Cs-IJk, Ae-110m, Aç-111, 1-131, Sr-89

and Sr-90. Transported by helium coolant as a free atoms /mole­

cules/ or adsorbod on graphita dust surfaces tlioy can reach surfa­

ce of primary curcuit components. Graphite dust has a profound

influence on fission products behaviour in the primary curcuit

/" 1 _7 but tliis problem will not he consider here.

During the plate-out process at high temperature besides the

adsorption and desorption important part start to play such pheno­

mena as solubility and diffusion into balk of the materials /~2_7. Understanding of the procosscs involve in a transport from Reliability of predictions of any model depends on general a gas to a surface phase and to a bulle solide is crucial to pre­ knowledge of the procosscs involved clarity of physical meaning dict the plate-out results. of chosen parameters. Simple corclation relations shall not be

undervaluod/somo of their, are extr4«id.y poworful and useful e.g. Inspito of the groat progress which lias been madо in sur- Arrhcnius relation/ but one should remember that quite often faco soience during last two decades /mainly due to new power­ such relations have a very narrow range of application. ful experimental methods/ this field is still in a situation

reminding that of the physics of bulk condensed matter at the Surface and undersurfacc layer. beginning of this century.

Unfortunally there are not theory of inelastic and roactive Surface /or more general intorfaco/ uso to be consider as

processes at surfaces evon for well define monocrystal metal a boundary condition for solid, liquid or gas piiaso but this

surfaces. The only way to deal with this problem is to deve­ is a very inadequate approach. Surface, duo to its role and

lop models using phenoraenological approach but even in this properties /often quite different from a bulk of matter/ has

case it is necessary to find methods for reducing the number to bo consider as new quasi-two-dimensional phase of matter.

of parameters i;o make the modol tractable without sacrificing Ilogion undezTieath of a surface has a disturbed structure

essential physics and chemistry. Difficult task is to over­ due to surface phenomena. This region plays an important role

come problems associated with the tremendous diversity in tiine- in many procosses e.g. in mass transport from a surface to

—scales exhibited by platc-out phenomena, an inital trapping a bulk of matter. This problem is still not well understood

event is occuring in less than a picosecond followed by proce­ and most of transport models dont take this transitional re­

sses /o.g. dissociative adsorption/ of microsecond time-scale gion into account. It is probably good approximation for a tran­

and at the end diffusion time scale is of tho order of days or sport which is not far from to be stationary, in other case it

years. In such situation the competition between different may cause a problem.

mechanisms involved in plate-out phonomena may be very sensi­ Most of our knowledge of surface phenomena to date has boen

tive on some parameters /sometimes oven difficult to control/. obtain for a narrow class of motals /mainly tungsten and other

It is worth-while to have this in mind making sensitivity study refractory metals/ in a from of monocrystal iifith a very speci­

for any model. fic surfaces/ e.g. LEED need flat, smoth sample and FU-l - fine needle/ and in an ultrahigh vacuum. It is clear that measur­ diffusion is large so even Tor very short lifortime of adsor­ ements of this type cannot be made directly on surfaces opera­ bed atom it has a chance to find appropriate sido for chomi- ting under the conditions used in practice. Dut on the other cal adsorption on tho surface. In this wya the influence of hand working with "idealized" systems in which member of varia­ point defects on desorption rate at high temperature may be bles has been drasticly reduced, one has a chance to make pro­ diminish too /5,7 J. gress in understanding the complex interactions taking place In the process of the diffusion of adsorbed atoms into in the surface region. a metal, channels of fast propagation may be of importance.

Progress in understanding the surface enhanced Raman efect Characteristic feature of grain boundary diffusion is that can offer a new powerful method of surface study under normal it have a wide spectrum of transport channels with very diffe­ operational conditions. rent activation energy. It seems that channels with higher

At the surface of an alloy segregation of -components may activation energy are "faster". This leads to roughly a com­ occur giving a vary different chemical composition at the mon variation of grain boundary diffusion coefficients as surface than in the bulk of a material. This segregation a function of the ratio of temperature to the fusion tempe­ depends on temperature, composition of the operating atmos­ ra ture £8j. phere and may change during the operation time. Uelium impurities Normal technological surface on a raieroscale consists area of a very different surfaces. This structure is not Chemical impurities in the ooolant helium of HTR, even only difficult to control but even to qualify. On top of very minute, nay be very important for plate-out results. that in industrial practice one never is dealing with a puro Simplest to estimate and probably the less impo'rtant are motal surface. the chemical reactions in the gas phase between fission

Different area on a suoh surface may have completly diffo­ products and impurities. Far more important are tho influ­ rent properties /e.g. sticking probability, accommodation ences of the chemical impuritiss on properties of surfaoe coefficients, coefficient of surface diffusion/ £ 3,^,5 J. and transitional region.

But for higher temperature and low fractional coverage ove­ Impurities affoct struotures and chomical compositions rall properties of technological surfaces may effectively of surface layers and often grain boundaries. For HTR-he- less differ than for low temperature. Tills may happen main- lium impurities a very important characteristic is the ra­ ly due to the"that in such conditions coefficient of surface tio of oxidation to carbonization potentials /still it is E7 not quite clear what value Tor this ratio one should ex­ References pect in HTR PH plants/. £ 1 J G.Ivons and R.Gilli, The First Year of Operation at

To sum up: 950°C in the AVR Power Station in Gas-Cooled Reactors

with Emphasis on Advanced Systems, Vol.1 1./ The only possible description of plate-out prosesses /Proceedings of a Symposium, Jülich, October is to develop pheno'menological models. 13-17 1975/, XAEA, Vienna 1976. 2./ Parameters used in such models should have clear physi­

cal meaning. This enable to measure them in independent f 2 J N.lniotakis et al., Fission Product Transport in High-

experiments /which may ba carry out with strongly rodu- -Teraperature Reactors in Gas-Cooled Reactors with Emphasis

ced number of variables/ and to estimate the range of on Advanced Systems, Vol. I /Proceedings of a Symposium,

application of the model /and define meaningful range Jülich, 13-17 October 1975/, IAEA, Vienna, 1976.

of variation of the parameters e.g. for sensitivity Г 3 J N.V.Roberts and C.S.McKee, Chemistry of the Metal-Gas

study/. Interface, Clarendon Press, Oxford, 1978.

3»/ Reliable predictions of the plate-out must base on firm £kj F.O.Goodman and H.Y.Uachman, Dynamics of Gas-Surface

data concerning surfaces and their changes expected du­ Scattering, Academic Press, New York, 1976.

ring the whole operation time of components /induce by £*5_7 D.Menzel, Desoption Phenomena in Interactions on Metal

chemical envionment, ageing, decontaminations, creep, Surfaces, /R.Goner, ed./, Springer-Verlag, Berlin, 1975»

fatigne and shocks/. Г в J K.2dansky and Z.Sroubek, J.Phys. F,6_ /1976/ L 205,

£l J S.Yu.Davydov, Fizika Tverdogo Tela, J9. /1977/ 1<+1S.

/"87 J.C.M.Hwang and R.¥.Balluffi, Scripta Metallurgica,

12 /1973/ 709. 1. Description of the reference plant

FISSION PRODUCT BEHAVIOUR IN THE PRIMARY CIRCUIT OF AN HTR The PNP-500 plant is planed as a demonstration plant for nuclear process

C.-B. von der DECKENi N. INIOTAKIS heat application. Kernforschungsanlage Jülich GmbH.

Jülich The main characteristics of this plant are: Federal Republic of Germany

pebble bed core with OTTO-fuel management; Introduction о о He-heating temperature intervall 290 С - 950 С; з The knowledge of fission product behaviour in the primary circuit of a mean power density 4. 0 MW/m ;

High Temperature Reactor (HTR) is an essential requirement for the integrated systera in a prestressed concrete vessel; estimations of the availability of the reactor plant in normal operation, three main cooling loops. One for hydro-coal-gasification(HKV), of the hazards to personnel during inspection and repair and of the two for coal gasification with water vapour (WKV-loop); potential danger to the environment from severe accidents. He-mass flow: HKV-loop 73.8 kg/sec and

On the basis of the theoretical and experimental results obtained at the WKV-loop 36. 9 kg/sec each;

"Institute for Reactor Components" of the KFA Jülich /1/, /2/ the transport- He-pressure: 40 bar. and deposition behaviour of the fission- and activation products in the In Fig. 1 the primary circuit of the reference plant is shown schematical­ primary circuit of the PNP-500 reference plant has been investigated ly. The arrows are giving the direction of flow during normal operation. thoroughly. The He leaves the core (1) at a temperature of 950 °C passes through the bottom reflector (2) to the HKV- and WKV-lcops. In the HKV-loop the He Specially work had been done to quantify the uncertainties of the investi­ flows through the hot pipe (3), the reversed flow section (4) to the steam gations and to calculate or estimate the dose rate level at different reformer at a temperature of ca. 700 °C, through a gas pipe (6) to the components of the primary cooling circuit. The contamination and the steam superheater (7), the evaporizer (3) and the economizer (9). From dose rate level in the inspection gap in the reactor pressure vessel is here the He flows at 290 °C through a gas pipe (10) to the blower (11) and discussed in detail. through another gas pipe (12) back to the outside of the economizer. The

For these investigations in particular the surface structure and the He is passing the outside of the steam generator (13), a gas pipe (14), the composition of the material, the chemical state of the fission products outside of the steam reformer (15) and the cold gas pipe (16) to the armulus in the cooling gas, the composition of the cooling gas and the influence space. Here the gas is mixed with the gas from the other loops. 30 Ta of of dust on the transport- and deposition behaviour of the fission products the gas flows through the inspection gap and 70 through the cooling gap have been taken into account. The investigations have been limited to the between the side reflector and the thermal shield. Finally the gas is led nuclides Ag-110 m; Cs-134 and Cs-137. back to the core through the top reflector. In the \VKV-loops the He flows 3. Basic assumptions through the hot pipe (3) and the reversed flow section to the Це-Ые-heat

exchanger (5), the outside of the heat exchanger (6), a gas pipe (7) and to The investigations have been carried out on the basis of the physical model the blower (8). Through a cold gas pipe (9) the gas is passing also to the developed by Iniotakis and the computer programmes "PATRAS" and annulus space. "PATRAS-S". The model gives a phenomenological description оГ the

The following materials are foreseen for the different components: transport and plate-out of the fission- and activation products /4/, /5/, /6/.

carbon stone for the hot pipe insulation; The transport and plate-out behaviour of the fission products depends very

Incoloy 802 or In-617 for the steam reformer and much on the flow and temperature conditions in the primary circuit and the

the He-He-heat exchanger repectively; manner of their interaction with the component materials. Specially from:

Incoloy 800 for the steam generator; the chemical state of the fission products; for the low temperature components ferritic material will the composition of the cooling gas; be used, for example 15 Mo 3; the structure and composition of the surface of the

materials and the materials themselves; the material for the blower has not been selected yet. the growing velocity of oxide layers;

the possible evaporation of the materials of high temperatures. 2. Fission product release from the core

For the fission product release from the core the reference values for In addition the presence of dust in the primary circuit plays an important

BISO-particles, highly enriched have been selected. They are / 3 / role.

Cs - 137: R = 150 Ci/a The parameter which have to be used for the calculations can be classified Cs - 134: É = 120 Ci/a as follows: Ag - 110 m: R = 500 Ci/a 1. Operation parameters

2. Fission product specific parameters In general, but especially in case of rëil 'r.ïng atmosphere over the operation a) for the transport in the gas phase time of 30 years condensations effects cannot be excluded for Cs and Ag. b) for the interaction with the surfaces This means one has also to know the total partial pressure for Cs and Ag. c) for the transport (diffusion) into the material Based on estimations for the total partial pressures for Cs three times d) for the interaction with the dust the partial pressure of Cs-137 and for Ag, 100 times He partial pressure 3. Dust specific parameters of Ag-110 m were used for the calculations. Ag-107 gives the main a) for the transport in the gas phase contribution. b) for the interaction with the surfaces , „ 22 Is 000 Based on the results of our own experiments, using the reported results 1.6 • 10 • e 000 1 RT from literature and specially the specific literature on dust, we did find 22 15 000 our model and the computer programmes justified. Furthermore we could = 8 . 10 2 RT evaluate most of the parameters with a sufficient accuracy.

p? 7 000 8 • 10 . e " 0co3 RT For some of the parameters which are important for the calculation it is not possible today and may be also in the future to specify them sufficient­ Concerning the grain size of the dust the partical radius ly for a plant in praxis, that is the structure and composition of the surface had been varied between r =0.01-10 лип. Also mono layers of the materials and the grain size distribution and amount of dust. P ' dispersity had been assumed. In addition the solubility 0œ(the saturation concentration of the fission products in the bulk of the material in /Jatoms/cm3_7) could not be evaluated The amount of dust had been estimated to be 900 kg in with a sufficient accuracy up to now. 30 years of operation, taking the mean value of 30 kg/a.

Both cases with and without dust had been envisaged. To quantify the uncertainties of the calculations resulting from the lack of knowledge of these data sensitivity analysis had been carried out, where Concerning the composition of the cooling gas the two in some cases the variation of the parameters had been extended even to cases non-oxidizing and oxidizing atmosphere had been values outside the expected areas. taken. In the case of the oxidzing atmosphere a pre-

oxidation of the component surfaces had been assumed. For the variation limits the following assumption had been made:

The roughness coefficient (the ratio of the real to the

geometrical surface area) had been varied between 4. Results of the sensitivity analysis

К = 10 - 1 000. For realistic industrial products one

should expect К = 50 - 100 under non-oxidizing atmos­ All the reported results are from calculations of the HKV-loop.

pheric conditions and К = 100 - 1 000 for oxidizing In Fig. 2 the influence of the roughness coefficient is shown. An operation conditions. time of 30 years, an oxidizing cooling gas and no dust had been assumed. The adhesional distance у , which corresponds referring to /4/, 2

/5/,/6/ to the height of the micro roughness had been On the ordinate the Cs-activity in juCi/cm is given. On the abszissa the о position along the HKV-loop is marked starting with the hot gas pipe (3)

= varied between У0 50 - 1 000 A. Here mono dispersity and ending with the annulus space (17). The numbers referring to the was assumed in spite of the fact that in reality there is description of the circuit in chapter 1. In addition the surface temperature a statistical distribution. along the loop is given in the diagram. For the solubility the following values had been used. One can see from Fig. 2 that the influence of the roughness coefficient is diffusion processes in prevention of local saturations. This is also-the fact decreasing with increasing temperature and increasing value of the rough­ at high temperatures in the case of oxidizing atmosphere. ness coefficient. At high values of temperature and roughness the deposition In the case of Ag-110 m one can read from Fig. 5 that the influence of the of Cs tends to be independent from these parameters. atmospheric conditions is most important for the economizer.

This result is plausible for Cs. At small values of roughness and low In Fig. 6 the influence of the dust is shown. Curve 1 réfères to the dust temperatures the degree of coverage tends to be high. At a given limit the free system. Curve 2 gives the deposition of Cs-137 on the surfaces cohesional forces cannot be neglected compared to the adhesional forces, assuming there is dust in the system. Curve 3 represents the amount of condensation effects occur and influencing the desorption kinetic. Cs-137 which is sticking on the dust particles which are themselves be о For Ag the roughness coefficient does not influence the deposition deposited on the surfaces. In this case an adhesional distance у = 300 A because the cohesion energy of Ag exceeds the adhesion energy with the and a particle radius of r^ = 1 yum had been chosen. surface material. This result is not shown in the Fig. It may be only One can see in Fig. 6 that there is a considerable influence of the dust on mentioned here. the plate-out.

In Fig. 3 the influence of the solubility on the plate-out behaviour of Cs-137 The shown diagrams. Fig. 2 - Fig. 6, are examples demonstrating is demonstrated. Again an operation time of 30 years, oxidizing gas atmos­ typical results from the extended investigations. Within the scope of this phere and no dust is assumed. As one can read from these diagram the in­ paper it is not possible to present all the results from this work in detail. fluence of the solubility is decreasing with increasing temperature and in­ But some additional general results should be mentioned in qualitative creasing values of solubility. Again this result seems to be plausible form: because the increase of the diffusion constant and solubility with increas­ The dust has the strongest influence on the plate-out ing temperature prevents local saturation can occur. behaviour compared with the other parameters.

In Fig. 4 and 5 the influences of the composition of the cooling gas on the In general the presence of dust in the system deposition behaviour of Cs-137 and Ag-110 m is shown. 30 years operation reduces the influences of the other parameters. time and no dust had been assumed. The reason for this is the considerable reduction

The influence of the gas atmosphere decreases with temperature. This can of the concentration of free fission product atoms in be understood by the following considerations. The oxidized layers on the the gas phase. Condensation effects in the cold part surface of the material which exist in oxidizing atmosphere serve as a of the circuit are reduced or eliminated. diffusion barrier for Cs. The diffusion constant of Cs in oxide is several orders of magnitude smaller compare to that in the metal. This is result­ The influence of dust can be neglected if the particle ing in the case of more oxidizing atmosphere in spite of the faster radius exceeds r^ > 5 yum. Neglecting the dust leads in general to pessimistic Reference case 3

results with respect to contamination of the inspection As case 1, but non-oxidizing atmosphere. gap and the blowers.

Reference case 4 In the WKV-loops the influence of the roughness, the

solubility and the composition of the gas is very small As case 23 but non-oxidizing atmosphere. or negligible depending on the special design of

the heat exchanger. In Fig. 7, Tab. 1, the influence of the solubility, roughness for the different

cases is given. One can read from this table that within the expected area

5. Estimation of the maximum level of contamination for the roughness coefficient K= 50 - 100 for the cases 3 and 4 and

К = 1 00 - 1 000 for the cases 1 and 2 all values are within a factor of two.

To estimate the maximum level of contamination four different cases, The expected value for the solubility for Cs is about equal to 0. reference cases, have been calculated. The contamination and dose rate In Fig. 3, Tab. 2, the expected contamination of the inspection gap is given. in the inspection gap is used as an example. The reference cases have The contribution of Cs-137, Cs-134 and Ag-110 m for the different cases been chosen in a manner that the influence of the different parameters and are shown. The expected values are based on the assumption tha the value the influence of the materials are shown on the one side, on the other side for the solubility of Ag is equal to ф ^. For the roughness coefficient the maximum level of contamination can be estimated. To be in the К = 50 for non-oxidizing atmosphere and К = 100 for oxidizing atmosphere pessimistic side the influence of the dust is neglected in these calculations. was assumed. As one can see the main contribution for the contamination

The reference cases are defined: originates from the Ag.

Reference case 1 In Fig. 8, Tab. 3. the corresponding dose rates are given which are

calculated for the center of the inspection gas. Again the Ag is dominating. The material of the steam reformer, the steam generator and the heat exchanger is Incoloy-800; the other components are made out of 15 Mo 3 These values are pessimistic because the influence of the dust has been with the exception of the blower. There is oxidizing gas atmosphere and neglected. the surfaces of the components are preoxidized. Summary

Reference case 2 The accuracy calculating the distribution of fission products which are

deposited on the surfaces in a primary cooling gas circuit of a high As case 1, but In-617 as the material for the steam reformer, the steam temperature reactor depends on the knowledge of the value of a large generator and the heat exchanger. number of parameters. Most of the parameters estimate by a large number of experiments in the meantime. For some of the parameters /6/ C.-B. von der Decken, N. Iniotakis, F.P.O. Ashworth 73

the knowledge is poor (solubility) or they cannot be defined well enough A Model for the Description of Fission Product Behaviour in the Primary Circuit of a High Temperature Reactor for a practical plant (roughness, composition of the surface, grain size Conference on Gas Chemistry in Nuclear Reactors and and amount of dust). To get a better knowledge of the uncertainties Large Industrial Plant, resulting from the uncertainties of these parameters, sensitivity stuides Salford, UK, April 1980, Preprints have been made and demonstrated in some examples within this paper.

Furthermore the maximum value of the contamination and dose rate level in the inspection gap of a PNP-500 Prototype plant has been estimated. И-1 Соге-ггдют

References

/1/ N. Iniotakis, J. Malinowski, K.H. Münchow

Initial results of investigations into fission products deposition in in-pile experiments Gas Circulctar

L Gcs Drtulctcr Nuclear Engineering and Design 34 (1975), 169-180

/2/ N. Iniotakis, J. Malinowski, H. Gottaut, K.H. Münchow Flow distribution in primary coolant IЙЗ 60 circuit during normal operation 11/01 Results from plate-out investigations Fig. 1

IAEA-meeting, Jülich, December 2.-4., 1980 reactor operation 30 jeers /3/ K. Röllig oitfaing rjrioory 52s в..)Л- private communication

/4/ N. Iniotakis

Dissertation (not published)

/5/ N. Iniotakis, C.-B. von der Decken

The influence of dust on the behaviour of fission products in high temperature reactors

ENS/ANS Topical Meeting on Nuclear Power Reactor Safety, Brüssel, 16.-19.10.1978 Proceedings, Vol. 3, 2336 - 2347 mtluence of roughness factor x on deposition of Cs-37 183 73 1PNP-500. HKV-Locp.malenal of 5S and SG Inroîcy S02) В2Г.ГЛ. Fig. 2 rractor operation 30 years _ particle radius rp • 1 ym î| «itnai re5c.nl to dust

» •..«•»"»-so. уо«300д 21 m rpgom to dJSt

• B-O0,--™- 1°С r<>C 1033 3le. .i.«"i-Jf- 1X0 --' -n Г 900

• BOO - 600 '600 130

•TOO I I г |l 200 ZOO

0 10 20 30 40 50 60 70 90 90 И ПО "0 10 20 30 40 50 И 70 80 90 100 110 Position m [ml № Position n Im] *•

influence ol solubility Ф„ on deposition of Cs-137 IBB 80 inlluence of dusl on deposition ol Cs-Ш (PNP-500. IRB 60 (PNP-5uO.HKV-l.oop, cose ol relerence 1) HKV-Loop, operation time 30 years) 11/12 C9 Fig. 3 Fig. 6

reoclor operation 30 years

...!•«• ,-Jif TABLE 1: INFLUENCE OF SOLUBILITY ROUGHNESS FACTC« К AND плинии, он CS-137-«NT AKIN АЛО* OF THE

INSPECTION GAP.* CONTRIBUTION OF THE H7K-L00.P IN Cl AFTER 30 YEAAS. 1) поп oiidiung (х - 50 )

21 enduing (x>100 I CASE 1 CASE 3 CASE 2 Cue 1 1000 к Й„ ! B_ ! «- 2 »- 3 «- 2 ». 3 «- 2 3- 3

800 lo 1.33 1.00 0.13 0.77 0.63 0.39 1.» o.5a 0.31 0.16

•600 5o 0.63 0.33 0.27 0.33 0.29 0.26 0.59 0.27 0.11 0.33

loo 0.19 о.Зо 0.21 0.27 0.25 0.21 0.16 0.31 •all temperature •WO 0.21 о.Зо

looo 0.32 0.23 0.21 0.22 0.21 0.21 0.31 0.21 0.27 o.U TOO

u_ n < IT ft 7 1 « о » о 1—i—I h£AÄLV EXPECTED: • loo - looo AJOUT AND 0„2 0 10 20 - 30 ТТГТ7Т40 50 Г60 70 BO 90 KM 110 HINTS: RANGE К REFERENCE CASES: Position |m] • CASE 1 АЮ 2 FOR К - loo AND 0œ ^ influence ol composition ol primary gas on deposition CASE 5 AMD ^ FOR К - 50 АШ 0^ 2 ol Cs-137 (PNP-500.HKV-loop. cose ol reference I and 3 ) Fig. 7 Fig. 4

TAILE 2: CONTAMINATION OF THE INSPECTION UAP AFTER 3a TEARS IN CI

Ost Cs-Ш Cs-1» Aî-llon

1 0.11 o.oS 0.79 г 0.62 о.оЗ 0.79 S 0.37 o.o5 0.57 С 0.52 о.оЗ 0.57

3: DOSE WIE I* THÉ INSPECTION CAP AFTER ÎO TEARS IN PWEn/RT

CASE Cs-157 Cs-131 Дс-Им TOTAL 0 » 20 30 10 50 to 70 BO 90 100 It 1 6 3 17 56 Posilion in |ml • 2 9 5 17 SI inlluence ol composition of pnmory gas on deposition IBS 79 3 5 J W « 1 ol Ag-I10m (PNP-500. HKV-Loopl 7 5 « 46 Fig. 8 02/1/15 Fig. 5 2. REGULATORY STANDARDS

DERIVATION OP CRITERIA FOR PRIMARY Circulating and plateout activity criteria in the HTGR-GT primary cir­ cuit are derived to conform with regulatory standards for radiation protec­ CIRCUIT ACTIVITY IN AN HTGR tion of occupational workers and the general public. Table 1 summarizes the applicable regulations. S.D. SU, A.W. В AR SELL General Atomic Company 3. CIRCULATING ACTIVITY CRITERIA San Diego, California The circulating activity criteria were derived from the following USA considerations: (1) requirements for personnel access to the containment building during normal operation, (2) containment accessibility following a ABSTRACT depressurization accident, and (3) offsite doses to the general public dur­ ing normal operation and depressurization accidents. The bases, methodol­ ogy, and procedure used in deriving the circulating activity criteria are This paper derives specific criteria for the circulating and plateout described below. activity in the primary cfrcuit for a 2170-MW(t) high temperature gas-cooled reactor-gas turbine (HTGR-GT) plant. Results show that for a design basis, 3.1. Assumptions and Bases (1) the circulating activity should be limited to 14,000 Ci Kr-88 (a prin­ cipal nuclide) to meet both offsite dose and containment access constraint The assumptions and bases used to derive the circulating activity during normal operation and depressurization accidents, and (2) the plateout criteria are summarized below: inventories for those important nuclides affecting shutdown maintenance should not exceed 10,000 Ci Ag-llOm, 45,000 Ci Cs-134 and 130,000 Ci Cs-137. 1. Fower level: 102Z of 2170 HW(t) - 2213 MW(t) This paper presents bases and methodology for deriving such criteria and 2. PCRV leak rate: 0.01Z/day (or 3.65Z'/yr) compares them with light water reactors. 3. PCRV decontamination factor for iodines and particulates: 100 4. Open containment purge rate: 0.5 vol/h. 1. INTRODUCTION 5. Closed containment normal leak rate: O.lZ/day plus semi-annual purge for refueling and maintenance The radioactivity level in the primary circuit directly impacts the 6. Free containment volume: 8.0 x 10^ m"' design, operation, maintenance, and safety analysis of nuclear power plants. 7. Principal nuclide: Kr-88 For a high temperature gas-cooled reactor (HTGR), the primary-circuit activ­ 8. Nuclide mix in circulating activity: see Table 2 ity results from fission product release from the core and subsequent trans­ 9. Reference site boundaries and meteorology: For a hypothetical port and plateout in the circuit. In contrast with light water reactor site less favorable than 90Z of existing LWR sites in the U.S.A. (LWR), the HTGR primary circuit is virtually free of corrosion products. The criteria were specified for the design activity. The expected Because the primary-circuit activity has such significant impact, cri­ activity provides the ALARA margin for normal operation. The difference teria must be developed, which specify the allowable activity levels. Spec­ between design and expected activity is a factor of 4 for gases. ifying the activity criteria is an important initial step in plant design and optimization. This paper presents the bases and methodology for deriv­ ing such activity criteria for a 2170-MW(t) HTGR-gas turbine (HTGR-GT) 3.2. Methodology and Analysis plant. The containment access requirements and offsite dose limits were con­ The HTGR-GT activity criteria correspond to exposures permitted by reg­ sidered separately to establish the circulating activity criteria, as ulatory standards and meet ALARA (as low as reasonably achievable) margins. discussed below. These criteria include circulating activity in the primary coolant and plateout activity on the helium-wetted surfaces. The circulating activity 3.2.1. Containment Access. Containment access criteria consider both criteria were derived mainly from containment access requirements and off- normal operation and depressurization accidents. The activity criteria were site dose limits during normal operation and depressurization accidents. first derived for normal containment access, then applied to the accident The criteria for the plateout activity were determined from shutdown condition to assess the post-accident containment accessibility. maintenance considerations.

This paper compares the activity levels between the HTGR and the LWR. All. tables and one figure are grouped at the end of this paper. The containment access evaluation calculated the gasborne activity Figure 1 depicts the calculated annual offsite whole body dose to the concentration in the containment atmosphere due to primary coolant leakage public as a function of the amount of Kr-88 in the primary circuit for both through the prestressed concrete reactor vessel (PCRV) and determined the open and closed containment. The maximum allowable circulating activity for corresponding dose rate. The activity calculation «as performed only for an open containment can he inferred to correspond to 14,000 Ci of Kr-88 to noble gases and their important daughters, since the PCRV decontamination meet annual offisite dose limits. Offsite .loses for a closed containment would reduce the concentrations of iodines and particulates in the contain­ are exceedingly low. ment to negligible levels. The gasborne activity was calculated for each important nuclide at steady state. Accidents involving release of primary circuit activity are primarily those in which the primary coolant is depressurized or blown down to the In calculating the gaseous dose rate in the containment, the realistic containment due to penetration closure failure or opening of the PCRV safety approach determined the direct dose rate from noble gases for a confined relief valve (inadvertent or due to steam Ingress). containment volume using the point-kernel integration computer code PATH (Ref. 1). This approach appears reasonable, since inhalation of noble gases Upon helium blowdown to the containment, safety-related closure valves is insignificant to internal exposures. For the metallic daughter products, isolate the containment with high Q99-99Z) reliability. Thus, direct re­ the conservative maximum permissible concentration (MPC) approach was used lease through a nonisolated containment is not a design basis event. How­ to account for internal and external exposures. For conversion purposes, ever, analysis of the consequences of this Class 9 event is of interest to one MPC for a given nuclide is regarded equivalent to 2.5 rarem/h. demonstrate the inherent safety margin in the primary circuit activity cri­ teria. Fifteen minutes of decay were credited to account for the helium Based upon the nuclide mix in Table 2 for the circulating activity in blowdown time, release from a nonisolated containment, and transport to the primary helium, the Kr-88/Rb-88 chain was found to contribute about 902 of site boundary. For an HTGR gasborne activity mix where the Kr-88 level is the total gaseous dose rate in the containment, regardless of the contain­ 14,000 Ci, the amount of equivalent* Kr-88 at the site boundary after 15 min ment purge rate. With a dose rate of 2.5 mrem/h corresponding to a 40 h/wk decay would be 20,000 Ci. The resulting 2-h site boundary whole body dose access, Fig. 1 shows the allowable access time in open and closed contain­ would be 20 rem, just within 10CFR100 limits. ment as a function of Kr-88 inventory in the primary coolant. To allow a 40 h/wk containment access during normal operation, Kr-88 activity in primary The possible need to evacuate the public in such a Class 9 event was coolant should not exceed 14,000 Ci for open containment. This criterion also considered. In accordance with KUREG-0610, the criterion for a general will limit the allowable access time in closed containment to 11 h/wk. emergency requiring evacuation is atmospheric release of more than 10-* CI of 1-131 or 106 of Xe-133 equivalent (LWR specific). Considering relative bio­ Using a circulating Kr-88 inventory of 14,000 CI and the nuclide mix in logical hazards, the 106 Ci of Xe-133 is equivalent to 2.2 x 104 CI of Kr- Table 2, the containment gamma dose rates were estimated as a function of 88. The above 20,000 CI of equivalent Kr-88 would not constitute a site elapsed time following depressurization of the primary coolant. This esti­ emergency with standby for evacuation. mate considered both rapid and slow depressurization accidents and neglected possible liftoff, or reentrainment, of plateout activity. Results indicate For a design basis isolation of the containment, the leakage rate to that limited access to the containment would be possible at about 15 h after atmosphere is 0.1Z/day. This implies that only one part in 12,000 of the a rapid or slow depressurization event. The use of the expected circulating activity released to the containment would contribute to the controlling 2-h activity would lower the dose rate by a factor of 4 and permit even earlier site boundary dose. Thus, design basis depressurization accidents are not post-accident access to the containment. important in establishing the primary circuit activity criteria.

4. PLATEOUT ACTIVITY CRITERIA 3.2.2. Offsite Doses. Limits on primary circuit activity were evaluated first for long-term radiological exposure of the public during normal HTGR The plateout activity criteria are essentially governed by shutdown operation. The single dominant exposure pathway assumed was slow leakage of maintenance considerations. Important factors influencing the criteria in­ gaseous activity and high pressure helium from the PCRV to the containment, clude maintenance philosophy (contact versus remote), decontamination fea­ then to the atmosphere. sibility, shielding" availability, and personnel acr-ess requirements. To establish appropriate criteria, both planned and unplanned maintenance of Annual doses at the boundaries of the hypothetical reference site were the primary system components need to be evaluated case by case. For the calculated using the TDAC computer code (Ref. 2). These doses were based on HTGR-GT plant, the major components in the primary system consist of the annual average meteorology and site boundary distances (Ref. 3). Initial turbomachines, heat exchangers (recuperators and precoolers), control results showed that Kr-88 was the dominant nuclide contributing to the most valves, etc. Details for deriving the plateout activity criteria follow. important offsite dose, namely, whole body gamma exposure. The inhalation thyroid dose due to radioiodine release was orders of magnitude less impor­ Gamma energy of each nuclide normalized to that of Kr-88; product of tant. Therefore, the parametric results were keyed to Kr-88. normalized energy times nuclide inventory summed over all gasborne nuclides. permanent shielding (e.g., cask) design is based upon the design activity 4.1. Plateout Source Assumptions for conservatism. A dose rate limit of 50 mrem/h has been established for the on-site contact maintenance. The plateout activity criteria were derived in terms of the 40-yr in­ ventory in the primary circuit for those important nuclides affecting shut­ With the expected plateout level, the unshielded dose rate in the com­ down maintenance. Such nuclides include Ag-110m, Cs-134, Cs-137/Ba-137m. pressor bearing cavity should be fairly low, since much activity has plated Other nuclides were neglected for one or more of the following reasons: out on the components further upstream. In the turbine bearing cavity, the short half life, no gamma emission, and relatively low release from the preferential plateout of Ag-110m requires shielding for personnel protection core. The criteria are intended as a design basis. The expected activity during shutdown maintenance access. is lower than the design activity by a factor of 4 for iodines and telluium and by a factor of 10 for metals. The turbomachine handling cask requires shielding to reduce the dose rate from the design plateout activity to 20 mrem/h. The physical cask For component maintenance evaluations, the local plateout activity was specification allows sufficient shielding so that the cask shielding re­ determined by considering the plateout distribution for a given nuclide, quirements will not dictate the plateout activity criteria. component operating (or service) time, and shutdown time prior to maintenance. A set of curves developed for the various nuclides related the Turbomachine disassembly will be accomplished in an on-site service plateout activity distributions to the uniform plateout activity in the facility after 100-day decay. Manual operations may be required to facil­ primary circuit. The uniform plateout activity is referred to as uniform itate the disassembly. To enable contact operations, the turbomachine distribution of the plateout inventory on all the helium-wetted surfaces in plateout activity was assumed to be decontaminated by a factor of 100. For the primary circuit. The total plateout surface area for the 2170-KW(t) simplicity, the same decontamination factor (DF) was applied to all nu­ HTGR-GT plant was taken as 1.6 x lo" cm2. clides. A DF of 100 is recognized to be difficult for Ag-110m, because of its possible diffusion into metals. If a DF of 100 is impractical, Ag-110m In the HTGR-GT plant, Ag-110m tends to preferentially plate out on the release will be further reduced or remote maintenance will be used. turbine, and cesium plateout is greatest near the recuperator inlet. The distribution factor (defined as the ratio of local plateout to uniform To meet the criterion of 50 mrem/h for contact maintenance after decon­ plateout) for Ag-110m on the turbine is about 90; for cesiums at the recu­ tamination, the average expected plateout activity on the turbine after 6-yr perator inlet, the distribution factor is about 50. Hence, the plateout operation with 100-day decay needs to be limited to 42 uCi/cm2 (before de­ criteria for Ag-110m and cesiums are influenced by the maintenance on the contamination). With a distribution factor of ~90 for Ag-110m on the tur­ turbine and recuperator. bine, this activity level is translated into a design plateout inventory of ~10,000 Ci Ag-110m for 40-yr operation and no decay.

4.2. Maintenance Considerations The limit of 10,000 CI Ag-110m inventory will permit on-site contact The major primary system components requiring planned or unplanned maintenance and personnel access to the turbine cavity with some shielding. maintenance include the turbomachines, heat exchangers, and control valves. Although Ag-110m is a principal turbomachine contributor, other nuclides The turbomachine is the controlling component for the Ag-110m plateout cri­ (particularly cesiums) should also be controlled so that their collective terion. The turbine and heat exchanger maintenance similarly establishes contribution is relatively small. To meet this objective, the permissible the cesium criteria. design plateout inventories for cesiums should be restricted to 45,000 Ci Cs-134 and 130,000 Ci Cs-137 (40-yr operation and no decay). Note that the 4.2.1. Turbomachine. The principal turbomachine areas of concern include higher inventories for cesiums than for Ag-110m will not significantly im­ maintenance, removal, and disassembly. Routine maintenance and in-service pact the turbomachine maintenance, when the plateout distributions are taken Inspection (ISI) on the turbomachine after reactor shutdown require person­ into account.' nel access to the turbine and the compressor bearing cavities. At regular 6-yr intervals, the turbomachine will be removed from the PCRV cavity for 4.2.2. Heat Exchangers. The heat exchangers in each HTCR-GT loop consist inspection and blade replacement. A shielded handling cask transports the of a recuperator and a precooler. Planned maintenance of the heat exchang­ turbomachine to an on-site service facility where the turbomachine will be ers generally relates to routine surveillance and ISI. Unplanned mainte­ cooled, decontaminated, and disassembled. nance includes tube plugging and component replacement. Maintenance on the precooler is not expected to influence the plateout activity criteria, since Table 3 itemizes the bases for the turbomachine evaluation, consistent the expected plateout level for the precooler is fairly low. Maintenance on with the maintenance access requirements. In general, the expected plateout the recuperator is affected by the preferential plateout of cesiums near the activity determines personnel accessibility during maintenance, whereas the inlet. With the design inventories of 45,000 Ci Cs-134 and 130,000 Ci Cs-137 accident exposures dictate the allowable circulating activity in the primary established by considering Curbomachine maintenance, any maintenance or ISI coolant. The circulating activity criteria should be specified as 14,000 Ci at the recuperator inlet must be remote. In this case, remote maintenance of Kr-88 (principal nuclide) for a design basis to simultaneously satisfy may be more practical and feasible than contact or semlremote maintenance, public offsite dose limits and containment access requirements during normal which requires a substantial reduction in the cesium plateout. operation. To provide an ALARA margin, the core should be designed so that the expected Kr-88 release is a factor of 4 less than the 14,000 Ci design 5. COMPARISONS WITH LIGHT WATER REACTORS basis.

To provide perspective, the following section compares the expected As established from the HTGR-CT primary-system component maintenance activity criteria recommended for the HTGR-GT vith LWR experience. assessments, the following plateout activity criteria are recommended for a design basis: 5.1. Circulating Activity Ag-110m - 10,000 Ci Table 4 compares the expected circulating activities between the Cs-134 - 45,000 Ci HTGR-GT and the LWR. References 4 and 5 give pressurized water reactor Cs-137 = 130,000 Ci (PWR) and boiling vater reactor (BWR) data, respectively. All concentra­ tions correspond to appropriate reactor temperatures and pressures during These figures refer to the 40-yr plateout inventory in the primary circuit normal operation. without decay. The expected pla'.pout for the above nuclides should be a factor of 10 lower. Among noble gases, the principal HTCR nuclide is Kr-88, whereas the PWR coolant contains mainly Xe-133, due to the differences in fuel temperature, By comparison, the criteria for the HTGR-GT expected activity are about release rate, and cleanup rate. Hoble gases from the BWR are vented to the the same order of magnitude as the expected activity for LWRs. off-gas system with little activity circulating in the coolant. Although Kr-88 is more biologically hazardous than Xe-133, the PWRs offer no advan­ ACKNOWLEDGMENT tage over the HTGR in containment access, because the PWRs leak primary coolant at a rate 100 times faster than the HTGR. Furthermore, the longer This work was supported by the U.S. Department of Energy under Contract half life for Xe-133 makes the offsice doses for PWRs less favorable than DE-AT03-76ET35300. HTGRs.

5.2. Plateout Activity REFERENCES

Table 5 compares the expected plateout activities between LWRs and 1. Clarke, S. S., and B. A. Engholm, "PATH -A Highly Flexible General HTGRs at particular locations in the primary loop. The activities are given Purpose Gamma Shielding Program," General Atomic Report GA-9908, only for nuclides with long half lives, such as cobalts, cesiums, and Ag- December 10, 1969. 110m. The HTGR-GT plateout represents expected criteria. Table 5 also in­ cludes the Ref. 6 data for the HTGR-steam cycle (HTGR-SC). The LWR informa­ 2. Buckley, D. W-, "TDAC: An Analytical Computer Program to Calculate the tion was taken from Refs. 7, 8, and 9. The operating time for each case is Time-Dependent Radiological Effects of Radionuclide Release," General about 3 yr. Atomic unpublished data, 1976.

The isotopic composition for plateout activity differs distinctly 3. Bunch, D., K. Murphy, and J. Keyes, "Demographic Statistics Pertaining between LWRs and HTGRs. The HTGR plateout arises from contamination with to Nuclear Power Reactor Sites," U.S. Nuclear Regulatory Commission fission products, whereas corrosion products are responsible for the LWR Report NUREG-0348 (draft), December 1977. plateout (or crud deposit). In the HTGR-CT turbine region, the principal nuclide is Ag-110m, which emits hard gammas. The plateout activity for the 4. "Stone and Webster Engineering Corp. PWR Reference Nuclear Power Plant HTGR-GT turbine appears to be greater than that for LWRs. The plateout for Safety Analysis Report, SWESSAR-P1," Docket-STN-50495-5, June 21, 1974. the HTCR-SC is about the same order of magnitude as for LWRs. 5. "General Electric Standards, Safety Analysis Report, BWR/6," 6. CONCLUSIONS Docket-STN-50447-50, Vol. 5, 1975.

Results of the analyses-for the HTGR-GT containment access and offsite 6. Engholm, B. A., and S. Su, "Shielding Design Report for the 900 MW(e) doses show that the radiation exposures during normal operation rather than HTGR Steam-Cycle Reference Plant," General Atomic unpublished data, June 15, 1979. TABLE 1 7. Crotzer, M. E., "Measurements of Radionuclide Surface Activities in a Pressurized Water Reactor," Trans. Am. Nucl. Soc. 38, 603 (1978). U. S. REGULATIONS FOR RADIATION PROTECTION

8. Shaw, R. A., "Getting at the Source: Reducing Radiation Fields," Nucl. EXPOSURE CONDITION REGULATION LIMIT Tech. 44, 97 (1979). TYPE

SREM/YRW.B.1*1 9. "Decontamination and Decommissioning of Nuclear Facilites," M. M. 10CFR20 OCCUPATIONAL MPC (Ы Osterhout (ed.), Plenum Press, New York, 1980, pp. 665-693. NORMAL REG GUIDE 1.1 ALARA OPERATION S MREM/YR W.B. PUBLIC I0CFRS0. 15 MREM/YR AT ЕАВМ APP.I THYROID OCCUPATIONAL 1KFR20 } REM W.B. PUBLIC AT EAB Zi REM W.B. FOR SUING lOCfBIOC ACCIDENTS CRITERIA 300 REM THYROID

PUBLIC > IIs Ci Xt-IMOR EVACUATION NUREG-0610 > IB3 Ci 1-131

OPEN CONTAINMENT (i) WHOLE BODY OOSE. (b) MAXIMUM PERMISSIBLE CONCENTRATION. CL0SE0 CONTAINMENT (c) EXCLUSION AREA BOUNDARY. 60 3.65 X/YR PCRV LEAK RATE

Of FSJTE 00SJJ.IMIT 50 CONTAINMENNT ACCESS И CRITERION = ново ад} TABLE 2 40 ANNUAL RELATIVE NUCLIDE MIX IN CIRCULATING ACTIVITY CONTAINMENT OFFSITE/ \\ OFFSITE ACCESS TIME 00SET0 (H/W) DOSEv, l \ PUBLIC (MREM/YR) NUCLIDE M RELATIVE ACTIVITY

Kr-83 m 0.43 Ki-BSm 0.53 Kr-87 0.67 Kr-BB 1.00 Kr-89 0.33 Xe-133m 0.02 Xe-133 0.35 Xe-13Sm 0.25 Xe-135 0.62 5 10 15 20 Xe-137 0.14 CIRCULATING ACTIVITY (THOUSANDS OF CURIES Ki 88) Xe-138 0.27

1. Flg. Parametric results of offsite doses and containment access dur­ M DAUGHTERS NOT LISTED ing normal operation (b) RELATIVE TO Kr-88 ON THE BASIS OF CURIES TABLE 3 TABLE 5 BASES FOR TURBOMACHINE EVALUATION HTGR/LWR PLATEOUT COMPARISON (EXPECTED ACTIVITY. -3-YEAR OPERATION. NO DECAY)

DOSE RATE OPERATING DECAY CRITERION PLATEOUT PLATEOUT ITEM TIME (YR) TIME (DAYS) (MREM/H) TYPE REACTOR TYPE AND LOCATION NUCLIDE ACTIVITY OCi/CM?) ACCESS TO BEARING CAVITIES 40 10 20 EXPECTED HTGR-GT. TURBINE MIOPOINT АВ-ИОгп 55 REMOVAL CASK 6 ia 20 DESIGN Cs-134 4.5 ON-SITE CONTACT MAINTENANCE 6 100 50 EXPECTED СИ 37 1.4 HTGR-SC.STEAM GENERATOR Cs-134 10 Cs-137 3.5

PWR. STEAM GENERATOR С 0-58 17-46 TABLE 4 Co-60 9-13 COMPARISON OF EXPECTED CIRCULATING ACTIVITIES BWR, RECIRCULATION PIPE С 0-58 1-2 Co-60 5-20

CONCENTRATION (/iCi/CM3)

ISOTOPE PWR BWR HTGR-GT

N•16 130.0 50.0 —

CORROSION PRODUCTS 0.06 0.06 —

FISSION PRODUCTS NOBLE GASES 9.0 — 5.0 HALOGENS 2.4 0.59 0.4 METALLICS 3.0 1.4 0.6

(a) AT APPROPRIATE REACTOR TEMPERATURES AND PRESSURES DURING OPERATION. The course of the accident as far as the flow- and temperature distribution

is concerned had been investigated over a time period of 50 sec in detail

THE INFLUENCE OP DUST ON THE HAZARD POTENTIAL OF by G. Fritsching /2/. As the results of this work are the basis for our A DEPRESSURIZATION ACCIDENT OF A HIGH TEMPERATURE REACTOR fission product study, they are described shortly. N. INIOTAKIS, C.-B. von der DECKEN Kernforschungsanlage Jülich GrribH. The flow distribution which appears already after the first second of the Jülich accident is shown schematically in Fig. 1. The leak was assumed to occur Federal Republic of Germany at the heat exchanger I (HX - I). The arrows are representing the direction

of now, the marks (+) (increase), (-) (decrease), (o) (no change) are de­ Introduction monstrating the change in flow velocity compared to normal operation

The possible hazard to the environment in case of accidents of a high conditions.

temperature reactor depends on the behaviour of the fission and activation As one can see the gas velocity is decreasing during the course of the products in the primary cooling circuit of the reactor. The behaviour of accident in the hot'pipe of the HKV-loop, in the steam reformer and the the fission products is influenced considerably by the presence of dust in steam generator. In the core region there is no change and there is an the system. Extended work has been done and is going on in the Institute increase in both of the WKV-loops. In the outer space of the HX - I the forReactor Components to investigate the consequences of severe accidents. direction of flow is changed and the velocity increases by a factor of about A large number of accident sequences are under investigation. 2.4. In the HX - I the velocity in the cold part for some hundredth of a

The paper describes as an example the investigations which have been second is higher by a factor of max. 3.5 compared with normal operation.

carried out to get qualitative and quantitative results about the release of The mean increase for the total heat exchanger HX - I is a factor 2.0. non noble gas fission products into the containment as a consequence of a The situation in the HX - II is as follows: Average increase in the HX - II

depressurization accident. The presence of dust in the primary cooling factor 1.2. In the outside space of HX - II for some hundredth of a second

circuit of the reactor is specially taken into account. factor max. 3. 30 % of the gas which is flowing from HX - II to the blower

is passing directly to HX - 1. 1. Description of the reference plant

3. Principles of calculation and principal assumptions The PNP-500 reference plant is described in the previous paper of this

conference /1/. We are referring to this. The investigations were carried out on the basis of the miotakis model,

using the codes PATRAS-S and PATRAS - LF /3/, /4/, /5/. 2. Description of the accident

As principal sources of dust in the pebble-bed reactor were identified: For the investigated accident a leak of 500 cm was assumed at the closure the abrasion of fuel elements and of the side-reflector and the scaling of of the pressure vessel at the He-He-heat exchanger in one of the WKV- loops. metal-oxides from corroded metallic surfaces. 4. General remarks This means that the dust consists out of graphite and metal oxide. The melting or weakening temperatures of these materials are much higher The interaction of dust particles with surfaces of the components is mainly than the maximum gas temperature in the primary cooling circuit of 950 °C. influenced by the Van der Waals - forces. The adhesional distance v is As in addition these dust materials do not contain any component which has " о determined by the unknown micro-roughness of the surfaces which may be a melting point within the given temperature range, it can be excluded that о the reason for sticking of the dust particles on the surfaces could be thermal about 200 - 1 000 A for realistic surfaces. Therefore the micro-roughness melting processes. beside the particle size is also used as a parameter in these calculations.

The sticking of the dust on the surfaces is depending on surface temperature, From measurements in the AVR-reactor coolant the amount of dust can be wall material, flow conditions, composition of the gas and micro-roughness. pessimistically estimated for a PNP-500 plant to be about 900 kg of graphite dust in 30 years reactor operation. For the amount of metal oxide Decisive for the sticking of the dust particles is their motion behaviour dust there is also an experimental result available from the AVR-reactor within the range of influence of the Van der Waals-forces. The sticking to be 30 % of the graphite dust. For the calculations we have used this value, probability of a dust particles by collision with the surface is determined although this value is probably depending on the conditions in the specific only by the energetic state of the particle directly before and after the plant. collision. The question if the particle is remaining of the surface or not

is influenced by the energetic state of gas stream (shear forces) and the With respect to the grain size distribution of the dust particles there are surface temperature /3/, /4/. only a few informations available from dust filter experiments on the AVR- reactor. But these results cannot be used for our calculations because A dust particle can adhere to the wall, if the energy criteria - as well as agglomeration occured. Therefore the grain size and the grain size the flow - criteria is fulfilled. That means, that a dust particle adheres to distribution were used as parameters. The investigations have been per­ the wall if: formed with regard as well to monodisperse as to polydisperse dust.

Experiments reported by Rumpf have shown that for dust of comperable immediately after the collision with the peaks of the nature the tendency for sticking is increasing with decreasing grain size micro-rough surface his kinetic energy is smaller of the dust /6/. For example, experiments with limestone powder in a flow than the binding energy due to the Van der Walls - parallel to steel surfaces at 20 - 50 m/sec velocity gave the results that forces (energy criteria) and, only particles with a diameter d^ 2 ^um did stick /6/. For the calculations the shearing forces due to the flow, acting on the within this paper the range of variation of the particle radius is limited to particle, -are smaller than the frictional force 10 yum maximum. This is in good agreement with the experience from the between the particle and the wall (flow criteria). AVR-reactor. If the first criteria is not fulfilled the dust particles escape after collision results from dust-fixed Cs-137 on the surfaces of the primary

out of the range of Van der Waals-forces. circuit?

If the second criteria is not fulfilled, the dust particles are carried away In Fig. 3 the influence of the grain size on the amount of dust which is from the wall with the gas flow. deposited on the surfaces of WKV-loop is shown for different values of

adhesional distance у 5. Results

The deposition of dust is limited for larger particle radius by the described 5.1 Case of monodisperse dust sticking criteria and to the smaller particle radius by desorption.

о With the assumption of monodisperse dust the investigations begun with the At the flow conditions, which are considerated, for у £ 200 A the energy behaviour of dust and its influence on the transport- and deposition behaviour о ° criteria prevails, for yQ ^ 250 A on the other side the flow criterion. of the fission products in the primary cooling circuit. Fig. 2 demonstrates

the deposition profile over the length of the WKV-loop of PNP-500 plant. The maximal total amount which is deposited on the surfaces of WKV-loop

The profile is drawn for different particle sizes. A micro-roughness of within the lifetime of the plant is estimated by this means to be 160 kg which y^ = 300 A and a non-oxidizing gas atmosphere were assumed. is about 18 % of the totaly produced dust in this plant. The deposition of the dust is limited to particle radia between about 0. 05 - 3.2 ju. These results

It can be seen that the dust is sticking on the different components very were obtained for normal operation condition. selectively. A comparison of the curves for different particle sizes makes

obvious, that the sticking tendency of the dust decreases with particle size, At depressurization accident especially larger dust particles may be torn which is in good agreement with experimental experience reported by off by the gas stream (flow criteria). With the present accident conditions

Rumpf /6/. In addition one can read out of the diagram that the amount of and by assumption of y^ = 300 A, dust particles may be torn off by the gas stream, if their radius belongs to the hatched region of Fig. 3. The figure dust on the surfaces of the heat exchanger is dominating. shows the amount of dust, which is torn off maximally by the gas stream,

Investigations as described before have been extended for different micro- to be about 160 kg. This value is very pessimistic, in particular as mono­ roughnesses taking the particle radius as a parameter and have been used disperse dust had been assumed. to answer the following questions- The activity of Cs-137, which is fixed on the primary circuit deposited

1. What total amount of dust could possibly stuck on the surface dust, is influenced by particle radius and adhesional distance extensively

of the primary circuit of PNP-500 plant? in the same manner, as the dust itself. As a comparison in Fig. 4 the

influence of particle radius and micro-roughness is demonstrated. Similar 2. What total amount of activity, which potentially could be trans­ to the amount of dust, the activity of Cs-137 remained constant in a vast ported into the containment in case of a depressurization accident, field of variations of particle radius and adhesional distance. It is obvious that the total amount of Cs-137 activity which is bound on the Fig. 6 shows an integral plot of the mass distribution for different degrees dust, deposited on the surface of the primary circuit, is very small. The of fineness of the dust. activity is ca. 1. Ci max. Within the same time period the Cs-137 activity Moreover the influence of the particle spectrum on the different fraction of which is released from the core into the primary circuit equals to about dust is given. One fraction contains the amount of dust which cannot stick 3 300 Ci. the surfaces under normal operation conditions (region C). The second

At accident conditions Cs is released with the dust only if the radius of fraction contains the dust particles which can stick under normal operation dust particles belongs to the hatched region. The figure shows the activity (region A + B). The third fraction finally represences the amount of dust of Cs-137, which may be released to be at maximum 0. 7 Ci. which will be blown off from the surface during the accident (region B).

Fig. 7 shows the influence of the particle spectrum on the activity of Cs-137,

5.2 Case of polydisperse dust which is

bonded to the dust (A, ) under normal conditions The dust in the primary circuit of the HTGR plants exists generally as poly- b

released into the containment by blown off udst fractions (Ar) disperse. With regard to the presentation at the preceding chapter only one in case of accident. part of the dust (finer dust particles) is to be expected to deposit surfaces of the primary circuit and only a part of the deposited dust is expected to The initial activity of the dust particles, that means the activity of Cs-137, be torn off by the gas flow during the accident. The amount of dust and the which is firmly bonded to the dust in the moment of its formation has been activity of Cs-137, which may be released into the containment, therefore varied. The curves refer to a starting activity of 22 yuCi/gr (index 2) is essentially lower, than in the case of monodisperse dust and essentially respectively 90 yuCi/gr (index 1). depends on the particle spectrum. As the particle spectrum is not known, By this for the observed conditions the release of Cs-137 into the contain­ this spectrum has been varied. ment is evident to reach its maximal value, when the mass distribution of

Following the work done by Walkenhorst /7/ the exponential function has dust shows it maximum for rp 2 yum. been applied to our calculation. In Fig. 5 the applied function is defined for Furtheron, the investigations led to the result that the release of activity is mass distribution by equations 1 and 2. The distribution parameters and principally caused by that loop, which shows leakage. A are correlated to radius r and dust concentration m (mass/volume) by P 1 equ. 3 and 4. Where r is the radius of distribution maximum. Moreover the investigations lead to a large number of results, some of P which will be presented in the following, without going into details. The particle spectrum is defined by the parameters or r^. Lower values

for r correspond to higher degree of fineness of the dust. Neglecting the contribution of the gas stream's dead water regions, the

following situation for the release of Cs-137 results respectively: In the frame of parameter uncertainties the upper limit produced. This deposited dust lies mainly in the

for the activity of Cs-137 released results to be about heat exchanger.

0. 8 Ci. The expected value might amount to 0. 05 - 0. 2 Ci. for the accidents conditions described the upper

The desorption contribution to the relase of Cs-137 limit for the activity of Cs-137 released is about

depends on the special design of the heat exchanger. 0.8 Ci. The expected value might amount to

This makes up about 15 yuCi for a heat exchanger of 0.05 - 0.2 Ci.

HELIX-design, 800 щС\ for U-tube design respectively, the release of activity is principally caused by that the free activity of the cooling gas comprised. loop, which shows leakage.

The dust bound activity of the cooling gas is about 40 juCi. References The most important uncertainties result out of the

insufficient knowledge on the interaction parameters /1/ C.-B. von der Decken, N. Iniotakis

between Cs-137 and dust, particle spectrum and Fission product behaviour in the primary circuit of a HTR

micro-roughness. IAEA-meeting, Jülich, December 2. -4., 1980

/2/ G. Fritsching

Summary private communication

/3/ N. Iniotakis The influence of dust on the plate-out behaviour of fission products is Dissertation (not published) analysed and consequences for large depressurization accidents are /4/ N. Iniotakis, C.-B. von der Decken estimated. The influence of dust on the behaviour of fission products in high temperature reactors The following are the salient conclusions: ENS/ANS Topical Meeting on Nuclear Power Reactor Safety, Brüssel, 16.-19.10.1978 dust is deposited selectively on certain components. Proceedings, Vol. 3, 2336 - 2347 for the flow and temperature conditions described /5/ C. -B. von der Decken, N. Iniotakis, F. P. О. Ashworth the radius of the particles which can be deposited A Model for the Description of Fission Product Behaviour in lies in the range 0. 05 - 3. 2 yum. the Primary Circuit of a High Temperature Reactor

Conference on Gas Chemistry in Nuclear Reactors and Large the maximum fraction of the dust deposited on the Industrial Plant surfaces of WKV-loop is about 18 % of the total dust Salford, UK, April 1980, Preprints /6/ H. Rumpf opera! on time X years operitcn time 30 years

Über das Ansetzen von Teilchen an festen Wandungen 'Ш-loop WKV-loop

VDI-Berichte Bd. 6 (1955), S. 17/28 у-ЗООД ,У,-ИД /7/ W. Walkenhorst _ 10' /Tri Das Massenverhältnis des nach verschiedenen Trenn­ / funktionen abgeschiedenen Feinstaubes in Abhängigkeit S КГ- von der Staubfreiheit I Staub-Reinhalt. Luft 37 (1977), Nr. 1, 16-19 "О 1. И*

idV 10° 10' га" t" Pcrtcle redus r,l yml - Particle radius rBluml——

Influence Ol gram spectrum on IBB 80 Influence of oran soectnin ISS 30 dust deposited nPNP-ртюту OTdustbiu* Cs-Ш ocnwiy tor «Gnous rncro- circut tar varuns rncro nur/resses 11ЛЗ 1 rouennesses П/П

Fig. 3 Fig. 4

Flow distribution in primary coolant IRB ВО circuit during Q depressurisation occident 11/01 Fig. 1 Definition:

Porlicie radius achesicna! distance y. 300 A 1) dm=P(rp)drp — rp . С,39 un

3 ar гэ • 10 um 2) P(rp)=Arp e" P

Гр . 20 pm 3) from P(fp)=Max= •'200 OD

•1005 4) from ^P(rp)drp=mt

sao with mt=concentration of dust in primary gas • 500 Tp=radius of distribution maximum

• 400 rp=particle radius

Ю 50 60 IRB B0 Position Im] - Grain size distribution deposition of dust en PNP-5nD-pnmory circuit 11/08 158 80 Fig. 5 öfter 30 yeors. WKV-loop НПО Fig. 2 MODELLING OP PLATE-OUT UNDER GAS-COOLED REACTOR (OCR) ACCIDENT CONDITIONS

A.R. TAIG UKAEA Safety & Reliability Directorate Culcheth, Warrington UK

Summary

The importance of plate-out in mitigating consequences of gas-cooled reactor accidents, and its place in assessung these consequences, are discussed. The data requirements of a plate-out modelling program are discussed, and a brief description is given of parallel work programs on thermal/hydraulic reactor behaviour and fuel modelling, both of which will гр [И provide inputs to the plate-out program under development. The represen­ tation of a GCR system used in SRD studies is presented, and the equations Integral grain size us во governing iodine adsorption, desorption and transport round the circuit ere distributions tu os derived. The status of SRD's plate-out program is described, and the type of sensitivity studies to be undertaken with the partially-developed computer program in order to identify the most useful lines for future research is discussed.

INTRODUCTION

A typical scheme for the assessment of consequences of Gas Cooled Reactor (GCR) accidents is outlined in figure 1. The basic questions which such an assessment must answer are as follows:

1. How much activity will accidents release to the environment?;

and 2. What will be the consequences of such releases?

This paper concerns primarily the former question, ie. the defini­ tion of source terms for accidental releases from GCRs.

A major difference between water-cooled and gas-cooled reactors in general is that whereas the former tend to rely on a secondary contain­ ment to instigate the consequences of accidents involving serious fuel damage (eg. TMI), the latter utilise the capacity of primary circuit surfaces to retain gas-borne radionuclides. In order to obtain source terms for releases from such reactors, the extent of primary circuit retention must be quantified to an acceptable degree of accuracy. This Work at SRD currently centres on the Advanced Gas-Cooled Reactor (AGR) in turn implies that, in order to extrapolate to primary circuit condi­ concept, of which a simple primary circuit schematic is given in figure 3. tions not encountered in normal operation of reactors, models of plate­ The reactor is CO2 cooled and graphite moderated, with a contra-flow boiler out processes must be developed, validated, and incorporated into programs system; the massive graphite moderator is cooled by the diversion through to calculate the extent of radionuclide retention. This paper describes it of a substantial portion of the 'cool' gas from the boilers prior to the work being performed in this field at the Safety and Reliability its being passed through the fuel channels. Operating experience with Directorate of the United Kingdom Atomic Energy Authority. these reactors suggests that the boiler units are responsible for about half of the plate-out which is observed when monitoring circuit contami­ nation; accordingly, as a first stage towards modelling plate-out in the 1. INPUTS TO A PLATE-OUT MODELLING CODE AGR, a very simple reactor model, ignoring plate-out elsewhere in the circuit, has been employed. The model is illustrated in figure 4. Both physical and chemical plate-out mechanisms, involving deposition of particulate and vapour phase species, operate to remove activity from The boiler is split into six pipe sections, corresponding to different coolant gas. The kinetics of both processes and the extent to which they tube bank arrangements and different tube materials ranging from austenitic remove activity depend on reactor geometry; gas flows; mass transfer stainless steel at the top to a 1% Cr, J% Mo steel at the bottom. The rest processes; the nature of materials released from the fuel; and the path­ of the reactor is regarded simply as a well-mixed space into which fission ways by which activity is released from the reactor. Rates of adsorption products are released from the fuel. Work on extension of the model towards and desorption of gases will also show a strong dependence on the tempera­ a system sucn as that shown in figure 3 is currently in progress. ture and nature of the surface on which adsorption takes place. The para­ meters required as input to a plate-out modelling code are summarised in figure 2. 3. IODINE BEHAVIOUR

In the case of involatile nucldes, the chemistry of surface inter­ Observations made in the Windscale prototype AGR^ 1 ^ suggested that action is immaterial; their saturated vapour pressures are so tiny that iodine plated out with a half-time of a few minutes, until eventually some they will rapidly enter the solid phase, either by condensation or by sort of steady state was reached, with the gas-borne iodine level reduced plate-out at a rüte governed solely by the kinetics of mass transfer to by a factor of around 1000. Low-temperature experiments gave slower plate­ reactor surfaces. At the other extreme, the inert gases will not plate out half-times, and, with sampling experiments revealing the apparent out at all. Most effort needs to be (and has been in the past) devoted to presence of two chemical forms of iodine, it was for a long time assumed the so-called 'volatile' species such as iodine and caesium, which plate that iodine existed in the circuit in two forms, one of which plated out out at a rate slower than that limited only by mass transfer. This paper readily and one of which did not. In this work, it is assumed at present deals with the plate-out of gaseous iodine; future work will give separate that iodine behaviour may be explained satisfactorily on the basis of a consideration to caesium and less volatile species. single species with simple first-order adsorption/desorption kinetics;

kf I , . —^ I , , . , (1) 2. REACTOR DESCRIPTION (gas) — (surface) r In the modelling work in progress at SRD, a reactor is regarded as a series of interconnected pipes characterised by homogeneous distribution by choosing suitable temperature-dependent rate constants and k^ of chemical and flow parameters down their length. Thus, along each pipe, the flow area and nature of surfaces are assumed not to vary, while the Values of the rate constants as a function of temperature have to be volume of gas and surface area available for deposition of radionuclides obtained by fitting to observed data. Such an exercise leaves considerable are distributed uniformly. The gas pressure, and the temperatures both of uncertainty as to the values of the parameters appropriate for use in the the surfaces and of the bulk gas, are assumed to vary smoothly along the AGR, particularly since the variety of surface materials complicates the length of each pipe, and are to be defined as a function of time and data fitting process. At present, the seme rate constants are used for all length along the various pipes. Solution of the differential equations boiler tube materials, and are being adjusted to give the most realistic governing flow and plate-out within a pipe is accomplished by splitting interpretation of observed iodine behaviour. Uncertainty in the surface the pipe into a number of compartments within which temperature and chemistry parameters is almost certainly, at present, the limiting factor pressure are assumed to be homogeneous. After calculation of rate determining the accuracy of the plate-out model. constants and mass transfer coefficients in each compartment, the equations for all compartments are solved simultaneously to obtain gas After a best fit of the parameters presently used to describe surface and surface-borne concentrations, plate-out factors and quantities re­ chemistry has been obtained, sensitivity studies will be performed in order leased from the reactor as a function of time. to assess the importance of improving the surface chemistry data. Possible options for making such improvements include collection of more reactor where к is a mass transfer coefficient. Eliminating С , we have data and the setting up of an active-loop to make more direct measurements о of rate constants. k.k к

4. EQUATIONS GOVERNING IODINE PLATE-OUT IN A PIPE

The equations governing iodine plate-out in a homogeneous box or k.k, Q к Q compartment have been derived as follows: A £ ( J£ - _E • _£ ) (4) V к +k V к, A W r ft Let J be the flux of iodine (Ci s * ) from gas to the surface;

-3 = d Q d Q s = - ё С be the concentration of iodine in the bulk gas (Ci m ) dt dt

Cq be that in the gas layer immediately adjacent to the surface; If this compartment is one of several linked together to form one of the pipes which make up the reactor primary circuit, there will be additional terms to be added to the derivatives of gas-borne concentrations to allow С be that on the surface (Ci m ); s for flow along the pipe. If subscript i defines one of the above quantities in the i'n compartment of the pipe, then 2 A be the area of surface (m ) ; d Q 3 8i = - J. + F. ,. С - F. С V the volume of gas (m ); dt i i-r g._1 i ч

Qg the quantity of iodine in the gas;

where F£ is the volume flow rate of coolant out of the i compartment. It Qg the quantity of iodine on the surface; is easier, in fact, to define concentrations in terms of mass, since the mass fluw rate U of coolant does not vary with temperature and pressure along the pipe as does the volume flow rate. Then, if M£ is the mass of coolant tn where all quantities are defined within the compartment or box considered. in the i compartment,

Q Q J. + U , si-l - si^ (5) J (kf . q = At c - kr . cs } (2) UU — — J. ' U/ i— 1 — X. N

dT \ 1 (ir. , in )

with k£ and kr defined as above. This flux must match that to the surface x-l X from the bulk gas; hence The set of equations (5), with j£ and -g^- si as given in equation (4), has been adapted for solution using the FACSIMILE package developed at AERE, Harwell. 6. THERMAL-HYDRAULIC BEHAVIOUR SO 5. MASS TRANSFER COEFFICIENTS

Transfer of heat and of molecular matter through a fluid have been It is clearly necessary to obtain a description of the variation with recognised as analogous for some time (3), and the analogy has been used time of gas and surface temperatures, gas pressures and flow rates around the primary circuit under accident conditions, since the rate and extent many times to determine mass transfer coefficients(^)t in the course of reactor design, much effort is devoted to optimising boiler performance, of plate-out depends critically on these parameters. For this reason, and and, as part of this process, it is usual to obtain heat transfer co­ because of the intrinsic value of a knowledge of the response of the efficients in the form: reactor system to accident conditions, generic AEA work on gas-cooled A A reactors includes the development of a computer code to model AGR behaviour Nu = A Re Pr (6) under accident conditions. The code will be designed to give output suitable for use both in fuel modelling and in plate-out studies. where Nu, the Nusselt number, is proportional to the heat transfer co­ The types of accident condition which might involve sufficient depar­ efficient; Re, the Reynolds number, is characteristic of the gas flow ture from normal operating conditions to demand a new modelling effort properties; and Pr, the Prandtl number, is related to the thermal might be summarised under three headings: properties of the gas. Values of the parameters A0, A^ and A2 have been 1. obtained for the various boiler sections shown in figure 4. Using Extra heat inputs (ie. reactivity insertions) these same parameters, the mass transfer coefficient is given by an exactly analogous equation: 2. Loss of heat removal capacity

A A 3. Loss of primary circuit integrity. Sh = A Re 1 Sc 2 (7) Although it is impossible at present to make quantitative estimates of о AGR response to these types of condition, an important qualitative feature of this reactor which is worth noting is that the massive graphite where Sh, the Sherwood number, is given by moderator represents a very large heat sink, and the primary circuit should thus retain at least some surfaces sufficiently cool to give significant iodine retention even in very severe heat-up situations. Sh = к D _f (8) D 7. ACTIVITY RELEASED FROM FUEL where к is the mass transfer coefficient as in equation (4) above Another parallel AEA program of generic work, on fuel modelling, exists not only to help improve fuel performance, but also to provide fuel failure is the flow diameter of the pipe, and criteria and estimates of fission product release from fuel under accident conditions. Given a defined accident sequence and fuel temperature D the diffusion coefficient, in this case, of iodine or transient, calculations of pin internal pressure due to fission product methyl iodide in CO^. release, in conjunction with failure criteria based on the response of fuel cans to thermal and mechanical stresses, will be used to predict times of failure of pins of various fuel burn-up from various parts of the core. The Schmidt number, Sc , is equal to -Щ- , where V is the viscosity and The corresponding gas gap inventories of fission products might then be P the density of the carrier gas. used as a source term for "instantaneous" release of activity, which would be followed by a prolonged release as fuel overheated and possibly oxidised Appreciable effort is involved in calculating mass transfer coefficients on exposure to the coolant. appropriate to each compartment of each of the pipes which constitute a reactor. The flow parameters - linear fluid velocity, density, and viscosity In connection with the AGR, fuel modelling codes used in studies on - are all dependent on both temperature and pressure, as indeed is the dif­ normal operational behaviour are based on semi-empirical data, and tend to fusion coefficient of one gas in another. For accidents in which circuit be able to cope only with very slow temperature transients. While this may temperatures and pressures depart from their normal operational values, the not hamper their performance for the slow heat-ups which might arise, for parameters defining the derivatives of gas and surface-borne concentrations example, from the loss of decay heat removal capacity following reactor (ie. those in equations (4) and (5) above) must therefore be input as trip, it would render them quite incapable of modelling fuel response to functions of time. The FACSIMILE package allows this task to be performed rapid transients such as might arise, for example, in reactivity accidents. with minimal effort. In such cases, a different approach, based on the models developed in (iv) Blowdown system : Gas-cooled reactors will in connection with the fast reactor safety programme, seems to be the best general be equipped with plant designed to clean up the option at present. coolant and maintain its chemical composition at the desired level. In the current AGR designs, this plant, which involves filter and iodine (charcoal bed) removal systems, has the capacity to have discharged through 8. PATHWAYS FOR ACTIVITY RELEASE it the entire contents of the reactor system, achieving decontamination factors in excess of 1000 for most radio­ It does not necessarily follow that the presence of extensive gas- nuclides but the inert gases. This is obviously an ideal borne contamination of the primary circuit will lead to a large release way of dealing with serious contamination of the primary of activity to the environment. It is first necessary for the carrier circuits; its one disadvantage is that the rate of gas to escape from the reactor system. Some generalised pathways by discharge is very small. which gas might escape from gas-cooled reactors are described below.

(i) Normal Leakage ; If the primary circuit integrity 9. STATUS OF SRD PLATE-OUT WORK is retained throughout an accident, activity release will be confined to the normal leakage of contaminated gas, A simple computer program modelling plate-out in a multi-compartment British experience with CO2 - cooled reactors suggests reactor boiler has been constructed and tested for the AGR. Reasonable Shat up to a few per cent of the reactor gas inventory duplication of observed iodine behaviour under normal operating conditions may leak per day all or mostly via the reactor pressure has been achieved. Sensitivity studies are currently being performed to vessel penetrations. In British reactors, and presumably test the influence on activity release to atmosphere of the following in any other future designs, this gas is or would be parameters. collected by a contaminated ventilation system and cleaned by passage through a filter system prior to discharge to the atmosphere. The AGR system is designed to give (i) Values of the rate constants for iodine adsorption excellent retention of particulate matter; the extent to on and desorption from reactor surfaces; which it might retain gas-borne iodine is a matter which merits further consideration. (ii) Assumptions concerning reactor behaviour in accident (ii) Depressurisations : Consideration has to be given conditions; eg. temperature profiles during depressurisation; to low-probability events which might lead to a serious drop in reactor pressure, possibly revealing defects present in a small proportion of fuel pins. Much greater (iii) Duration of release from fuel to coolant (ie. retention proportional release of activity from the primary circuit of iodine in the fuel or within fuel pins). would be anticipated in these circumstances, since signi­ ficant coolant gas release may occur before the fission products have reached adsorption/desorption equilibrium Work is currently in progress to extend the "boiler only" plate out - ie. before the gas is fully decontaminated. Modelling model to a whole AGR; the main change involved is the introduction of a plate-out during a depressurisation is complicated by large graphite/C02 interface, which will require corresponding adsorption/ the rapid change in pressure with time, plus, presumably, desorption kinetic data. The next priority in actual program development changing temperature profiles around the circuit. work will be to include models of caesium and particulate activity behaviour Attention must then turn to validation studies, among which the most impor­ (iii) Relief Valves : British gas-cooled reactors - and, tant will involve the models of surface reaction kinetics for iodine. again, presumably any other future designs, are equipped Validation of these models will, it is hoped, be obtained both by using the •with relief valves which discharge coolant via a filter computer program to perform calculations for reactors such as the Windscale to atmosphere in the event of reactor overpressurisation. prototype AGR, for which a considerable body of information on circuit Such relief valves represent a possible additional, low activity has been obtained, and, should the sensitivity studies justify the probability release pathway for any activity present in work, by 'loop'-type experiments aimed at making as nearly as possible the primary coolant. direct measurements of adsorption/desorption rate constants on reactor steel and graphite over the range of temperatures and pressures anticipated in reactor accidents. 10. ACKNOWLEDGEMENTS

The author wishes to acknowledge helpful discussions with staff of the Accident definition Nuclear Power Company, the electricity supply boards and other UKAEA establishments during the course of the work presented in this paper. 1 event tr Accident progression

11. REFERENCES 1 thermal-

1. COLLINS R D & HILLARY J J : "Some experiments relating to the behaviour of gas-borne iodine" TRG Report 983 (W). Reactor response

2. CHANCE E M et al : "FACSIMILE - a computer program for flow and I fuel modelling / failure criteria chemistry simulation, and general initial value problems" AERE Report R8775. Activity release from fuel 3. e.g. JAKOB M : Heat Transfer, vol.1, Ch.28. Wiley & Sons Ltd, 1949. I plate-out model mg 4. e.g. LAWLER M J : "Deposition of fission products in a primary heat exchanger of the Dragon HTR", SRD report R35 part 1. Activity released to environment I consequence modelling

RADIOLOGICAL CONSEQUENCES

ASSESSMENT OF GCR ACCIDENT CONSEQUENCES

FIGURE 1 93

> Temperature & Vault flow data reheater

Vault Release pathways superheater 1 main Activity input Fuel Graphite Reactor boiler from fuel channels channels core & vaults • économiser PLATE OUT CODE t decay heat boiler j

tailpipes ACTIVITY RELEASED Vault TO ENVIRONMENT <

INPUTS TO A PLATE-OUT SIMPLE SCEMATIC OF AGR SCHEMATIC FOR PLATE-OUT

MODELLING CODE PRIMARY CIRCUIT on BOILERS ONLY

FIGURE 2 FIGURE 3 FIGURE U There is evidence that metal iodide was formed for essentially all the iodine released during the Three-Mlle Island (TMI-2) plant accident (Ref. SAFETY RESEARCH ON IODINE PLATE-OUT 1). Measurements of 1-131 concentration in primary coolant water indicated DURING POSTULATED HTGR CORE HEATUP EVENTS that the iodine released from the fuel was retained in the primary coolant and the containment sump water. This is attributed to the formation and dissolution of metal iodide. Only trivial amounts of radioiodine were A.W. 3ARSELL, O.P. CHAWLA, C.G. HOOT released to the atmosphere or trapped by the building exhaust filters. General Atomic Company The importance of possible metal iodide formation, particularly iron San Diego, California iodide (Fe^), during postulated high-temperature gas-cooled reactor (HTCR) USA core heatup events has been recognized in HTGR probabilistic risk assess­ ments (PRAs) (Refs. 2 and 3). Such events are initiated from loss of main loop cooling coupled with failure of auxiliary system cooling, and they lead to release of volatile fission products in the core. Research on iodine plateout under core heatup conditions was recommended in early work on HTGR ABSTRACT PRA (Ref. 4). A combined analytical and experimental effort has been pur­ sued regarding plateout behavior of iodine on metal surfaces. A plateout computer program, PADLOC, has been developed (Ref. 5) which solves mass In support of probabilistic risk assesment (PRA) studies on the high- transport and adsorption equations for a nuclide species along defined mul­ temperature gas-cooled reactor (HTGR), an experimental program was conducted tiple flow paths and surfaces. The solution considers a time-dependent for iodine plateout on HTGR primary circuit metals during core heatup condi­ radionuclide source, radioactive decay, surface adsorption, desorption, and tions. Metal iodine formation and adsorption characteristics were measured surface heating by decay heat of plated-out fission products. Application primarily for mild steel and to a limited extent for Incoloy 800 and other of this program to core heatup scenarios formed the basis for an updated alloys. Pseudoisopiestic tests indicated quantitative formation of less evaluation of iodine time-dependent release to the containment for Phase II r volatile and water soluble iodides, Fel2 ° Crlj, during core heatup condi­ risk assessment of the HTGR (Ref. 2). tions. The rate of formation of Fe^ was limited by mass transfer at tem­ peratures above 570°K and was proportional to the partial pressure of io­ This program discusses the experimental work performed under the HTGR dine. The rate of iodide formation on chrome-nickel alloys appeared to be Safety P.esearch Program to provide input and correlation data for the plate­ temperature sensitive, indicating slower reaction kinetics. The iodides out analysis. The emphasis of the tests is on the possible formation of preferentially plated out on surfaces at 520 to 620 K. Fel2 under core heatup conditions and its plateout behavior.

in Plateout tests were also performed for Fel2 helium carrier gas 2. PSEUDOISOPIESTIC TESTS flowing over mild steel or quartz surfaces over which a temperature gradient was maintained. PADLOC computer program correlations of the plateout pro­ The predominance of the chemical reaction process which forms iodides file based on the Fel2 vapor pressure assumed in the PRA studies were In shifted the emphasis of this experimental program to studying the control­ fair agreement. The temperature at which most of the plateout occurred was ling iodide formation mechanism. The well known pseudoisopiestic static from 620 to 700 K, depending on the partial pressure of the Fel2 tested. technique was used to provide the iodine adsorption data. This technique has proven useful in obtaining equilibrium isotherm data when adsorption 1• INTRODUCTION processes are the only transport mechanism. However, as shown below, the iodine - mild steel system was more complex in that iodine transport in­ For many years, radioiodine release during hypothetical loss of cooling volved chemical formation and transport of iodide species. Hild steel was events has been considered to be a major contributor to assessed public risk chosen since it comprises a relatively large surface area of the primary and is a prime consideration in licensing of nuclear reactor sites. This is circuit internals. However, there are a number of components made from reflected in the iodine release fractions (25Z of the inventory available other alloys. Limited scoping experiments were performed of iodine plateout for leakage from containment) specified in U.S. Regulatory Guides 1.3 and on alloy candidates such as Incoloy 800, T-22 (2-1/4Z Cr - 1Z Mo steel), 1.4. In addition, high iodine release fractions have been assigned in many stainless steel 304, and Hastelloy X. WASH-1400 accident sequences. High iodine release is closely tied to the assumption of elemental iodine behavior. 2.1. Experimental Procedure

In the static tests, iodine transport was achieved by natural evapora­ tion of iodine at known temperatures in an L-shaped evacuated quartz tube Work supported by Department of Energy Contract ГЕ-АТ03-76ЕТ35300. (Ref. 6). The specimen was held over a 51-mm-long isothermal zone at one end of the long horizontal portion of the tube, while the iodine crystals core heatup accident. Moreover, the rate appeared to be (tagged with 1-131 isotope) were held at the other end in the vertical por­ independent of temperature. tion. A steep temperature gradient separated the specimen and iodine source chambers. Figure 1 shows the test arrangement. Based on these results, emphasis iras placed on understanding the mechanism of Fel2 formation and plateout. Figure 2 gives the ceasuxed Specimens were made from strips of test alloy foils (-Ю.05 mm thick) reaction rate data for temperatures of 1100, 1000, 700, and 600 К and shows rolled to give a coil (18 mm in diameter and 50.8 mm wide) that filled the that over a wide range of temperatures and iodine burdens, the overall rate isothermal zone. Except for being degreased and cleaned, the alloy strips of Fel2 formation is proportional to the iodine burden. Thus, the reaction were used in the as-received state. The specimens were of a known geometric kinetics are fast enough above 700 К that mass transport is limiting. area and surface roughness. The iodine vapor pressure was calculated from the iodine source temperature (Ref. 7), and the iodine burden at the speci­ Considering only one iodide species, nanely Feig, the following men was calculated after appropriate corrections for the transpiration mechanism was developed to predict the eventual gas phase concentration of

(Knudsen) effect and thermal decomposition of molecular iodine. Fel2:

+ 2I + + The specimen temperature was varied from 1100 to 400 К and the iodine »*(.) (8) b(e) * ^2(3) % - « burden from 10-^ to 102 Pa. Up to 200 mg of iodine crystals were used in

P each test. When iodide formaton and its subsequent depositon were observed, (FeI2)g the rate of buildup of iodide deposit was measured in situ by gamma spec­ (2) troscopy until all the iodine was used up. The iodide deposit was water soluble. At the end of a "test, the iodide deposit was quantitatively ana­ lyzed for iodine and metal content by gamma spectroscopy and atomic absorption, respectively. Available thermodynamic data can be used to calculate the values of

the equilibrium constant Kp as well as Pj and Pj . Therefore, ?сет2, }.

ia the 2.2. Test Data the overall partial pressure of Fel2 systeif'can he predicted." wa The predicted Ppei_. . s in excellent agreement with the calculated

r Static tests on iodine and mild steel with specimen temperatures of saturation pressure §f Fel2 f° the experimentally neasured depositions 1100, 1000, 700, and 600 К and an iodine burden of >10~5 Pa revealed that temperature. iodine retention on the mild steel surface by a simple adsorption mechanism was relatively insignificant compard with its reaction with the iron con­ Iodine transport data for mild steel temperatures of 500, 400, and stituent of the mild steel alloy. In the latter reaction, a less volatile 300 К were also measured. The tests at 500 К indicated a significant reac­ deposit was formed vhich condensed along a temperature gradient on cooler tion rate and gave no evidence of adsorption equilibrium, even after several, parts of the tube. Chemical analysis of the deposit revealed it to be Fel2> hundred hours and total iodine loss at 81 Pa of iodine burden, as shown in The transformation of iodine to Fel2 in the presence of mild steel had the Fig. 3. Analysis of the deposit showed it to be FeXj. The iodide Condensed! following characteristics: on the specimen itself. The tests at mild steel temperatures of ADO and 300 К did not show a significant reaction rate even after 600 to SCO beers 1. The formation of Fel2 deposit continued until all the iodine was of testing. Iodine retention by an adsorption oechanlsa vas observed only used up. at temperatures below 400 K.

2. The rate of iodide deposition measured in_ situ was constant during The tests with other HTGR candidate alloys showed similar behavior in each test and was correlated with the rate of iodine loss from the that the iodine reacted with the constituents of the alloy. It is cctahle source. that, the alloys containing high concentrations of chroaiua and nickel showed the formation and redeposition of Crl2 and NÜ2 in addition to Felj- 3. The temperature of eventual Fel2 deposition depended on test Table 1 presents the data from these scoping tests. conditions and ranged from 520 to 620 K. It is apparent that under high iodine partial pressures, such as those 4. The deposition of Fel2 during tests at specimen temperatures predicted for a core heatup accident, removal of iodine by fomatioa arrrt above 700 К produced a reddish brown ring with a sharp edge at redeposition of Fel2 is a very important eechanisn for iodine retention. the high-temperature end and a diffused mass toward the low- The metal iodides are less volatile than elenental iodine. However, there

can temperature end. are conditions under which significant decomposition of Fel2 result in reprecipitation of elemental iron. In addition, the presence of chronica 5. The rate of iodide formation was directly proportional to the and nickel in a substrate alloy results in the formation of their respective

iodine burden over the wide range of conditions predicted for the iodides (Crl2 and N1I2) under similar conditions. 3. DYNAMIC TESTS The dynamic test correlations are in good agreement with the lecatica and temperature at which the peak of the plateout profile occurred. Cepend— Testing under dynamic (flowing helium) conditions was conducted to ing on the partial pressure of FeI->, this peak ranged froœ 650 to TOO I. provide verification and correlation data for the PADLOC program. Iodine As expected, there was a general tread of higher partial pressures resulting transport was achieved by imposition of a helium stream with a low flow rate in higher peak deposition temperatures- Thus, the codel correctly predicted in a tube to simulate the convective flow conditions during core heatup. the location of the plateout distribution peak, Ifcwever, the data Indicate The I2 uas exposed to mild steel to form Felj at high temperature, and the movement of some of the adsorbed Fel2 along the surface ie tbe direction! of deposition characteristics of the Fel2 were measured for fixed isothermal or lower temperatures. It is possible that this occurred Airing the later time temperature gradient conditions for comparison with PADLOC predictions. period, only when flow was caintained after all Fel-> was depleted freer the Figure 4 depicts the test arrangement. Pure helium flow at 50 co^ per source. This phenomenon was not simulated in the FADLOC model- minute was maintained. A U-tube containing the tagged iodine crystal source Comparable Fel2 plateout profile data obtained frara tests at Згас&Ьатеш controlled the desired vapor pressure by means of a constant temperature National Laboratory (5SL) (Ref. S) were correlated, and the results were bath. An in situ gamma counting unit [Nal (Tl) crystal] for determining similar (see Fig. 7). In addition, the narrow gaasa scam resoXutîoa in the plateout profiles along the 1.5-m-long and 1.62-cm i.d. quartz tube was BNL tests confirmed the narrow plateout profile at earlier times- provided. Unreacted iodine was trapped by activated charcoal downstream. In a typical test, 11 coupons of 5 x 5 cm mild steel foil were positioned in 4. SUMMARY the flow tube. Each coupon was spread along the inside circumference of the quartz tube to form a tubular shape. The pseudoisopiestic tests provided data for analytical FADLCC COMPUTER code predictions of the fondation sad adsorption characteristics of iodine Twelve dynamic tests were performed. The test parameters included the species during core heatup conditions of surface temperature {370 to IQ7Q> £} amount and temperature of iodine source material which controlled the iodine and iodine vapor pressure (10""* to 100 Fa). These static plateoet tests partial pressure available for plateout. Five different temperature showed that iodine undergoes rapid reaction with tee iron constituent of profiles were tested, from room temperature to 1100, 900, 700, 400, and mild steel (such as In upper plemn elements or thermal barrier cover 294 K, as shown in Fig. 5. The time of the test varied, with each test plates) to form FeI->- The rate of formation of Felj was measured as a func­ running until all the iodine was depleted from the source. Some of the tion of iodine loading, and it was observed that Fel> condensed alcag; the tests ran beyond that limit with only helium flow. Tests 1 and 2 were run adsorption tube surface over the 520 to 620 К temperature range. without metal foils lining the quartz tube; the other ten tests used mild steel foil coupons. The dynamic tests showed that the analytical node! correctly predicted the location of the plateout distribution peak (&53 to 7QÖ K, depending 00

A PADLOC computer program model was constructed for the dynamic test iodine partial pressure), assuming that all the I2 was concerted to Fel2 conditions. As illustrated in Fig. 5, the model consisted of 12 consecutive the high-temperature part of the furnace and the F"el2 deposited along branch sections, each with 4 nodes, over which the measured temperature pro­ the tube according to its vapor pressure. Initially, the plateout profile file was imposed. The Felj Inlet partial pressure as a function of time and appeared to be narrow. Later, the data indicated movement ci some of the helium flow rate was specified. The model calculated the Fel7 partial pres­ Felj plateout along the surface; this effect was net simulated by the sure down the tube over the test time. Where this pressure exceeded the analytical model - saturation vapor pressure of Fel2» mo<^e^ calculated the vapor deposition on the walls according to the calculated mass transfer coefficient. The safety research on iodine plateout has demonstrated the hlghi probability that iodine will fora less volatile Iren iodide during ETGa core Figure 6 gives an example of the calculated plateout profile for mild heatup conditions. The availability of cooler (<7CQ KJ metal surfaces dur­ steel (test 3, upper temperature profile in Fig. 5) compared with the raw ing main core heatup sequences should provide significant retention: of Fel2- data of the gamma scans and post-test counting of iodine loading on fail Further work needs to be dace on the effect of an osidfnlng environment 00 sections. The calculated profile at 13 h is narrow, with virtually all the retention of Felj. plateout within a 1-1/2-cm-wide band at around 700 K. Unfortunately, neither the foil sectioning nor gamma scan data were capable of such a fine REFERENCES resolution. The combination of gamma detector opening and distance from the tube was such that oblique radiaton was picked up from as far away as 6 cm 1. Stratton, V. S., A.. F. Kalinanskas, and О. O. Caaipbell, letter to> 5HC from the centerline of the detector. A computer program is being written to Chairman J. Aheaae, dated August 14. 1950. deduce the actual narrow profile given the raw data and the known resolution 2. "HTGR Accident Initiation and Frogression Analysis Status otepart, Ehase function of the detector; however, the results of this program are not yet II Risk Assessment," DOE Report A-A1500O, General Atomic Company, Ajrft available. 1975- 3. "Sicherheitsstudie fur HTR-Konzepte under Deutschen Standortbedin­ gungen," Kernforschungsanlage/Instltut für Sicherheitsforschung and Gesellschaft für Reaktorslcherhelt, Draft Report, July 1980. 4. ""HTGR Accident Initiation and Progression Analysis Status Report, Volumes IV and VI," ERDA Report CA-A13617, General Atomic Company, Decenber 1975, January 1976. 5. Hudrltech, U. W., and P. D. Smith, "PADLOC, a One-Dimensional Computer Program for Calculating Coolant and Plateout Fission Product Concentrations," DOE Report CA-A14401, General Atomic Company, November 1977. 6. Hllttead, C. E., W. E. Bell, and J. H. Norman, "Deposition of Iodine on Low Chrora. Alloy Steel," Nucl. Appl. 2» 361 (1969). 7. Cevry, H. T., and L. J. Clllepsi, "The Calculation of Normal Vapor Pressure froa the Data of the Cas Current Method Particularly In the Case of Iodine," Phys. Rev. «0. 269 (1932). 8. "Reactor Safety Research Program, Quarterly Progress Report, January 1 to March 31, 1978," NRC Report BÎO.-NUREG-5082O. Brookhaven national Laboratory, April 1978.

SEALED EVACUATION PORTS GRADED SPECIMEN 11» QUARTZ TUBE

St MM ISOTHERMAL ZONE FURNACE LEAD / ««NIFREEZER SHIELDING SHIEIOEO 0 SHIELDED GAMMA GAMMA SCINTILLATION DETECTOR SPECTROMETER

Fig. 1. Iodine adsorption apparatus (static tests) ft МО» IЙ TOC 11 77 «Mrtfc

QUARTZ FLOW TU8£ v SPECIMEW

THERMOCOUPLE МОЛ. OLSUATIA» ^JtfUirr «Г IKMOIM)

IODIDE DEPOSIT

SHIELOEO DETECTOR

TO ATMOSPHERE; THROUGH FLAAL CHARCOAL TRAP IHOISHCMU

Fig. 4. Test arrangement for dynamic plateout tests

» i * ь » ~ ю i* к ifc ia да zr г»

Fig;. 6. Calculated alaCeomt grcfile far test 5b>. 3 ccmnaxedi vi tit raw data measoreaests

I A Bill QATA\;R£F Ц

6 10 1 I

CENTER OF FUKHACC ITI» (ca) I 0 20 43 S3 13 MO

Fig. 5. PADLOC model for dynamic plateout experiments DQTAKCE AIQHIS TU3E (OQ Fig. 7. Correlation of Fel^ plateaat profile TABLE 1 PLATE-OUT MEASUREMENTS AND ББСШГАМ1КАТТШ DATA ON IODINE TRANSPORT AND REACTIONS a OP A COMPONENT OP THE AVR REACTOR AT JÜLICH IN THE PRESENCE OF DIFFERENT HTGR ALLOYS( )

J. HANULIK, H. SCHMIED, J. WAHL Atomic Absorption Keraforschungsanläge Jülich. GmbH. Analyses of Total Iodide Deposit Iodine Jülich Spécimen Iodine (mg) Reaction Temperature Transported Rate (К) (mg) Fe Ni Cr (ug/s) Federal Republic of Germany

Incoloy 800 1. Introduction 1000 195.1 23.5 0.25 14.5 6.2 As a result of a water ingress on the 13.5.197S, cue to a steaa iooo(b) 100.0 12.0 0.2 8.0 2.8 850 192.6 34.0 (с) (с) 1.7 generator damage, various components of the AVR had to he re­ 700 196.5 19.4 10.4 8.4 0.4 paired. The gas blower had to he removed ООоОз the icop.

Hastelloy X Thus accessible surface materials, which had been in contact: 1000 196.5 17.6 (с) (с) 0.71 700 37.74 3.1 3.7 1.4 1.6 x 10~2 for several years with the helium used as coolicg——ediuzi were

available for research. This opportunity was used to get a T-22 юоо(ъ) 193.2 40.8 -0 ~0 8-9 survey of the deposition of fission and activation procccrts; in 1000(b) 200 44.0 -0 -0 4.2 -0 2.4 the primary circuit of a gas cooled high temperature reactor. 700(b) 186.8 41.5 -0

304 stainless steel In addition it was attempted to reduce the personal radiation 1000 195.7 28.0 -0 12.6 4.9 700 191.7 23.0 -0 9.6 0.6 exposure during the necessary maintenance work by decontami­

nation treatment. (a)lncoloy 800, Hastelloy X, T-22, and 304 stainless steel at various temperatures. Specimens are metal foil with a 129.03-cm^ geometric area. The data show the formation of This report refers for work performed within the coopération cf nickel and chromium iodides from alloys containing these elements. Rate data were calculated from tracer analysis. The B3C Aktiengesellschaft Brown 3overi & Cle., Baden iodine pressure was 100 Pa. Brown Boveri & Cie. AG, Mannheim (b)Smaller specimen with geometric area of 12.9 cm2 used. Eidg. Institut fur Seaktorf orschung, Wûrenlingen (c)Not measured. Gebrüder Sulzer AG, Winterthur Hochtemperatur-Reaktorbau GmbH, Köln Xernforschungsanlage Jülich GmbH, Jülich Nuclear-Chemie und Metallurgie GmbH, Wolfgang Schweizerische Aluminium AG, Zürich

on the development programme for nuclear power stations with higb. tempera­ ture reactor and helium turbine (НКГ), which is sponsored by the Federal Republic of Germany, the state Kordrhein-Äestralen and the Swiss Gcvermmen; 100 In the following, the surface sampling technics and the decon­ Experimental Methods

tamination work are described. The results are discussed. In the following the sampling methods and measurement results at The AVR, a high temperature reactor with Helium-cooling is the blower-wheel, at the wheelcap and at the blower entry are working since 1967. Picture 1 shows the primary circuit, which presented. is enclosed by a external and internal reactor containment. The main primary components are the cooling gas blower, the core Picture 2 gives a schematic presentation of the information depth with its graphitic structures and the steam generator. Helium of the various sampling techniques in order to measure the con­ flows at gas pressure of 10.8 bar through the core from bottom tamination. to top and is thereby heated from 270°C to a maximum mean gas temperature of 950°C. Dosimetry The plate-out of the fission and activation products on metallic Thermoluminescence dosimetry components is of interest with respect to maintenance work on

primary circuits in future high temperature reactor systems. Thermoluminescence dosimeters were fixed close to the surface

Therefore it was decided to conduct these analyses on a larger in order to measure the y- and ß-surface doses of the vheel.

scale than AVR needs. Groups from BBC Baden and EIR Wiirenlingen One was located in the capsule itself, the other on the capsule- were involved. Complementary methods were used as far as possible. surface , by which the capsule was pressed, on the vheel—surface.

By virtue of the screening effect of the capsule naterial for It was decided to perform the work jointly. The following work the ß-radiation (slight back radiation), it has been possible load was planned: to separate the y-dosis from the total dose and to assess the

- Visual examination with photographic documentation ß-dose.

- Measurements of the activity plated-out on the surfaces by Direct measurements different methods

- Decontamination studies at the wheelcap by using decontamination At the same time dose measurements were done with cose—measuring

solutions, developed for the HHT-project. instruments.

These experiments had to be performed without disturbing the actual Contamination measurements maintenance work. Furthermore all experiments had to be performed in such a way that the functions of the components are not affec­ The measurement of the contamination was performed by five (I-VÎ ted. The decontamination of the wheel and the remaining blower independent methods. parts (without the wheelcap) had to be carried out by AVR per­ sonnel using their own equipment and technique. Wipe and scratch specimens (I,II) direct current hand drilling machine with cental polishing

heads.- The polishing head is covered with a small amount of a

Previously existing,but not unequivocal filter experiments havo special grease which collects dust activity. Thereafter a coa- shown that the solid fission products appear to be fixed in taminated surface of 10 - 100 mm"* (depending on the surrace, the cold part of the primary circuit. surface layers, etc.) can be polished away and all the material

collected. The penetration depth into the metallic surface To examine the dust contaminated wheel-surface and blower varies from 20 - 50 pm. entry,it was decided to take wipe and scratch specimens in order to analyse loose and strongly fixed contamination sepa­ The samples were analysed by y-spectrometry aud the absolute rately. activity was assessed".

Wipe specimens : According to many experiments in other reactor systems г»^*

in laboratory experiments tolerances of + 20 % for the final By wiping with a soft felt cloth a certain surface is removed. result of the surface specific activity are estimated. The felt cloth may be used in the dry state or moistened with Photo 1 shows the sampling technique by surface nicjropolisü—g_ a liquid.The loosely fixed contamination can be thus determined.

Electrolytic sampling technic (IV) Scratch specimens:

With a special tool a definitive surface is scratched clean Photo 2 shows the electrolytic sampling technique

without affecting the metallic basic material. The scratched- This method is mainly suitable for;

off material, which contains the loosely fixed and some of the - taking samples at inaccessible locations since it is re­

surface contamination is collected and measured. The analysed motely controlled and for, 2 surfaces were about 10 cm . - step-wise layer removal to determine concentration profiles

(for instance diffusion profiles).

Description of the micropolishing sampling technic (III) Through direct current a certain amount of material, depen­

ding on the current and tine, is stripped away arc absorbée By the micropolishing sampling technique it is possible to remove by a small sponge. The small sponge is then dissolved anà the surface contamination of fission and activation products the activity determined by y-spe4rtrometry. For control pur­ and to measure its nuclid composition in the primary and secondary poses the alloying elements were analysed. Two methods were usedl circuits of reactor systems. Because of the simple tool needed,

samples can be taken in a quick and prepared action, directly a) Electrolytic single samples

at the component in strongly contaminated areas or in a compo­ A small sponge was exposed for a certain time at a definite nent. The samples were taken by using an electric miniature location. This method was used to determine the ягт.гаттяяТ! The decontamination of the wheelcap and of the bJLcÄ-er>fceeicap> activity at the blower entry and at the blower wheel. At the blower entry the sample taking instrument had to be brought into place by means of a 4 m holder. It was held Due to the fact that decontamination rooms of KFA. were rat

there by magnets during the sampling time. always available, the work hat to be carried out under unfa­

vourable conditions. Some investigations had to be left oat. b) Cassette method

This method was used in the case of metallic clean surfaces Shortly after the first decontamination tests it became apparent

of low-alloyed steel and for KANIGEN-nickel plated steel. that only a decontamination cf 3 (TU>-sieasuremen.tsî cocJLd he о As electrolyte,a solution of 10 % oxalic acid in water was 2 achieved. This even with hot water at SQ C, which contained used. The current density was 0.1 A/cm . Dekopan and had been squirted on the wheelcap under high

This solution was employed to avoid harmful residues. pressure.

Photo 2 shows the electrolytic sampling by the cassette A two-stage deco-procedure, developed for the EHT-project, vhf rrr

method at the wheelcap. is close to the well known APCITKGX-procedure was then csed.

The contamination of the dissolved layer was determined This procedure consists of an alkaline, oxidative, prelimicary by analysing the activity of the used deco-medium (V). treatment with subsequent, slightly acidic, reducing arac

forming after-treatment. Experiments on the nickel layer The working conditions in the hot workshop were very uirffa

Thickness of the layer rable for a chemical decontamination as it was only possible

to work by wearing air-fed protecting suits, which Ifmited

On the untreated blower the thickness of the KANIGEN-nicke1— the working time to 2 hours. Furthermore, the technical in­ layer at the wheel was measured to be 30 -70 um. stallations were insufficient or were completely lacking (as

e.g. lifting instruments, pumps, bath-heating).

Pore and crack testing Photo 3 shows the blower-wheel before decontamination t:

Photo 4 shows the blower-wheel after decontamination treat This so called Ferryl-test consists of fastening a filter paper, moistened with a liquid reagent on the surface to be Results examined. At certain locations, where iron or low alloyed steel are moistened, Berlin-blue is formed, which is recogni­ zable by its colour on the filter paper. As the wheel was After a lay-off period of the blower of more than 300 days rusty due to the water which had entered, this method could only the contamination was found to be mainly due to ionglived be performed after the decontamination treatment. fission products. Longlived activation products were found Picture 1 Schematic description to a lesser extent. From the fission products Sr-90 was found of the primary circuit to be the isotope with the greatest activity followed by of the AVR Cs-137 and Cs-134 with some Ag-110m. During the maintenance Steam generator work the Cs-isotopes were largely responsable for the radi­ ation exposure of the personnel. The finger exposure and the danger of incorporation were mainly due to Sr-90. Short lived isotopes that can contribute to the radiation level had de­ cayed away (Picture 3). Keflector

The decontamination of the blower wheel was first started Core with conventional technique such as the use of high pressure water to which decontamination products were added. The y- radiation was reduced by a factor о Г 2.1, the ß-radiation Blower by a factor of 3.1. Further use of the decontamination solu­ tion Ox-24 and R-6 increased the decontamination factor to a mean value of 5.6 for the y-radiation and 5.8 for the ß-radi­ ation. These values were measured with thermoluminescent dosimeters. The values agreed with y-spectroscopic measure­ ments (Table 1, Picture 4, 5).

The decontamination of the wheelcap provided a reduction of the ß-radiation of 900 and of 56 for the •y-radiation. The dosis rates were measured with thermoluminescent dosimeters.

The material depth removed had a mean value of 0.5 um. Gamma- spectroscopic measurements of the caesium isotopes gave de­ contamination factors between 60 and 90 (Table 1).

Operational constraints did not permit the optimum use of the chemical decontamination procedures.

Picture 2

Schematic presentation of the information depth of the various surface sampling techniques for meaning contamination. A-Electrolytic sample, B-Wipe sample, C-Scratch sample, D-Micropolishing sampling technic, E-Dissolved layer by decontamination. Picture 4 Points of contamination measurements by means of

thermoluminescence dosimetry at the blower-wheel.

Picture 3 Wheel contamination. Mean activity according to three

independent methods. Calculated as follows: activity/

volume=activity/(treated surface x removal depth)

1 from the analysis of the already used deco-mediums

2 from the electrolytic samples

3 by means of the mlcropolish sample-technique.

Picture 5 Activity profile of sample at the wheel after

decontamination treatment, take up with electro­

lytic sampling technique. Table 1 Dose rate and contamination at the wheel and at the cap, untreated and after decon­ tamination treatment.

2 2 [pCi/cm ]untreated [pCi/cm ] after decontamination Decont.-factor [-]

Micro- Direct TLD * Micro- Direct TLD from Direct from TLD

polishing measure [rad/h] polishing measure measure

Ag-110m 0.97 0.82 0.02 0.005 160

Cs-134 5.8 6.5 0.04 0.1 63

Cs-137 10.1 12.0 ca p Y = 2.2 0.1 0.16 75 DF =56 Y 1—1 Eu-154 0.45 0.69 0.05 0.005 140 CJ S = 187 DF = 900 p л Eu-155 0.37 0.34 0.04 0.001 340

Co-60 0.11 0.22 0.01 0.002 110 Sr-90 - - 0.37 - -

Ag-110m 0.06 >1 и Cs-134 0.51 и JJ •H с •и 0) Cs-137 0.92

Eu-154 0.09

Eu-155 0.03 Blowe r

Co-60 0.01 Electro :

Sr-90 8.05

Micro- Electrolytic TLD Electrolytic TLD Decont-factor polishing from TLD Wheel Picture 3 Picture 3 Y=18 Picture 5 ?= 3.2 DF у = 5.6 6=168 6 =29.4 DF g = 5.8 Pict.4

* TLD = Thermoluminescence dosimetry Photo 2. Electrolytic sample - taking at the wheelcap. Mat the entent or penetration o? active material, rntcp the contaminated

surface can be either predicted; or measured; (1) ("i-g:.. 1 ');.. Thus* the- depth-, EVALUATION OP A DECONTAMINATION MODEL to which, the surface layer shouîd be removed ta achieve the desired TevftTi

D.W.T. RIPPIN of decontamination can be- determined'., finderstandi.ng: о* the decontamination: Swiss Föderal Institute of Technology process will: allow, the détermination jf the most appropriate conditions, far?-

Zurich the ccntroTTect removal of the surface1 layer- to. the' repaired; depth, а-t. any,'

Switzerland sclae. of operation.. If a mathematical' mc.de Ti of the process, can be developed which describes, with, acceptable а-ссцОчсу the behaviour, of the- process, as, a> J. HANULIJC, E. SCHENXER, C. ULLRICH Swiss Federal Institute of Technology function of the design, and: operating; conditions.,, such. a> mcde.Ti w.i.T.TI be- au Wtlrenlingen considerable assistance in the interpretation of laboratory, data: and; the.' Switzerland subsequent prcDlems. of scale-up. desijn and: operation..

Г4TR0DUCTtOM For certain materials.. laboratory decontamination tests appear- ta contradict,

the above suggestion that the amount if remaining- activité' i;ss л unique- function: In the scale-up of a laboratory decontamination process difficulties, arise of the depth ta which, the surface has, been removed.. Äcttvisty prof-hies, measured due to the timited understanding of the mechanisms controlling, the process. in the laboratory for- different types, of decontamination, fT.ui'i show,1 that, whani

Me same amount of material has been removed' from the- surface- as measured; This paper contains some initial proposals which may contribute to the gravi metrically, différent decontamination-; factors can, be? obtained {^щ- 24). quantitative understanding of the chemical and physical factors which This behaviour could be- explained by, tasteОО-£ТесОтепОЖcrcsccpe; studies; tin influence decontaaination operations. which. Me formatier; of layers, of soli i product on the- topi of Me ceccntamrnatad

me:aT, surfaces, was observed. These layers, were- interpreted as being undesüred The basic concept of decontamination by wet chemical treatment is that the solid by-products of the reaction. It was further observed xm the tafcairitary activity near the surface boundary is removed as the surface itself is that, at the sore depth af removal,, ai sharp: increase: im йе decnn.taminatiojr, dissolved away by the decontaminating fluid. Earlier work has demonstrated factor can, be ofcta.ir.ed by appropriate' changa of the decc—medium: frenr

acid to alkali r.e) (Fiç. 2b}. In fact* fer a r.eçî.ig~bty small fuirther remeval. This report refers for work perfcnr-od within the cocje:d:ic.i cf of the surface at the Tirrt of what is çravi.iietrrcaTTy measurable,, the BBC Aktiengesellschaft Brown Boveri i Cie., Baden.

Brown Boveri i Cie. AG, Mannhein decontamination, factor can be increased by up tc a factor cf ten-.. It is Eidg. Institut fur Reaktorforsch-ing, würenlir.geri suggested that after dissentier cf the active rrucJefces at the surface Gebrüder Sulzer AG, Winterthur being treated, are held same possibly by Hochtemperatur-Reaktorbau GmbH, Köln they back in way, adsairgttQR, ам Kernforschungsar.lage Jülich GmbH, Jülich the by-product layer. The solid, by-prcducts fcnrîrç this layer are assumed Nuclear-Chemie und Metallurgie GmbH, "Wolfgang to arise from sice reactions accompanying the desired oaf- reaction which: Schweizerische Aluminium AG, Zürich reraves the metallic surface. Tne sucden. chance in residual actuuty сад on the development programme for nuclear power stations with high tempera­ ture reactor and helium turbine (HHT), which is sponsored by the Federal then be ascribed to the removal by the new decc-neciuirc of the аЬсче scHxd: Republic of Germany, the state Nordrhein-Westfalen and the Swiss Government. by-product layer. It has also Ьеегт found that chem-scat ce

This improved performance may also be ascribed to the removal of the by-product the main reaction their stoichiometry must also be described. The stoichio­

layer. A similar effect can be achieved by chemical treatment in several steps metric equations serve to calculate the component and material balances

in sequence with different deco-media. These changes of deco-medium serve to around the reactor and conversely experimentally measured balances should be

dissolve the by-product layers already built-up. used to check and correct proposed stoichiometric relationships.

Gersral features required in a mathematical model to describe a fluid-solid 2. Chemical Thermodynamics

reaction will be discussed, and initial work will be presented with a simple Thermodynamic data for the proposed reactions ray also be available frora model which has had some success in describing the observed laboratory literature sources (e.g. 2, 3). The equilibrium constants deternined frora behaviour. these data indicate the theoretical feasibility of proposed reactions for

the corrosive removal.of the contaminated material. From the equilibrium GENERAL REQUIREMENTS FOR A DECONTAMINATION MODEL constants the equilibrium concentrations in the system can be determined as

The dissolution of the contaminated surface takes place by means of a chemical functions of temperature, feed concentrations and other relevant variables

reaction between the surface material and components of the deco-medium. and provide an upper limit to the attainable conversion.

However, the rate of this reaction may be limited by the rate at which reactants and products are transported to and from the surface. In order to model the 3. Chemical Kinetics

system both the chemical reactions and all other significant operations at The rate of progress of a reaction depends on the local concentrations or near the phase boundary between solid and fluid material must be described of the reacting materials and the temperature. These functional dependencies quanti tati vely. must usually be determined experimentally for the relevant reactions. It is

commonly assumed that the rate constant of the reaction has an Arrhenius

CHARACTERIZATION OF THE CHEMICAL REACTIONS dependence upon temperature and that the reaction rate depends on the local

reactant concentration to a simple power (one or two). However, when summary 1. Stoichiometry expressions are used to approximately represent more complex underlying reacti

A simple stoichiometric equation can be used to summarize the overall schemes more complex concentration dependencies may be called for. reaction taking place. Such equations are commonly available from the

literature (e.g. 2). A typical example is given for the solution of a CHARACTERIZATION OF THE TRANSPORT OF MATERIAL AND AMD FROH THE REACTION SITE chromium oxide surface layer with alkaline potassium permanganate

(AP - Citrox Process). The decontamination process involves interaction between the liquid and solid

Mn04" + 2 H20 + 3 e" -» Mn02 + 4 OH" phases:

2 Cr(0H)3 + 5 OH" -v CK>4 " + 4 H20 + 3 e" Deco-medium (liquid) + contaminated surface (solid) =ф Cr QH ( b * \ Cr2°3 + I H2° ==• products (dissolved) + byproducts (may be partly solid).

2 СОЖ + 2 OH" + 2МпО/ - 2 CrO. " + 2 Mn02 + H20 This interaction comprises some or all of the following steps: An indication of the role of the diffusional resistance to transport from

the bulk of the fluid to the surface can be obtained by changing the degree (1) Diffusion of the reactive components in the decontaminating fluid to the phase boundary. of agitation of the fluid. As the agitation rate is increased the diffusional

resistance to transport will be reduced and if this resistance plays a signi­ (2) Activities at the phase boundary: ficant part in controlling the progress of the reaction, the overall reaction (a) adsorption of the reactive components (b) chemical reaction between the reactive components rate will increase. If the agitation rate can be increased until there is no and contaminated surface, as discussed above. further increase in the overall reaction rate, this indicates that a level of (c) desorption of the reaction products (d) dissolution of the soluble reaction products agitation has been reached at which the diffusional resistance is no longer in the solution significant in comparison with other processes contributing to the control of (e) possible deposition of a layer of undesired solid by-products on the treated surface the reaction.

(3) Diffusion of soluble reaction products from the phase boundary into the solution. Responses of an experimental laboratory decontamin?rion process to changes

of temperature and agitation rate are illustrated in Fig. 3. There is a Since the above steps take place in series the overall progress of the continuing increase of the overall reaction rate with increased stirring reaction is limited by the slowest step. indicating that even at the upper limit of stirring rate used the fluid

phase diffusional resistance is still significant in comparison with the QUALITATIVE ASSESSMENT OF THE DECONTAMINATION PROCESS other rate controlling processes. The effect of temperature on the overall reaction rate is significant but sufficient evidence is not yet apparent

An initial assessment of the relative importance of the various stages in in this case to locate a break point where one controlling mechanism takes

the overall process can be made by observing the effects of changing conditions over from another. on the decontamination performance.

A SIMPLE MODEL OF DECONTAMINATION Changing temperature allows a discrimination between stages having different

sensitivities to temperature. Generally chemical reactions are more sensi­ Qualitative features of the decontamination process indicate that a model

tive to temperature change than the accompanying transport processes. Thus, must include at least a description of the chemical kinetics, the transport

at a low temperature where.the chemical reaction rate is slow this may of reactants to the surface and, where appropriate, the effect of the accumu­

control the overall rate of progress of the reaction. The chemical reaction lated solid by-product film (i.e. stages 1, 2b and 2e of those discussed

rate may increase rather rapidly with rising temperature until it is above). The success of the simplest form of model will show whether the

greater than the overall rate of the transport processes. Further increase effects of the other stages also need to be considered.

of temperature will result in the overall rate of conversion increasing

more slowly, corresponding to the increase in the rate of the transport The simplest model incorporating the three features discussed above assumes

processes. Thus, a breakpoint in the curve of overall reaction rate against that the reaction proceeds at the surface with a velocity proportional to

temperature will be an indication of a transition of the control of the the concentration of the fluid reactant there, that the rate of transport

overall reaction rate from one stage of the process to another. of reactant through the fluid is proportional to the concentration difference 110 across a fluid film divided by its thickness, which depends upon the degree To examine the validity of the model more closely the experitrantally observed

of agitation, and that the rate of transport through the solid by-product values of с or s should be compared directly with values predicted by the

film is proportional to the concentration difference across it divided by model when the best values of the parameters in the integrated forms of Eqns.

its thickness, which is assumed proportional to the amount of material A4 or A4' have been fitted by a non-linear regression program. For a good

which has reacted in the main reaction. model the discrepancies should be of the same order of magnitude as the

experimental measurement error. Since the analytical solution of the

For such a model, the rate of change with time of reactant concentration differential equation is implicit in the concentration it will probably

or of the amount of solid dissolved is derived in the Appendix (Eqns. A4 be more convenient to integrate Eqn. A4 or A4' numerically to obtain the

and A41). It is clear from this expression that, as expected, the overall concentration explicitly as a function of time. Non-linear regression

rate of progress of the reaction will be dominated by the slowest of the programs are available for parameter fitting of models expressed in the

three stages - surface reaction, diffusion through the by-product film and form of differential equations (4).

diffusion through the liquid film. A similar equation for a second-order surface reaction is also given (Eqn. A7).

CONCLUSIONS The differential equation (A4 or A41) can be integrated analytically and expressed in a linear form (Eqns. A6 and A6^). This can be used as an initial In further work the simple models already proposed will be further developed check on the plausibility of this simple model. When measured values of с or s by corroarison with additional laboratory and pilot plant data. The experimental are available as a function of time they can be inserted, together with known program will be designed to identify those stages which are critical for the constant values of and c0 or s0, into the equation. If the model is correct progress of the overall reaction and the means by which these stages can cost the plotted results should lie on a straight line, the slope and intercept effectively be controlled. The understanding of the cost îcportant stages of of this line giving respectively the value of and [щ + TJT). the system will facilitate the design, operation and control of full-scale

decontamination processes. When the degree of agitation is changed at constant temperature the terra Я-

u 1 ags0 Appendix should change but r— and -к— should remain the same. Thus, the model Ki ьр Derivation of Model Equations predicts a line of the same slope but different intercept.

It is assumed that reactant from the bulk liquid phiss diffuses throuçh a This type of behaviour is borne cut by the experimental results plotted in liquid film and the solid by-product filn before reacting at the surface the required linear form in Fig. 4. The validity of the model cannot be tested statistically from this transformed data.

However, it seems clear that the inclusion of a term in the model for the resistance of the by-product film is of benefit. A model with only fluid film resistance and a first-order reaction at the surface would give a constant value when plotted on these coordinates, corresponding to a = 0. If it is assured that accumulation in the liquid and solid films and at These equations can also be written in terns af fee a^cctt at scttîcf

the surface can be neglected, the rate of decrease of reactant in the liquid fron the surface s(noIes/area} by the suhstitutiec с = =^ (s -s]; '«теОе

4 a _SA _Y__ds s0 - s phase (I) can be equated to the flow through each of the liquid (II) and Co " V о SÄ dt I_ . . X =È£ С—"J solid (III) films and to the reactant consumed at the surface (IV).

- of = "V at ' AQ " AD' ^ = AD ^ = Ar(cs> v с . I II III IV Vs Dc -In -la

The simplest assumption for the reaction rate expression is that the reaction

in

h1 D' 0 Vdç 1^=5—"И" h*^(-g§ t^>l c»!

A ct - r^Cc0-c> s - The amount of material deposited in the solid film is assumed to be proportional

to the amount of reactant that has reacted.

Nomenclature

aV(cQ-c) = A x p (A3)

Substitution for x in (A2) A area of contaminated surface Q V dc molar concentration: of reactant irr, tulk Tîçuic prrase (A4) A dt " J_ + A i aV(co-c) co initial value of с k] 0' D Ao c* molar concentration of reactant a* c^.er surface cf scttd fe^-crcxaict fS molar concentration of reactar.t at solid surface Eqn. A4 can be integrated analytically to give cs D diffusion coefficient in sollid by-product, film

A . .1 6. . с 1 gVcQ , , с с (AS) Q' diffusion coefficient in îicutd; fiîra Y t - - (О- • ÖT) In — • ïï -jJl (- in — • — - 1) reaction velocity constant, ffrst-crcsr reactior, 10 0 0 kl AS can be further re-arranged to produce a linear form kZ reaction velocity constant, seccnc-crcer reaction: N number of coles of reactarrt ir. bulk liquid prase

Q flow rate of reactant (roles/area} to-arcs soïic surface !-T-+ïïTCa (1-T^-> (A6) -In "1 " u ^ -In r(cs) rate of reaction (noles/аОэа of surface} as a furrcrtcit cf co co surface concentration af reactar.t s amount of solid removed from surface (moles/area)

sQ amount of solid that would be removed from the surface (moles/area)

if all the initial reactant of concentration cQ were used up.

V volume of bulk liquid phase

x thickness of solid by-product film a ratio of moles deposited as solid film to moles reacted at surface

ß moles of liquid reacting per mole of solid removed from the surface

6 thickness of liquid film p density of solid film (moles/vol)

ABTRCGTTEFE IN I pml

VERGLEICH ZWISCHEN THEORETISCH BERECHNETEN UND DEN VON EIR EXPERIMENTELL ERMITTELTEN DEKG-FCIITCREN VON CS-137 OUT NIMOCOS» 713 LC REFERENCES

Fig. 1: Predicted and measured values of the penetration of 1. INIOTAKIS, N.. VON DER DECKEN, C.B., HANULIK, J., SCHENKER, E. active material into the contaminated surface Experimentelle Ergebnisse und Einflussgrössen bei der Dekontami- nationsproblematik von HTR-Werkstoffen. Atomforum 1980, Berlin.

2. Handbook of Chemistry and Physics, CRC.

3. BARIN, I., KNACKE, 0. Thermochemical Properties of inorganic substances. Springer-Verlag Berlin 1973.

4. KLAUS, R., RIPPIN, D.W.T. A new flexible and easy-to-use general-purpose regression program for handling a variety of single- and multi-response situations. 12th European Symposium on Computer Applications in Chemical Engineering CACE '79, Montreux, April 1979. to be published in Computers and Chemical Engineering.

Fig. 2a: Decontamination factor (Df) Fig. 2b: Change in decontamination as function of depth of factor with change in surface removed with different deco-fluid deco-fluids (a and s) Fig. 3: Amount of surface removed in laboratory experiment a function Fig. 4: Plot of linearized equation A6 for model with first-order reaction and of temperature and stirring speed n. transfer resistance by liquid film and growing solid by-product film Effects of the coolant on the materials

On material specimens originated from high temperature helium cooled reac­ DECONTAMINATION AND HIGH TEMPERATURE MATERIALS tors (see table 1), we studied the kinetics and morphology of the surface E. SCHENKER, G. ULLRICH, J. HANULIK, layers, the plate-out of fission products, the effects of fission products W.B. WAEBER, K.H. WIEDEMANN on the surface layers and on the bulk material, the building in and sub­

Swiss Federal Institute for Reactor Research sequent fixing of the fission products in the surface layers, the fission Wurenlingen product diffusion behaviour (i.e.caesium and silver isotopes), the changes Switzerland in the chemical and mechanical properties of the materials caused by fission

Abstract product diffusion, and the hot gas corrosion.

Our R +D-programme on development of decontamination methods from primary A helium loop is available for out-of-pile experiments on the hot gas

circuit of helium cooled high temperature reactors comprises of the follo­ corrosion of high temperature materials. The most important details are wing major points: found in table 2 and a diagram is given in Fig. 1

- Effects of fission products and other helium impurities on the materials, The behaviour of caesium and silver in contact with Nimonic 86, Inconel 625,

Nimocast 713 LC and TZM in vacuum and in helium from 700 - 1050°C and until - Development and optimization of the chemical decontamination solutions 5500 h will be carried out. The first results concerning diffusion behaviour and also the conditioning of the used solutions, from tests done in vacuum have been reported (7). - Materials research in support of the programme.

Some highlights on the programme and on the results were presented. Development and optimization of the decontamination process

The total programme and first results were reported earlier (1) - (6). We are developing chemical solutions for removing the contaminated surface Therefore we limit this presentation to only some major points: layers (8), (9). We study the reaction of the solutions on both the affec­

- Effects of fission products and other helium impurities on the mate­ ted material and the bulk. Of particular interest and importance is the

rials, degree of corrosive attack. We study also the technological aspects of the

procedure (application, equipments), so the тсЛе of action of the solution - Development and optimization of the chemical decontamination solutions by different application methods (soaking, spraying) in pilot plant tests and also the conditioning of the used solutions, on large work pieces, and by adjusting the chemical activity and recycling - Materials research in support of the programme. of the decontamination solutions. We are examining methods to concentrate

and solidify the contaminated solutions in order to safely store them. This work has been performed as part of the development programme for a This work includes investigations to separate non active components. nuclear power station with a high temperature reactor and high power helium A great deal of effort has gone into carrying over laboratory results to turbine (HHT) carried out in collaboration with the following firms: Brown, a technical operation and out of these a decontamination model was developed Boveri & Cie. AG, Hochtemperatur-Reaktorbau, Kernforschungsanlage Jülich, which has been already reported (10). NUKEM GmbH, Swiss Federal Institute for Reactor Research and other Swiss partners. If considerable diffusion of fission products into the material occur and decontamination solutions which cause such corrosion are eliminated during long time at service (11), we must remove not only the surface in decontamination development stadium using short time experiments. layer, but also a certain amount of the bulk material (possibly up to Materials can, with their oxidised surfaces and carburised layers suffer

200 um). Therefore, experience gives the best method which allows metallic accelerated selective attack on their grain boundaries even when decon­ layers to be removed evenly and controlably. taminated with a milder decontamination solution. The selective attack

may favour crack growth when the part is put back into service and

possibly rest solution from decontamination solutions remains in the Materials research crack tip (Fig. 3).

A large part of our work is done in the field of material research in When one thinks that structural changes such as welding or thermal ageing support of the decontamination programme. It is necessary to have a large leads to totally other reactions then these types of experiment (after material testing programme in order that the turbine fabricator can suitable thermal treatments) are very important. guarantee reliable material properties before and after decontamination In addition these experiments, as acceptance trials, will be completed treatments. This is product surveillance in effect. If the mechanical with standardised high end low cycle fatigue experiments for completeness. testing shows that the decontamination procedure is not favourable we have to find one that is. When we can demonstrate that we have found a The detection of eventual hydrogen embrittlement in turbine components safe decontamination procedure we must prove it as such to the relevant caused by the decontamination solutions will be studied using creep tests 4 authorities responsible for Nuclear Power Safety. of up to 10 hours at working temperature.

The fatigue and the creep properties in this framework are important.

Sharp cracks and their slow growth will be evaluated. The main quantity Acknowledgment of interest in this test is the stress intensity factor К which is well known in fracture mechanics. As is known, with this quantity it is The authors wish to thank Dr. ?. Tipping for helpful discussion amd trans­ possible (Fig. 2) to lay down two asymptotic limiting values with the lation of the paper. smallest and largest growth of crack for each chance in load da/dM.

References Below К is the region where an existing crack will not grow and any о formation of a new crack is not Dossible. Above К is the region where (1) Schenker E., Loevenschuss H., Ullrich G.. Wiedemann 5.E. о Elaboration d'une méthode de àêcontamir.aticn du circuit premaïre è*um a fracture due to overloading is possible. réacteur à haute temperature refroidi à i"Helium (EST) IAEA-SM-2G0/4, Vienna 1975 A material delivered with these known properties can drastically be (2) Ullrich G., Schenker E. affected when chemical attack occurs. Crack formation and growth can Jahrestaçung Kerntechnik "5Э, 25. - 27.3.1553, Serlim, p. 352-55 be accelerated and even the value for К may be considerably reduced. с (3) Hanulik J., Schenker E. Stresses may be only due to internal distortion and they can initiate Anforderungen an das Dekcntaminaticnsmittel aus chemischer Sicht stress corrosion cracking. The last mentioned case is the most severe (4) Wiedemann К.H. Table I Samples Korrosionsprobleme bei der Entwicklung von Dekontaminationsmitteln ibid. p. 856-59

(5) Hanulik J. 1 !' Bestimmung von Aktivitätsprofilen in kontaminierten Reaktorkreisläufen Reactor Experiment Allcvs * Т'-р Te=eratura ï ibid. p. 860-63 *

(6) Schenker E., Hanulik J. DRAGON EPOS A, 3, C, СО ; -v. ~oo \ Plate-out von Spalt- und Aktivierungsprocokten auf Hochtemperatur­ werkstoffen im Reaktor DRAGON ibid. p. 864-67 E?05 A. 3. С 120 j 7C0 i' "*2CCC

(7) Gras D.J., Limon A., Waeber W.B., Brown J.D. Quantitative Caesium and Silver depth profiles in high temperature KFA Experimental i ! Molybdenum- and Nickel-base alloys Duct j ^ TOO : i i' 10. Kolloquium über metallkundliche Analysen, 3.-5.11.1930, Wien j ; (will appear in Microchimica Acta) Fuel Eleaent

(B) Schenker E., Ullrich G., Loewenschuss H., Hanulik J. Assemblies Contamination and decontamination of Incoloy 600 samples fron the KFA F?C 19/21 А. В. С 120 j SSO - 9-5Q i ; experimental duct of the DRAGON reactor Alloy 800, proc. Petten int. conference, Amsterdam 1978, p. 347-53 i \ AVR VAMPIR I А var. I var. 2Q (9) Ullrich G., Schenker E. 1 Mechanical properties of Incoloy 800 from the KFA experinental duct j — at 700°C and 480 d in the DRAGON reactor before and after decontami­ nation

ibid. p. 357-62 *Alloys: A Stainless steels, e.ç. AISI 316, Iccolcy SCO

(10) Rippin D.W.T., Hanulik J., Schenker E., Ullrich G. 3 Nickel alloys, e.ç. Niscnic» Iricciïel, Eizccaat TL3 1С Evaluation of a decontamination model Jülich, 2. - 4.12.1980 С TZM

(11) Iniotakis N., von der Decken C.B., Hanulik J., Schenker E. Experimentelle Ergebnisse und Einflussgrössen bei der Dekcntaminations— Problematik von HTR-Werkstoffen Jahrestagung Kerntechnik '80, 25.-27.3.1980, Berlin, p. 633-6

Table 2 Technical data or the EIR Helicn-Ioco

a) Basic loop

Gas Eelitn

Pressure 2 — 3 bar

Gas purification 2-5 rsor=ai. =~/b b) Test section I Hotgas corrosion

Corrosion ovens 30

Temperature 800 - 1200 К

Helium flow (per oven) 10 - 1000 normal litres/h 2 Hot metal surface (without specimens) ca. 3500 cm Material Inconel 600 ttu A/F (Area/Flow) 12 - 1200 sec/cm • 4 Available usuable diameter 90 mm tin nji S=> M Available usuable length 500 mm -в^— c) Test section II HCF and LCF < <

HCF: 1 Push-pull puiser

Dynamic loading + 3 kN rj1 -rj1 »rj! Frequency 25 Hz

Temperature 1200 К max. „AIS SI* , ' 4Ï* SIT

1 Push-pull puiser

I I -2bar ~ Dynamic load +1,5 kN î Lx u

50 Hz Frequency PIÎ «3 PU Temperature 1200 К max.

LCF: Dynamic load 200 kN

Temperature 1200 К max. -x- d) Test section III

Fretting corrosion experiment in a heat exchanger model (SULZER).

A Pipe to gas Chromatograph M Magnetic valves

D Flow indicator H Helium purification unit

E Evacuate P Pressure measurements

F Furnaces for hot gas corrosion R Control valves

G Graphite furnace S Cleaning flow valves

I Injection of impurities T Temperature measurements

К Compressor w cooling water MAINENANCE CONCEPT OF THE GAS TURBINE IN A 1640 MW DIRECT CYCLE HTR

H. SCHMIED, H, KARAUS Brown, Boveri & Company, Ltd. Baden

Federal Republic of Germany

К. RÖLLIG Hochtemperatur-Reaktorbau GmbH Mannheim

Fig• 2 Amplitude of the stress intensity Federal Republic of Germany

E. SCHENKER Swiss Federal Institute for Reactor Research Wurenlingen Switzerland 1. Introduction and Constraints

1.1 Structure of the Power Generation System and its Influence on the

Maintenance of the Turbomachine

The description which follow refer to a helium-cooled high-temperature reac­ tor with spherical fuel elements and a thermal output of 1640 MW. The cycle is intercooled, and the helium flows through the plant components built in­ to the presstressed concrete reactor vessel (PCRV) in the following sequen­ ce:

Low-pressure compressor (LPC), intercooler, high-pressure compressor (HPC), recuperator, reactor core, turbine (T), recuperator, precooler. The reactor pressure vessel is made of prestressed concrete. The LPC, HPC and T are ar­ ranged in the turbomachine (TM) as shown in Fig. 3. The energy is transmit­ ted from the turbomachine to the generator, which is located outside of the reactor vessel, by means of an intermediate shaft. Those parts, including the turbomachine, which come into contact with the cooling gas are all con­ Fig. 3 Incoloy 800 from the KFA Experimental Duct, exposed in DRAGON reactor taminated with radioactivity during operation, the extent in each case being mainly dependent upon the temperature. for 480 d at 700°C. The extend of the intercristalline attack as re­ vealed by the decontamination procedure: The fissure has been opened For the maintenance of the turbomachine, a major inspection is planned at further by tensile testing (oxalic acid electrolytic etching, 200 x) . six-year intervals. This overhaul involves disassembling and dismantling the turbomachine. Two minor inspections, at two-year intervals, are planned bet­ The transport of the fission products from the core outlet to the surfaces ween the major inspections. The minor inspection involves inspection of the of the turbine is described by the following mechanisms: bearings, for which purpose the turbomachine does not have to be disassemb­ - Convective material transport in the flowing cooling medium led and dismantled. - Mass transfer to the surfaces due to the flow-dynamic boundary layer Since the TM is installed in a helium atmosphere in the PCRV and because of - Adsorption-desorption equilibrium between the gaseous phase and surfaces the radioactive contamination which is present, it is necessary to plan in - Penetration of the fission products in the wall material advance suitable equipment such as airlocks, remote control systems and de­ contamination equipment, etc. Official regulations covering the handling of The deposition activities of Cs-137, Cs-134 and Ag-110 m of the turbine radioactive materials must be adhered to. The economical aspect necessitates blading after 6 full-power years, calculated with the expected values for knowledge of power plant shut-down times and capital costs. The maintenance the release, are plotted in Fig. 1 in the axial direction from the turbine planners investigate given circumstances and develop concepts for the sequen­ inlet to the outlet. On average, the axial profile of Cs-137 lies above ce of work to be carried out. Planning of the civil works and also the de­ the Cs-134 by a factor of approx. 2. The Ag-110 m values are lower than tailed development of the components must take full account of these concepts the Cs-activities by more than 3 orders of magnitude. in order to ensure favourable conditions in which the maintenance work can be efficiently carried out. Buildings and Equipment for the Major Turbomachine Inspection This paper presents in summarized form the sequence in which the maintenance work is to be carried out for both the major and minor inspections. The shut­ A specially equipped building is necessary for maintaining the turboma­ down time and the collective does were estimated to this end. Costs have not chine. The equipment installed in this building is also used for main­ yet been determined. Details of the concept are given and significant results tenance of other plant components. Figure 2 shows the plan view of the from the areas dealing with release, deposition and decontamination have been maintenance building, which adjoins the fuel store. For the major in­ appropriate adapted. spection the turbomachine has to be moved from the helium atmosphere of the reactor pressure vessel through the airlock to the air atmosphere. 1.2 Contamination of the Turbomachine To this end, a transport tunnel for the turbomachine leads through the fuel store, which lies between the reactor protection building and the During operation the turbomachine is contaminated by radioactive depositions maintenance building. This tunnel connects the airlock to the dismant­ resulting from the flow of cooling gas. After the six-year inspection inter­ ling cell. val the dominant depositions are the relatively mobile nuclides Cs-137, Cs-134 and Ag-110 m. These are preferably transported in atomic form to The maintenance building lies in a controlled area and contains the fol­ the surfaces over which the gas flows, where they absorb and penetrate lowing three zones: the wall materials. Moreover, dust-trapped activities are also deposited in joints, gaps and dead spaces in the turbomachine. Despite the dust col­ - Zone with personnel access without special precautionary measures lections contributing only slightly to the overall contamination, special . Erection room measures should be taken when carrying out maintenance to prevent aerosol . Operating corridors for the hot cells (dismantling cell, decontami­ activities. Determination of the surface activities of the gas turbine are nation cell, airlock) based on the calculation of the fission product release from the reactor . Blading inspection and decontamination room (control room for the core. The basis for this was the reference design of the equilibrium core blading inspection and decontamination box) for spherical fuel elements with highly enriched Th, U-fuel particles with a coating of pyrocarbon/silicon carbide. The release rates (expect values) - Zone with personnel access with special precautionary measures such as calculated for Cs-137, Cs-134 and Ag-110 m, are as follows: room monitoring and ventilation . Hot workshop

- Zone with restricted personnel access Nuclides Provision for personnel access only for the purpose of trouble-shooting Cs-137 Cs-134 Ag-110 m and providing appropriate precautionary measures are taken. Release rate . Dismantling cell (Ci/full-power year 78 82 0.54 The dismantling cell has shielding walls with lead-glass windows and of operation)* is equipped with manipulators, optical sensors and sound transmission equipment. The turbomachine enters through a door, travelling on rails.

* 1 Ci=3,7-10

Power plant shutdown tine* = 12 days (3-shift operation)

Minor Inspection (for positions of components see Fig. 3) Collective dose = 0.9 man-Оеш (design value)

The minor inspection is carried out at two-year intervals between the major inspections. Inspection of the turbomachine should be performed, as far as 4. Major Inspection (for position of components see Fig. 3) possible, without disassembling and dismantling the machine. The main ob­ ject of the inspection is to check the bearings. The intermediate shaft can The major inspection takes place at six-year intervals, and involves dis­ also be inspected. assembling and disn-antling the turbonachine, and the inspection of the in­ dividual parts- It also includes the inspection of the TK connections and Sequence for the Minor Inspection mountings in the PCRV cavity.

- Reduction of pressure in the primary circuit 4.1 Disassembly of the Turbonachine The reactor must be taken out of operation and allowed to cool off. The helium inventory of the primary circuit is to be reduced until the pres­ - Reduction of pressure in pritrary circuit: (see Section 3.1) sure is slightly below the ambient pressure. - Opening the TH cavity: The TM cavity (7) is dismantled. After connecting - The entry shafts (9) are opened and the access areas are entered. the TH airlock and replacing the heliua between the diaphragms seal (S) The entry shafts (9) to the access areas (2 and 6) are closed on the un­ and the peripheral seal (1) by air, the diaphragn seal is opened (7?t air­ derside of the reactor pressure vessel by two covers (15 and 16). After lock remains closed). removing the outer cover (15) an airlock is connected and the inner cover - Opening the TM entry shafts: see Section 3 (16) is subsequently removed. The entry shaft interconnections (10) com- - prise gangway bellows, whose connection to the turbomachine casing is remote-controlled. The helium in the access rooms (2 and 6) and in the * Calculated frora the disconnection of the generator frca the power svsten to reconnection of the generator to the power systea. - Disconnection of tne 7*t connections, supply lines and roto* coupling; the The components, a-e mcved by remote' control: tarnt the dismarr.tlinc ceüli to 131 intermediate pipes supplyin-j gas to the T*t, the TK mountin the TM. The doors tc the transfer room are closed. The air in the transfer (Outer casing, compressor sections., stationary blade сапОте»- ami rotcr after disassembly of the turbine biasing;) room and in the space tetween tiie TM peripheral seal and the transfer door is replaced by helium, after wnich the peripheral seal and the inner - Steam blasting, possibly with, the addition: of chemicals transfer door are opened. The traction machine pulls the TM into the trans­ fer room. The innter door is closed again and the helium in the airlock is - Intermediate cleaning in the iecjntamfnatfen. cell replaced by air. The outer transfer door is r.ow opened end the ТГ! is moved - Drying with hot air into the dismantling room. A decontamination, factor of 3 and a residual lecse surface COT.tantiaaticrri etf

2 Dismantling the Turbomachine _< IC~3 u Ci/cr.2 may be expected.

Dismantling of the turbcmachine mainly takes place in the dismantling cell and is carried out by remote control. The access roons (2 and 6) are re­ .3.2 Treatment of Parts Exposed tc Hot Sas entered for the purpose of manually disconnecting the inner flanged con­ nections. The remote-controlled dismantling requires and exact analysis In the case of parts which are exposed te- het cas, the radxcactîve ссл.са- of the operating steps, a design which takes account of convenient hand­ mination is also diffused inta layers clcse to the surface, ami these must ling, the development of special dismantling devices and thorough testing also be partly rencved. before being put into use under nuclear conditions for the first time- Three variants for the disassembly of the turbine blading were investi­ Treatment of Turbine Inlet Casing; and Diffusor gated:

- Variant A - Steam blasting of the components, Remote-controlled disassembly of the turbine blading - Spraying with an oxidizing alkaline solution, at SCI^C far rsraat cf o.vÈdgs.

- Variant В and С - Rinsing of the components. Manual dismantling of the turbine blading after pre-decontamination, em­ - Cleaning of the decontamination; cell. ploying personal protection gear such as auxiliary shields, lead-glass - Spraying with a slightly acidic, buffered salutier, containing substances windows, gripping tongs, etc. Variants 3 and С differ in the pre-decon­ with a reducting and complex foraine action, at 53°^. tamination process used. - Rinsing with hot water.

4.3 Decontamination of the Turbomachine components - Cleaning of the decontamination cell.

For the purpose of the treatment, the components are divided into the fol­ - Steam blasting of the components with adr.ix.ture cf inhibitors. lowing groups: - Drying with hot air. - Components exposed to cold gas Decontamination factors of 200 and a loose surface residual contamination - Components exposed to hot gas <^ 10-3 y Ci/cm2 may be expected. . Turbine inlet casing and diffusor . Turbine blading (including heat shields) Treatment of the Turbine Blading (including heat shields) the cabin the components and installations in the TM cavity can be ob­ served through lead-glass windows and by means of optical sensors. From The blade roots are not heavily contaminated, since these lie in the rela­ the inspection cabin manipulators can be used to replace and adjust parts tively pure flow of cooling gas and have operating temperatures of only in the TM cavity. When the TM cavity inspection has ended the inspection approx. 500°C. cabin is dismatled, so that the TM airlock is available again for the re­ assembly of the TM. The sequence for the disassembly, inspection and decontamination of the turbine blades was investigated for three different variants (A,S,C), and a comparaison of these is given in Table 1. 4.6 Mounting and Commissioning the TM

The TM is mounted and connected in reverse order to that used for the dis­ assembly. The position of the TM corresponds to the adjusted TM-supports in .4 Inspection and Reassembly of the Turbomachine (TM) the cavity. Final adjustments to the shaft train are carried out on the ge­ nerator side. The cavity cover and entry shaft covers are to be subject to Inspection of the TM components: All TM components are subjected to a a pressure test and leakage test. visual inspection before the TM is reassembled, for the purpose of de­ tecting possible deformations, wear and cracks. Parts or sections which have to satisfy exacting requirements are measured. Depending on the 4.7 Duration of Major Inspection and Collective Dose part concerned, checks on cracks are carried out by means of the dye penetration test, magnetic particle test or utrasonic test methods. The overall maintenance stratégie is based on a flexible use of decon­ tamination, shielding and remote handling techniques and in the use of The dismantled turbine blades are subject to a special inspection pro­ sensitive spare parts. cedure in that they are tested in the contaminated state in a special blade inspection box. In the case of the demonstration plant this is particularly relevant since it is desired that the state of the blading Лп analysis of the operating sequence provided the following results: remain the same for the purpose of a better assessment.

The turbine outlet casing is not decontaminated and remains in the dis­ Var. A Var. В Var. С mantling cell until reassembly, a visual inspection being carried out through the lead-ylas windows and with the aid of optical sensors. Power plant shutdown in days for 3-shift The turbomachine is reassembled manually in the assembly shop. Depending 50* 45 60 on the residual contamination, auxiliary shielding devices may be required operation for the assembly, e.g. for mounting the turbine blades. The gas turbine outlet casing is the last part to be mounted on the turbomachine in the Collective dose in dismantling cell. To shorten the time required for the overhaul, the tur­ man-rem (design 18.5 33.5 19.5 bine blading and the diffusor are alternately replaced by duplicates. values)** These are then ready within the shortest possible time for re-use after the major inspection has been carried out (decontamination, inspection, maintenance).

.5 Inspection of the Turbomachine Connections and Mountings in the TM * Details of the time for Var.A are shown in the paragraph (Fig. 4) Cavity of the Reactor Pressure Vessel in rough form.

The installations and components of the TM cavity are inspected. To this ** The expected value for the dose rate were multiplied by the factor end a cavity inspection shielded personal carrier is built into the TM 2 to take account of the remaining nuclides and again by the safety airlock, The cabin is coupled to the personnel access of the TM airlock factor 5. The product corresponds to the design value of the dose by a withdrawable bellows gangway, thus forming an escape route which rate. has an air atmosphere. The cabin leads to helium-filled tunnels. From Table 1 Decontamination of the turbine blading

Vâr. В Var. С

Ho pre-decontamination Pre-decontamination of the mounted pre-decontamination of the mounted

turbine blading without removal of turbine blading with partial removal

material of material

- Steam blasting with admixture of - This decontamination sequence" is

wetting agents and inhibitors for being developed by the HR.

the final treatment It is intended to remove пТх. KO^jm

- Drying with hot air from the TZH blade foils and only the

oxide layer from the Nimocast 713 LC blade

foils. After this treatment of the turbine

blading, the entire rotor and turbine

blade carrier are subjected to a high-

pressure steam blast.

Decontamination factors of 3 and Decontamination factors for the blade foils Fig. 1 Deposition activities on the gas turbine after a residual loose surface contami­ of between 45 and 1500 (depending on the blade

nation of & 10"^Ci/cm^ may be row) and a residual loose surface contamination 6 full-power years -3 ? expected. of£10 ytiCi/cnr may be expected.

Remote-controlled disassembly Manual disassembly of the turbine blading, employing auxiliary shielding

of the turbine blading and personal protection gear

Oecontamination of the disassembled TZH turbine blades

- The blade foils are treated in an oxidizing alkaline The blades are treated in an oxidizing alkaline

solution until the specified degree of contamination, solution until a max. of 5^m of material has

e.g. decontamination factor 200, is attained (removal been removed from the entire surface (incl.

» UO^m). This is followed by treatment of the blade blade root).

roots for a shorter time (corresponding to a removal

of max. 5^m).

- Rinsing Rinsing

- Drying I Drying

Dccontamination of tne disassembled Ninocast 713 LC blades and heat shields

Removal of the oxide layers in an oxidizing alkaline solution

at 80°C

Electrolytic removal of nateria) from the blade foils until the specified contamination factor is attained.

Electrolytic removal of material from blade roots (max. 5yan)

Rinsing

Drying Fig. 2 Layout of the maintenance building 124

Fie. 3 Vertical section of the turberachîne 13 и i» COOLANT CESHISTRY 0? THE ÂD7ASC3D ! CARBON БЮХПЕ COGLSD RSACTCB m R.L. FAIRCLOTH, E.S. KCSSTQQD, E.A. FSICS О «rî-jte vu~t A3RE XI Earrfell

r-tar t_>: itjv аггч- UK иг lorrabictwa I I I I I1 I I I iTTTH The targe scale provJucccr of eiertricirj by urariuni ässicn: lus beea achievied in the СшвеЗ! Kmgdcro. т*г Ш1 exdusively by reactors «.hic!: are gas. cooled ar^rrcderatedby grar^rite-Irraisway meusecf4irairium:wiÉi «л***1 шгТЫ; of ТЧ SÄ crose to the natural isotonic content was pessäSe- Ortee ce choice of graphite as. naxhntcr haii beem made 1>тЛГй Uà*\f*r OF then the selection of a suitable gas в? transport beat front the cere to ie steam generating: ецшгяпеас was. limited ar.d. in fait, only two- bave bee? iiiecccei as. basics suitable chemicall and aiirinr grcpenies^ namely О» FIPL со**» helium and carbon dioxide. Tbe erst of Oese has Йе disadvantage cf beng aipeashe bue bas. ai irijrir heat Cfurqiftt] of pria, circuit. prtbVj't and Ink*)- t**tl on ВРи со***** transfer capability and is fursiarnertaiiy irenu its reacnvir» being controlled enere« by the lend! cf impurities; such as hydrogen and water. 4 ith tbe closure of be OECD Dragoc High: Тепрегашге Reactrr Project 1UJ interest in telium coding for nudear plant bas faded im tbe OC although: dev elerment wot*, is continuing; in: otler countries, cotabty is Germany and tie USA. The ahemarKe codant, carbon dioxide, wiact ischeap- but chemically reactive is used in tbe first gene.-.iricc Ma-gneji power stancus- and in; ас Cociioercjai: Fig. -\ Eargraph for the time required for the major TM inspection Var.A Advanced Gas Coded Reactor (CAGRi design h tbe Vagrrox stances. mr;rjl агагаиш; merer is dad: in an: alioy of magnesium a^ii iunaniun (&ucc »'iùob зесагге Vagr-jx is ibsiveibamiriie cooiancgas. eacrges from ±s core at a tersrerarare around. 550-"C- & öe iok bigbi? ratei CACS diesgi the fue£ consiss of enricbej uraniiira irt r>e гогз of ic.iide ел^еа ст. stair-'ess steeî; ne gas ouie; tsniperacire et öe areis increased to arourvi 600X.

.^lihcu-îb carbon dioxide is a rdarrvdy stabie Kibsarrce its bebaviour im tbe high: energy апйлптг. en^irocmer-t of the reactor core is ccinjiex- Ai=oii aï materials, umiergc change where es^osei to> high: ecersy radiaricn ari carbon cioxide is ce exception- to this r^Ie. A-tecuga esseac'aily static irr. rie icacïaccir. Йеи short lived species are produce! w Ьть'Ь. oq coilisca with tbec bun tbe prebbm *as егТесглй; sotted b; caeäseeeer? tàac the addinoQ of methane to the coolant at levels of cp to a tew hundred parts per nriiBra: ggcù± gcotSa^t a; dramatic reduction in the tnoderaror oxitxioti rate. Such addiii.'ns cf CKthace increase tbe pc^ssibiäri of producing carboaaceoLS déposas on hot surfaces wither the circuit and this is partircicrt secccs whee tbe fuel pins are involved since the resulting; toss cf bean transfer capabfiity can: bad to 5ic£ pia overheatings Tbe coolant strategy of the CAGRSv tiererora. concerns the choice of carbon «Sowie based cociü3s it teens of additions of carbon raoroxide and methane, and aiso witer. so as to provide a sacsractocy глосегаалг Efè ».hilCT iSf amf rVnmwcr; r>f rkV nf nirfvyi^nr^ .-y-4>gfn-r- г^»1тэтг frfv^t ^rrt» S*t*." pins. гот /э уг

—S> The purpose of this paper is to provide an insight into the current thinking about the nature of the complex methane destruction rates coupled with energy balance cocstderariens tnc_a:e that these are the prirrrvy chemistry associated with the CAGR coolant and how this chemistry influences the rate of deposition onto radiolytic steps and that methane breaks down by reacting with these specie:- rather than by the «frect actioa the fuel pin surfaces and the rate of graphite moderator oxidation. The application of these ideas to the of radiation. prediction of the behaviour of a CAGR core with particular reference to the calculation of coolant The core pressure of a CAGR isJ I bar. and oxygen atoms are quickly relaxed a tier groend state

composition within the porous moderator structure at points remote from the surface, is outlined and the use 3 (0(" P)) by collisions. (0( P)) reacts with CO to reform CO;, and with hydrocarbons (inducing сеглазг) to of all this information to define a satisfactory range of coolant composition is also described, f ^ , produce free radicals (incuding methyl). . • &>»Cf CO^" undergoes solvation very rapidly with more CO. and CO. to form clustered ions (ssch as Coolant chemistry and production of carbonaceous deposit C704" and C3 О Г )- These can react with water to form hydratsd dustered ions, and with methane and other

During the steady running of a CAGR, only four gases in the coolant are controlled . CO, and CH4 are hydrocarbons to produce CH^ and CH,CO~. among other spedes.Thehydrated ions and those containing added, to balance losses by leakage and radiolytic destruction respectively; and CO and H.O are removed by combination with oxygen and drying. In very simple terms, an equilibrium is maintained against the reaction: CO react more slowly with methane, resulting in the fan that water and CO reduce the rate of methane destruction; water also reduces the tendency to form deposit. The acyHum ion СНГСО can be hydrated to

3C02 + CH, - 4CO + 2H,0 form acetic acid which has been observed in large quantities in the driers of the prototype AGR at Wmdscale A typical coolant may contain I%CO and löfvpmCH, (vpm = volume parts per million), along with (WAGR).

300vpmH2O. In addition hydrogen is present at around 200vpm, formed by the water-gas shift reaction:

H20 + CO s H, + COj Once methyl radicals are formed, they build more complex hydrocarbon spedes by a variety of reactions Radiation converts these basic constituents into a host of organic compounds in amounts from IOvpm (for that can be categorised as hydrogen atom transfers, radical thermal decompositions, radical ethane) downwards. There is no limit to their number and increasing the detection sensitivity increases the disproportionanonAecombinarions and radical additions. variety seen. Some are created from contaminants such as lubricating oil and organic debris introduced during Hydrogen atom transfers are reversible, and the rates of the fi Tward and reverse reactionsar e not usually the construction of the reactor, but most are built from methane by radiation-induced reactions. well known although the equilibrium constants can be calculated from the thermodynamic properties of the Under the influence of radiation, the coolant mixture can interact with metal surfaces, in particular fuel reactants. The two most important in CAGR coolants are: element cladding, to produce deposit consisting of filamentso f carbon; these may pose a hazard by restricting CHj + C;H» & CH. + с.н_. the (low of heat from the uranium dioxide to the coolant gas causing the fuel to overheat both during normal

operation or under certain fault conditions. Three processes are in balance: the firsti s the actual deposition С H, + С,Н4 si CH4 + C-Hj process where labile hydrocarbon species (probably vinyl radicals) stick to the surface and degrade to carbon. The second is radiolytic oxidation, where activated species (probably COj or related species) gasify deposit. Fortunately, quantitative application of the chemical scheme is not particularly sensitive to the rates of Thirdly, CO, can be thermally activated to oxidise carbon. The firstan d third of these processes are catalysed these reactions. by transition metals, and in particular the components of the fuel clad. Each process will be considered in The main radical thermal decomposition is: turn, beginning with deposition. C,Hj - C.H,+ H- Carbon on fuel pins is mostly derived from methane, and this has been demonstrated by tracer experiments with "C; however, there is also evidence that CO can contribute, and this will be discussed later. The fact that ethylene is produced in an activated process explains the general observation thai, while the The dilute organic mixtures which comprise the CAGR coolant will deposit carbon onto metal surfaces if alkane content of reactor coolants does not change with increasing gas temperatures, the ethylene content heated sufficiently, but not at temperatures attained in a CAGR. However, the radiation causes carbon to 2 rises dramatically. deposit very gradually (at rates of a few mg/cm /year), with a speed that varies with the gas composition, fuel can temperature, gas velocity and radiation intensity; in the absence of radiation very little deposit forms. The majority of the radicals in the coolant are lost by disproportionation and recombination, and an essential step in the production of hydrocarbons from methane is another recombination: The ultimate source of the carbon is methane, but it undergoes several gas phase transformations before reaching the surface. In the absence of methane no deposit forms, but the deposition rate cannot be correlated CH, + CH, - C:H6 simply with the methane destruction rate, and furthermore, less than 5% of the methane destroyed ends up as carbon on the metal surfaces. The route which methane and radiation energy travel to produce carbon can be For completeness, reactions are included where free radicals add to ethylene to produce larger radicals. represented by: These are unimportant in the description of deposition, although they provide a route for the formation of the large hydrocarbons observed. energy-CO,+ -C 0+ — CH' — C,H — C.H' -C,H. -C,H,' -deposit Two sorts of surface reaction can also occur: catalytic hydrogénation or hydrogenolysis of stable The gas phase chemistry is complex and interlocking and is represented schematically in Figs 1 and 2. The hydrocarbons, and the deposition reaction itself. main forms of ionising radiation in a CAGR core are gamma, electrons and fast neutrons, and gamma is most Whatever hydrocarbon is added to coolant mixtures, it reacts to form the whole spectrum of species from important from a chemistry standpoint. Its interaction with CO. is well understood', and causes ionisation to methane upwards, but the free radical chemistry alone is inadequate to predict this. However, if hydrogen CO* in various electronic states, and dissociation to CO and O, also in various forms. Measurements of reacts catalytically on metal surfaces with species like ethane and ethylene, the correct gas mixtures can form. The existence of these reactions absorption of energy by the gas to form positive ionic clusters as described in the previous section. Certain of these ions have sufficient energy to gasify carbon to produce carbon monoxide, which is also a product of C2H4 + H2 — CjH4 CO, radiolysis.The rapid recombination of the oxidising species and CO is responsible for the apparent

radiation stability of C02. Radiolytic corrosion takes place mainly within the graphite pores rather than at the C2H6 + H, - 2CH4 exposed surfaces because the oxidising species are very short lived and_ at reactor pressures, they may only is tentative, and the same effect might also be produced by the reaction of hydrogen with free radicals. At travel a few microns from their point of production before being neutralised. Carbon monoxide, therefore, acts present neither alternative seems preferable, although catalysis implies that the gas chemistry is sensitive to as a getter for the oxidising species and, as a result, serves as an inhibitor of the graphite oxidation process. In the metal surfaces present, and this could explain the large variations in organic concentrations in some fact a level of around 1% has been found to give the required level of inhibition for the early Magnox reactors. experimental facilities. In the case of the CAGR design the degree of inhibition afforded by CO on its own is insufficient, however, The formation of carbon on surfaces is undoubtedly catalytic, even when initiated by radiation; for to provide a satisfactory reduction in the rate of moderator attack and the addition of methane to the coolant instance, very little forms on copper or silica, and large amounts form on nickel dispersed on alumina. This to give additional inhibition is necessary. Methane serves as a getter for the short lived oxidising species but. means that deposit may be controlled by surface treatments. On normal fuel cladding, the hydrocarbon more importantly, these gas phase interactions produce a carbonaceous deposit on the internal pore surfaces species adsorbs on small metal particles where it is dehydrogenatcd to carbon which dissolves in the metal. of the graphite which behaves in a sacrificial manner, being oxidised in preference to the underlying graphite. Provided the dehydrogenation is exothermic carbon migrates through the particle and precipitates on the This process of carbonaceous deposit production is closely related to that described in the previous section cooler surfaces. Here it extrudes a filamentwit h the particle lying at its base, or rising on its tip. The which leads to deposit being formed on fuel pin surfaces although, because of the considerable differences in quantitative chemistry of this process is unknown, but it can be described as diffusion to the surface with a flow characteristics and temperature distribution in the two cases, the more detailed chemistry is unlikely to be fixed proportion of species producing carbon; the proportion depends on the surface material. Ethylene, the same. For simplicity, however, the scheme of the reaction leading to the corrosion of the moderator may acetylene and vinyl radicals are all capable of depositing, but experiments show that the deposit precursor be represented by does not survive long away from radiation. This narrows the choice to vinyl and acetylene; however calculations show that acetylene is produced much too slowly to account for the measured deposits. CO, - CO + (O)

These chemical ideas have been applied to experiments, in the Materials Testing Reactor DIDO at Harwell2, designed to study the phenomenon of carbonaceous deposition from carbon dioxide based coolants. where (O) is the short lived active species. If this species is produced sufficiently close to the moderator Some 100 experiments have been carried out and the observed trends and the corresponding model surface reaction with a carbon atom occurs predictions are summarised in Tables 1 and 2. The major difference lies in the CO dependence, and there are (O) + С - CO three possible explanations. Firstly, the extra carbon may be produced from the direct radiolysis of CO, which is known to form C,0, (carbon suboxide) which will polymerise and degrade to carbon on surfaces. Secondly, leading to the overall reaction it may be a purely thermal process where CO disproportionates to carbon and CO, (the Boudouard reaction), catalysed by the spinel oxide layer on the fuel cladding. Thirdly, it may involve interacton of vinyl with CO to CO, + С - 2CO produce acryloyl radicals It follows from this simplified description that the rate of radiolytic oxidation of the graphite will be CjH, + COs CJHJCO proportional to the flux of oxidising species to the surface of the pore walls and hence is relatively insensitive to temperature but is approximately proportional to radiation intensity and gas density. It will also be which can survive longer in the gas phase resulting in more being able to reach the surface and produce critically dependent upon the pore size distribution since only active species produced within a certain deposit. There is no agreement yet on which route actually occurs in practice. distance of the pore wall are able to reach the pore surface. Considering a single pore it follows that, if this A further important factor affecting the behaviour of deposits is their removal by cither thermal or pore is small enough, inhibition of the oxidation process by either methane or carbon monoxide is not possible 1 radiolytic oxidation. Thermal oxidation has an activation energy of 300—400 kJ.mole" , but the rate, itself, since all reactive species reach the surface. For a larger pore both CO and CH4 are acting as getters for the depends on the history of the deposit. When the carbon has formed from gases containing relatively high active species and also products of the CH, decomposition reach the pore walls and produce sacrificial concentrations of CO, thermal oxidation dominates the situation above 700°C and deposit disappears. carbon. For very large pores a net gain in weight is achievable. However, carbon formed in low CO gases is resistant to oxidation and carbon can still form at temperatures Analysis of the radiolytic oxidation of graphite in terms of single pores is clearly a gross simplification up to 800°C. Carbon monoxide also exerts an effect on the rate of oxidation for all deposits. Radiolytic bearing in mind the known complexity of the graphite structure. It does however serve to explain all the oxidation is temperature independent, and only appears appreciable (by comparison with the other two observed features of the process5. processes) at low temperatures. It is the same reaction as moderator graphite oxidation, but is much more rapid since the deposit has a large number of small pores (see next Section). Carbon monoxide also has an Since the volume of a pore determines the number of active species produced within it, it could be assumed important inhibiting effect on the rate. that the pore growth rate will increase with time in an exponential manner. However, because the small pores are more reactive than the larger pores it follows that the pore size distribution will change during the process Graphite corrosion processes of graphite burn-off. This will result in an increase in the effectiveness of inhibition and a departure from the exponential rate of growth. Models have been developed based on these ideas and are used to predict the way It may be stated at the outset that thermal oxidation processes are insignificant at the operating in which the rate of moderator oxidation is expected to change with time1. Some parts of the graphite structure 3 temperature of magnox and AGR moderators. The AGR moderator is maintained below 450 C, i.e. well might be expected to be difficult to inhibit; in particular closed pores with narrow entrances. This results in the below the fuel channel outlet gas temperatures, by allowing part of the coolant returning from the heat fact that the reaction rate cannot be reduced below a certain minimum rate (the terminal rate) even in the exchangers to flow down the outside of a graphite sleeve which supports the hot fuel and through interstitial presence of large concentrations of inhibitors. passages between the moderator blocks. At the bottom of the fuel channel it combines with the main flow which passes upwards over the fuel within the sleeve. Corrosion of the moderator is, therefore, caused solely An important difference in the relative efficiency of CO and CH, is a consequence of their points of production and introduction. Carbon monoxide, although less effective as an inhibitor than CH4, is always by oxidising species that have been produced from the radiolysis of C02, the primary process being the present throughout the graphite structure as a result of the radiolytic oxidation process. Methane, on the other of its deficiencies, been able to provide a reasonably good representation of the WAGR situation but its hand, must be introduced to the pore structure from the outside and the depth to which it may penetrate application to CAGRs must be approached with caution, involving as it does, a change in pressure (20atm О depends upon its rates of diffusion and radiolytic breakdown. To improve the accessibility of the moderator, 40atm) and other gas conditions (temperature and flow rate). gas access holes have been drilled in the bricks; the possibility also exists of increasing the rate of penetration The interactions of the coolant composition boundaries are shown, in schematic form, in Fig. 6. Two other by imposing a permeable flow upon the system. This, so called, ventilation process may be achieved by constraints are included in addition to those of moderator life and deposition rate. These are, firstly, that the inducing a small pressure drop across the moderator bricks. rate of methane destruction must not be too high since the removal of the water produced in the process may In order to provide information which the reactor designer can use to predict the life time of the moderator not be possible with the existing coolant control plant. This puts a limit on composition in the low CO and

many experiments have been carried out, in a wide variety of irradiation facilities, to measure the rates of high CH4 direction. The other constraint arises from the fact that CH4 is produced radiolytically in the core of oxidation of graphite specimens. Since these specimens were of various sizes, some ventilated, others not, it the reactor and the equilibrium level reached (ie when its production rate equals its destruction rate) represents

has proved to be necessary to devise a model to calculate the mean gas composition, in terms of CO andС H4, the lower limit achievable; this tends to increase with increasing CO. The result of an analysis of this kind is to within each specimen9. The oxidation rate of each specimen could then be related to the calculated gas define a coolant 'window* which will serve to provide a satisfactory reactor performance. composition and this analysis led to the reaction rate/compositional plot shown in Fig. 3. This figure In conclusion it is worth nUting that the CAGR reactors at Hinkley Baint have now operated for the demonstrates clearly how the addition of methane is relatively more beneficial at low CO concentrations and equivalent of around 600 full power days and have a current coolant composition of 1% CO and 165vpm how the terminal rate is achieved even with high concentrations of inhibitor. The model is based on the CH4. No deposition is occurring on the fuel pins and graphite samples taken from the core indicate that the diffusion, permeable flow and radiolytic destruction of methane and the diffusion and permeable flow of water rate of oxidation is in agreement with predictions. There is, therefore, every confidence that the CAGRs now- and CO which both arise from the methane destruction process. The production of CO from the oxidation of operational or being commissioned wfli help satisfy the electrical needs of the United Kingdom for the next the moderator is also taken into account. The rate of methane destruction is a function, not only of the quarter of a century. methane concentration, but also of the levels of water and CO, a feature which is also incorporated into the model. The calculated variation of the gas phase composition with distance from the surface of an Reference] experimental graphite assembly irradiated in WAGR is shown in Fig. 4; it will be noticed that a considerable 1. С Willis and A W Boyd, Int. J. Radiât. Phys. Chem. 8 71 (1976) reduction in theС H4 concentration occurs across the assembly although this is somewhat compensated for, in 2. M J Bennett, R L Faircloth, R J Firth, M R Houfton, К S Norwood and H A Prior, 'Gas chemistry in inhibition terms, by the concomitant increase in CO level. There is, therefore, a considerable degree of nuclear reactors and large industrial plant-' Proceedings of a conference held at the University of self-inhibition. In the calculation of the in-pore gas phase composition it is assumed that the graphite structure Salford, 21-24 April, 1980. p81. Heyden 1980 can be defined by two bulk properties, namely open pore volume and diffusivity. The use of bulk properties in 3. J Shennan, ibid. p98. this way is obviously a simplification of the real situation but a knowledge of the variation of these parameters 4. J Best, ibid. pMl. as burn-off proceeds also permits the calculation of how the degree of inhibition may change during the life 5. R L Faircloth, ibid, pi48. time of the reactor'. These ideas.in conjunction with the graphite oxidation concepts described above have led 6. A Blanchard, ibid, p 134. to the development of a comprehensive model for the prediction of CAGR moderator life times taking into account the effect of changes in pore size distribution and gas transport properties. Because of the importance of the methane diffusion model, both in its application to the analysis of experiments and to the prediction of whole core behaviour, experiments using composite graphite specimens, Table 1: Deposition trends in DIDO giving a 10cm path length, have been carried out in the DIDO reactor. The reaction rate profile along the length of the assembly of specimens was related to the corresponding calculated gas composition and the results from the three experiments carried out to date are shown in Fig. 5. The continuous curves in this figure are derived directly from Fig. 3 and the agreement between the experimental points and the curves gives response considerable confidence as to the validity of the model. The small discrepancies, particularly at the low CH, parameter end of the upper curve in Fig. 5 are, in fact, largely explicable by the variation of the CO concentration, from time cc rime point to point, the trend being for the CO level to increase as the CH, concentration is reduced. radiation dose rate ccD Coolant strategy gas pressure ce p but confused. The coolant strategy for the CAGRs depends,therefore, upon the suitable choice of coolant composition gas temperature t until 650°C, then i which will give a satisfactory life to the moderator whilst not causing any significant level of deposit to form on the fuel pins. Ideally, the understanding of the basic chemistry leading to moderator corrosion and fuel pin volume flow rate profile changes rr. „Oft deposition should provide methods of prediction which would make the choice of coolant relatively CH4 mole fraction straightforward. Although, as indicated in the previous section, this is probably true for the case of the moderator the situation is less satisfactory for deposition. The chemistry in this case is now reasonably well CO mole fraction CO rr -0.6 understood; its application to the turbulent gas situation of a power reactor has just started but a complete HjO mole fraction description of the complex processes involved may be difficult to achieve. As asserted earlier the processes are very sensitive to catalytic influences and hence to metal surface conditions. The life times of certain key can material dependent species may be very temperature dependent and hence afiected to a large extent by the detailed properties of the boundary layer. For this reason the predictions which can be made at the present time are based entirely upon the observed behaviour in WAGR on which a rather empirical model has been based. This has, in spite н,со;

hv + CHj Table 2: Model response •> со,

C2HJ,C2HJ parameter response сго;- CH,CO*, C,HJCO-. с,н,со- time oc time

radiation dose rate ccD'3 1 H]C,0J* gas pressure ccp gas temperature /Ш* = 70 kJ-mole- сн'\ i volume flow rate sligbt changes

CH4 mole fraction г/ CH, *сн. сн,со*, с,н,со*. с,н,со* / CO mole fraction xco

+ Q o —•C^HJ.CJHJ H20 mole fraction X HTO can material dependent CHJCO*. C,H,CO*. с.н,со-

H.ÇA*

4- C2HJ,C2HJ •*—— СзОд;

СН,СО*. с,н,со\ с,н)со**у

CH3 н,с,о;

CH,CO*, C,H,CO*. C,H,CO- CH. он

нгс,о4*

C.H„ C.H. C.H, : deposit <=Л ОН CHj denotes important species с1нт denotes by-product СИ, СНО

CjH C,2"6H " C2HS C2H4. 3

СН3 denotes important species deposit

CjH, denotes product of minor reaction CH, оно denotes by-product

Figure 2: Hydrocarbon chemistry Figure 1: Production of free radicals C.H. C,H, CA FIG.3. Calculated Oxidation Rate vs Methane for Various Carbon Monoxide Concentrations. 131

Q I I 1 I I I I I J 0 100 200 300 400 500 600 700 600

CH4 Conctntration (vpm)

Flo.5.RESULTS FROM LONG DIFFUSION PATH EXPERIMENT The compressor which provided gas flow for this purification system «as 132 located downstream of the entire system and was lubricated by oil. PRIMARY COOLANT CHEMISTRY OP THE PEACH BOTTOM AND 1.2. Fort St. Vrain PORT ST. VRAIN HIGH TEMPERATURE GAS-COOLED REACTORS The Fort St. Vrain HTCR is a 330-HW(e) plant. The core outlet temper­ R.D. BURNETTE, N.L. BALDWIN ature is 1050 K, and helium pressure is 4.8 MPa. Its primary coolant system is based on the same fundamental principles as the Peach Bottom reactor. General Atomic Company- However, the Fort St. Vrain reactor does incorporate a number of new design San Diego, California features, the most prominent of which are (1) a prestressed concrete reactor vessel (PCRV); (2) once-through modular steam generators with integral USA superheaters and reheaters; (3) four steam-driven axial flow helium circu­ lators with water-lubricated bearings; (4) prismatic fuel elements with im­ proved fuel particles; (5) fibrous ceramic insulation adjacent to the PCRV liner; and (6) lack of a fuel element purge system. ABSTRACT Fission product and chemical impurity control is provided by a 12Z/h The chemical impurities in the primary coolants of the Peach Bottom and bypass purification stream consisting of (1) a high-temperature-activated Fort St. Vrain reactors are discussed. The impurity mixtures in the two carbon bed and a sintered stainless steel filter which removes condensable plants were quite different because the sources of the impurities were dif­ fission products and particulates; (2) a molecular sieve bed which removes ferent. In the Peach Bottom reactor, the impurities were dominated by H2 H2O al,d COj; (3) a liquid-nitrogen-cooled carbon bed which removes xenon, and CH4» which are decomposition products of oil. In the Fort St. Vrain krypton, CO, N2. and CH^; and (4) a hot titanium sponge bed which removes H2 reactor, there were high levels of CO, CO2» and HjO. Although oil ingress at and tritium. Peach Bottom created carbon deposits on virtually all surfaces, its effect on reactor operation was negligible. Slow outgassing of water from the 2. COOLANT CHEMISTRY DATA thermal insulation at Fort St. Vrain caused delays in reactor startup. The overall graphite oxidation in both plants was negligible. 2.1. Peach Bottom

1. PRIMARY COOLANT SYSTEM The Peach Bottom reactor was started up in January 1967, and it achieved full power in June 1967. The coolant impurities during rise to 1.1. Peach Bottom power and steady state are given in Table 1. The steady values persisted throughout the life of the plant (with the exception of occasional transient The primary helium coolant system in the 40-MW(e) Peach Bottom high- hydrocarbon ingresses) until shutdown on October 31, 1974. Only one reactor temperature gas-cooled reactor (HTGR) consisted of a reactor pressure ves­ shutdown and rare power reductions were due to excessive hydrocarbon impur­ sel, two steam generators, two oil-lubricated helium compressors, piping, ities, and reactor startups were seldom delayed because of impurity outgass­ and auxiliary equipment. The steam generators and helium compressors were ing. The technical specifications for full-power operation were 10 ppmv of arranged to form two parallel loops for circulation of the helium coolant at CO, 2 ppmv of CO2, and 2 ppmv of СНд. References 1 and 2 present the Peach a pressure of 2.4 MPa. The core outlet temperature vas nominally 1000 K. Bottom impurity data.

Within the reactor, approximately 802/h of the main helium stream vas 2.1.1. Moisture Measurements. Except during startup, moisture was not de­ withdrawn from the main coolant system through the tubular fuel elements. tected in the Peach Bottom helium coolant by ordinary instrumentation. The This purge stream went to the low-temperature trapping system for removal of moisture levels during operation of core 2 were inferred from measurements fission products. In addition, helium was withdrawn from the steam gener­ of gaseous tritium in the primary circuit. The tritium monitors distin­

Н Н 2 2 ator tube sheet baffles and was purified by passage through a chemical puri­ guished the HT species from ТО, and by assuming that the HT/ ТО and H /H 0 fication system (Cu/CuO catalyst bed followed by molecular sieve beds) which ratios were equal, the H20 concentration could be calculated from measure­

Н 2 Н ц removed gaseous contaminants such as H2O, H,, CO, and COj- The total helium ments of HT, ТО, and H - Figure 1 plots HT/ ТО versus Р and shows that PH Н flow rate through this system was about 20Z/h. The purified helium then over a wide range of , the HT/ ТО ratio at steady state was 20. There­ joined the fuel element purge stream for gaseous fission product removal. fore, at 10 ppm H2, the moisture concentration was estimated to be 0.5 ppmv. During hydrogen injection (Section 4), the HT/НТО ratio increased. These data are in close agreement with those obtained at Dragon at steady state Work supported by Department of Energy Contract DE-AT03-76ET35300. and during H2 injection (Ref. 3). 2.1.2. Oil Ingress. The first suggestion of a possible oil leak Into the Water ingress before and during initial nuclear startup has been primary circuit was the observation of persistent H2 and СНд impurities. frequent and sometimes large. When ingress was large, it was thought that Toward the end of core 1 operation, there were occasional high hydrocarbon water entered the fibrous ceramic insulation and condensed on the water- concentrations (consisting primarily of up to 100 ppmv of СЦд in the helium cooled PCRV liner. Dry-out prior to restart was often quite slow because coolant). This transient source of hydrocarbon proved to be oil ingress of the slow diffusion of water out of the insulation materials. Once the from an oil demister/filter which removed oil vapor and oil mist from the reactor attained high steady power, dry-out of the system was rapid, and discharge of the helium compressor in the purification system. The filter the purification system reduced the impurities to acceptable levels. was saturated with oil, and it was speculated that on occasion, liquid oil or mist was injected into the reactor. This source undoubtedly also con­ Table 2 (Refs. 6, 7) summarizes the coolant impurity data obtained to tributed to steady oil vapor ingress, since the piping downstream of the date. Caseous impurities are given at each power level, since the reactor filter/demister contained liquid oil with a finite vapor pressure. This was brought to power in stages. The concentrations listed were obtained at steady source could have added 10 to 100 kg of oil vapor per year to the the beginning and end of each time or power Increment. In general, the con­ primary circuit. centrations were higher at the beginning of each power increment, because of increased outgassing at each temperature increase. Another source of oil ingress may have been the oil—lubricated main compressors. A small amount of back diffusion of oil vapor past the All impurities decreased (with the possible exception of CO) as the helium-buffered labyrinth seals could have contributed to continuous reactor was brought to higher powers and temperatures, indicating the com­ low-level oil ingress. bined effects of graphite outgassing and decreasing amounts of water vapor leaving the insulation materials. In general, whenever moisture was measur­ The final conclusive evidence for oil ingress was the carbon deposit able (indicating ingress), the H2/H20 ratio was low, indicating low overall

(up to 0.2 mm thick) which coated virtually all the primary circuit metallic oxidation reactivity of the core. The persistent concentration of C02 indi­ surfaces. The carbon scales were layered, indicating possible periodic cates possible contributions from the radiolytic shift reaction H2O + CO ingress. The deposits had no discernible effect on the heat exchange = H2 + C02. properties of the steam generators nor on the metallurgy of the underlying The detection limit of the dew point meters for H2O is about 1 ppmv. A structures. Chemical analysis of the deposits revealed that they contained single measurement of moisture was made at 50% power using the HT/НТО method carbon (80% to 100%), iron (2% to 3%), and traces of chromium and sulfur described above. The calculated P^O was 0-2 ppmv, indicating that dry-out (Ref. A). Approximately 80% of the cesium and strontium plateout activity of the primary circuit was indeed possible. on the steam generator tubes was associated with the carbon deposits (Ref.

5). The density of the carbon deposits was estimated to be 1.0 g/cm^ and the average thickness 0.05 mm, making the total amount of carbon deposits The anticipated consequences of moisture ingress into the Fort St. about 100 kg, which was consistent with the estimated steady-state ingress Vrain HTCR are (1) oxidation of graphite structures, (2) hydrolysis of rate. exposed carbide fuel particles, and (3) oxidation of metallic surfaces.

Transient high levels of oil vapor or hydrocarbons in the helium The degree of graphite oxidation to date has been minimal because the coolant contributed to early failure of the moisture monitor cells. These impurity concentrations have been reduced to low levels for reactor opera­ cells were the P205~coated electrolytic type (similar to Goldsmith hygrom­ tion at high temperature. For example, whenever the core outlet temperature eter cells). The hydrocarbons coated the cells with resinous deposits, is >922 K, the maximum allowed impurity concentration is 10 ppm total oxi­ which caused their response time to increase. Cells with excessive response dants, which includes the impurities CO + C02 + H20- For temperatures below times were replaced. 922 K, the limit is based on a moisture content which is allowed to increase with decreasing temperature. Thus, the reactor operator must bring the 2.2. Fort St. Vraln reactor to power slowly and in stages to ensure that the impurity limit is not exceeded. As a result of this limit, most of the moisture exposure has The coolant impurity mixture at Fort St. Vrain is dominated by H^O and been at relatively .low temperatures, where graphite oxidation is not a an problem. the graphite oxidation reaction products CO, C02. d H2- The major source of the water is typically the water-lubricated bearings of the main circu­ lators. Hater is normally prevented from leaking up the shaft hy labyrinth The total amount of oxidation to date may be inferred from the quantity seals and a buffer helium system. During startup and shutdown, the pressure of gaseous carbon species removed by the purification system (Table 3). An -4 balance between the buffer helium and the primary circuit pressure is some­ overall oxidation of 1.4 x 10 fraction after 174 effective full-power days times upset, allowing water to flow into the reactor. (EFPD) can be calculated by dividing the number of moles of gasified carbon (CO + CO2) by the number of moles of carbon in the core and lower reflector. This calculation is conservative because it assumes that all gaseous carbon Bottom was performed while the reactor was at 552 power. Hydrogen was injec­ is from oxidation rather than outgassing. Outgassing of the entire Fort ted at 3.35 liters/min In two pulse tests of 3.5 and 2.5 h. A steady-state St. Vrain graphite inventory, i.e., 5.7 x 105 kg, could account for a large test at 1.4 liters/min which lasted for 72 h was also accomplished. At the part of the gaseous carbon observed to date. end of the steady run, the reactor was brought to 8051 power. Figure 2 pre­ sents data from these tests. The data indicate that the concentration of

Hydrolysis of exposed carbide fuel particles would cause a large methane in Peach Bottom was proportional to (PH2) suggesting that high increase in the rate of fission gas release (R/B) and could therefore be concentrations of hydrogen can be tolerated without concern about high easily detected by routine monitoring of the circulating activity in the methane concentrations. These results are in general agreement with those primary circuit. The overall core R/B has remained relatively constant and of the Dragon injection tests (Ref. 3). very low; hence, in-service fuel failure to date appears negligible.

The H2/N20 ratio in the primary circuit has remained quite low during 5. TRITIUM BEHAVIOR almost all reactor operation. Thus, the metallic components in the primary circuit are expected to have oxide films. This Is undoubtedly beneficial The tritium concentrations in the primary coolant at Peach Bottom and because it precludes corrosion processes such as carbon deposition and car- Fort St. Vrain have been measured by grab sample and continuous monitor de­ burization and possible sulfidation reactions. vices. Table 4 gives the results of the tritium measurements. The tritium concentration in the Peach Bottom primary circuit increased by a factor of 3. PURIFICATION SYSTEM PERIORMANCE 10 during core 2 operation. The initial concentration vas roughly equiva­ lent to that expected from He-3 activation (using He-3/He-4 = 2 x 10~^). 3.1. Peach Bottom The increase in the tritium released to the primary coolant reflects the re­ lease from (1) fuel (ternary fission), which is influenced by increased fuel temperatures toward end of life; (2) control materials (B-10 activation); The Peach Bottom helium purification system operated efficiently and (3) graphite (Li-6 activation). The total amount of tritium produced in throughout the life of the reactor. It maintained the circulating activity core 2 during 897 EFPD was about 1900 Ci. The total amount released to the of xenon and krypton at less than 1 Ci and the chemical impurities at the primary coolant during 3 yr of operation was about 600 Ci, for an overall levels In Table 1. Table 3 lists the total impurities removed by the sys­ release fraction of 0.32. tem during core 2 operation. H2 was the major species removed. The overall

core oxidation was <3 x 10~^t assuming that CO and CO2 were generated by graphite oxidation. Near the end of core 1 life (452 EFPD), the CuO cata­ A total of 307 Ci of tritium as НТО was removed from Fort St. Vrain by lyst bed was no longer effective in oxidizing hydrogen, although it still the purification system after 101 EFPD. Another unknown amount vas removed oxidized CO. In_ situ regeneration of the catalyst was accomplished by by the titanium getters, which operated sporadically. During this iniital adding oxygen to the bed inlet. period, most of the tritium was in the form of НТО because of the low Н^ЛЦО ratio. The 307 Ci of НТО is roughly equal to the amount produced by He-3 ~ 3.2. Fort St. Vrain activation.

The Fort St. Vrain purification system has performed well to date, especially considering the large amounts of water removed by the dryer beds 6. CONCLUSIONS (Table 3). An exception to the excellent performance of this system is the The following conclusions can be drawn fron the Peach Bottom and Fort titanium H2 getter beds, which have, been in service only sporadically. These beds are located downstream of the liquid-nitrogen-cooled carbon beds St. Vrain coolant chemistry data: and are therefore in theory exposed to only pure dry helium and R^. In practice, however, the beds have occasionally become contaminated with 1. HTCR systems can be designed and operated with low concentrations nitrogen, which reduces the H2 removal efficiency. This has not caused a of chemical impurities in the primary circuit. large increase in H2 In the primary circuit, and it appears that the excess hydrogen may be absorbed by the graphite components. 2. The concept of bypass purification has been proven.

3. High methane concentration as a consequence of high hydrogen 4. HYDROGEN INJECTION EXPERIMENT concentration is not expected. There was concern that in future HTGRs used for process heat, hydrogen diffusion from the process side to the helium side could cause high levels 4. Using currently available grades of nuclear graphite, outgassing of graphite components is a transient source of impurities during of hydrogen in the primary coolant. It was conjectured that the hydrogen startup. Long-term graphite outgassing is an insignificant source would react with graphite to form high concentrations of methane, which is of gaseous impurities at steady state. an undesirable carburizing agent. Hencs, a hydrogen Injection test at Peach Q> Ш 5. With the exception of contamination of the moisture monitor cells, small amounts of oil Ingress have little effect on reactor operation. o

6. The Fort St. Vrain coolant Impurities have had no measurable effect on fuel, graphite or metal component performance to date. CP

REFERENCES О чсаидь лиги»» wwurajs ф "ramm*- чщи.гая; 1. Burnette, R. D., et al.» "Chemical Impurities In the Helium Coolant at CKT1WL3US the Peach Bottom KTCR," General Atomic Report, Gulf-CA-AI0809, August 23, 1971. —с— 2. Sche'fel, W. J., N. L. Baldwin, and R. U. Tomlln, "Operating History- ta на Report for the Peach Bottom HTCR, Volume I: Reactor Operating History." ERDA Report CA-A13907, Ceneral Atomic Company, August 31, 1976.

Figure 1 - ЕХ/ЕТО vs P_ in РеасЬ Battant 3. Carlyle, K., and С V.-Klnsey, "Results of the Injection of Impurities ^2 Into the Dragon Reactor Primary Coolant," Dragon Project Report 544, June 1969.

4. "Metallurgical Examination of Primary Circuit Components Froo the Peach Bottom HTCR," DOE Report CA-A14506, Ceneral Atomic Company, February 1978. О SSÏ ГЕН* • IKS UNO 5. Baldwin, N. L., B. L. Norman, and W. E. Bell, "Radiochemical О PULSE tMECTlW. ssî KV« Examination of Peach Bottom HTCR Component Samples," DOE Report 5гелг» TIESRRIO». 5SS ИЛА СЛ-Л14495, General Atomic Company, August 1978. REIF К Kussa ТО ÎCÏ slri«s

6. "HTCR Fuels and Core Development Program Quarterly Progress Report for -V- the Period Ending February 28, 1978," DOE Report CA-A14863, General Atomic Company, March 1978.

7. "HTCR Generic Technology Program, Fuels and Core Development Quarterly Progress Report for the Period Ending August 31, 1978," DOE Report GA-A15093, General Atomic Company, September 1978.

IDC

Figure 2. Results of hydrogen iajectioa expertrv-its at Peach Sottas 136

TABLE 1 TABLE 2 GASEOUS IMPURITY CONCENTRATIONS IN PEACH BOTTOM PRIMARY COOLANT IMPURITIES DURING RISE TO POWER AT FORT ST. VRAIN DURING INITIAL APPROACH TO POWER AND AFTER 4600 HOURS AT FULL POWER

Reactor Outlet Gaseous Contaminant(ppmv) Concentration (ppav) Power Power Temperature

2 2 CO н2о Date H N сн4 co2 4 (*> Date (2) (K) H2 н2о CO co2 CH н2/н2о

Core L July 3-6, 1976 2 490 2-15 240-70 0.2 2-1.8 0.6 0.01-0.2 July 24-28, 1976 11 590 45-35 50-25 4-2 6-1 5-3 0.9-1.4 1- 1-67 0 55 1-22-67 0 7.0 0 0 0.3 5 July 28-30, 1976 21 700 40-90 180-80 2-4 4-10 3-6 0.2-1 1-24-67 2 — 15 0.05 0.05 0.5 July 30 - Aug 2, 1976 26 785 85-30 140-76 4-3 6-10 6-2 0.6-0.4 1-25-67 5 — 0.08 0.05 0.4 —5 1-26-67 .3 — 1—5 0.05 0.6 0.3 Dec 10 - Jan 9, 1977 28 895 10 40-4 1.5-3 3-5-1 0-5-0.6 0.25-0.25 1-27-67 1 — 16 0.09 0.5 0.3 —7 Sept 16 - Oct 24, 1977 38 895 3-2 3 1-28-67 5 — 27 0.4 0.4 0.5 3 1-29-67 20 — 50 1.8 1.8 1.7 5 Oct 29-31, 1977 50 945 3 0.2(b) 6-5 1-1.5 0.4-0.2 15 1-30-67 30 — 40 1.0 4.1 1.8 2 — Apr 28 - May 4, 1978 65 980 5-4 <1 10-7 3-2 0.8-0.4 >5 12- 9-67 30 3 0.8 0.7 0.5 <0.05 <1 Dec 10 - Jan 27, 1979 63 945 2-7 <1 1-3 0.5-1 0.2-0.1 >3 12-12-67 100 9 0.5 0.6 0.5 <0.05 <1

Core 2 Limit of detection of dew point moisture monitors — 1 ppmv. 1971-1974 100 10 0.5 1.0 0.5 <0.05 -0.5 'н20 calculated from НТО, HT, and H2 measurements: HjO - SîEE5A^5z2_.

Average steady-state values.

H20 calculated from (НТО/HT) x H2- TABLE 3 IMPURITIES REMOVED BY THE FORT ST. VRAIN AND PEACH BOTTOM PURIFICATION SYSTEMS CHEMICAL REACTIONS DURING NUCLEAR DRYING OP THE Total Concentration (g-mole) AVR PRIMARY CIRCUIT FOLLOWING A WATER INGRESS Fort St. Vrain^3) Peach Bottom (b)

H2O 12,260(c> l,260(d> R. NIEDER, К. VEY Arbeitsgemeinschaft Versuchsreaktor AVR H2 2,475 6,900 Jülich e f CO l,310( > 340< > Federal Republic of Germany f) C02 1.040(e) <120( 250 CH4 240 1 Introduction

Total С (CO + C02 + CH4) 2,590 590-710 In May 1978, after the AVR had been operating for more than

Total О (CO + 2C02) 3,390 340-580 10 years, a small leak was discovered in the superheater of

Total H2 (H2 + 2CH4) 2,955 7,400 the steam generator, the cause of which has still not been (a) 0Z-65Z power through May 1978 (174 EFPD). identified. When the reactor was shut down a volume of some (b) Based on average impurity data, core 2 operation, 897 EFPD. 30 tons of water entered the primary circuit through this tiny (c) Includes moisture removed during power operation. leak over a period of four days. Figure 1 shows a cross-section (d) Estimated from HT/НТО measurements. (e) of the AVR reactor. The broken line indicates the maximum Fraction of Fort St. Vrain core graphite oxidized (~1 yr) - 2350 T 1.7 x 10' = 1.4 x 10-4. Includes fuel element aan d water level. The course of the interruption, the repair and lower reflector graphite, but neglects contribution- of outgassing. maintenance work carried out and the technical details of ^Fraction of Peach Bottom core 2 graphite oxidized (~3 6 -4 restarting the reactor after a period of 15 months for repairs yr) = 460 * 1.6 x 10 < 3 x 10 . Includes fuel element and lower reflector graphite, but neglects contribution of outgassing. have been reported on in detail (Ref. /1/, /2/, /3/). Ref. /4/

contains a list of the activity measurements of the water. TABLE 4 TRITIUM DATA FROM FORT ST. VRAIN AND PEACH BOTTOM

The greater part of the water, approximately 25 tons, was Peach Bottom Fort St . Vrain drained through a valve in the fuel element handling system Primary Secondary Primary Secondary Circuit Circuit Circuit Circuit and pumped into a vessel. During the repair phase the primary 3 3 3 3 Year (yCi/cm He) (pCi/cm H20) (yCi/cm He) (uCi/cm H20) circuit was repeatedly evacuated, thus allowing another 2 tons 1 1 x 10-5 10-« 5 x 10-5 2 x 10-4 of water to be disposed of. In the evacuation process the inner 2 5 x 10-5 3 x 10-4 -- — 3 1 x 10-4 3 x 10"4 — — reactor vessel was heated up externally. Pressure was reduced to 1 mbar in the primary circuit by the vacuum pump. This ponding to a dew point of + 24°C, to prevent the moisture 138 paper will describe how the remaining moisture and other impurities condensing in the colder parts of the circuit. After the were removed during the start-up phase, dealing also with the moisture had been sufficiently reduced the H^O and C0o chemical reactions in the primary circuit during this period. concentrations in later operating stages at temperatures over 500°C had to be kept low enough for the corrosion rate of the 2 Nuclear Drying fuel elements to reraain below the specified tolerance level.

The removal of residual moisture from graphite by nuclear To meet both requirements the core had to be heated up heating is a common procedure in HTRs during initial start-up slowly so that the corresponding concentration limite of processes and after water ingresses. The surplus water is the impurities were not reached. The maximum, gas température either drained off by the ordinary helium purification plant was fixed by the licensing authorities at 650°C fcr the first or additionally by a special start-up filter with a relatively drying period from August 1979 to January 1930. The température high gas purification constant. A barium oxide filter was used was not allowed to exceed 850°C in the second period. at the AVR reactor in the initial start-up phase to remove water from the carbon brick insulation of the core. After the last The drying process was rendered difficult by numerous technical accident a new filter was constructed containing 120 kg problems. Gas tube sections and components had to be usee,

molecular sieve; the purification constant of this filter was that either could not be dried at all or could only be dried —1 —1 0.35 h > compared with that of 0.04 h in the normal helium partly during the repair period. Consequently, ever new

purification plant. small amounts of water entered "the core. Part of the liquid nitrogen bed of the helium purification plant froze on

Figure 2 shows the arrangement of the purification plant in the account of too much CO2» making it necessary to bypass the gas circuits. In the molecular sieve filter moisture only was plant. Another problem was the very high nitrogen concentrations drained to begin with, but later, carbon dioxide too, when another that occurred as a result of a leak in the pneuma"tic control type of molecular sieve was used. All helium impurities were system of the helium purification plant; some of tire valves removed in the normal helium purification plant. in the plant are pneumatically controlled with nitrogen. Minor air ingresses were also detected. For all these reasons it

In nuclear drying operations the moisture in the helium should was very difficult to gain an accurate picture of the not be allowed to exceed an initial limit of 30 mbar, corres­ time/behaviour relationship of the helium impurity concentrations. It is possible that other reactions take place in addition 3 Impurity Concentrations to those mentioned above. The water shift reaction is also

The progression of the impurity concentrations can be seen influenced by radiation effects. in Figure 3. The average gas outlet temperature is also plotted.

Thermal reactor power has not been included, being obviously The interpretation of the CO and CO^ curves also gives rise of minor importance for the desorption processes and the to some problems, as the desorption of graphite increases vith chemical reactions. temperature. Furthermore, both oxides are corrosion products

from the graphite water reaction and can also occur as a

The rL^O concentration reveals a maximum concentration of result of minor air ingresses. It can only ce deduced fron 26,000 ^ubar ( = 4 300 vpm) which decreases steadily. At the CO the curves that the qq~ ratio depends on the temerauJire: end of the first nuclear drying period, the beginning of in a characteristic canner. According to other neasureaents January 1980, the coolant humidity was still 100 yubar, while at the AVR an apparent activation energy of between 4Э and at the end of the second drying period at the end of 50 kcal/mol was found for the Eoudouarc reaction. October 1980, the detectable hygrometer limit was reached with a value of <10 yubar. The water shift reaction

+ C02 ^ HgO + CO The hydrogen concentration runs parallel to the water concentration within a scatter area and is influenced by which hydrogen injection experiments have shovn to occur desorption and corrosion processes. It is rather difficult under reactor conditions, clearly only influences the to explain the curves in detail, because different chemical concentrations to a negligible extent. reactions contribute to the concentrations:

4 Methane - Desorption of graphite High methane concentrations also occurred in the first fev - С + H20 H2 + CO

s weeks of the first drying period, parallel to the high hydrogen - С + 2 H20 C02 + 2 H2 concentrations (Figure 4). The гахтNUR; value vas 0.75 пЪаг - CO + Ho0 —* C0o + H0 =125 vpm. A major ingress of nitrogen fron September onvarcs, after which no more methane was measured, clearly shows that 6 Activity neasureaents the methane concentration was not caused by the desorption The гаге gas and tritium activity of the coolant gas was of graphite but by the radiolytic reaction measured in the drying periods besides the inactive impurities.

С + 2 H2 л^,> CHk Figure 5 shows the total У-activities of the fission rare gases Xe 133, Xe 135, Xe 135 m, Xe 137, Xe 133, Xe 139, Kr 85 a, It is known from reactor experiments (Ref /5/) that the Kr 37, Kr 83, Kr 89, Kr 90, Kr 91 for the initial drying formation of methane under HTR conditions is inhibited by period. The lower curve shows the values that were measured large nitrogen concentrations. directly. The activities depend on reactor power and temperature. Those in the upper curve have been normalised 5 Results of the Drying Processes to 45 Ш and 950°C to allow better comparison with earlier Some individual impurity values are compiled in Table 1, val-es and between one set of data and another. Both curves, supplementing representation of the individual concentration the normalised representation in particular, reveal a distinct curves. A similar impurity level to that found prior to the drop in the rare gas activities. Similar patterns were also accident can be recognised for October 1980. Table 2 shows observed in the period from 19&8 to 1970 during initial reactor the total volume of impurities removed. In the two nuclear start-up. It is still not possible to find a clear explanation drying periods approximately 100 kg of water was cleared of this interesting phenomenon. There are, however, indications as compared to the 27 tons pumped off or evacuated during that the rare gas activities are influenced by inactive the repair period. impurities, e.g. by displacement effects, but other causes are also conceivable. Disregarding minor ingresses ol air and considering that the CO and CO2 impurities occurred as a result of graphite-water Normalised activity towards the end of the first drying period reactions, an equivalent volume of water of 121 kg, likewise was 136 Ci. Total activities measured 128 Ci at a gas temperature cleared by nuclear drying, can be calculated from these two of 950°C prior to the accident in April 1978. The two values impurities. The amount of corroded graphite of some 51 kg concur well, bearing in mind that there are similar fluctuations corresponding to the 121 kg of water was spread over an during trouble-free normal operation and furthermore, that extremely large area of the core on account of the relatively slight discrepancies may arise as a result of normalisation. low corrosion temperatures. Normalised activity during the second nuclear drying period information as regards the design of future high-temperature

diminished in a similar fashion to that in the first period. reactors. The activity fell to roughly one third of the initial value

at the beginning of the drying period. At the end of this References period the total activity measured still amounted to 108 Ci. /1/ C. Marnet, Dampferzeugerschaden am AVR-Kraftwerk mit Hochtemperaturreaktor, VGB-Kraftwerkstechnik The tritium measurements in the primary circuit were relatively high during both drying periods. Tritium is evidently released /2/ К. Krüger, W. Dering, Steam Generator Leakage in the AVR Nuclear Power Station, by the action of hydrogen and water in the gas phase; it is ENS-Conference, Hamburg, 1979 found in the core graphite in a chemically adsorbed condition. Similar exchange reactions were also observed in earlier /3/ K. Krüger, G. Ivens, Die Wiederinbetriebnahme des AVR-Kemkraftwerks nach einem Uampferzeugerschaden, operating states and reactor experiments. Reaktortagung des Deutschen Atomforums, Berlin 1980

7 Conclusions /4/ J. Wahl, W. Jacobsen, Activities and their Origin The experience gained as a result of the accident is that no in Leakage Water following the Incident on the AVR Reactor, ENS-Conference, Hamburg, 1979 great damage was caused to the AVR by the massive water ingress.

The loss of graphite through corrosion that occurred during /5/ R. Nieder, Die Reaktion Graphit-Wasserstoff unter nuclear drying was extremely low, as by far the greatest part den Bedingungen eines Hochtemperaturreaktors, Carbon-Conference Baden-Baden, 1976 of the water was removed by drying. The measurements of rare gas activity were the same at the end of both drying periods as before the accident. It may therefore be assumed that the fuel particles were not damaged. This assumption is also borne out by subsequent examinations. The repair and maintenance and the drying periods could all have been reduced if technical facilities such as a start-up purification Älter and an adequately dimensioned normal helium purification plant had been available. The accident, however, has provided valuable outer reactor vesse 142 Feb'78 Aug 21/79 Dec 12/79 Oct 8/80 950'C 525'C 650'C 850°C inner reactor vessel End2.Nucl. Normal Beg.tNucJ. Endl.Nucl. steam generator Operation Dry. Dry. Dry. core H20 «5 26.000 200 «10 C02 2 3.700 700 12 biological shield Nat 400 800 650 со 600 fuel discharge tube Н2 120 3.500 1.100 200 СН4 1 500 • 100 6 circulator N2 400 26.000 125.000 200 waterlevel fuel discharge system Helium Impurities (/ubar) in the H6-L2 AVR AVR Primary Circuit Table 1 shut down rod

Rg.1 Primary System ot the AVR-Reactor otter the Ingress ot Water Water tank 25.1 to H2O Evacuation 1.9 to H2O Gas Purifica­ 1. Nuclear Drvinq. Auq.9.1979-Jan. L. 1980 7i kg H2O tion Facility Molecular Sieve Filter < 6.1 kg CO2 3.3 kg H2O À 70m3/h kg I N Helium Purification Plant < i.7,2 CO2 1.2 kg H2 Gas Cooler 0.7 kg CHi Reactor ZNuclear Drying.May 5-0ct.71980 Vessel 31 kg Molecular Sieve Filter < H2O 2 kg CO2 Dust Filter > Gas-pre- 2 kg H2O cleaning 27 kg CO2 Helium Purification Plant < 72 kg CO 1.1 kg H2 Start-Up- 0.5 kg CH4 Purification

H6 - L2 Fig.2 AVR Removed Impurities Table 2 600mVh Scheme of the Gas Purifi­ cation Facility

PREDICTIONS ON AN HTR COOLANT COMPOSITION AFTER OPERATIONAL EXPERIENCE WITH EXPERIMENTAL REACTORS ftk MW R. NIEDER IXlJTfl Arbeitsgemeinschaft Versuchsreaktor AVR Jülich

Federal Republic of Germany

SM Introduction

Long-term operational experience of the HTR experimental reactors TOO Dragon (1966 - 1975), Peach Bottom (1967 - 1974) and AVR normalised oct^vity

K6MW|h.9SC°CI (since 1967) has yielded a large number of common quantitative and qualitative results about the sources and behaviour of Ci 2G0 helium impurities in the primary circuits. Additional infor­ mation has also been obtained from experiments made at the

100 _/"\^V Л measuremeosured activity three reactors. The results at the AVR are particularly in­ teresting because the gas outlet temperature can be varied

Aug 79 Sept 79 Od 79 Nov 79 Dec 79 May80 Junt80 July60 Aug60 SeptÊû Oct ЬС Shut Gown) from 770° С to 950° С when the reactor power is kept constant. period Hence they can be studied according to the temperature dependence Thermal Power{Pi.i).GasTemp.№aG)ond Noble Gas AktivitylSAj-.) H6-L2 AVR Fig 5 during Nuclear Drying of all chemical reactions. It should be possible to apply the results from the operating measurements and experiments made at the reactors, in particular the interrelation of the impurity concentrations, to future reactors. The absolute values of these impurity concentrations are obtained first and foremost by the corresponding helium purification constants. Sources of the helium impurities statements can be made for nitrogen and hydrogen from

Experience has shown that the helium in the primary circuit radiochemical measurements if the concentration profiles of the reactors is constantly contaminated during normal operation, of C-14 and H-3 are measured in graphite samples taken from mostly in very different ways, by the ingress of impurities, the core, e.g. in discharged fuel elements. C-14 is produced water and air in particular. An early assumption, namely from N-14 by an n, p reaction. These measurements indicate that frequent tiny leaks in the steam generator which that in the AVR less than 0.1 % of the hydrogen and nitrogen occurred during normal operation constituted a basic cause present in the primary circuit occurs in the gas phase. of the impurities in the primary circuit, has proved incorrect. Rather, it is an important operational result There is no analogue radiochemical method for oxygen. However, that the steam generators were mainly dense. Any leaks it may be concluded from desorption measurements of carbon which did occur led to significant water ingresses, followed monoxide on reactor graphite that can be compared to similar immediately by a shut-down of the reactor. hydrogen and nitrogen measurements that the adsorption behaviour of oxygen resembles that of hydrogen and nitrogen.

Small, but not unimportant amounts of water and air adsorbed Accordingly, the volume of oxygen chemically adsorbed on the on graphite were found to have entered the core in fresh core graphite should move in a higher order of magnitude to fuel elements in pebble bed reactors. that measured in the gas phase.

A new realisation is that the major part of the hydrogen These chemical adsorption mechanisms are not yet fully occurring during normal operation originates from a diffusion understood. In all probability, radiation effects also play process in the secondary circuit. a part. This would account for the very high hydrogen concen­ tration of 50 to 100 ppm in the graphite, found in graphite

Adsorption on graphite samples taken from the AVR.

Despite the relatively high graphite temperatures of A large number of the chemically adsorbed impurities almost approximately 300° С to 1100°C, considerable amounts of gas were chemically adsorbed on graphite (Ref. /1/). Quantitative certainly remain permanently adsorbed and play no further 148 part in chemical reactions. The adsorbed gases found near It is interesting to note that the values for the three the surface, however, significantly influence the impurities reactors are approximately the same although the water sources in the circuit. are different on account of the different reactor designs.

- Adsorbed gases are released with every rise in An all-important reason for the very low water concentrations temperature, particularly, for example, during is the great chemical reactivity of the core, i.e. the high start-up conditions. An increase in all helium temperature of 1000°C causes graphite to react rapidly with impurities is observed. Some of the impurities any water present. Hence the concentration of the reaction released are removed by the helium purification plant. products H£ and CO must be observed in addition to the Others are readsorbed. The state of equilibrium coolant humidity when leaks are to be recognised. With reference to the AVR reactor Figure 1 shows the CO and is normally only achieved after some hours; after concentrations plotted against one another over a certain water or air ingress accidents, only after a long period at a constant temperature in a leak-free operating period. phase. The CO values are clearly seen to be independent of - All chemical reactions are influenced by the adsorbed the H2 values. Figure 2 shows analogue measuring values, gases. Exchange reactions also occur. Adsorbed plotted in the same way, in an operating phase with small tritium, for example, is displaced from the graphite quantities of water in the circuit. Here, in contrast to surfaces by surplus hydrogen in the gas phase. Figure 1, a linear dependence can be identified.

- The release of absorbed gases, even if only slight, Still referring to the AVR reactor, the low water concentration must also be expected in steady state conditions of approximately 1.5 i^ubar present in normal operation can where there are no ingresses of water or air. be accounted for by the loading of fresh fuel elements in Hence there will always be a certain basic amount which small amounts of water are always adsorbed. Some 100 mg of impurities. H2O is adsorbed in addition to other impurities (CO, Hg, Water in a fuel sphere; these are released in the hot core. When

As mentioned in the introduction, very slight water concentrations 60 fuel elements are loaded daily there is an ingress rate are detected when the experimental reactors are operating of approximately 6 g per day. The water released is "transformed normally. The values measured are in the region of 1.5 - 1 /Ubar. according to the chemical reactivity of the core. The water concentration can be calculated thus The ingress rate derived from this reaction is, however, lower

by comparison. A water formation rate of mol h-1 was

W measured for a water concentration of 10 vpm in experiments H2O performed in the Dragon reactor (Ref /3/)- Ii this measured

value is applied to the AVR reactor, taking into consideration 46 M H20 in the circuit g the different reactor power, a ^ = 2.3 times greater water N H20 ' water ingress rate gh formation rate of 1.84 - 10"° mol/h, which corresponds to

w approximately 0.8 g Н20/рег day, is obtained. This value must W -1 H2O ' helium purification constant h be set in relation to the some 6 g Н20/рег day which enters

cC : integral chemical reactivity h the circuit through fresh fuel elements.

Long-term operational experience of the three reactors with Rc The helium purification constant of the AVR primary circuit the low water concentrations measured gives reason to believe is oC = 0.04 h , the integral chemical reactivity for a gas that the H20 concentrations in the primary circuits of outlet temperature of 850°C is 2.3 h~1 (Ref. /2/). A daily future reactors will produce similarly low values. addition of 60 fuel elements produces a stationary water concentration of approximately 0.8 ^ubar. Another, constant, Hydrogen source of water is obtained by the radiolytic reaction As mentioned at the beginning, the relatively high hydrogen

concentrations can be accounted for by diffusion from the H2 + С (0) } н2о + С steam-water circuit. According to the Schikorr reaction The expression in brackets is intended to indicate that the

Fe 4 oxygen chemically adsorbed on graphite reacts with the 4 H20 + 3 Fe » 3°4 + ^ diffused hydrogen. As already mentioned in the section hydrogen is produced in steam and diffused comparatively entitled "Adsorption on graphite", appreciable amounts of easily in statu nascendi. oxygen are also sorbed on graphite, given the relatively high temperatures in an HTR core. The process is enhanced by the catalytic influence of hydrazin. The hydrogen concentrations in the helium circuits depend on the gas purification constants. The following table dissociation of hydrogen which constitutes the first stage in

gives average hydrogen concentrations normalised with the the formation of methane, is not effected directly by radiation

corresponding purification constants. but by excited helium atoms. Probably the most important of the various excited states possible is that of metastable Hp concentration Gas purification H~ ingress rate constant He (2^S). Metastable helium reacts with hydrogen in xhe -1 -1 /Ubar h /Ubar h following way

3 + He (2 S) + H2 » He + H + H + e Dragon 2 0.5 1 AVR 100 0.04 4 It is interesting to observe that other gases, e.g. nitrogen Peach Bottom 200 0.2 40 (Ref. /4/) also react with metastable helium. A sufficient excess of nitrogen, for instance, prevents the formation of The high values obtained at Peach Bottom are attributable to methane. frequent minor ingresses of oil. The steam temperature,

40°C higher than at the AVR, was probably another factor. Furthermore, ths inhibition effect of nitrogen indicates that Hydrogen concentrations in the primary circuits of future the number of metastable helium awins in a considered system HTRs will be determined in the first instance by the tolerable is restricted. Consequently there is no further methane tritium concentrations. Tritium can only be removed by the formation above a certain hydrogen limit when the hydrogen helium purification plant in conjunction with hydrogen. concentration increases. In Figure 3, with reference to the AVR reactor, hydrogen and methane concentrations measured Methane at a gas temperature of (850°C) are plotted against one

Unless accidents such as oil ingresses occur, methane is another. The СНд concentration can be seen to rise in proportion

formed in a radiolytic reaction according to the general to the H2 concentration up to an H2 limit of 20 vpm. The CH^ expression does not progress beyond this limit.

С + 2 H2 > CH4 The radiolytic formation of methane which bears no relation

The reaction mechanism has not yet been fully explained. It to temperature when regarded in isolation, is in direct is a known fact, however, that radiolytic excitation or the contrast to thermal decomposition. Accordingly, methane 143 concentration is still dependent on temperature. The CH^ Carbon monoxide and carbon dioxide

concentrations of other HTRs can thus be forecast from the Carbon monoxide concentraticms depend basically cm minor respective concentration present and the temperature, ingresses of air, the degassing of fuel elements anci core provided the H2 limit is not exceeded. With gas temperatures graphite and also on water which may seep im. Whether the

of between 700°C and 800°C an H2/CH^ ratio of approximately betw latter has happened can be deduced, as already mentioned:, front 10 to AO was measured at the three experimental reactors. This hydrogen and carbon monoxide measurements. result may be regarded as satisfactory given the existing

differences between the three reactors and the extent of Few predictions can be macs for other ETHs abotrt the

measuring accuracy. The greatest number of H2 and CH^ values, fo frequency and extent cf air ingresses. this depending; which the ^/CH ratio at a particular temperature can be primarily on the reactor design. Assumptions may, Ьс»етге_г» specified more precisely with better statistical data, was be aade from the operating results of the experimental measured at the AVR reactor. Only at the AVR was it possible to reactors. Average CO concentrations, normalized with the obtain measured values at higher temperatures. appropriate helium purification constants» are compilée in

Hp the table belcw. At a gas temperature of 950°C an — ratio of 99.6 1 3.1 was

identified at the AVR as the mean value of a large number of Concentration Helium purification Ingress rate const ratios measurements. The ^/CH established at the different ubar h~T ,ub2r hT1 temperatures show a clear relationship. AVR 500 O.OA 20 Ho Dragon 10 0.5 5 Figure 4 shows ratios at the AVR in an Arrhenius plot Peach Bottom 10 0.2 2 as a function of temperature. The activation energy of

methane decomposition of 72 kcal/mol derived from the measured It can be seen that the ingress rate is greatest at the values is borne out satisfactorily by the relevant literature. Hp AVR. This is due to the constant loading and. unloading cf fuel

The ~ ratios established at the AVR are thus confirmed and elements. The smaller share of air is adsorbed on the graphite

should therefore be applicable to other reactors. of the new fuel elements while the greater part enters the circuit during loading and unloading itself. It should be account of С 14 formation and the possible nitricing of heat

possible to reduce this share in future reactors by exchanger netals. Impurities mist therefore be lisited for

constructional measures. these reasons.

Nitrogen enters the core steadily, above all through fsn""1"1. The carbon dioxide concentration depends directly on the ingresses of air. An additional source at the Dragon reactor carbon monoxide concentration, following the Boudouard was its special mode of operation. Whenever there were equilibrium. The table below contains a list of the relatively long shut-down periods the primary circuit vas C ^C0? ratios measured at different temperatures in the AVR. filled with nitrogen. Small amounts of nitrogen regained sorbed in the core following the replacement of nitrogen by Mean gas outlet с^/со temperature 2 helium; these were released during subsequent operation. A °C further source of nitrogen impurities may be leaks in the

700 4.7 helium purification plant; the low temperature bed is cooled 750 6.4 with liquid nitrogen. 800 15.2 850 40 900 80 It is interesting to note that the amounts of nitrogen entering 950 190 the primary circuit per time unit were of the same order of

If the measured values are indicated on an Arrhenius plot magnitude in all three reactors, as the following table shows. (Fig. 5), an apparent activation energy of 44 kcal/mol is obtained for the Boudouard reaction under reactor conditions. N- concentration Purification No Ingress rate constant

This tallies with the values found in literature on the subject LUBAR H~1 /UBAR H~1 Ref /5/.

Nitrogen AVR 250 0.04 10 Dragon 5 0.5 2.5 Nitrogen is of minor interest as regards the reactions of Peach Bottom 15 0.2 5 graphite to gas. Nitrogen concentrations high enough to inhibit the formation of methane are not expected to occur during The relatively high value at the AVR is chiefly attributable, normal operation. By contrast, nitrogen is important on as mentioned in the previous section, to air ingresses H2 occurring during handling of the fuel elements. Such high N2 [vpm] j i ingress rates should be avoided in future reactors by appropriate 50- constructional measures. Values such as those achieved at

Dragon and Peach Bottom are expected to be obtained at future HTRs.

Conclusion 30-

Evaluation of the operating experience of the three experimental 20- reactors AVR, Dragon and Peach Bottom produces good concurrence of the sources and behaviour of the helium impurities in the respective primary circuits. Most of the differences can be accounted for satisfactorily. It should be possible to apply the greater part of the results to future HTRs.

References

/1/ R. Nieder, The Role of Adsorption Processes in Gas Graphite Reactions in High Temperature Gas-Cooled Reactors, Gas Chemistry in Nuclear Reactors (Salford), Heyden, London, 1980

/2/ L. Stolz, F. L. Werner, Kohlenstofftransport in den Hochtemperaturreaktoren THTR und AVR, Atom- wirtschaft, 1968, 99-103

/3/ C. W. Keep et al, An Interim Report on the Analysis of Injection Experiments in the Dragon Reactor, Dragon-Report DPTN/441, 1974

/4/ R. Nieder, Die Reaktion Graphit-Wasserstoff unter 50 100 150 Ivpm] 200 CO den Bedingungen eines HTR, Carbon-Conference 1976, Baden-Baden 2 Hydrogen and Carbon monoxide Concentrations April 1978 тЛю=950°С /5/ E. Wicke, Dragon Report DPR 112, 1962 F\qM Activation Energy of Methane De - composition 2. Tritium Production 153

TRITIUM BEHAVIOUR IN AN HTR-SYSTEM In nuclear reactors tritium is produced by ternary fission BASED ON AVR-EXPERIENCE and by several neutron reactions predominantly with light elements. The specific features of high-temperature gas- W. STEINVIARZ cooled reactors permit a reduction of the great variety Gesellschaft für Hochtemperatur- of potential tritium forming reactions to only a few which reaktor Technik are compiled in the following table 1 : Bergisch Gladbach Federal Republic of Germany Table 1 : HTR-relevant tritium production conditions H.D. RÖHRIG Kernforschungsanlage Jülich GmbH. Jülich material isotonic abundance dominant reactions

Federal Republic of Germany helium: He-3 ca. 10~^ from air He3(n,p)

7 R. NIEDER ca. 10 from nat.gas Arbeitsgemeinschaft Versuchsreaktor AVR Jülich fuel: U-233.U-235 dependent fron temarv fission burn-UD Federal Republic of Germany

core and reflector 6 7.42-10 Li In .00 ? 1 . Introduction graphite: Li-6

The hydrogen isotope tritium is of particular interest as _г 10 criticality control 1Q к,.,п B (n.2ot)T related to high-temperature gas-cooled reactors (HTR), t ,U system: B-10 '^• " B10(n,c< )bi7(n,noc)T because it is the only radionuclide which can permeate from the primary circuit through heat exchanger walls into secondary circuits and there cause an unwanted contamination. Nevertheless, for the AVR-reactor it is difficult to cal­ The AVR-reactor started its operation in December 1967 culate the tritium production, because approved data ca and since then numerous measurements have been made in the the lithium content of the different graphite structures helium coolant and on the steam side. Moreover, a series do not exist. This may be explained by the fact for the of in-pile and laboratory experiments has been performed time when this reactor was constructed and initially aimed at a better understanding of tritium behaviour in operated the radionuclide tritium was paid only a miner a HTR. attention. Another reason is that about 10 years ago The view we have got by these studies shall be discussed the analytical techniques for determining extrenely low in this paper. m lithium impurities in graphite were not as much developped Table 2: Calculated tritium production rates at AVK-reactor as today. Very recently the detection methods have been as a function of Li-content and operation periods improved to a standard which allows lithium analyses in the ppb-range. years Tritium production (Ci) £ for different Li-contents (ppm) Another difficulty in the tritium calculation procedure for the operation F.E.:0.033 F.E. : 0.5 F-E. s 1 AVR-reactor is that the fuel element types have frequently been R. :0.28 R. : 1 R. Ï 1 changed for testing purposes. 1963+1969 2546 9862 12178 Rough estimates on tritium production rates have already shown 1970 1221 4193 4733 the relative importance of the Li-6 activation reaction which 1971 911 3002 3369 1972 765 2443 272Э even increases if a small HTR-system with a relatively high 1973 - 778 2450 2777 graphite inventory is considered. 1974 564 1716 1ЭЗЭ 1975 634 1893 2143 For the AVR-reactor this inventory amounts to: 1976 592 1643 1797 ca. 160 t so-called Kohlestein 1977 357 1053 1236

ca. 60 t nuclear grade graphite, F.E.г fuel element graphite these both used as neutron reflector and thermal insolation P.- : reflector graphite and Kohlestein materials; and

ca. 20 t matrix graphite in form of fuel pepples.

The abundant use of Kohlestein in the reflector region with its relatively high lithium content considerably contributes to the total tritium production, although in its majority it In fig. 1 the tritium production rate has been broken down with respect to the sources Li-6, He-3, and ternary is exposed to only a low neutron dose. fission, and drawn as a function of AVR operating time. This fact is clearly demonstrated by the next table where the The main source, from Li-6, decreases on the whole with results of a parametric calculation of tritium production time, as a consequence of lithium burn-up in the reflector rates in the AVR-reactor during 1968-1977 are compiled. graphite. But even after 10 years of reactor operating this Changes in any operating conditions influencing the tritium is the predominant tritium source. source term are thereby taken into consideration /1/.

The difference between the first and the second column A comparison of these calculations with measured data reflects the variation in the assumed Li-level of the and a general assessment will be carried out after a few reflector, whereas a comparison of column 2 with column 3 features have been discussed which are typical for shows that the Li-impurity of the fuel element graphite helium-cooled HTR-systems. doesn't matter much. Indeed, the tritium specific activity in the helium coolant

has been measured to be in the range of 10 Ci/cm1 (STP) which means an inventory in the cooling gas being only a few Curies. This compared with the removal paths for gaseous tritium, gas purification and permeation, on one hand (see table 3), and the total production rate as shown by table 2 on the other hand, reveals that the majority of the tritium must have been retained by the graphite. This will be discussed in more detail in chapter 5. Unfortunately not all graphite components are accessible for measurements, but valuable information is delivered by discharged pebbles. As an example figure 2 shows the tritium activity depth profile of a graphite pebble taken from AVR-core in the middle of 1975.

J 10 —i 1 1 1 1 1 3 p Ci H 68 70 72 7^ 76 78 g graphite

Fig Л : Tritium source terms in the AVR with a assumed 300- Li-content of 0.5 ppm for the core graphite and 1 ppm for the reflector graphite 200-

3. Interactions of Hydrogen and Tritium with Reactor Graphite

Evidences on hydrogen and tritium behaviour in the primary 100- system of a HTR may be derived from routine cooiing-gas measurements and particular gas injection experiments at

the AVR-system in connection with tritium profile and - "T" "T T inventory determinations at discharged pebbles. It is 10 20 30 known that graphite materials have a considerable sorption 0ЕРТН [mm] capability because of their porosity and liability to

chemical surface interactions. For the AVR-reactor with Fig. 2: Tritium activity depth profile of a graphite pebble its great graphite and Kohlestein inventory it is therefore taken from AVR-core in the middle of 1975 expected that bonding effects play an essential role for the tritium balance. It should be observed, that there is a steep activity measured Ha-concentration in the primary circuit has been decline near the surface, and that the average specific drawn after a pulse injection of hydrogen gas. The dotted

activity is relatively high. Additional measurements have line indicates the theoretical decrease of Ha-concentration revealed that there is no difference between moderator by means of the cooling-gas purification plant with a pebbles and fuel-containing pebbles. This observation gives purification time-constant of c<= 0.04 h-1. In fact, to understand that tritium from ternary fission is however, the H,-concentration declines much faster because practically retained. of graphite sorption.

It is also obvious to explain the measured tritium profile by tritium sorption from the gas phase rather than by a high internal production rate and a more or less complete retention within the pebbles. Otherwise the assumed Li- content in the pebbles must be unreasonably high, but even so the whole balance would not significantly be influenced. In order to understand tritium sorption on graphite, laboratory measurements have been performed exposing matrix graphite samples to a tritium atmosphere at elevated temperatures /2/. These charging experiments have revealed, that after a time-dependent period conforming to an Elovich-formalism a steady-state was reached according to a Langrauir-adsorption isotherm. For AVR-reactor conditions one can derive that the sorption capacity of hot graphite is only in the range of a few^Ci/g or a few 100 Ci as a whole. These values seem to be contradictory with the measurements at the discharged pebbles. In fact they are not because at the AVR-reactor the pebbles are removed at the cold end. This means that they have been kept suffi­ ciently long at cold-gas temperatures before they are dis­ charged. The same temperature is prevailing for an essential part of the reflector, so that a considerable sorption capability of this part of the graphite structures is to be expected. These conditions have been evidenced by a great deal of gas-injection experiments and by the unintended water Fig. 3: H,-injection experiment, 29.10.1980 ingress accident. In figure 3 the time-dependence of the Another important effect for the tritium sorption equilibrium A quantitative evaluation of the data delivered a linear graphite which may be attributed more to the hot zones "relationship between the exchanged amounts being 0.6 Ci in the AVR-reactor is demonstrated by another hydrogen tritium per mol hydrogen injected, no matter in what injection experiment. Figure 4 shows the tritium and chemical composition /3/. hydrogen concentration values measured in the primary helium on the same time-axis. The good agreement of the time- From the end of march 1978 an increase of hydrogen dependence leads to the conclusion that a very fast exchange concentration was detected in the primary helium, which of hydrogen isotopes has taken place until a new wasn't initially understood. In addition after shutting- sorption equilibrium has been attained. down the reactor in the middle of May 1978 an ingress of about 30 t water occurred. Both incidents found their Tritium 10"' Ci explanation in a steam generator leak. During the period cm3 [STP] of the enhanced hydrogen level about 5000 Ci tritium 200- were removed via gas purification plant. In the penetrated 180- water, which had accumulated in the lower parts df the 160- reactorj another 2900 Ci tritium were measured. Thus a 140- total of about 7900 Ci tritium were incidentally released. 120- 100 This quantity must have been set free from the cold 80 reflector graphite. As expected, the above mentioned linear 60- relationship between the hydrogen and tritium concentration 40- in the primary helium was no longer valid. 20- 0 4. Tritium Permeation into the Steam Circuit i г 1—Г The high mobility of hydrogen in metals enables tritium vpm н2 to permeate from the primary circuit through the heat 100- exchanger walls into the steam circuit. The permeation 80- rate, however, is not as high as could be derived from 60- the permeation coefficient for ferritic materials at the 40- temperatures considered. This reduction in permeation is 20- due to the barrier effect of a corrosion scale (most 0- T I—I I I I—i—i—i—I—I—Г" probably magnetite) which is formed on the steam—side. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 On the other hand the tritium permeation is impeded by TIME [h] a surplus of hydrogen (H^) in the helium as compared

to tritium. This H2-level in turn is influencée by the Fig. 4; Tritium and hydrogen concentration values during corrosion reaction at the heat exchanger walls, but there a hydrogen-injection experiment, 8.12.1971 158 are still other sources, e.g. outgassing and a chemical The tritium sinks: radioactive decay and helium leakage reaction of moisture at the hot graphite. can thereby be neglected. In table 3 the relevant data for balancing tritium in the AVR-reactor are compiled. The permeation impeding effects have been analyzed using the hydrogen and tritium data measured in the coolant and the tritium concentration measured in the steam /4/. Table 3: Measured data on AVR-tritium release The barrier effect of the oxide scale could be described by a permeation reduction factor of about 20 as compared

to quasi bare surface wall conditions. As for hydrogen operational thermal helium Tritium release /Ci/ to: year energy outlet interference with tritium permeation it could be shown /мка/ tempera­ gas purifi­ secondary pebble totals at least for normal operation conditions that the tritium ture /°С/ cation plant circuit discharge permeation rate obeyed the relationship:

1968 6 581 700 80 30 20 130

55 р р 1969 10 103 740 70 40 20 130 T ^ нт^ ^ н2 1970 13 108 725 865 50 100 1 015 1971 13 045 780 2 015 80 160 2 255 (0 : tritium perm, rate; P„„ „ : partial pressures of HT,T,) 1 ni , ij 1972 12 936 820 1 110 70 270 1 450

which can theoretically be derived when the mass balance law 1973 15 167 825 580 90 300 970 is introduced into Richardson's permeation equation. 1974 11 980 900 1 030 100 380 1 510 The practical consequences from the permeation impeding 1975 14 725 910 1 590 70 800 2 460

effects will be discussed in the following chapters. 1976 15 595 925 1 005 80 700 1 785 1977 8 722 840 280 40 420 740 Jan.-March 1978 5 278 925 260 25 380 665 5. Tritium Distribution in the AVR-Reactor

While the allocation of tritium to direct sinks (helium puri­ normal operating release ^ 8 885 675 3 550 13 110 fication plant, secondary circuit, pebble discharge) can be determined by concentration measurements, this is not possible Apr.-May 4 970 for the tritium which is stored in the graphitic reflector 1978 583 930 7 870 structures. It cannot be deduced either from the tritium (accidental release) 2 900 generation data, because of the uncertainties in the Li-6- source as was above-mentioned (see chapter 2). 20 980 Only an attempt can be made to draw conclusions on the actual tritium balance from a comparison of the available release rates with the parameter field for tritium production. The importance of gas purification and pebble discharge as the The high production rates of the first reactor operation dominant tritium release paths is clearly demonstrated, whereas years havenot found expression in the release rates. It can the permeation path doesn't account. One further recognizes thus be concluded that the reflector must have acted as an that the released quantities initially increase with time until effective tritium buffer, until lateron a steady state must a steady-state is approached. Since the gas outlet temperatures have been attained. This tritium inventory stored in the have been raised in the same period one could be tempted reflector represents quite an essential activity potential to attribute this to the release rates, but in fact the tendency during fault conditions. This has been evidenced by the is by no means regular. release of the 7900 Ci tritium during the water ingress

The total release rate can now be compared with the calculated accident. production rates. This is illustrated by figure 5. One can further derive from table 3 that an assumption of only 0.28 ppm Li in the reflector graphite leads to a deficit in the tritium balance. From figure 5 it is even more probe.ble that an assumed Li-content of 1 ppm is not yet [Ci] sufficient. First measurements with original Kohlestein 10' samples yielded indeed a Li-concentration of 4 ppm /5/. TOTAL PRODUCTION

6. AVR-Tritium Balancing from a Radiological Viewpoint

Balancing tritium is of general importance for nuclear plants with respect to the radiation exposure of the environment. 103 By the AVR-reactor the sewage path is charged with ^100 Ci tritium per year via the elutriation of the steam circuit. From a power viewpoint this value is absolutely within the scope of power plants. The release is performed via the Decontamination Department of the Kernforschungsanlage 102 Jülich. The tritium release into the exhaust air, if the reactor is duely operated, can be estimated to only a few Curies per year which is insignificant as compared to the emitted radioactive noble gases.

10 -—i 1 1 1 1 1 Approximately 100 % of the tritium activity removed by the 68 70 7 2 74 76 78 purification plant is collected in form of highly active tritiated water and stored in tanks. Fig. 5; Comparison of the tritium production rate and the A similar procedure is applied for the discharged pebbles tritium release rate for the AVR-reactor which take away a considerable amount of sorbed tritium as was shown by table 3. 160 For the reflector structures one can expect that the /3/ : R. Nieder, Freisetzung von Tritium im Primär­ tritium inventory stored there will be available for waste H.D. Röhrig, kreislauf eines Hochtemperaturreaktors disposal after the final shut-down of the reactor. R. Wischnewski, bei Störfällen; P.G. Fischer: KTG-Fachtagung "Spaltproduktfreisetzung bei Reaktorstörfallen"; Karlsruhe, 1./2. Juni 1976

7. Summary and Conclusions /4/ : W. Steinwarz: Report on The tritium balance of the AVR-reactor is essentially governed Seminar über Wasserstoff- und Tritium­ by the abundant use of graphite in the core. For the tritium verhalten in Hochtemperaturreaktoren, source term this is particularly true because of the use of 8. März 1978 Kohlestein with its relatively high lithium impurity level /5/ : VGB Untersuchungsbericht 910/79, 19.1.1979 as reflector material. It should be recognized that this is more a feature of the AVR-reactor than of big power HTRs with a modern design. The importance of graphite components being on coldgas temperatures for tritium sorption should be the same for all HTR types. It is recommended to investi­ gate this effect by further experimental studies.

8. References

/1/ : H.J. Cordewiner: Numerische Berechnung des Tritium- Verhaltens von Kugelhaufenreaktoren am Beispiel des AVR-Reaktors; Report Jül-1607 (July 1979)

/2/ : V. Malka, Investigations on Sorption and

H.D. Röhrig, Diffusion of Tritium in HTGR-Graphite; R. Hecker: Proc. ANS-Conf. on Tritium Technology in Fission, Fusion and Isotopic Appl. Dayton, Apr. 29 - May 1, 1980 161

I4.45 Session 2: Fission product plate-out II 4. AGENDA OF THE MEETING Chairman: 0. Baba

"Remarks on possibilities and limitations Chairman: C.B. von der Becken of theoretical approach to plate—out problems" E. Obryk

15.15 Coffee break Tuesday, December 2

09.CO Opening I5.45 "Fission product behaviour in the primary circuit of an HTR" О9.3О Session 1: Fission product plate-out I C.-B. von der Decken, N. Iniotakis Chairman: R.J. Blanchard 16.15 "Derivation of criteria for primary circuit activity in "In-pile helium loop "Comédie" an HTGR" J.J. Abassin, R.J. Blanchard, J. Gentil S.D. Su, A.W. Barseil

10.00 "Out-of-pile helium loop for liftoff experiments" 16.45 "The influence of dust on the hazard potential of a depressur­ R.J. Blanchard, A. Bros, J. Gentil ization accident of a high temperature reactor" N. Iniotakis. C.B. von der Decken IO.3O "Erperimental facilities for plate-out investigations and future work" I9.CO IAEA cocktail reception, KFA-JUlich K.H. KUnchow, H. Dederichs, H. Iniotakis, B. Sackmann

11.00 Coffee break

II.30 Session 1: Continuation Wednesday, December 3 12.00 "Results from plate—out investigations" N. Iniotakis, J. Malinowski, H. Gottaut, K.H. KUnchow O9.OO Session 2: Continuation 12.00 "Fission product plate-out study using in-pile loop OGL-1" 0. Baba "Modelling of plate-out under gas-cooled reactor (GOR) accident conditions" I2.3O "Fission product behaviour in the peach bottom and A. Taig Port St. Vrain HTCR's" D.L. Hanson, H.L. Baldwin, D.E. Strong 09.3° "Safety research on iodine plate—out during postulated HTCR core heatup events" 13.00 Lunch A.H. Barsell, O.P. Chawla, CG. Hoot

10.00 3: I4.I5 Session 1: Continuation Session Decontamination of activity Chairman: E. Obryk "Iodine Sorption and Desorption from low-alloy steel and graphite" "Plate-out measurements and decontamination of a component of R. Wichner, M.F. Osborne, H.A. Lorenz, R.B. Briggs the AVR reactor at Jülich" j. Hanulik, H. Schmied, J. Wahl 162

10.30 Coffee break 10.00 "Chemical reactions during nuclear drying of the AVR primary circuit following a water ingress" 11.00 Session 3? Continuation R. Nieder, К. Vey "Evaluation of a decontamination model" IO.3O Coffee break D.W.T. Rippin, J. Hanulik, E. Schenker, G. Ullrich 11.00 Session 4: Continuation 11.30 "Decontamination and high temperature materials" E. Schenker, С. Ullrich, J. Hanulik, W.B. Waeher, K.H. Wiedermann "Predictions on an HTR coolant composition after operational experience with experimental reactors" 12.00 "Maintenance concept of the gas-turbine in a 1640 MW direct cycle R. Nieder HTR" H. Schmied, H. Karaus, К. RBilig, F. Schenker II.30 "Tritium behaviour in an HTR-system based on AVR-er- perience" 13.ОО Lunch W. Steinwarz, H.D. Röhrig, R. Nieder

14.00 Visit Tour, KFA-JUlich I2.3O Lunch 17.00 Departure to Aachen I4.OO Round table discussion Visit of cathedral Chairman: C.B. von der Decken — Summaries by the session chairmen Dinner — Discussion

— Recommendations and conclusions

16.ЗО Closing Thursday, December 4

09.00 Sension 4! Coolant Chemistry Chairman : й. Hi ebner

O9.3O "Coolant chemistry of the advanced carbon dioxide cooled reactor" Й. Faircloth, K.S. Norwood, H.A. Prior

09.30 "Primary coolant chemistry at Peach Bottom and Ibrt St. Vrain HTGRs" R.D. Burnette, N.L. Baldwin m

5. LIST OF PARTICIPANTS

W. KATSCHER SERTRAN SÄCKÄWS FRANCE KFA, Institut TIIR Nukleare Sicherheitsfarschunc KFA, Institut rur Reaktsröauslisnertts Postfach 1913 Pcsrfacn 1313 D-5170 Jülich D-517Q Jülich Jean Jack ABASSIN CEN/Grenoble Er* in SÄLTHESEN DMG/SER KFA, Projektträgerschaft Hochteiaperatur-Reaktcr Waas VEÏ 85 X - 38 041 Grenoble Cedex Postfach 1913 Arbeits seseinschaft Versuchsreaktor £A35J D-5170 Jülich KFÄ-Gelitce Postfach 1513 René BLANCHARD 3-517Q Jülich CEN/Grenoble Herbert KRCHN_ 85 X - 38 041 Grenoble Cedex KFA, Institut rur Nukleare Sicherheitsforschunc

Postfach 1913 Hans-Jülrcert SCÎLESISŒR D-5170 Jülich L4TE5ATCM = GKT Jean GENTIL Fri ecr f ch-^ert-Str. Service des Piles D-3C60 Sercisch-Slacbach I CEN/Grenoble Herbert DEDERICHS 85 X - 38 041 Grenoble Cedex KFA, Institut rür Reaktorbaueleiïente Postfach 1913 ?terjika SEIF _

D-5170 Jülich 1 Gesellschaft rur Kccntsrçsratur Seaktcrtschnik пей £GH1T) Marcel ROBIN Friedrî ch-Ebert-Str. Commissariat à l'Energie Atomique (CEA) Q-5050 Ssrçîsch-Slacbèch 1 Division d'Etude et de Development Horst GOTTAUT des Reacteurs KFA, Projekt und Entwicklungsgeneinschaft HTR Mr. CHRIST Boite postale No. 2 Postfach 1913 Kochtacperatur-Seakarfeau GcbK (КЗ} F-91190 Gif-Sur-Yvette D-1570 Jülich Postfach 5360 Gottlieb-Qaialer-Str. S D-5SOO Mannheia 1 Nicolaos INIOTAKIS GERMANY, FED. REPUBLIC KFA, Institut für Reaktorbauelewente Postfach 1913 Jerry MALINCKSKI D-1570 Jülich KFA, Institut rür RsaktcruauaîeEer.ts Peter BIEDERMANN Postfach 1913 D-5170 Jülich KFA, Institut für Reaktorbauelemente Peter KUBASCHEWSKI Postfach 1913 Hochtemperatur-Reaktorbau GmbH (HR3) D-5170 Jülich Gottlieb-Daimler-Str. 8 Rudolf NIEDER D-6800 Mannheim Arbeitsgeaeinschaft Versuchsreaktor AVS, Claus-Benedict von der DECKEN KF.A-Gel är.ce KFA, Institut für Reaktorbauelemente Rainer MOORMAN Postfach 1913 D-5170 Jülich Postfach 1913 KFA, Institut für Nukleare Sicherheitsforschtmg D-5170 Jülich Postfach. 1913 D-5170 JUlich GERMANY, FED. REPUBLIC SWITZERLAND Anthony TAIG UKAEA Safety & Reliability Directorate Wigshaw Lane Heinz-Dieter RÖHRIG Culcheth near Warrington Jozef HANULIK Kernforschungsanlage Jülich Cheshire Swiss Federal Institute for Reactor Research Postfach 1913 CH-5303 Wurenlingen Institut für Reaktorentwicklung d-5170 Jülich David RIPPIN USA Swiss Federal Institute of Technology Klaus ROLL IG CH-8092 Zürich Hochtemperatur-Reaktorbau GmbH (HRB) David Lee HANSON Postfach 5360 General Atomic Comoany Gottlieb-Daimler-Str. 8 Erhard SCHENKER P.O.Box 81 608 D-6800 Mannheim 1 Swiss Federal Institute for Reactor Research San Diego, CA. 921 31 CH-5303 Wurenlingen

ITALY Henri SCHMIED Robert WICHNER Brown, Boveri & Cie. Oak Ridge National Laboratory 5401 Baden P.O.Box X Oak Ridge, TN 37 030 G. BASSO Dipartimento Reattori Termici U К Comitato Nazional per l'Energia Nucleare Viale Regina Margherita, 125 1-00198 Rome Reginald FAIRCLOTH AERE, Harwell, Didcot I A Oxon 0X11 ORA JAPAN J. XDPITZ Scientific Secretary of the Inter­ Reginald Robert GALLIE national Working Group on Osamu BABA UKAEA, WNPDL, Windscale Gas-cooled Reactors Japan Atomic Energy Research Inst. Sellafield, Seascale Oarai Research Establishment Cumbria International Atonic Hiergy Agency (IAEA) Narita-cho Wagramerstrasse 5 Oarai-machi P.O.Box 100 A-1400 Vienna Higashi Ibarakigun Joseph Jackson HILLARi* Austria Ibaraki-ken UKAEA, WNPDL, Windscale Sellafield, Seascale Cumbria POLAND

Peter Andrew SOLARI Edward OBRYK NPC (Risley) Limited Institute of Nuclear Physics Warrington Road Radzikowskiego 152 Risley, Warrington, Cheshire 31-342 Krakow WA 3 6BZ