o ~ ORNL MASTER COPY ORNL-3144 %C/\ F C-84 — Reactors—Special Feat Aircraft Reactors AEC RESEARCH AND DEVELOPMENT REPORT M=-3679 (24th ed.) AIRCRAFT NUCLEAR PROPULSIONPROJECT cay ke b EAT ad g rn. J SEMIANNUAL PROGRESS REPORT AR [_‘. Eili 8-15-77 Y FOR PERIOD ENDING APRIL 30, 1961 i w} n 6 . - L . I ; " . baro f& 7 ; e \r\ \ " If;;_‘l . b (.\f., $2 3 ] ey 5 I‘E\‘: \‘I\ »-;,h Tmmed b A = e .- % . et % i L . T ¢ .."‘S‘“" \ = e b 'AETWM’HI‘:E:S‘ N o T oac - B .p : n SR S OAK RIDGE NATIONAL LABORATORY = fl operated by S UNION CARBIDE CORPORATION o ' for the s U.S. ATOMIC ENERGY COMMISSION @y 3 sND L 3 "LEGAL NOTICE This report was prepared as an account of Government sponsored work. Neither the United States, “nor the Commission, nor any person 0ct|ng on behalf of the Commlssnon'. . . ‘ "A. Makes any warranty or representutuon, ‘expressed or |mpl|ed wnth respect to the occuracy, . completeness, or usefulness ‘of the information contained in thls report, or ‘that the use of any |nformat|on, upporatus method or process_dlsclosed in. th|5 report moy not mfrmge : prwutely owned rights; ar L - . ‘ ' B. Assumes any liabilities with respect to the use of, or for domoges resulting from tl"le use of“ ' any mformcltlon apparatus, method, or process dnselosed in this report. ‘ . As used in the above, ‘person actang ‘on behalf of ‘the Comm|ssuon" includes. any. employee or: contractor of the CommISslon, or employee of such contractor, to the extent thut such employee - or contractor of the Commlsslon, or employee of such _controcter prepares “disseminates, or provides access to, c|ny mformutlon pursuant to l'ns employment or contract wn?h ‘the Comm|ssnon,' or his employment wnth such contractor. - & ~ ORI — 3B/ S (Report Number) ANP CLASSIFICATION REVIEW FORM 5732; Report Title: Author: Please indicate in square below classification determination / / Classification confirmed as in accordance with CG-RAN-4 topic / / Classification downgraded to ” in accordance with CG-RAN-4 topic / / Declassification 1s contingent on (Explain) / iZ/ Contains no Restricted Data nor other Defense Information and may be declassified in accordance with CG-RAN-4 Guide Topic S, 3,/ s 2—,# 6 ! 6K gey” A e £ 1972 C/ISignature Pate OBNL-31H4 ERRATA p. 10 (1) In equation for ¢7(E,2), "dr" is-replaced by dr." p. 141 (1) In equations flm?¢n(E,2) and Sn(E;fi), "Sn(E,fiU"is replaced by "S'(E,&)." ! 3 (_/: (ZiiH -2 (£") —g-(fl——))dg ] ) IiH (2) "éxp { is replaced by 2 | | | - 91 nex'p [ f(ZilH g'g;?—l - Z‘.r(g') > dfl':|.l" S LiH p. 147 e l n 2nDsiny’ " 1 " (1) TIn the equation, Tl replaced by p. 148 (1) TIn Table 12.13, "P(E,8,D)" is replaced by "P(E,6,D)." ORNL-3144 C-84 — Reactors—Special Features of Aircraft Reactors M-3679 (24th ed.) This document consists of 182 pages. Copy (7? of 231 copies. Series A. Contract No. W-7405-eng-26 ATRCRAFT NUCLEAR PROPULSION PROJECT SEMIANNUAL PROGRESS REPORT Q for Period Ending April 30, 1961 otaff \* Oak Ridge National Laboratory “ \3? ?' T S Date Issued &wfigl JUL 11 1961 i, . T . & e . . - NS e : TR e e Voeo T e OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee operated by . UNION CARBIDE CORPORATION for the U. S. ATOMIC ENERGY COMMISSION 3 FOREWORD The ORNL-ANP program primarily provides research and development support in reactor materials, shielding, and reactor engineering to organizations engaged in the development of air-cooled and liquid-metal- cooled reactors for aircraft propulsion. Most of the work described here is basic to or in direct support of investigations at Pratt & Whitney Aircraft Division, United Aireraft Corporation, and General Electric Company, Aircraft Nuclear Propulsion Department. i, 41 1 CONTENTS QUMM AR Y ottt i it e ettt st ettt aneeneseeesotesaoeessesananansennaes xix PART 1. MATERIALS RESEARCH AND ENGINEERING 1. MATERIALS COMPATIBILITY vttt ittt et vosnnsnenenssnnoenonnens 3 Oxidation of Columbium Alloys at Low Pressures ........... Corrosion Studies for Determifiing the Mechanism by Which Lithium Attacks Columbium .......c.iti it iineenienenss S . Effect of CGrain Orientation .......eeeeeeeenneerennennn 9 Effect Oof Time ..o i ittt e ittt it te i et e nnnnnnennanas e 9 . Effect of Lithium Purity .......citiieiiiininrnnnnssns 12 Cofrosion Studies on Cb—1% Zr Alloy in Lithium ........... 13 Effect of Oxygen Additions to Cb—1% Zr Alloy on Its Room-Temperature Tensile Properties .............00... 13 Effect of Heat Treatment of Welds .......... ... ..., 14 Dissimilar-Metal Mass-Transfer Studies in Cb—1% Zr Alloy- NaK-Type 316 Stainless Steel Systems ........cciivnienesn 15 Effect of Carbon and Nitrogen Pickup on Tensile P OpEr Lie S i it e it i s e e e e i5 Effect of Stress on the Rate of Carbon and Nitrogen PiCKUD tvi i i e e e e 18 Vapor Pressure of NaK (43.7 wt % X) in the Temperature Range 1520 1o 1832 % vttt it it ittt teteneenneneennnnnns 20 Potassium Compatibility Studies .....cciiiiieteenenennnas 21 Refluxing Potassium Capsule Tests ........cciiiivnnn. 21 | Boiling-Potassium Loop Tests ...... ... 23 . Purification of Potassium Metal: Methods of Reducing the OxyEen Contbent ..ttt ettt ettt tee et 24 Determination of Oxygen in Potassium .........0cciiienens 24 2. AGING STUDIES OF COLUMBIUM-BASE ALIOYS ....ieiiiivinnnen.. 28 Aging of Wrought Material ......c.ccieiriieenenennnnennnnns 28 Aging of Welded Material .......ccviuieieinniennrerarannnns 31 Internal Friction of Cb—1% Zr ALIOY i eevvrvreeraeennonann 36 3. MECHANICAL PROPERTIES INVESTIGATIONS o veveevnvernennnnn. 42 Columbium Alloy Tube-Burst TeSTS +.vvv it rrvrenreeranencens 42 vi ALLOY PREPARATION v i ittt et e s eenesoenasnonsnsssassssssnsanss Melting of High-Purity Columbium and Columbium-Zirconium Alloys ..eeee... Gt e et et e e et es et e ettt Addition of Nitrogen to Columbium and Columbium-Zirconium AlLOYS it e e it e et eres e Electron-Beam Melting of Columbium-Hafnium Alloys ......... FABRICATION STUDIES 4 ivieunsvenesesnesnonsnescssosnanasasss Braziné of Columbium .....ccviieenniennns Ceereesaesaas e Fabrication of Clad Tubing .vvveernrennenss ee et e e eeenaan BERYLLIUM OXIDE RESEARCH 4t veeviururnenineroonnreacesaanans Preparation of Refractory Oxides from Molten Salts ........ Preparation and Calcination of Beryllium Oxalate - Monohydrate ........ciiiiiiiiiiiiii i it i . Purification of Beryllium Compounds by Solvent Extraction .. Phase Relationships in BeO-Metal'Oxide Systems ............ The BeO-Cal System ...ttt iineteteneenernneeneeannanns The BeO0-MgO, Be0O-CeOp, BeO-Zr0O, Systems .........c.0.... ENGINEERING AND HEAT-TRANSFER STUDIES .....¢.iveierrnnennnns Boiling-Potassium Heat-Transfer Experiment ................ Thermal Conductivity of Lithium .......ccciiiiiiienennn. RADIATION EFFECTS v i it ie it tesenenesennsnnsascannansenens Radiation Effects on Columbium-Zirconium Alloy ............ Radiation Effects on Stainless Steel and Inconel .......... Beryllium Oxide Irradiation Studies ......civiiriniinnnnns ' PART 2. SHIELDING RESEARCH DEVELOPMENT OF REACTORS FOR SHIELDING RESEARCH .......... Tower Shielding Reactor IT (TSR=II) .+.'vvvirruvrnneenn. Ceeee Reactor Mechanical System ...ttt it i e i Reactor Contrqls ..................................... Flow Measurements and Fuel Plate Temperatures .......... Nuclear Measurements ereeies i e e e e e e s e s ee e 46 46 © 48 50 51 51 54, 56 56 5'7 58 59 60 61 67 67 72 77 77 81 86 93 93 93 95 95 97 o & &} R —— DEVELOPMENT OF RADIATION DETECTION EQUIPMENT ............. . 99 Gamfia-Ray.SpectrOSCOPy ..................................... 99 The Model IV Gamma-Ray Spectrometer .........civevviuvens 99 Unscrambling of Continuous Scintillation Spectra ........ 100 NeUtron SpPECHIOSCODY v vt v vt e tneeenenenneeesenneancessennns 113 Use-of Silicon Surface-Barrier Counters in Fast-Neutron Detection and SpectroSCOPY v it i e iintonttnnnereoenennss 113 A TFast-Pulse Integrator for Dosimetry . .......iiiiiieennnnn. 121 BASIC SHIEIDING STUDIES . eteennvuernneennoresnanoannnnonnaens 123 Energy and Angular Distributions of Neutrons Emerging from Planar Surfaces of Diffusing Media ...........ciiiiiiaenn. 123 Experimental Verification of a Geometrical Shielding Transformation . .....ii ittt teninernenoonesnsonsonsas 127 APPLIED SHIEIDING ittt et ioeasntntataenenensaenanasosnnnas 129 Preanalysis of Pratt & Whitney Divided-Shield Experiment B TOF ittt ittt ittt i s e e e e 129 Radiation SOUTCEes .. .iiii i it ininiinnrnnrns che e 131 Attenuation and Transport ProCessSes .....vieervrersasannas 131 Future Calculations .......c....n ST P 139 Mathematical Description of Calculation ......coveivueenen 139 Preliminary Results of the Pratt & Whitney Divided-Shield Experiment at TOF ...ttt ittt ittt iistnstarsarneeensnenas 148 S A vii A e wes eleweg T e vt . TR v DI R [N © o ) 2 et A s) ANP PROJECT SEMIANNUAL PROGRESS REPORT SUMMARY Part 1. Materials Research and Engineering 1. Materials Compatibility The oxidation rates of several commercial columbium alloys and numer- cus experimental alloys were measured in low-pressure oxygen and in low- pressure air at temperatures up to 1200°C. In most cases, it was observed that alloying had a detrimental effect on the oxidationlresistance, which is the reverse of the effect that is observed when the alloys are tested at atmospheric pressure. Products from the corrosion of columbium by lithium were shown to occur as a transgranular precipitate on certain crystallographic planes. The depth to which the corrosion product is found was determined to be a function of grain orientation. Although the depth of attack was found to be independent of exposure time between 1 and 500 hr, oxygen was con- tinually leached from the columbium by the lithium and the specimens were weakened by this depletion. The addition of 2 wt % Li,0 to lithium re- duced the depth of attack in oxygen-contaminated columbium to below that resulting from exposure to pure lithium. This result is in line with other evidence that lithium reacts with the oxygen contamination in the columbium and that the driving force for lithium to produce lithium oxide is of consequence in the corrosion process. An addition of LisN to the lithium did not reduce the amount of corrosion. Additions of up to 1500 ppm of oxygen were made at 1830°F to Cb—1% Zr alloy. The oxygen contamination incrzased the room-temperature tensile strength and decreased the ductility of the alloy. When the alloy was exposed to lithium in this condition it was corroded, and a loss of ten- sile strength occurred. A high-temperature heat treatment in vacuum at 2910°F, however, nullified the strengthening effect of oxygen and made the alloy again corrosion resistant. The resistance of Cb—1% Zr alloy welds to corrosion by lithium was also shown to be strongly influenced by heat treatments. Oxygen additions at 1830°F to the alloy welds made them very susceptible to corrosion by lithium, but the effect of additions made at 2190°F was very small. Heat treatment at 2370°F completely eliminated corrosion by lithium for oxygen concentrations up to 2500 ppm in Cb—1% Zr welds. All the specimens pre- pared by welding oxygen-contaminated Cb—1% Zr alloy and testing in the as-welded condition demonstrated a high susceptibility to corrosion. It was concluded that the corrosion resistance of the Cb—1% Zr alloy welds is strongly influenced by the .mode and/or distribution of the oxygen. » It was found that CbC and Cb,yN layers approximately 1 mil in thick- ness formed on the surface of columbium or Cb—1% Zr alloy‘speciméns when = s they were exposed to NaK in a type 316 stainless steel container during isothermal tests at 1700°F for 500 hr. It has also been determined that the carbon remains mostly in the surface layers, while the nitrogen com- pletely penetrates the specimen. This transfer of carbon and nitrogen results in an increase in the tensile strength and a reduction in ductility of Cb—1% Zr. When the brittle surface layers were removed in recent ex- reriments, these same effects on mechanical properties were still observed. Since the transfer of carbon was limited to the surface layers, it was concluded that nitrogen is the pfincipal element responsible for the ob- served changes in the mechanical properties. In tests on unalloyed colum- bium, stresses up to 70% of yield strength did not increase the rate of pickup of either carbon or nitrogen, thus a major rate-controlling step in the transfer process appears to be the one which supplies carbon and nitrogen to the alloy surface. The vapor pressure of NaK was measured in the temperature range 1520 to,l830°F. The data were compared with those obtained by calculation using Raoult'’s law and empirical equations for vapor pressuré as a func- tion of temperature for the pure species. Type 310 stainless steel and Inconel were tested in refluxing po- tassium systems at 1600°F for 1000 hr. TFabrication was completed and the tests were begun on boiling-potassium loop systems constructed of . Inconel and Haynes Alloy No. 25. These systems will be operated for. i (» i ulRNE., 3000 hr at a boiler temperature of 1600°F. Two batches (10 and 90 kg) of commerclal potassium contalning approximately 300 ppm oxygen were puri- fied by hot-gettering and cold-trapping treatments. The oxygen content was lowered to approximately 25 ppm by these treatments. 2. Aging Studies of Columbium-Base Alloys Evidence was accumulated which indicated that the aging reactions which occur in Cb—1% Zr alloy are due principally to the oxygén in the alloy. It was observed that the tendency to undergo the aging reaction increased as the annealing temperature was increased in tests of speci- mens with the same oxygen content and that only alloys of intermediate- oxygen-content aged when annealed at temperatures of 1600 and 1800°C. Increasing the carbon and nitrogen impurities in the alloy did not cause the alloy to respond to aging under the conditions studied. The use of a Rockwell-B hardness test to evaluate the aging be- havior of Cb—1% Zr welds was investigated. It was found that the nu- merical values of hardness are of dubious value as absolute indicators of aging behavior, since some points associated with brittle and border- line behavior have lower hardness values than points associated with ductile behavior. The hardness test, however, has value as an indicator of trends and provides useful information to assist in the analysis of the aging behavior. Various postweld annealing treatments were studied as possible meth-. ods of circumventing the aging reaction. Annealing from 1 to 3 hr in the temperature range of 1900 to 2200°F is known to be effective in pre- venting aging. This annealing treatment caused a heavy precipitate to appear as a network throughout the matrix, the distribution of which pre- vents it from acting as a strengthening agent. The results indicate that even with high oxygen contents in the welds, appropriate postweld anneal- ing can prevent aging. An apparatus to measure internal friction has been built and placed in operation. This eguipment will.be used to study the role of oxygen in aging of columbium-base alloys. Preliminary measurements were obtained with columbium and the Cb—1% Zr alloy specimens. —— 3 3. Mechanical Properties Investigations Tube-burst data on Cb—1% Zr alloy specimens were obtained at 1800°F to provide control data for evaluation of results from in-pile experi- ments. The data are summarized and compared with those of other investi- gators. 4. Alloy Preparation Methods were studied for the preparation of columbium and columbium alloys with closely controlled compoéitions of both the alloying addi- tions and impurity elements. In this study the eleétron beam, inert-gas- shielded tungsten arc (noficonsumable-electrode arc), and consumable- electrode vacuum arc melting processes have been investigated. Emphasis has been on the first two of these melting processes. Electron-beam melting was demonstrated to be very effective for pre- paring columbium of high purity with respect to interstitial impurities, oxygen, nitrogen, and carbon. However, the addition of alloying elementé more volatile than columbium, such as zirconium, was difficult to control. Studies were made to determine the capabilities of the arc-melting furnaces with respect to sizes of heats of columbium that can be melted and to the purity which can be maintained in the metal. Preliminary cali- bration curves were established for the controlled addition of nitrogen to columbium by arc melting under various partial pressures of nitrogen and argon. 5. "'TFabrication Studies Experiments showed that 12 refractory-metal-base brazing alloys which have been developed readily wet and flow on colgmbium;to-columbium T-Jjoints. No visual evidence of fillet cracking was seen on these joints. Five of these alloys also appeared promising for joining columbium to stainless steel, but Jjoint designs should be utilized which minimize the effects of thermal expansion coefficient differences. Columbium tubing was clad with type 304 stainless steel or type 446 stainless steel by co-drawing tubes at foom temperafure and annmealing at about 1000°C for 2 hr to obtain a bond. Copper was used at the interface xii of the composite to serve as a bonding metal as well as a diffusion bar- rier to inhibit the formation of a brittle reaction layer. Preliminary evaluation of these composites indicated that producing metallurgical bonding of the cladding and the tube by cold forming and subsequent heat treatment is not promising. 6. Beryllium Oxide Research The steam precipitation of beryllium oxide from melts containing lithium and beryllium fluorides is being developed as a direct route to the formation of crystalline material without the necessity of a cal- cination step. Tests at 800°C made with helium as the carrier gas for the steam gave products which were well crystallized but were contami- nated with corrosion products from the metal container. Additional tests were in progress in which hydrogen is being used as the carrier gas to suppress the corrosion of the container. Two approaches were studied in efforts to obtain crystals of U0, coated uniformly with BeO: 1in one, preformed crystals of UO, of less than 20 p in diameter were suspended in a fluoride melt during the stéam precipitation of BeO; in the other, consecutive precipitations of U0, and BeO from the same melt were attempted, with the initial precipita-. tion being very rapid to obtain small particles of UO, and the subse- quent precipitation being very slow to obtain uniform coating with BeO. Hot filtration of an aqueous slurry of beryllium oxide monohydrate, at a temperature above that at which the trihydrate can form, gave a product almost completely free from traces of the trihydrate. Calcina- tion of the product gave a BeO product of higher purity than that usually obtained from the trihydrate. Exceptionally pure beryllium hydroxide was produced, first, in a l-g quantity and, later, in a 30-g batch, by the solvent extraction process previously under development. The material was considerably better than that available for use as analytical standards, and additional quantities are being prepared for use as improved standards. The only reportable impurity in the 30-g batch, as determined by spectrographic techniques, was 5 ppm maghesium. xiii & Investigations of phase relationships in BeO-metal oxide systems were continued. Data were obtained for the binary systems Be0O-CaO, BeO- MgO, BeO-Ce0,, and BeO-Zr0O,. The effects of cooling rates on the forma- tion of the intermediate compound Caj;Be30s; in the Be0-Ca0 system were in- vestigated, and the information developed was used as a guide for the successful growth of single crystals of the phase. The eutectic tempera- ture and composition were determined to be 1860 * 10°C and 69 * 2 mole % BeO in the BeO-Mg0 system; 1890 + 20°C and 63 * 3 mole % BeO in the BeO- Ce0, system; and 2045 # 10°C and 58.7 * 2 mole % BeO in the Be0-Zr0, system. 7. Engineering and Heat-Transfer Studies Construction of an apparatus for studying forced-flow boiling with potassium has been completed. A temporary boiler section was installed so that cleanup and the determination of the operational characteristics could be accomplished while awaiting completion of the final boiler as- sembly. Preliminary heat-transfer data were obtained for potassium in the laminar-flow regime; the results are in reasonable agreement with data on mercury and lead-bismuth eutectic in the same flow range. Measurements of the thermal conductivity of molten lithium were con- tinued using a modified version of the axial heat flow apparatus. The apparatus changes allowed better definition of the heat flow and tempera- ture distributions and resulted in some improvement in experimental pre- cision. The most recent set of measurements show less variation with temperature than did the earlier results; thus, the conductivity varied from about 27 Btu/hr-ft-°F at 750°F to 29 Btu/hr-ft-°F at 1500°F in con- trast with a variation from 20 to 40 Btu/hr.ft.°F for the previous meas- urements over the same temperature range. The results are in good agree- ment with published results. 8. Radiation Effects In-pile tube-burst tests at 1800°F were conducted on specimens of Cb—1% Zr alloy in the poolside facility of the ORR. The rather high stresses used in the experiment caused all specimens to rupture within xiv s‘:“"“uflllllllllll 60 hr., The results indicated that the in-pile rupture strength was about 10% less than the out-of-pile strength. Longer times to rupture and higher radiation doses will be obtained with the use of lower stresses in future experiments. Additional in-pile tube-burst experiments were completed on Inconel and stainless steel. Serious physical damage to BeO which had received fast-neutron dos- ages of the order of 1021 neutrons/cmz'in the ETR was revealed during postirradiation examination of test specimens. The damaged specimens con- "stituted a portion of the 57 BeO pellets 1 in. in length and 0.4 to 0.8 in., in diameter that were contained in 16 capsules distributed among five separate irradiation assemblies. A comprehensive range of exposures was achieved at a significant variety of temperatures. Damage observed in the BeO that received the higher dosages varied from minute cracks to gross fracture and disintegration into powder. The BeO has been recovered from the capsules of all five assemblies. The disassembly and inspection of the last three assemblies was carried out in the Battelle Memorial Institute hot cell facility. Visual in- spection, macrophotography, and'physical dimension measurements of the specimens have been completed. Metallographic examination was completed for the specimens used in the early experiments, and exploratory gas analyses and x-ray diffraction studies have been performed. A major por- tion of the thermal flux dosimetry analysis was completed, and corre- sponding fast flux determinations are under way. A survey of the tem- perature data and composition changes of the thermal-barrier gas is being made to evaluate thermal conductivity changes observed during the progress of the irradiations. Part 2. Shielding Research 9. Development of Reactors for Shielding Research Upon completion of the critical experiments with the TSR-II, the reactor was disassembled, all the temporary eguipment which had been in- stalled for the critical experiments was removed, and the reactor was S xv reassembled for operation. As a result of the shakedown runs of the com- plete system which continued through December 1960, some components had to be modified before the reactor was placed in operation in January and subsequently operated at a 100-kw power level. Minor design changes are contemplated to provide more reliable operation. 10. Development of Reactors for Shielding Research Gamma-Ray Spectroscopy. During shakedown tests with the BSF model IV gamma-ray spectrometer it became apparent that the spectrometer must be positioned at least 5 ft from the reactor surface to reduce the gamma- ray background to an acceptable level. As a result, measurements of the radiations leéving the surface of the reactor will have to be made by attaching a 5-ft-long air-filled cone to the front of the spectrometer shield. The large composite NaI(Tl) crystal obtained for use with the spectrometer has been tested further and appears to be satisfactory. The finite limit imposed upon the output of a gamma-ray scintilla- tion spectrometer by the number of channels available in a multichannel analyzer prevents exact determination of a continuous gamma-ray spectrum. This difficulty is equal in importance to the better-known problems of- fered by statistical distributions of ‘observed counts and the lack of exéct knowledge of spectrometer responses. The problem is currently being studied by arbitrarily assuming an exact knowledge of the number of cdunts produced at given energies and an exact knowledge of a spec- trometer response in order to determine how closely a contimious spec- trum can be reproduced. Single parameters that are dependent upon the spectrum, such as gamma-ray dose, can be written as linear combinations of spectrometer response functions if the cross sections governing these parameters are indeed exact combinations of a finite number of response functions. This is not usually true, but the assumption of "nonnegativity," that is, that the particle spectrum is everywhere equal to or greater than zero and that the energy range lies between zero and infinity, permits reasonable estimates to be made. Upper and lower bounds can be determined in this fashion for wvarious parameters. A method borrowed from the mathe- matics of economics and scheduling, called "linear programing', for which xvi computer codes already exist, can be used to attack the problem of setting the narrowest possible upper and lower bounds for a given parameter. For multiple-parameter problems, Jjoint bounds may be set if the correlations are known. Approximate methods for determining the "smallest allowed re- gion" have been found, based on considération of the allowed regions as rectangles of given length, width, and center coordinates, or an irregular bodies with given centroids and moments of inertia. Neutron Spectroscopy. The possible use of a neutron-sensitive semi- conductor detector for neutron spectroscopy has been investigated further with some resulting modifications in the design. From the tests conducted thus far it is concluded that the full width at half maximum is relatively constant, of the order of 300 kev, regardless of the incident neutron energy. The presence of gamma rays will produce a broadening of the neu- tron peak, but measurements have been made in gamma-ray fields up to ~200 r/hr without appreciably altering the response. Further, the device produces a pulse height that is a linear function of incident neutron energy. The results of calculations indicate that the foreground-to- background ratio will be about 5 to 1 when the counters are used to meas-} ure a fission spectrum, and a preliminary experiment with the TSR-II has substantiated this prediction. The effects of further variations in the design are being studied; specifically, the effects of increasing the area of the counters and substituting Li® metal for Li®F as the neutron- sensitive material. A Fast-Pulse Integrator for Dosimetry. When the A-1A amplifier used during dosimetry measurements at the Lid Tank Facility was replaced with a DD-2 amplifier which had a known dead time of 2.6 usec, it became de- sirable to develop an integrator which had a resolving time within the same interval so that counting loss corrections could be calculated. As a result, a new integrator with a total resolution of 2.4 psec was placed in operation early this year. It consists of two sections, each having four discriminator stages coupled together as a true binary scale of 16. The output pulses from both sections are fed to a single decade scaler. xvii 11. . Basic Shielding Studies Bnergy and Angular Distributions of Neutrons Emerging from Planar Surfaces of Diffusing Media. The investigation of the energy and angular distributions of neutrons emerging from planar surfacesgs of diffusion media has now included NDA moments method calculations of the energy spec- tra of neutrons from a point source in an infinite medium of LiH for dis- tances from the source of 11.5, 23, and 34 g/cm2 and for the four source ~energies 16.3, 1.48, 0.12, and 0.01 Mev. The results indicate that low- energy eguilibrium is reached at all thicknesses for all input energies employed, which 1s in agreement with earlier Monte Carlo calculations. In addition, measurements have been made at the BSF of the angular dis- tributions of 4.95-ev, l.44-ev, and subcadmium neutrons leaking from a 4=in., -thick LiH slab, the results of which appear to agree with calcula- tions performed with the NDA NIOBE Code. It was originally planned that energy spectral measurements would be made with a neutron chopper facility to be built at the BSF, but the withdrawal of project support has cancelled the construction of the facility. Some experimental data which can be compared with computed energy spectra for LiH will be obtained, however, from a series of measurements at the Linear Accelerator Facility of General Atomic, San Diego, in early May 1961. Experimental Verification of a Geometrical Shielding Transformation. The Lid Tank Shielding Facility is currently involved in a program to verify a geometrical transformation concept that the axial dose rate from a large source plate in a homogeneous medium can be inferred from the dose rates from a small disk source. During the program the source plate size is being varied by inserting cadmium irises of several different diameters between the source plate and the incident beam of neutrons from the ORNL Graphite Reactor. Thus far, only the measurements with a 7-in.- diam iris have been completed. 12. Applied Shielding The preanalysis for the Pratt & Whitney divided-shield experiment at the TSF has been completed. The shield design consists of a highly asymmetric reactor shield surrounding the TSR-II and separated from the xviii #: sgf . I TSF compartmentalized cylindrical crew shield by approximately 64 ft. The reactor shield incorporates two advanced shielding materials; that is, lithium hydride as the neutron shield and depleted uranium as the gamma-ray shadow shield. The radiation sources which were considered important and calculable included fission neutrons from the core, prompt- fission and fission-product decay gamma rays from the core and shadow shield, and capture gamma rays from the reactor and reactor shield. The attenuation and transport_processes were divided into three categories: determination of the spectra of radiation leaking from the reactor shield; the subsequent scattering of this radiation in air; and the attenuation by the crew compartment of radiation both unscattered and scattered in air. ' Experimentation with the Pratt & Whitney divided shield mockup at the TSF is under way, and some preliminary results are available. Dur- ing the course of the experiments the optimization of the reactor shield is being checked by use of a water-filled "patch" tank constructed so that it can be moved around the shield in the horizontal midplane. Also, the relative effectiveness of water, transformer oil, and a mixture of transformer o0il and an organic boron compound as the crew compartment shield material is being investigated. Most of the measurements com- pleted to date consist in measurements of the dose rates and fluxes in air around the reactor shield, although some measurements have also been made inside the crew shield. Additional in-air and crew compartment measurements are still to be made, as well as a number of spectral meas- urements. S xix A s - mr TR R TR R Sy ST S30E ;&‘,:?fii’ - iR . * eI R - i PART 1. MATERTALS RESEARCH AND ENGINEERING 1. MATERTALS COMPATTIBILITY Oxidation of Columbium Alloys at Low Pressures H. Inouye The reactivity of columbium with oxygen places severe restrictions on the use of columbium at high temperatures because of scaling and em- brittlement by internal oxidation. Through alloying, it has been possible to reduce the reaction rates as much as a factor of 100 at atmospheric pressure. Furthermore, the reaction rates of unalloyed columbium can be reduced by a factor of 400 by reducing the oxygen pressure to 1 X 10~% mm Hg in the temperature range 850 to 1200°C.1*2 Studies are being made to determine whether alloying will have the same effect at low oxygen pres- sures as at atmospheric pressure. In the previous progress report,3 the general effect of alloying additions on the reaction rates was reported for exposure ét 1000°C to oxygen at a pressure of 5 X 104 mm Hg and at 1200°C to air at a pressure of 5 x 10™% mm Hg. This study has been extended to include rate studies at 850°C, as well as screening tests on additional compositions at 1000 and 1200°C. In general, it has been found that when the alloying element forms an oxide more stable than the oxides of columbium, low concentrations of the element reduce the reaction rate in low-pressure oxygen below that for unalloyed columbium. However, as the concentration of the alloying element is increased the reaction rate also increases. This behavior is shown in Fig. 1.1 for zirconium additions and in Fig. 1.2 for aluminum additions. ©Similar results have also been observed for additions of cerium, beryllium, hafnium, and titanium. The difference between the oxidation rate of high-purity columbium and that of commercial-purity columbium is about that observed for low levels of the alloying elements (Fig. 1.3). | 1ANP Semiann. Prog. Rep. April 30, 1960, ORNL-2942, pp. 3—7. 2H. Inouye, The Scaling of Columbium in Air, ORNL-1565, Aug. 29,. 1956, JANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, pp. 36. UNCL ASSIFIED ORNL~-LR-DWG 57251R 5.6 1000°C 4.8 4 0, PRESSURE: 5x10” *mm Hg ?2'40 5% Zr > TC 1400 min X = < HIGH- PURITY Cb "f 2.4 T % 1070 Zr 2 16 0.8 0 o 100 200 300 400 TIME (min) Fig. 1.1. Oxidation Rates of Columbium-Zirconium Alloys at 1000°C in Low-Pressure Oxygen. UNCLASSIFIED ORNL-LR-DWG 57252R 3.2 ! 1000°C ‘ 58 0, PRESSURE: 5x10~% mm Hg _6_,£LCb 2.4 o + / - L~ 2.5% Al o ek g //"‘—— ~. 20 L Y L o E /fi/d ] z 16 P ' (] 1.5 o]o Al £\t o 1.2 _"fl___fl_,.—— 0.5% Al L flé.——-—f"’" * 0.8 - ® 0-4 { 0 0 100 200 300 400 TIME {min} Fig. 1.2. Oxidation Rates of Columbium-Aluminum Alloys at 1000°C in Low-Pressure Oxygen. UNCLASSIFIED ORNL-LR-DWG 57250R 3.0 ’ t000°C -4 0, PRESSURE:5x10 mmHg —".-—. o—* / o £ 2.0 - -a_ // ELECTRON-BEAM-MELTED Cb E ® =z o/ . S 0'/ LO—" / - ° . { }dww w /. \ 2 10 . : . 79 r o ¢ 7—45”0”64i\\‘ARC—CAST Cb / (COMMERCIAL) ® 2 | o O 100 200 300 400 TIME {min) Fig. 1.3. Oxidation Rates of Unalloyed Columbium at 1000°C in Low- Pressure Oxygen. Inasmuch as the oxidation characteristics of an alloy are dependent to a large extent on the type of oxide formed on the metal, it was ex- pected that a very high concentration of alloying material might prove beneficial, since the probability of forming a different oxide would be greatér. However, as shown in Figs. 1.4 and 1.5, the higher levels of elemental additions proved to be detrimental to the oxidation resistance for the specific test conditions used. Tests of another class of alloying elements in columbium showed little, if any, change in the oxidation rate as the amount of alloying element was increased. The elements added were tin, molybdenum, tungsten, and palladium., In these tests the added element formed oxides less stable than the oxides of columbium. The rate curves for tin additions are shown in Fig. 1.6. The oxidation rates of columbium in low-pressure air at 1200°C were greatly different from the oxidation rates in okxygen, as shown in Fig. 1.7. In oxygen, the reaction rates fiere linear from the beginning of the test; however, in air, an incubation period of about 6 hr was required before a UNCLASSIFIED ORNL-LR-DWG 57249R 850°C 0, PRESSURE: 1x10™% mm Hg 3.0 — 2.5 ‘0'4’31:10% Mo-10% Ti- BAL.Cb — c/ a E / [ ~ £ 2.0 F-48:15% W-5% Mo- 0.7 % Zr - BAL. Cb— 3 A-—fiL-_—--AflA--‘}'—'A" - e AT B = [ %— e ot I 1.5 o o o = 1.0 / . / “/ F-82:33% Ta-0.7% Zr—BAL.Cb / < / — 0.5 " 4 “—i‘_—.. / ._..--' ! et UNALLOYED Cb —. .-l— 0 Se—oT® t I 0 100 200 300 400 TIME (min)} Fig. 1l.4. Oxidation Rates of Commercial Columbium Alloys at 850°C in Low-Pressure Oxygen. measurable amount of oxidation occurred. Furthermore, the oxidation rates for equivalent oxygen pressures were significantly lower when the oxygen contained nitrogen (from air). Numerous investigations have shown that the nitrogen contamination of columbium oxidized in air is slight, and it appears that the observed difference might be due to a "poisoning" of the columbium surface or the formation of a nitride layer. _ Since nitrogen has an important effect on the oxidation rate in low- pressure air, it would be expected that elements which form nitrides in a nitrogen-containing oxygen atmosphere would show different behavior in a nitrogen-free oxygen atmosphere. In tests of this hypothesis, it was found, as shown in Fig. 1.8, that increasing the zirconium content reduced the oxidation rate in a nitrogen-containing atmosphere; whereas, in a nitrogen~free atmosphere the reverse effect was observed. This study has shown that alloying additions do not reduce the oxida- tion rates of columbium at low pressures in the same proportion as at UNCLASSIFIED ORNL-LR-DWG 57253R . 1000°C ) —~4 i 0, PRESSURE: 5x10° " mm Hg /’U/ 5 / — / £ 70% Cb-25%Ti—5% Al ALLOY { E 4 // = = o 804-HNO3-H20 (11-11-28-50 vol %). 500X (Confidential with caption) of 1 to 500 hr.’ The specimens continue to lose weight, however, as a function of time. This can be related to the oxygen being removed from the specimen. The weight-loss data and the oxygen-loss data are compared in Fig. 1.10. Hardness measurements were made from one surface to the other along the cross section of each specimen, and these data are plotted in Fig. 1.11. These results indicate that although intergranular attack occurs very rapidly and does not increase with time, an oxygen gradient is established across the specimen, and leaching of oxXygen by the lithium continues until a minimum level is reached. This level was approximately 250 ppm for the conditions of this test. The columbium is considerably softened and therefore weakened by this depletion in oxygen, but there is no change in the metallographic appearance of the specimen. The 250- and 500-hr test specimens showed slightly less attack than the specimens tested shorter times. ©Since the depth of attack in 5ANP Semiasnn. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 9. 10 UNCLASSIFIED ORNL-LR-DWG 57001R —-1000 WEIGHT LOSS —-800 OXYGEN LCSS -600 OXYGEN LEVELS OF Cb SPECIMENS —-400 PRICR TO EXPOSURE = {000 -1500 ppm SPECIMEN THICKNESS = 0.043in. LOSS (ppm) -200 1 2 5 10 20 50 100 200 500 TIME OF EXPOSURE (hr) Fig. 1,10. Weight Loss and Oxygen Loss of Columbium Specimens as a Function of Time of Exposure to Lithium at 1500°F. UNCLASSIFIED ORNL-LR-DWG 57002 170 1hr 160 /f\ 150 . ® 2hr \ , At o 130 o} N 3hr N 120 T‘\ 0 ! /e 1o 'HARDNESS,DPH (200g LOAD) 100 \ / 16 hr \ 90 / ; \ / 20 \ \ 5 L] g T 0.040in. - 70 ‘ : EDGE OF CENTER EDGE OF SPECIMEN SPECIMEN Fig. 1.11. BHardness Profiles Across Columbium Specimens Exposed to Lithium as a Function of the Time of Exposure. 11 polycrystalline specimens varies from grain to grain, the determination of the effect of time as a test variable could best be evaluated in specimens cut from a single crystal. ©Opecimens of this type will be exposed to lithium for 1, 100, and 1000 hr. Effect of Lithium Purity From previous work® it appears that oxygen or nitrogen in lithium does not tend to promote corrosion of columbium if the oxygen concentra- tion of the columbium is low. In the earlier experiments there was no indiéation as to what the effect might be i1f the oxygen concentration of the columbium were high. Columbium specimens to which approximately 1000 ppm O, had been added were therefore exposed 100 hr at 1500°F to Li, Ii + 2 wt % LisN, and Ii + 2 wt % Li,O. It was found that the columbium specimens exposed to lithium and the Ii + 2 wt % LisN were completely attacked, but that the columbium specimen exposed to Li + 2 wt % Li,O was attacked to a depth of only 1 mil. If the driving force for corrosion is the tendency for lithium to react with the oxygen in the columbium to form lithium oxide, then this driving force would be reduced if the oxygen content of the lithium were high. The re- sults of this test tend to support this supposition. Further tests will be conducted in an attempt to substantiate this conclusion. Several possible mechanisms could explain the results of the columbium- lithium corrosion studies which have been conducted thus far. One mecha- nism would require lithium diffusion into the columbium and the precipita- tion of lithium oxide or a complex lithium-columbium-oxygen compound as a corrosion product. The fact that corrosion occurs so rapidly and does not progress with time {(as determined by metallographic studies) does not favor a diffusion mechanism. An alternate mechanism assumes that the lithium reacts chemically with columbium oxide at the surface of the specimen which might be present either before exposure to lithium or might precipitate during the early stages of the test. As was mentioned earlier, 6ANP Semiann. Prog. Rep. March 31, 1959, ORNL-2711, p. 11. 12 no evidence has been found metallographically to indicate the presence of columbium oxide before testing. Corrosion Studies on Cb—1% Zr Alloy in Lithium J. R. DiStefano Effect of Oxygen Additions to Cb—l% Zr Alloy on Its Room-Temperature Tensile Properties A series of Cb—1% Zr alloy sheet tensile specimens were oxidized at 1830°F at an oxygen pressure of 5 X 1072 mm Hg. Following oxidation, one group of specimens was exposed to lithium for 100 hr at 1500°F Withofit further heat treatment. A second group was homogenized at 2910°F, and a third group was homogenized at 2910°F and then exposed to lithium at 1500°F for 100 hr. All these specimens and a group which received no further treatment after oxidation were tensile tested at room temperature. The results of>these tests are plot%ed in Fig. 1.12. The tensile strengths of the specimens tested in the "as-added" condition increased with in- creasing oxygen concentration and the ductility decreased (0% elongation for those specimens containing 1000 ppm 0, or greater). When oxidation 3 SRR (x11500) ORNL-LR~DWG S5TOOO0R | ] | OXYGEN ADDED AT 1830°:‘>_’__ —— ‘// L= LT 100 /{/// OXYGEN ADDED AT 1830°F AND TENSILE STRENGTH (psi) _J/’ /§XPOSEDT0 Li AT 1500°F FOR 100 hr ./ | _ 50 ~au OXYGEN ADDED AT 1830°F //HEAT TREATED AT 2910°F ,E; L _ o W T ! \ L il | “~OXYGEN ADDED AT {830°F I[HEAT TREATED AT 2910 °F AND EXPOSED O Li AT 1500°F FOR 100 fr 0 500 1000 1500 2000 2500 0, CONCENTRATION (ppm) Fig. 1.12. The Effect of Oxygen Additions, Heat Treatment, and Ex- posure to Lithium on the Room-Temperature Tensile Strength of Cv—1% Zr AMloy. 13 was followed by a 2910°F heat treatment, the tensile strength remained constant with increasing oxygen concentration, and there was no loss in ductility. Exposure to lithium following the oxidation treatment resulted in attack of the Cb—1% Zr alloy. The depth of attack increased as a func- tion of increasing oxygen concentration. The tensile strengths of these specimens decreased as a result of corrosion testing but femained above those of the heat-treated specimens. The corroded specimens were also more brittle. If the specimens with oxygen additions were heat treated at 2910°F before exposure to lithium, the tensile properties were found to be essentially the same as those prior to exposure, and no attack was observed. It is apparent from these results that the form and/or distribution of oxygen in the Cb—1% Zr alloy has a marked effect on both the strength and the corrosion resistance of this materisl. In order to predict the cor- rosion resistance of the alloy, it is necessary that both the oxygen con- tent and the thermal history of the material be known. Effect of Heat Treatment of Welds In order to establish the conditions under which corrosion of high- oxygen-content Cb—1% Zr alloy welds by lithium will occur, eight specimens which had reeeivedsv®rious pretest treatments were tested in lithium at 1500°F for 100 hr. The results of this test are summarized in Table 1.1. It may be seen from the table that the pretest treatment determined whether corrosion did or did not occur. When welding was the final treat- ment before exposure to lithium (weld Nos. 3 and 7), almost complete attack of the Cb—l% Zr alloy occurred. Also, as previously o'bserved,7 when oxygen was added to the welded alloy at 1830°F and no additional heat treatment was given, the Cb—1% Zr alloy was susceptible to corrosion by lithium. When the oxygen addition temperature was raised to 2190°F, the attack fol- lowing exposure to lithium was 2 mils or less and was found in only a few areas. Heat treating at 2370°F as a final treatment completely eliminated the attack. It appears that heat treating Cb—1% Zr alloy in the range 2190 to 2370°F stabilizes the oxygen and no corrosion occurs. 7ANP Semiann. Prog. Rep. Oct. 31, 1959, ORNL-2840, pp. 30-32. 14 Table 1.1. Effect of Oxygen-Contamination Temperature and Subsequent Heat Treatment on the Corrosion Resistance of Cb—1% Zr Alloy Welds to Lithium weld® Histor;yb Prior to Testing in Lithium 0, Added Attack No. for 100 hr at 1500°F (ppm) (mils) Welded; O, added at 1830°F 1700 19 Welded; O, added at 1830°F; heat-treated 1700 0 2 hr at 2370°F 3 O, added at 1830°F; welded 2500 25 O, added at 1830°F; welded; heat-treated 2500 0 2 hr at 2370°F Welded; O, added at 2190°F 1300 6 Welded; O, added at 2190°F; heat-treated . 1300 2 hr at 2370°F 7 O, added at 2190°F; welded 1800 25 8 '02 added at 2190°F; welded; heat-treated 1800 0 2 hr at 2370°F aInert—gas—shielded tungsten~arc welds in 0.060-in. sheet. bStarting material was annealed for 2 hr at 2910°F prior to any subsequent treatment. Specimens that were corroded had greater hardness than the uncorroded specimens both before and after the test in lithium. This indicates that the oxygen distribution which makes the alloy susceptible to corrosion also strengthens it appreciably. Further studies will be made to deter- mine whether there is a correlation between Cb—1% Zr alloys which have been observed to undergo an aging reaction under certain conditions and Cb—1% Zr alloys which are subject to corrosion by lithium. Dissimilar-Metal Mass~Transfer Studies in Cb—1% Zr Alloy-NaK-Type 316 Stainless Steel Systems J. R. DiStefano, E. E. Hoffman Effect of Carbon and Nitrogen Pickup on Tensile Properties It was observed?® previously that CbC and Cb,N layers approximately 0.001 in. thick formed on the surfaces of Cb—1% Zr alloy specimens exposed, 8ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 13. 15 to NaK in a type 316 stainless steel container at 1700°F for 500 hr. In these tests there was a type 316 stainless steel to Cb—1% Zr alloy surface area ratio of 10:1. It was also found that when these surface layers were removed the nitrogen concentration of the specimens remained high but the carbon concentration was only slightly higher than that of the starting material.® | | In order to evaluate the effects of the layers on the mechanical properties of the alloy, several static tests were conducted under the conditions deécribed above on alloy specimens annealed at 2190 and 2910°F. One set of specimens was electropolished to remove the surface layers that formed during the test. These specimens and specimens with the layers intact were then tensile tested at room temperature and at 1700°F. Control specimens were heat treated 500 hr in argon and were also tensile tested at room temperature and 1700°F. The results of these tests are summarized in Table 1.2._’ The interpretation of the behavior of the specimens that were annealed at 2910°F is complicated by the aging that occurred during the test period. The Specimens annealed at 2190°F did not age during the test period and therefore gave more reliable information regarding the effect of the carbon and nitrogen increase. Results of chemical analyses of the Cb—1% Zr alloy specimens annealed at 2190 and 2910°F are similar in that both indi- cate a high nitrogen_concentration after removal of the surface layers. The changes observed in tensile properties of specimens 11, 12, and 13 and of specimens 18, 19, 20, and 21 illustrate the strengthening effect of the high nitrogen concentration. After removal of the layers, the tensile strength remains high. These data are consistent with the chemi- cal analyses for nitrogen. An interesting effect was observed when the control specimens were heat treated in a type 316 stainless steel container. The Cb—1% Zr speci- mens picked up carbon and nitrogen, apparently from the stainless steel, although an inert-gas environment of argon separated the two metals. An increase in the tensile strength and a loss in ductility simil@r to those for the specimens exposed in the NaK system were observed. When the 16 specimens annealed at 2190°F were heat treated in a columbium container, however, no change in tensile properties was found. Table 1.2. Effect of Brittle Layers on Tensile Properties of Cb—1% Zr Alloy Tensile , Elongaticn Chemical Analysis Specimen Specimen History Test gi:ziéih in 2 in. (ppm) No. Temperature . Gage —_—_— (°F) (psi) (%) Ne G 0, H Specimens Annealed at 2910°F 1 Annealed 70 38 700 14.0 85 120 92 14 2 Annealed and heat treated® 70 79 000P 0 710 560 150 5 3 Annealed and heat treated® 70 65 000P 0 Not analyzed 4 Annealed and exposed® 70 74 000 0 1000 300 170 7 5 Annealed, exposed, and electro- 70 83 000 0 700 120 1S90 19 polished 6 Annealed 1700 25 000 10.5 85 120 92 14 Anmnealed and exposed 1700 €9 800 0 Not analyzed Annealed, exposed, and electro- 1700 58 300 0 - Not analyzed polished Specimens Annealed at 2190°F 9 Annealed 70 37 100 23.5 91 50 100 10 10 Annealed and heat treated® 70 69 500 14.0 830 960 250 19 11 Annealed and heat treated® 70 38 800 26.0 Not analyzed 12 Annealed and exposed® 70 88 500 13.5 920 1730 250 16 13 Annealed, exposed, and electro- 70 79 999 10.0 550 120 270 20 polished 14 Annealed 1700 31 000 13.5 Not analyzed 15 Annealed 1700 32 000 14.5 Not analyzed 16 Annealed and heat treated® 1700 42 000 8.5 Not analyzed 17 Annealed and heat treated?® 1700 39 500 11.5 Not analyzed 18 Annealed and heat treated® 1700 28 200 19.5 Not analyzed 19 Annealed and heat treated® 1700 27 900 18,0 Not analyzed 20 Annealed and exposed ‘ 1700 48 500 6.5 Not analyzed 21 Ammealed and exposed 1700 4'7 500 6.5 Not analyzed 22 Annealed, exposed, and electro- 1700 41 500 4.0 Not analyzed polished 23 Annealed, exposed, and electro- 1700 44 500 4.5 Not analyzed polished ®Heat treated 500 hr at 1700°F in type 316 stainless steel container with an argon atmos- phere. : b . . . - Specimen aged during heat treatment and picked up nitrogen and carbon from type 316 stain- less steel container. CExpose¢f§OO hr at 1700°F in (Cb—1% Zr)-NaK-type 316 stainless steel system; surface area . ratio of stainless steel to Cb—1% Zr alloy was 10:1. dElectropolished to remove brittle surface layers; approximately 2 mils removed. “Heat treated 500 hr at 1700°F in columbium container with an argon atmosphere. 17 Effect of Stress on the Rate of Carbon and Nitrogen Pickup In order to determine what effect, if any, stress might have on the rate of pickup of carbon and nitrogen by columbium in a columbium—INaK- type 316 stainless steel system, columbium specimens were tested in stressed and unstressed conditions in the same system. Columbium was used in these tests rather than Cb—1% Zr alloy because wrought-alloy ma- terial was not available in a form suitable for the fabrication of the desired specimens. On the basis of earlier tests? it was assumed that the mechanisms of carbon and nitrogen transfer to the alloy would not be appreciably different from those for transfer to columbium. Two 10-in. columbium rods were machined so that each had a 0.4-in.-diam section and a 0.3-in.-diam section of equal length. One of these specimens was placed in tension during the exposure while fhe other was suspended immediately ad jacent to it and was under no load. The test conditions and the results of chemical analyses of turnings taken from the specimens after test are summarized in Table 1.3. These data indicate that for the conditions of this test stress was not a factor in the rate of pickup of either carbon or nitrogen. Both the stressed and unstressed specimens picked up substantial amounts of carbon and nitrogen, and the stressed specimen actually showed less increase in these elements than the unstressed specimen. Although the nitrogen con- centration is higher at the surface, some diffusion into the specimen occurs, but the carbon increase is almost completely confined to the sur- face layers. Since no effect of stress was observed, it would appear either that stress does not affect the diffusion rate of nitrogen and carbon in the columbium or thét the limiting step in the transfer process at this tem- perature is the rate at which carbon and nitrogen are supplied to the columbium surface. That is, the diffusion rate of carbon and nitrogen in the type 316 stainless steel and/or the solubility of carbon and nitrogen in the NaK could be the major rate-controlling steps. ANP Semiann. Prog. Rep. April 30, 1960, ORNL~2942, p. 1l. 18 Table 1.,3. Chemical Analyses of Columbium Before and After Exposure to NaK in a Type 316 Stainless Steel Container for 500 hr at 1700°F Rod Depth of Turnings CPemical Analyses Diameter Exposure Condition® From Surface (ppm) (in.) ‘ (mils) N, o 0.563 As received, not exposed (b) 93 250 0.4 Stressed to 1700 psi 0~5 2000 1600 0.4 Stressed to 1700 psi 5-10 520 380 0.4 Stressed to 1700 psi 10-15 400 290 0.4 Stressed to 1700 psi 1520 420 280 0.3 Stressed to 3000 psi 0-5¢ 2400 1700 0.4 Unstressed 0—5 2500 2300 0.4 Unstressed 510 650 480 0.4 Unstressed 1015 600 250 0.4 Unstressed 15-20 640 420 0.3 Unstressed 05 2900 2500 0.3 Unstressed =10 1100 660 0.3 Unstressed 1015 680 480 0.3 Unstressed 1520 650 420 a’I'ype 316 stainless steel to columbium surface area ratio: 10:1. bBulk analyses; turnings on as-received rod presently being made. cSam.ples lost for remaining analyses. 10,11 o1 this dissimilar-metal These and previous isothermal tests system have shown that nitrogen is the principal element that increases the strength and reduces the ductility of Cb—1% Zr alloy in the system Cb~1% Zr alloy-NaK-type 316 stainless steel., Transfer of carbon is limited to a surface layer which has little or no effect on the mechanical proper- ties. Stress does not increase the rate of pickup of either carbon or nitrogen. The effect of the surface area ratio has not been measured in any detail; but, in a single test conducted with the columbium surface aresa greater than the type 316 stainless steel surface area, a CbC layer was found on the éurface of the type 316 stainless steel. No effort has thus far been made to determine the effect of specimen thickness, the effect of flowing NaK, the effect of thermal gradients, or 10ANP Semiann. Prog. Rep. Oct. 30, 1960, ORNL-3029, p. 13. 11ANP Semiann. Prog. Rep. April 30, 1960, ORNL-2942, p. 9. 19 the effect of reducing the nitrogen concentration of the stainless steel. These variables will be considered in future tests. With the data obtained thus far, 1t cannot be stated categorically that these materials are not compatible under the conditions described, but only that considerable changes in the mechanical properties of the Cb—1% Zr alloy can occur under the conditions of these tests. Vapor Pressure of NaK (43.7 wt % K) in the Temperature Range 1520 to 1832°F J. R. DiStefano The design of the apparatus used in the dissimilar metal tests de=- scribed above was such that it was necessary to know the vapor pressure of the NaK in order to specify accurately the stress on the columbium specimen. Experimental determinations of NaK vapor pressure were there- fore made in the temperature range 1520 to 1832°F. For these measurements, the NaK was heated in an evacuated stainless steel container until equi- librium was reached, and the pressure measured by a strain-gage type of pressure transmitter (Taylor Instrument Company's Transmitter Model 706 TN1103) was recorded.t? The container was well insulated to avoid re- fluxing in the system. Above 1830°F, large pressure changes were con- tinually observed, indicating that the system was in a nonequilibrium state; hence, the measurements were not continued above this temperature. The data obtained from this experiment, along with the data calcu- lated from Raoult's law and empirical equations for the vapor pressure 13 are given in Table 1.4. of pure Na and K as a function of temperature, The measured vapor pressure at the highest temperature (1273°K, 1830°F) was found to be much less than the calculated value, possibly as a re- sult of dimerization. 12y, R. Miller, High Temperature Pressure Transmitter Evaluation, pPp. 29-32, ORNL-2483, May 16, 1958. 13¢, B. Jackson (ed.), Liquid-Metals Handbook Sodium-NaK Supplement, Atomic Energy Commission, Department of Navy, Washington, D. C. (1955), p. 38. 20 Table 1.4. Vapor Pressure of NaK (56.3 Na—43.7 K, wt %) Temperature Pressure (mm Hg) °K -~ °F Calculated Measured 1100 1520 ' 758 97 1123 1562 024 1163 1173 1652 1388 1587 1223 1742 2014 _ 2016 1273 1832 2850 2420 Potassium Compatibility Studies D. H. Jansen, E, E, Hoffman Refluxing Potassgsium Capsule Tests The refluxing tests described previously14 are continuing in an ef- fort to evaluate the relative corrosion resistance of iron-, nickel-, cobalt~, and columbium-~base alloys in contact with refluxing potassium at temperatures in the range from 1600 to 2000°F. Sleeve-type inserts of the test materials have been used to line the walls of each test capsule. Weight changes observed on specimens from the condenser and boller regions of type 316 stainless steel {(iron-base) and a Haynes alloy No. 25 (cobalt- base) refluxing test system have been reported.15 Recent results of 1000-hr tests on Inconel and type 310 stainless steel are given in Fig. 1.13. The weight losses from dissolution in the condenser (vapor) region and the weight gains from deposition in the boiler (liquid) region were larger for type 310 stainless steel than for Inconel. This result is surprising in view of the fact that type 310 stainless steel has exhibited more resistance to mass transfer in pumped sodium systems than Inconel.l® The difference in corrosion resistance observed in these refluxing systems 14ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, pp. 14~15. 19ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 16. 16 ANP Semiann. Prog. Rep. Dec. 31, 1957, ORNL-2440, p. 147. \ 21 may have been influenced by the differences in condenser region tempera- tures for the two tests (Fig. 1.13). The lower condenser temperature (1535°F) in the Inconel test, as compared with the condenser temperature .of 1595°F for the type 310 stainless steel, suggests that less refluxing occurred in the Inconel test. Such a variation in test conditions would “tend to explain the ”appérent” superior corrosion resistance of the Inconel. Additional tests on these alloys at higher temperatures should indicate FLAT CAP SPECIMEN UNCLASSIFIED ORNL-LR-DWG 53497R R N R NN \/ SN % 1535°F UNCONEL)\/‘ § [N 1595°F (TYPE Q/ N 310 STAINLESS STEEL)igz \ |__TYPE 310 STAINLESS STEEL e N N — INCONEL \ \ _ N N N N VRN V1A £ N N = : N : - b @ N S g N N S VIR N N | N VI N N N N Vi1 | Vi A Y N ' V) N i;»‘— _4\§ \ \% Z3 NA 1IN - \ \1/ /§ —_— . N N N . . N N A ,~§§ N N £ N 1N - - 0 N 1 EN - . N N - N BN o NZER AN | 3 VA 3 N 181 ‘ / %f ] f\ . / §/ 7 f% / 1IN of §/ [ /% 1600°F N A T ) v N iy — | g -10 0 10 20 30 40 50 THERMOCOUPLE WELL —Y] |/ 2 ‘ WEIGHT CHANGE (mg/inS) ~MOVABLE THERMOCQUPLE Fig. 1.13. Test System Configuration and Weight-Change Results for Inconel and Type 310 Stainless Steel Specimens Tested in Refluxing Potas- sium for 1000 hr at a Boiler Temperature of 1600°F (871°C). 22 more conclusively which of these alloy systems (iron- or nickel-base) has the best corrosion resistance to potassium in refluxing systems. Boiling-Potassium Loop Tests An Inconel and a Haynes alloy No. 25 boiling-potassium loop have been fabricated. These loops are presently operating at a boiler temperature of 1600°F. The tests are scheduled for 3000 hr of operation at this tem- perature. ©Several modifications were made in these loops relative to the design previously described for the type 316 stainless steel test,17 as described below. | Machined, sleeve~type inserts similar to those used in the refluxing capsule testst? were used to.line the total length (6 ft) of the cold leg in order to obtain weight-change data, as well as metallographic data, in the regions where dissolution and deposition have been observed in similar previous tests on type 316 stainless steel. Two modifications have been made in the design of the boilers of the Inconel and Haynes alloy No. 25 systems. The vapor region of the boiler (2-in. sched.-40 pipe) has been lengthened by 12 in. to minimize the pos- sibility of liquid carryover to the condenser. The desired condition is that 1600°F saturated potassium vapor be delivered to the condenser. An additional boiler modification has been the incorporation of a sodium jacket contained in 4~in. sched.-40 pipe around the full length (3 ft) of the potassium boiler. The purpose of this jacket is to provide more uni- form heating to the potassium in the boiler, which should, in turn, tend to suppress the pressure and temperature instabilities observed in previous boiling-potassium loop tests. Another Haymes alloy No. 25 loop has been designed for operation at 1800°F. 1t is designed to permit quantitative weight-change determinations on the pipe wall in both the condenser and liquid regions of the cold leg. A Cb—1% Zr alloy loop for circulating boiling potassium at 2000°F is also being constructed. This loop will be operated in an inert-atmosphere chamber.: 17ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 19. 23 Purification of Potassium Metal: Methods of Reducing the Oxygen Content A, P. Litman, E. E, Hoffman A program for determining the most effective and practical method of removing oxygen from potassium has been initiated. Hot gettering, cold trapping, low-temperature filtration, and vacuum distillation are methods that will be investigated. With past purification experience with lithium as a basis, several batches (10 and 90 kg) of commercial potassium (Mines Safety Appliances) . were purified by hot gettering and subsequent cold trapping to produce a relatively pure material for use in compatibility and heat-transfer experi- | - ments. The results of the two purification experiments were quite similar, and therefore only the results for the 90-kg batch purification are given in Fig. 1.14. | | Oxygen determinations were made by the n-butyl bromide method on samples of potassium taken at 12 intervals during the 1350-hr purification experiment. The results indicated that the bulk of the oxygen was re- moved during gettering with & kg of titanium at 1200°F in the first 250 hr of the test. Only aAslight additional oxide reduction occurred during the subsequent 1100-hr cold-trapping treatment. The oxygen content of the as-recelved potassium was 270 ppm; 250 hr of gettering reduced this to 30 ppm; and 1100 hr of cold trapping yielded a product containing 25 ppm. Determination of Oxygen in Potassium J. C. White, G. Goldberg The mercury amalgamation technique and the butyl bromide method 8 are being evaluated to ascertain the most accurate and reproducible method for determining the oxygen content of potassium. The potassium in ques- tion is of high purity, containing oxygen in the order of less than 50 ppm. 187, C. White, W. J. Ross, and Robert Rowan, Jr., Anal. Chem. 26, 210 (1954). 24 GAS TYPE 347 STAINLESS STEEL TANK (12 in. OD) TITANIUM LINER 4-in. HOT AND COLD TRAP UNCLASSIFIED ORNL-LR-DWG 40851 R DISCHARGE PRESSURE GAUGE Xy in e — A TITANIUM SPONGE TITANIUM SCREEN Oxygen Content** Process Description Teml:;:mms 1;;:; (ppm) ’ Individual Analyses Average As Received 246, 268, 293 269 - Hot Gottarad at 650°C 650 50 26, 36, 38 33 Using 8.4-kg Ti Sponge 150 18, 50 36 . 650 250 20, 20, 20 20 100 33, 40, 65 46 250 32, 38, 40 v Cold Trapped* 400 49, 5% 50 Using B.4-kg Ti Sponge 80 550 20, 25, 50 32 45 700 19, 27, 28, 47, 57 3 100 18,27, 35 27 Cold Trapped* 200 16, 42 29 Using 8,4-kg Ti Sponge 35 300 14,19, 20 18 30 400 18, 33 28 *Melting Point K = 64°C **Analytical Method: n-butyl bramide (Ref, ORMNL- 1286} Fig. l.1l4. Potassium Purification Experiment. 3 25 Pepkowitz and Juddl® first used the amalgamation technique for the determination of éxygen in sodium. The method involves the repeated ex- traction of the metallic sodium with mercury; the mercury-insoluble sodium monoxide floats on the amalgam. When the extraction is completé, the sodium monoxide is dissolved in water and titrated with standard acid. With potassium, as with sodium, the assumption is made that the oxide is the only mercury-insoluble impurity that reacts with water to form the base. The amalgamation technigue is therefore also applicable to potas- sium metal analysis. The original amalgamation apparatus has been simplified, as shown in Fig. 1.15. The Jamesbury ball valve was modified so that the potassium sample could be transferred from an inert-atmosphere dry-box directly to the reaction vessel. The valve acts also as the reaction-vessel cap. The helium used is of high purity. In addition, an oxXygen getter and a liquid- nitrogen cold trap are used in series with the helium line. In operation, the reaction vessel is first evacuated and flamed to remove all moisture. Mercury, 20 to 30 em’, is admitted slowly to the reaction vessel, after which the system is pressurized with helium at 2 psia. The reaction vessel side-arm stopcock is closed and the Jamesbury valve is opened, allowing the potassium sample to drop into the mercury. When the potassium has reacted with the mercury, the resulting amalgam is drained, leaving approximately 1/2 in. of mercury above the lower stopcock to keep from losing the floating oxide. The reaction vessel is alter- nately evacuated, filled, pressurized, shaken, and drained until a water extraction of the last draining is neutral to phenolpthalein. The remain- ing mercury and the insoluble potassium oxide are then washed from the reaction vessel with water and the resulting base is titrated with standard acid. | A few samples of potassium were analyzed by the amalgamation method, but the results were inconclusive. Sampling techniques are being investi- gated whereby similar samples of potassium can be analyzed by both the 9L, P. Pepkowitz and W. C. Judd, Anal. Chem. 22, 1283 (1950). 26 amalgamation and butyl bromide methods to compare the sensitivities of both methods. UNCLASSIFIED ORNL—LR-DWG 58616 M—/ ._/\/\// STAINLESS STEEL SAMPLE RECEIVER MERCURY RESERVOIR 3/4-in. JAMESBURY VALVE HELIUM ;)i ¢ 0 . __ STAINLESS STEEL TAPERED JOINT TYGON TUBING VACUUM CONNECTION PYREX PIPE REACTION VESSEL T'ig. 1.15. Alkali Metal Amalgamation Apparatus. 27 2. AGING STUDIES OF COLUMBIUM-BASE ALLOYS Aging of Wrought Material D. 0. Hobson Data were presented in the previous report1 which indicated that oxygen in the Cb-1% Zr alloy inhibited the aging reaction. It was postu- lated that the solution and precipitation of zirconium carbides and/or nitrides were responsible for the aging reaction. In contrast, daté ob- tained more recently indicate that oxygen is the chief, and probably the sole, cause of the aging reaction under the conditions being studied. Thirteen heats of Cb—1% Zr alloy having oxygen contents ranging from 24 to 950 ppm have been examined. In addition, one heat contaminated to oxygen levels of 470 to 1200 ppm was examined. All specimens were aged following a solution anneal at 1600°C. With the exception of three heats containing only small quantities of oxygen (24 to 50 ppm), a definite cor- relation was found between the oxygen content and the ageability of the alloy. For a standardized anneal at 1600°C followed by an aging treat- ment at 927°C, six heats having up to 260 ppm oxygen showed an aging re- action. The remaining seven heats (320-950 ppm oxygen) showed no aging reaction. This behavior indicates that the solubility of the precipitating phase was affected by oxygen additions. To further investigate these effects, various heats were annealed at temperatures ranging from 1200 to 2000°C. Heat S8FW, which contained 260 ppm oxygen and which had shown aging after an anneal at 1600°C, was annealed at 1800°C and aged under identical conditions. As shown in Fig. 2.1, the aging reaction was accelerated. This behavior is similar to.that of an alloy in which one or both anneal- ing temperatures is under the solvus line of the precipitating phase, with the higher annealing temperature resulting in more solution of this phase. Heat S16EC, containing approximately 950 ppm oxygen, was annealed at 1600, 1800, and 2000°C and aged at 927°C. As shown in Fig. 2.2, the speci- mens annealed at 1600°C showed no aging response. The specimens annealed 1ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, pp. 24—25. 28 e UNCLASSIFIED ORNL-LR-DWG 56B803A {(x 103) | | © ANNEALED 2hr AT {800°C 90| ® ANNEALED 2hr AT 1600°C 80— ULTIMATE TENSILE STRENGTH ~— (8} (8) YIELD STRENGTH - . P / Y 60 (8) TENSILE AND YIELD STRENGTH {psi) v '/ o \ 25 50 100 200 500 750 AGING TIME AT 8i5°C (hr) Fig., 2.1. Results of Tensile Tests of Specimens of Colunibium- Zirconium Alloy Heat S8FW That Were Annealed 2 hr at 1600 and 1800°C and Aged at 815°C. at 1800°C showed some variations in hardness that might possibly be due to aging. The specimens annealed at 2000°C showed a definite aging peak after approximately 50 hr at 927°C. (The oxygen content of heat S16EC UNCLASSIFIED 75 ORNL-LR-DWG 56B05R A ANNEALED 2hr AT 1600°C ® ANNEALED 2hr AT 1800°C © ANNEALED 2hr AT 2000°C o wn 60 HARDNESS, Rznr &l w o O 45 0 16 25 50 100 200 500 AGING TIME AT 927°C (hr) Fig. 2.2. Results of Tensile Tests of Specimens of Columbilum- Zirconium Alloy Heat S16EC That Were Annealed 2 hr at 1600, 1800, and 2000°C and Aged at 927°C. was reduced from approximately 950 to approximately 850 ppm during annealing at 2000°C.) Specimens of heat S24WC, which contained 120 ppm oxygen and which aged after annealing at 1600°C, were contaminated with oxygen in amounts ranging fromv 470 to 1200 ppm and then annealed at 1600 and 1800°C and aged. None of the higher oxygen content speci- mens annealed at 1600°C showed any aging response other than a drop in hardness. However, after annealing at 1800°C, the specimens 29 with oxygen contents from 470 to 700 ppm showed definite increases in hardness, as indicated in Fig. 2.3. UNCLASSIFIED ORNL-LR-DWG 56806R 70 0, CONTENT 65 {ppm) ______-0- 470 530 ;;;;;32580____ 700 \ HARDNESS, R 37 wn o Q o # / ! ! » 1 - 1200 IF.\ _______--"__.—-—-O" \ ————|e . 45 1 et T $< 840 -~ T~-a ‘t4=:,r”" 800 40 0 25 50 75 100 125 150 AGING TIME AT 927°C (hr) Fig. 2.3. Bffect of Oxygen Con- tent on the Aging of Specimens of Columbium-Zirconium Alloy Heat S24WC-4 That Were Annealed 2 hr at 1800°C and Aged at 927°C. UNCLASSIFIED ORNL- I_.R- DWG 56807 2000 1900 ,/ AGING OCCURRED ”/,,/’ S 1800 . . ol =//./ A h i T 1700 = = -~ #1600 > o4 -k A A % ’,,/’//' b _ 1500 1 g 3 NO AGING OCCURRED uw 1400 & = =z 010 —o TEST NO. 1041 —707 ppm Op ] A TEST NO. 1009 -718 ppm O, i_ Z ul = 0.08 i o o \l o . A = 0.06 ; = T 0 0.20 [0 Q) L 0 o S 045 z 3 o Further, the improvements “ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, pp. 64—69. 72 UNCLASSIFIED ORNL -LR-DWG 58623 45 | KUTATELADZE ef ai(REF.7) ' A 40 | = =— CANEL (REF. 10) A ------ WEBBER efa/ (REF.9) A ® MILLER AND EWING {REF.5) g A ORNL (PREVIOUS) T A ORNL (CURRENT) x 35 > MELTING @ TEMPETATURE A — g - | - - 5 30 | - —— o - A A - ] ‘ Q -~ 4 | T el ”"- A a | sl E 25 1 ) -..T' uIJ | //4 ‘~~‘ — r ‘\‘ \5 20 | A L o | | 200 400 600 800 1000 1200 1400 1600 1800 TEMPERATURE (°F) Fig. 7.4. Summary of Data on the Thermal Conductivity of Lithium. have permitted a closer check on the accuracy of the measurements. The magnitude of the heat flow in the lithium sample can be better established through a comparison of the thermal fluxes in the upper and lower heat meters; the maximum deviation for the present investigation was less than 3%. An "in place" relative calibration of the thermocouples can be made at periodic intervals at temperatures up to 1200°F. One such check made following a measurement at 1490°F gave an average deviation of 0.45°F at 1200°F between nine thermocouples in the test piece. The better knowledge of the thermocouple locations in the guard cylinder allows closer esti- mation of the radial heat loss; in the current measurements this was found to be less than 1% of the axial heat flow in the sample region. By fixing the locations of the thermocouples in the thermowells more positively, it | has been possible to establish the interfacial temperatures wifih'greater 73 UNCLASSIFIED ORNL-LR-DWG 58255 GUARD HEATERS {4 2n.TYP | L0 | | ) < Y7 7 7 7 77T 77 | o ml L V0727773 NS W 7 \ L S\ ZS) - I % L — EXPANSION TANK | Al,05 REFRACTORY HEATER SUPPORT L MAIN HEATER L— TIE-DOWN ROD L— UPPER GUARD RING UPPER HEAT METER i - o S A A 7 A NS Y NP 7 7 7] X7z 7 7 A [ N N 7 P THERMOCOUPLE LOCATION PLANES SAMPLE CONTAINER |- SAMPLE GUARD RING LOWER HEAT METER / / LOWER GUARD RING ISOTHERMAL COPPER " COOLING PAD /SINK HEATER PLATE SINK COOLER PLATE NN N N NN NN NN = 7 Z Fig. 7.5. Modified Axial-Heat-Flow Thermal-Conductivity Apparatus. T4 UNCLASSIFIED ORNL-LR-DWG 58624 1100 \\\ @ ® 1050 \\ ® ® PRIMARY THERMOCOUPLES A SECONDARY THERMOCOUPLES 1000 ~ 950 - e \/A’_‘tNTERFACE_i'g F L " \ D < o 5 900 A a = \\i wl - \ \.\ 850 \/A’mTERFACE =0.4°F UPPER HEAT METER LITHIUM SAMPLE LOWER HEAT METER . (TYPE 347 —— -— (TYPE 347 —=f STAINLESS STEEL) \ STAINLESS'STEEL) 800 N N 750 \ Q\\ 0 {2 3 4 5 6 7 8 9 AXIAL POSITION (in.) Fig. 7.6. Typical Axial Temperature Profile, Lithium Thermal- Conductivity Study. 75 certainty. A typical axial temperature profile, shown in Fig. 7.6, indi- cates that the Interfacial resistances between the sample and the heat meters were quite small. This agrees well with other studies with alkaline liquid metals.”*® | The results of the recent measurements with lithium are given in Fig. 7.4 in comparison with earlier data and with the results of other investigators. It may be seen that the new data agree to within an average deviation of +1.4% with values reported by Kutateladze et al.? obtained by a method-of successive stationary states. (The results of Kutateladze et al. for potassium, sodium, and mercury compare closely with the data 8 for the same metals.) In contrast with the reported by Ewing and Miller data for sodium and potassium, and earlier results for lithium (Webber et al.?), the more recent data on the lithium conductivity show a positive dependence on temperature, that is, the conductivity increases with in- creasing temperature. For the present data, this variation is not as pronounced as with the earlier ORNL measurements or with the values calcu- lated at CANEL'? from electrical-conductivity measurements. While final énaiysis of these results with respect to establishing the experimental error is incomplete, it is expected that further corrections will have only a second-order effect. R. R. Miller and C. T. Ewing, Naval Research Leboratories, Washington, D. C., private communication. 6¢. T. Ewing, J. A. Grand, and R. R. Miller, The Thermal Conductivity of Sodium and Potassium, Naval Research Laboratory Report 3835 (August 1951 ). 7S. S. Kutateladze et al., Liquid-Metal Heat Transfer, pp. 2-3, Atomic Press, Moscow, 1958, translated by Consultants Bureau, New York, 1959. 8C. T. Ewing and R. R. Miller, Thermal Conductivities of Mercury and Sodium-Potassium Alloys, J. Phys. Chem. 59, 524 (1955). H. A. Webber et al., Determination of the Thermal Conductivity of Molten Lithium, Trans. Am. Soc. Mech. Engrs. 77, 97 (1955). 10yyclear Propulsion Program, Engineering Progress Report, Jan. 1, 1960 — Mar. 31, 1960, Pratt & Whitney Aircraft, CANEL Operation, Report PWAC-601, p. 92 (Secret) 76 8. RADIATION EFFECTS Radiation Effects on Columbium-Zirconium Alloy N. E. Hinkle, J. C. Zukas, J. W. Woods Two experimental assemblies containing tube~burst specimens of Cb—1% Zr alloy were irradiated in the poolside facility of the ORR. The tube- burst specimens in these assemblies were tested at 1800°F in a helium atmosphere. Because of early failure of the heaters, the specimens in the first experiment were not ruptured; Data on gamma heating rates were obtained, however, and were used to affect successful operation of the - second experiment. The second experimental assembly is shown in'Fig. 8.1 before the final enclosure was installed. This assembly contains eight specimens and two gettering trains. The specimens and associated pafts are shown in Fig. 8.2, before and after joining. Brazing and welding of tube-burst test assemblies are performed in the vacuum chamber of an electron-beam welding machine in order to eliminate contamination. The windings of the three-section furnaces shown in Fig. &.1 are of Nichrome V wire sheathed in stainless steel with MgO insulation. The insulated furnaces are wrapped with tantalum foil to obtain additional mass (for gamma heating) and thermal radiation shielding. | | . , In order to obtain reliable stress-rupture data for Cb—1% Zr alloy, it is very important that the tests be performed in = helium.atnosphere of the highest purity. Therefore a helium purification system was pro- vided. All organic materials were eliminated from the test system by using ceramic insulation on the thermocouple and power leads from the in-pile assembly to the out~of-pile junction box, and the experimental assembly was baked out at about 275°F in a vacuum. The assembly was then purged and evacuated a number of times and finally filled with helium. After insertion in the reactor, the experimenfal assembly was connected to a helium supply pénel by means of quick-disconnect fittings which es- sentially eliminated the possibility of admitting any air to the system. 77 84 UNCLASSIFIED PHOTO B2727 Pig. 8.1. Experimental Assembly for the Second Cb—1% Zr Alloy Stress-Rupture Experiment. A UNCLASSIFIED PHOTO 3542435 Fig. 8.2. Columbium—=Zirconium Alloy Tube-Burst Specimen and Associated Test Capsule Parts. The furnace shown is one of two types used in the first columbium experiment. form, platinum alloy windings, and Al,03 insulation. It consists of a tantalum The helium supply panel includes helium purification trains of charcoal at liquid nitrogen temperature and calcium at 600°C. The two end furnaces of the assembly (see Fig. 8.1) contain zirconium foil for scavenging the oxygen and nitrogen that may be released as the various materials outgas during the experiment. The specimens were tested at a rather high stress in the second experiment, and all ruptured within 60 hr. The data for these specimens and the comparative out-of-pile data from tests in vacuum are shown in Table 8.1 and in the graph of Fig. 8.3. The in-pile rupture strength was about 10% less than the out-of-pile strength. | It is planned to test two of the specimens in the next assembly at 1800°F to obtain longer times to rupture and to obtain data on the Table 8.1. Results of Stress-Rupture Tests of Cb—1% Zr Alloy Tubing at 1800°F . Irradiation at - Time to Sp;gj..men %;IS??S ' Rupture Rupture (Mwhr) | (hr) (a) 31 000 23 19-6 28 000 © 100 1.5 19-1 28 000 1900 3.8P (2) 27 500 149¢ 19-5 26 000 150 3.5 19-3 26 000 900 28 () 25 000 - 100 19-7 24 0002 1100 35 19-2 24 000 1700 55 18-7 23 000 5500 >160€ 19-8 22 500 870 27 19-4 22 500 1400 45 (2) 22 500 6950 a'Ou.‘t-of-pile test data. bSpecim.en stressed at about 1750°F for 57 hr before increasing the temperature to 1800°F. | cSpecim.en leaked and was repressurized intermittently during the test. dThe calculated minimum stress during the test was 19 000 psi. eSpecim.en did not rupture during the test. 80 UNCLASSIFIED ORNL-LR-DWG 55994R 50,000 40,000 _ — 2 TH 7 30,000 = TP = e Al L ~ [ H n :_ m e L Lol -.l—"--.._ '_:E fi = o IN-PILE, EXP. 1 [il° I \\"‘""“t 1 & 20,000 e IN-PILE, EXP 2 | 5 A 57 hr AT 1750°F & OUT-OF-PILE L 0 ) 10 100 1000 TIME TO RUPTURE (hr) Fig. 8.3. Stress Versus Time to Rupture for Cb—1% Zr Alloy (Pratt & Whitney Material) at 1800°F in Helium. The horizontal arrow denotes "not ruptured.'" The vertical arrow denotes a leaky specimen. remaining specimens at about 2000°F. This experiment will utilize the last of the material provided by Pratt & Whitney. A second batch of Cb—1% Z2r alloy, purchased by ORNL, 1s available for future experiments. Thé first experimental assembly was examined to determine the cause of the furnace failures and to observe the condition of the speéimens with respect to impurity contamination. It was found that excessive thermal expansion had bent the main structural element and obviously damaged the furnaces and the furnace leads. Possible, slight, surface contamination was observed and will be investigated in more detail. The initial postirradiation examination of the specimens used in the second experiment has shown that six of the specimens exhibited fractures similar to out-of-pile fractures. These six specimens were split open with a longitudinal crack. The specimen surfaces appear quite bright, indicating very little contamination. It is interesting to note that neither the thermocouples nor the furnaces were damaged during the blow- out type of rupture, although the furnaces were bent out of their normal shape. Radiation Effects on Stainless Steel and Inconel N. E. Hinkle, J., C. Zukas, J. W. Woods An experimental assembly was irradiated that contained ten tube- burst specimens of type 304 stainless steel. The steel specimens were 81 tested at 1500 and 1600°F in air. The results of these tests and compara- tive out-of-pile test data are given in Table 8.2. Results of earlier tests indicated that the time to rupture of this material was reduced Dby a factor of 2 by irradiation at 1500°F, and these data continue to show the same effect. Sufficient out-of-pile test data at 1600°F are not yet available for comparison with the in-pile test data. In a previous report,1 it was suggested that the presence of boron (specifically Blo) might be responsible for the reduced rupture strength lANP Semiann. Prog. Rep. March 31, 1959, ORNL-2711, p. 60. Table 8.2. Results of Stress~Rupture Tests of Type 304 Stainless Steel Tubing at 1500 and 1600°F in Air Test Specimen Stpess Irradiation Dose Time to Temperature T No (psi) at Rupture Rupture (°F) ' P (Mwhr) (hr) 1500 310 6300 5172 318 6300 %87% 537 6300 | 183 319 5800 6912 239 5250 10252 73 5250 1134 536 4700 857 240 4200 24,502 72 4200 1670 17-3 4000 36 900 g51p 17-10 3500 50 080 >1600¢ 17-1 3000 : 50 080 >1600¢ 17-9 3000 50 080 >1600¢ 1600 17-8 5000 3 750 49 360 3700 - 265 17-5 3000 13 500 414 17-6 2600 27 100 870 17-4 2200 43 900 1430 17-17 2200 47 400 >1505¢ “Specimens 6 in. long; remaining specimens 2 1/2 in. long. Specimen received 11 500 Mwhr of irradiation prior to stressing. cSpecim.en did not rupture. 82 of Inconel because of the formation of helium and lithium at the grain boundaries during neutron bombardment. In order to invéstigate this idea, the Laboratory purchased six heats of Inconel containing various amounts of boron (natural and isotopic).? Testing of these materials has begun with irradiation of two specimens of each of five of the heats. These tests are being conducted at 1500°F in an air atmosphere. The experimental‘ assenmblies were constructed the same as previous assemblies for testing standard heats of Inconel. The test.procedures were described previously.3 The experimental assembly presently being irradiated is shown in Fig. 8.4. The specimen test-parameters, the data obtained thus far, and the avail- able comparatifie out-of-pile test data are listed in Table 8.3, For 2ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 75. 3ANP Semiann. Prog. Rep. March 31, 1958, ORNL-2517, p. 58. Table 8.3. Rupture Characteristics of Inconel Tubing Containing Controlled Amounts of Boron in Tests at 1500°F in Air Heat Boron Content Specimen Stress Irradiation at Time to No (ppm) No (psi) Rupture Rupture ’ ; ’ (Mwhr) (hr) 0B <10 (a) 6000 155 (a) 5000 320 28-6 5000 4 250 140 (a) 4000 1250 28-2 2000 14 250 475 i (a) 3000 1950 108 100 28=-3 6000 .1 620 52 (a) 5000 580 28-8 5000 1 800 58 (a) 4000 1750 2B 20 (a) 5000 115 (a) 4000 880 (a) 4000 4200 4B11P 40 (a) 5000 105 28-9 4500 5 650 186 (a) 4000 250 28-5 3200 18 600 620 {(a) 3000 980 6B11P 0 28-4 5500 1 750 56 (a) 5000 610 28-10 4500 4 050 133 (a) 3800 1500 6B10° 60 (a) 5000 195 28-1 4500 2 450 80 (a) 4000 255 28-7 3500 6 930 229 aOu_t-of-pile test specimen. PBoron enriched to 98.5% B, ®Boron enriched to 95% B1O, 83 78 UNCLASSIFIED PHOTO 53278 ' 2 Oax WIDGE NATIONAL LABORATORY Fig. 8.4. Experimental Assembly for a Typical In-Pile Stainless Steel or Inconel Tube-Burst Ex- periment. The "Y" shaped tubing at the bottom left is part of the air-cooling manifold for the speci- men cooling system. comparison, Inconel. from the new data. Figs. 8.5 and 8.6 summarize the data on standard heats of No conclusions regarding the effect of boron can yet be drawn - A second experimental assembly is under construction, and it is hoped that with this additional data some conclusions can be drawn. 8000 — T T T T T T UNCLASSIFIED ORNL-LR-DWG 55999R — * L IRRADIATION TIME AND {) FRACTION- 6000 o ! oF g'° éfiRNUP : . . e " . : o-hfi‘ (ph——8—® .04 _. 4000 100 (7057~ o~ ¢ . 3 550 (30) T :::::::::: < 1000 (55 S T § o IN-PILE TEST | @ s IRRADIATED AT TEST TEMPERATURE \ fy 2000 PRIOR TO STRESSING » STRESSED FOR 48hr PRIOR TO NEUTRON | BOMBARDMENT . . s IRRADIATED AT TEST TEMPERATURE, POSTIRRADIATION TESTED e ° ¢ OUT-OF-PILE TEST | | |4J | |‘ [ 1 1000 P _ 30 50 100 200 500 1000 2000 3000 - TIME TO RUPTURE (hr) Fig. 8.5. Stress Versus Time to Rupture for Inconel (INCO Heat No. NX 8962) at 1500°F in Air. STRESS {psi) Fig. fication, UNCLASSIFIED ORNL-LR-DWG 56000R 8000 6000 4000 o IN-PILE TEST ® QUT-OF-PILE TEST 2000 1000 . - 10 20 50 100 - 200 © 500 1000 TIME TO RUPTURE (hr) 8.6. Stress Versus Time to Rupture for Inconel (CX-900 Spec1- INCO Heat No, NX 5757) at 1500°F in Air. ‘ 85 Beryllium Oxide Irradiation Studies R. P. Shields, J. E. Lee, Jr. Serious physical damage to BeO which had received fast-neutron dosages of the order of 102! neutrons/cm? in the ETR was revealed during post- irradiation examination of test specimens. The damaged specimens constitued a portion of the 57 BeO pellets, 1 in. in length and 0.4 to 0.8 in. in diameter, which were contained in 16 capsules distributed among five separate irradiation assemblies of the ORNL-41 test series. A compre- hensive rangé of exposures was achieved at a significant variety of tem- peratures. The BeO has been recovered from the capsules 6f all five assemblies. The disassembly and inspection df the last three assemblies was carried out in the BMI hot cell facility. Visual inspection, macrophotography, and physical dimension measurements of the specimens have been completed. Metallographic examination is complete for the specimens used in the early experiments and exploratory gas analyses and x-ray diffraction studies have been performed. A major portion of the thermal flux dosimetry analysis is completed, and corresponding fast flux determinations are under way. A survey of the temperature‘data'and thermal barrier gas composition changes is being made to evaluate thermal conductivity changes observed during the progress of the irradiations. Damage observed in the BeO that received the higher dosages varied from minute cracks to gross fracture and disintegration into powder. The first cracks were found to appear after an estimated irradiation dose of 5 to 8 X lOzo'neutrons/cm? (>1 Mev) in the temperature range‘of 700 to 1000°C, as shown by Fig. 8.7, which is a photograph of sample 3-52. The BeO samples disintegrated to powder or to easily crushed material when irradiated at 120°C to a dose of approximately 1 X 1021 neutrons/cm2 (>1 Mev). A sample of the disintegrated material found in capsule 5-7'ié shown in Fig. 8.8. The capsule cladding increased in diameter up to 3%, and, as shown, there was fracture or disintegration of the BeO. One cap- sule burst its cladding after an irradiation dose of 10! neutrons/cm? 86 UNCLASSIFIED RMG {7281 RMG {7282 RMG 17283 Fig. 8.7. Cracked BeO Pellet 3-52 That Was Ir- radiated to a Fast-Neutronh Dose of ~8 x 1029 Neutrons/cm? in the Temperature Range 700 to 1000°C. 5% UNCLASSIFIED PHOTO 54274 Fig. 8.8. Disintegrated and Frac- tured Material from BeQO Irradiation Cap- sule 5-7 After a Fast-Neutron Dose of 1 X 10?r Neutrons/cm? at 120°C. 4X (>1 Mev) at a temperature of 120°C. The partial view near one end of capsule 5-7 shown in Fig. 8.9 indicates the type of rupture which occurred in this unit. A summary of the specimen conditions and irradiation tem- peratures and an estimate of neutron flux are given in Fig. 8.10. Examinations and analyses of the specimens are continuing at both ORNL and BMI in order to gain as much quantitative information as possible regarding the specific nature of the observed damage and to determine the contributions of dosage and temperature. New information will be forth- coming after irradiation of a sixth experimental assembly now at the ETR. This unit contains 1.18-in.-diam specimens 3 in. long that are both UNCLASSIFIED PHOTO 654275 Fig. 8.9. Ruptured Cladding of Capsule 5-7. 4X 88 unclad and encapsulated. This experimental assembly is scheduled for removal from the reactor toward the end of the current fiscal year. The integrity of the BeO irradiated in these tests has been demon- strated to be less than had been anticipated. A need for further and ’ intensive study is indicated. UNCLASSIFIED ORNL-LR-DWG 56040A EXPERIMENT NUMBER NO. | 458°C \ (0) \ NO.3 [ (0) (0) SOUND SPECIMENS _ (1) FIRST CRACKS NO. 6 [ (2) GENERAL FRACTURES 950°C (3) POWDER PRESENT (2) @ l NO.7 [ NO.7 [ 120°C |2 140°C | 0 L (3) B a1 1783 2230 1343 10032 O 2.0 40 60 x10'4 EXPOSURE (Mwd) FAST FLUX (>1Mev) 203x10% 88x10% 1.4x10% 5.6x10% 4.95x10% IRRADIATION TIME (sec)” : Fig. 8.10. BeO Irradiation Conditions of ORNL-41 Series Experi- ments. PART 2. SHIELDING RESEARCH 9. DEVELOPMENT OF REACTORS FOR SHIELDING RESEARCH Tower Shielding Reactor II (TSR-II) L. B. Holland Upon completion of the previously described critical experiments,1'2 the TSR-II was disassembled, all the temporary equipment which had been installed for the critical experiments was removed, and the reactor was reassembled for operation. As a result of the shakedown runs of the com- plete system which continued through December 1960, some components had to be modified before the reactor was placed in operation in January and subsequently operated at a 100-kw power level. Minor design changes are still contemplated to provide more reliable operation. These changes are described below along with some results obtained during the shakedown runs. Reactor Mechanical System The changes in the control mechanism housing assembly were described previously,2 and a spare assembly is now being fabricated according to this design. The control mechanisms for the spare assembly are identical with those presently in the reactor; however, closer tolerances were held in the fabrication of the new mechanisms and more uniformity has been ob- served in the release and travel of the control rod under scram conditions on the test stand. The first mechanisms were placed in the reactor in November 1960 and have been operating in a satisfactory manner without attention. In an effort to help maintain a clean system, a 20 000-gal holdup tank has been added to the reactor system for demineralized water storage. Several filters have also been added at crucial points, and it is hoped that these filters, combined with a completely closed water system, will prevent the buildup of crud in the control mechanism lines. LANP Semiann. Prog. Rep. April 30, 1960, ORNL-2942, pp. 112-14. 2ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 82. 93 Two sections of 6-in. hose supplying cooling water to the reactor have failed under full-flow and full-pressure conditions. The failures were slow leaks which would not have Jjeopardized the reactor safety after high-power operation. The vendor has examined the damage and has recom- mended an improved flange design on replacement hose which will be de- livered in May. For the present the reactor is operating at reduced flow with temporary hoses in place. An operational error caused the collapse of two ionization chamber wells. The cQoling water pumps were started with the exit line from the reactor closed. This resulted in overpressurization in the core and failure of the chamber well., BSince the failure caused a pressure change to at- mospheric pressure, the chambers were not damaged. The chanbers have been temporarily insfalled in removable wells in the annular region and are op- erating satisfactorily. A tag-out procedure has been initiated to prevent a recurrence of this type of failure. During a scheduled shutdown in June, the three original chamber wells in the inner region will be replaced with removable heavy-walled wells. As previously described,? the TSR-II is suspended from a support platform so that the reactor can be rotated together with any shield for beam or shield mapping measurements. At first, considerable difficulty was encountered in rotating the reactor and shield. A hydraulic torque booster (which follows the motion of a positioning motor) is used to rotate the reactor. Originally the hydraulic pump which supplies power to the torque booster was remotely located, but it had to be mounted on the reac- .tor platform before it would operate correctly. Electrical cables to the reactor chambers and drive motors were too stiff to permit 360-deg rota- tion of the reactor. These cables have been stripped of the heavy insula- tion, and lighter insulation has been used to permit full 360-deg rotation for complete mapping of the leakage radiation from the Pratt & Whitney- designed shield (see chap. 12). JANP Semiann. Prog. Rep. Sept. 30, 1958, ORNL-2599, p. 186. 9% Reactor Controls For the most part, the difficulties in the reactor control system during the shakedown runs were concerned with minor adjustments or cable connector troubles. The notable exceptibn is the seat switch system which, because of its complexity, could not be kept in proper adjustment. The present system senses a pressure change when a control plate moves a leaf spring away from an orifice through which a dc pump is forcing a small flow of water. The system is further complicated because the pressure switch requires an air reference which equals the core water pressure at all flow rates and reactor elevations. Since the orifice and leaf spring part of the circult appear to op- erate satisfactorily, a flow-type sensing device has been designed and a prototype model tested. In the prototype modél the flow through the sensing device when the orifice is open lifts a éonducting ball from a set of contacts and thereby breaks a low-voltage dec circuit. The only auxiliaries required will be a dc¢ pump to force water through the orifice. This system will be independent of both the water flow rate throfigh the’ reactor and the reactor elevation. Flow Measurements and Fuel Plate Temperatures Measurements were made of the full flow through the reactor and of the flow distribution in the annular fuel eleménts under'fullelow and half-flow conditions. It was determined that the maxXimum flow that could be achieved with all the baffle plates in place in the reactor was 810 gpm with a pressure drop of 31 psi across the core. The reduction in total flow from the design flow of 1000 gpm required new studies of both the inner and annular regions. Since the optimum flow distribution in both the upper and lower fuel elements in the inner region under all flow conditions is nearly uniform, the reduction in total flow has no appreciable effect on the baffle design. Efforts to reduce the pressure drop through the upper fuel baffle plate met with no success. A long series of experimental measurements have provided a baffle design which provides uniform flow distribution. The baffle plate is being fabricated and will be tested during the month of May. 95 The water velocity outside the U237 .1o0aded spherical cover plates on the control mechanism housing varies with position; therefore, a calcula- tion has been made to determine whether the heat transfer coefficient at the point of lowest water velocity is sufficient to remove the heat gener- ated when the heat is removed from the external surface only. These calcu- lations were based on a uniform flow distribution in the upper and lower fuel elements for the measured total flow through the reactor and the cal- culated heat generation in the fuel cover plates after several hours at 5 Mw. For a heat generation rate of 101 000 Btu/hr.ft?, the surface tempera- ture of the cover plates at the point of lowest water flow was calculated : to be 217.5°F, which is well below the saturation temperature of 259°F for . a. core pressure of 20 psig. _ For the annular region it was necessary to measure the flow distribu- tion through the fuel elements and compare the measurements with the calcu- lated value of the flow required in each channel to keep the water below the saturation temperature for the minimum operating éore pressure of 20 psig.’ The baffle plate design for the annular elements was determined by flow measurements on a single plastic mockup of an annular fuel element which had flat rather than curved plates. The final determination of the flow distribution through the fuel plates of the annular fuel elements was made with all the fuel elements in place in the reactor. For full flow and half flow conditions, a salt solution was injécted into the system ahead of two conductivity probes which were separated by a known distance. Measurements of the time for the solution to pass between the probes gave a value ofrthe flow in each channel, Measurements were made in every other fuel element and in enough channels in these elements to determine the flow distribution as a function of channel number. The results corroborate the single element measurement and show that the flow is sufficient to keep the plate surface temperature in all channels below 225°F. The core pres- sure should always exceed 20 psi so that local boiling should not occur below 259°F. Data of Bernath® show that film boiling will not start unless “Chem. Eng. Prog. Symposium, Series No. 30, Vol. 56, pp. 95-116, 1960. | - 26 this value is exceeded by 60°F, which provides a safety margin of 94°F. There is, however, one channel in the annular elements where the velocity dropped below the value reguired, but the flow velocity in the adjacent channels is much higher than required, so the over-all effect is considered acceptable. The rather severe winter weather provided a good check of the freeze- up protection of the water system. For the most part the automatic de- vices protected the system adequately. Some wire heatefs and insulation were added where either the water velocity stays low or stagnates because of instrumentation. Nuclear Measurements The reactor was reassembled with the spherical cover plates on the control mechanism housing. The cover plates are loaded with 160 g of U235 (these plates were erroneously referred to as 116-g plates previ- ously5). Using these fuel plates limits the excess loading of the reac- tor to 0.97% £k/k when the reactor is water reflected and at 20°C. (The present operating limit is 1% Ak/k.) Attempts to measure the change in reactivity with loss of water reflection during the critical experiments showed a change of ~0.08% Ak/k.6 This was not a clean experiment, however, because the reactor was supported in the empty lead-water beam shield, and approximately l/2-in. water and some steel were present outside the pressure vessel. The reac- tor can now be operated in air with no water outside the pressure vessel. The excess loading drops 0.39% Ak/k to 0.58% Ak/k when the reactor is re- moved from the pool. As far as excess loading is concerned, the lithium hydride-uranium shield which was designed by Pratt & Whitney (see chap. 12) is the same as a water reflector. The absolute power calibration of the TSR-II has not hbeen made, but it is proposed that the measurement be made by comparing the heat output SANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 86. 6Tbid. o7 of the reactor with that of a 40-kfi sufimersible heater. The heater can be closely coupled to the reactor, and the temperature rise as a'function of time of the small amount of water in the pressure vessel can be com- pared for the reactor and the heater. The eiectrical input to the heater can be measured quite accurately, and corrections can be minimized by op- erating over a small temperature range where heat input due to pumping equals heat loss from the system. Experiments have been run to determine the effect of the regulating rod position on the reactor neutron and gamma-ray leakage. A change in leakage of 12.7% was observed for the full regulating rod travel. This change is due to the location of the ionization chambers with respect to the regulating rod. As a result of this variatioh it is necessary to keep the regulating rod position fixed. Thus far this has been accomplished by holding the system temperature constant. It is planned to examine the effect on leakage of the shim rod posi- tion. This will be accomplished by operating the reactor over a tempera- ture range with the regulating rod kept fixed. Measurements have been made of the linearity of the reactor power- 1evel-cdntrolling instruments. For these runs, three BF3; counters were used in addition to the reactor instruments. Preliminary runs show varia- tions of 10% in the linearity. OStudies are under way to improve the power level instrument so that such large corrections will not be regquired for all measurements. 98 10. DEVELOPMENT OF RADIATION DETECTION EQUIPMENT Gamma-Ray Spectroscopy The Model IV Gamma-Ray Spectrometer (G. T. Chapman) Considerable effort and time has been spent in investigating the ef~ fectiveness of the Model IV gamma-ray spectrometer shield and collimator.t A study of the backgrounds found with the spectrometer in a reactor radia- tion field has indicated that it is necessary to operate the spectrometer at distances no closer than 5 ft from the reactor to reduce the gamma-ray background within the shield. (Background is used here to mean spec- - trometer response with the lead collimator in the housing closed by a rotatable lead filler.) Consequently, a 5-ft-long air-filled cone has been attached to the front of the spectrometer shield to allow measure- ments of radiation leaving the surface of the reactor. It has also been shown that.the use of such a cone considerably reduces the number of de- graded gamma rays due to scattering in the lead collimator of the spec- trometer housing. The tests for optimizing the cone-collimator system have not been completed. The use of an air-filled cone permits the utilization of water as .additional shielding between the reactor and the spectrometer, and there- fore essentially all the remaining background measured in the spectrometer is due to radiocactivity in the components of the shield or crystal. Figure 10.1 shows the background measured under various conditions, with the energies of the gamma rays giving rise to the predominant peaks indicated, although the sources of these gamma rays have not been satisfactorily identified. It is evident that the background below about 4 Mev origi- nates from long-lived isotopes within the shield of the spectrometer, but the source of the high-energy gamma ray at about 6.87 Mev is as yet un- determined. The large composite NaIl{T1) crystall obtained for use with the spec- trometer has been tested further. This composite crystal consists of two 1ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 118. 99 UNCLASSIFIED 2-01-058-0-591 BACKGROUND ~ 16 hr AFTER LAST 10° COUNTS PER CHANNEL PER SECOND EXPOSURE TO REACTOR SPECTROMETER UNDER WATER, COLLIMATCGR CLOSED BACKGROUND DURING REACTCR OPERATION SPECTROMETER UNDER COLLIMATOR CLOSED & BACKGROUND ~ 4 hr AFTER LAST EXPOSURE TO REACTOR SPECTROMETER N AIR, COLLIMATOR CLOSED BACKGROUND ~ 4hr AFTER LAST EXPOSURE TO REACTOR 1072 SPECTROMETER IN AR, COLLIMATOR OPEN REACTOR POWER: 2 watts 3073 0 20 40 60 BO 100 CHANNEL NUMBER (RELATIVE PULSE HEIGHT) Fig. 10.1. ated with the Model IV Gamma-Ray Spectrometer. 100 Background Associ- crystals, one 9 in. in diameter and 5 in. long and the other 9 in. in diameter and 7 in. long, optically glued together. Based on previous measurements and calculations,? a l-in.-diam, 2-in.-deep well was drilled into the.crystal to reduce the "tail" of the pulse-height dis- tributions. Unlike the previous crystal with a conical end, this crystal has shown a uniform response to gamma-ray energies as great as - 6.1 Mev. response of the crystal to colli- Figure 10.2 shows the mated gamma rays with energies rang- ing from 0.511 to 2.754 Mev. As anticipated, the well has simpli- fied the Compton distribution in the "tail" of the higher energy dis- tributions.? The crystal has been in almost constant operation for approximately six months without noticeable deterioration of the glued interface. Unscrambling of Continuous Scintil- lation Spectra (W. R. Burrus) The problem of "unserambling" gamma-ray spectrometer data was in- troduced previously.2’% Difficulty in unscrambling results from the fact that while the actual spectra of interest are a superposition of 2Ibid., pp. 102-8. a continuous part and a nearly monoenergetic part, spectrometers always yield a finite set of numbers as a result of a measurement. In the simple *W. R. Burrus, Unscrambling Scintillation Spectrometer Data, IRE Trans. Nuclear Seci., NS-7(2-3), pp. 102-11, 1960. “ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 121. UNCLASSIFIED 2-01-058-0-590 100 A\ 0.511 Mev T 90 0.662 Mev 80 o Na’* DECAY & Na?? DECAY ® Cs'>7 pECAY o ) 3] Y o - L = 3 T 50 o a LJ ¢ | 2 4 ) E 40 > O o 30 7 1.277 Mev 1.368 Mev . \L £ 2.754 Mev 10 £ aj 'éf X&\kn Oggeoap 0 ¥ AOoYcRpnam 0o o 0 10 20 30 40 50 60 70 80 90 CHANNEL NUMBER (RELATIVE PULSE HEIGHT) Fig. 10.2. Response of the Large Composite NaI(T1l) Crystal to Col- limated Gamma Rays. 101 threshold foil "spectrometer," the set of numbers is just the counting rate of the foils after a certain amount of waiting time. For the NaI(T1l) scintillation spectrometer, the numbers are usually the number of counts in the channels of a multichannel analyzer. In order to understand the results of a spectrometer experiment, it is necessary to obtain the actual spectra from the measured set of values; that is, the measured set of values must be unscrambled. The solution to the unscrambling problem has proved unusually difficult and haé been a great stumbling block to otherwise straightforward experiments. The solu- tion is complicated by the fact that the number of counts measured actually has a statistical distribution, and the average cannot be determined “exactly by any experiment. In many of the previous attempts at unscramb- ling, it was found that the results were extremely sensitive to very small statistical fluctuations in the input data. The problem is further com- plicated by the difficulty of measuring the response of the spectrometer. | The statistical problem has been well recognized, however, even if not completely'resolved, but another basic problem has recently become apparent. The second problem is introduced by the finite nature of the spectrometer output. Since the desired radiation spectrum is continuous, it cannot be determined exactly by a finite number of points, thus with actual instruments which yield c,;, c2, ... c with n seldom larger than a few hundred, an uncertainty is present. This effect has usually been completely ighored, yet it may actually be of the same order of importance as the statistical uncertainty. Thus in this report, it is assumed that the exact average number of counts ¢y, ¢, ... ¢ is known and that the response of the spectrometer is known exactly. An effort is then made to see how closely the continuous spectrum g(e) can be determined. It is hoped that this treatment of the deterministic part of the unscrambling | problem will lay the ground work for a realistic and rigorous treatment of the complete probiem. Mathematical Formulation. The response of a scintillation spectrometer is related to the unknown spectrum by the equation 102 c, = J; K (e)gle)ae , J=1,2, .om (1) where c, = average number of counts obtained in the jth channel of the J spectrometer, ~K.(e) = response function of the spectrometer, that is, the average J fraction of incident particles of incident energy e which produce a count in channel j, g(e) = continuous spectrum, that is, g(e) de is the average number of particles between e and e + de. The gualitative appearance of the Kj(e) functions for a 3- by 3-in. single-crystal NaI(Tl) gamma-ray scintillation spectrometer with 50 chan- nels is shown in Fig. 10.3. Bounds of a Single Parameter. If the'quantities C1; Ca, «v. C from a scintillation spectrometer are to be used to determine a certain parameter which depends upon the spectrum, for example, the tissue dose, the dose can be expressed in terms of g(e) by the equation o= J o) ate) ae (2) UNCLASSIFIED 2-01-058-0-584 A [\ === —— / 7 — 0 0.2 0.4 0.6 0.8 1.0 ENERGY Fig. 10.3. The Response Functions K;(e) for a Few Illustrative Chan- nels of a 50-Channel NaI(Tl) Gamma-Ray Scintillation Spectrometer. 103 where p = tissue dose (in suitable units), w(e) dose absorption cross section, that is, the dose due to a single particle of energy e. For gamma radiation, w(e) has the well-known form consisting of contri- butions from the photoelectric effect, Compton effect, and pair produc- tion. The cross section is relatively smooth, with the exception of a few sharp "edges" in the lower kev energy range corresponding to photoelectric shell effects. | Now if it is desired to be cautious and to insist on using no informa- tion about g(e) other than that obtained from the scintillation spectrome- ter, the dose D can be obtained from ¢y, c3, ... c if the dose absorption cross section w(e) can be expressed as a linear combination of the response functions: n wie) = géi U Kj(e) . (3) Then J;°° i u, Kj(e) g(e) de p= J wie) ale) ae n - 2w LR Ge) ale) . - (4) 1 .t < o However, if the cross section is not an exact combination of n response functions, it is not possible to find the dose D at all. To understand this, consider that a linear combination of response functions can e found which nearly (but not quite) matches the dose absorption cross section. For example, it may be impossible to match w(e) perfectly in the neighbor- hood of an "edge." Thus the combination of response functions will be a 104 little high in certain places and a little low in others. It is then pos- sible to have a component in the gamma-ray spectrum which gives no counts (in any channel) but which does contribute to the dose. In other fields such components have sometimes been referred to as "invisible components."5 Mathematically, if any discrepancy exists, then a possible component of the spectrum may be found which is invisible [is orthogonal to all the Kj(e)] but which contributes to the dose [is not orthogonal to w(e)]. Thus, to continue the cautious point of view of not making any as- sumptions about g(e), it is usually not possible to determine a simple quantity such as dose. This, of course, is a very pessimistic viewpoint, which presupposes that several invisible components are present in the spectra which are of such large magnitude that estimates of dose are com~ pletely ruined. Of course, this probably is not true, and the artificial assumption that large invisible components are not present could be made. But a much less artificial assumption can be made which will give a reasonable estimate of dose. It is that ge) 20 , 0 (5) IN M IN 8 This "nonnegativity assumption" is physically justified in particle spectra, since a negative number of particles has no physical significance and its introduction into the formulation easily removes the difficulty of in- visiblé components. Although w(e) cannot be matched exactly with a combination of re- sponée functions, it is perhaps possible to find one combination which is always lower than w(e) and another combination which is always larger, as shown in Fig. 10.4. Then, it is possible to write o <1;‘r ut K. .(e) gle) ae (7) o & 3T or n noo L C,. ¥ b ENERGY ——tam ENERGY ——w ® ® & £ + + = > ¥ 5 M ENERGY —— ENERGY —w Fig. 10.7. Combinations of Kj(e) and K, (e) Which Bracket Wi(e), Wa(e), Wi(e) + Wa(e), and 3Wi(e) + Wo(e) from Above and Below. UNCLASSIFIED 2-01-058-0-589 BEST INDEPENDENT UPPER o RECTANGULAR HULL o 2 4 6 8 10 12 14 P Fig. 10.8. Allowed Region of Parameter Space for p; and ps. 8Spl+p2-<.l4 l6$3pl+p2-<_l8 These four inegualities determine the allowed values of p; and-pg. The allowed region is indicated in Fig. 10.8 by the shaded area. Solution of the Deterministic Problem. Even though the input data for the deterministic problem were exact, it may be seen that the results for p1, P2, ... P cannot be specified exactly. The most precise solution is a com- plete description of the allowed region (as in Fig. 10.8), but, in more complex multidimensional cases, a complete description is too cumbersome for practical ap- plications., One possibility would be just to give for a result the coordinates of a point near the center of the allowed region and to give the extreme values of each coordinate. This technique, how- ever, would-yield far less infor- mation than would be desired. For example, it may be seen from Fig. 10.8 that the combination of 3p1 + pp can be determined with much greater accuracy than either p; or ps separately, owing to the shape of the region. Thus an adequate description must con- tain some information about the shape of the allowed region. 111 Two possible descriptions are the following: 1. GCive the center and the principal axes (length and direction) of the smallest rectangular region which completely encloses the allowed region. _ 2. Give the centroid and the moments (including cross moments) of in- ertia of the allowed region, considering that it is a solid body with unit density. In terms of these descriptions, formulas can be developed which give the upper and lower bounds on any combination of the pi's by considering the geometrical properties of the description. | Although the determination of the smallest allowed region or of the smallest rectangular "hull" which enclosed the allowed region is a straight- forward mathematical problem, no satisfactory computational method for solving it has yet been found. Approximatermethods exist which give the approximately correct principal axes as the solution to an eigenvector problem. The details of these methods will be reported later. Conclusions. It has been shown that definite upper and lower bounds can be put upon parameters of a continuous spectrum of the type pi=.':wi(e) gle) de , i=1,2, ... m , by the use of a nonnegativity condition. The size and shape of the allowed region for the pi's depends upon how well the'Wi(e) functions can be bracketed between some linear combinations of the Kj(e) functions of the spectrometer. Approximate descriptions of the allowed region are suitable for practical applications and allow one to put an approximate bound on any linear combination of the parameters. The methods which have been dis- cussed could serve as a point of departure for the analysis of the statisti- cal part of the problem and will be iImportant in the development of improved methods for analyzing data obtained from crude spectrometers (such as thres- hold detectors) where the uncertainty due to the meager availability of suitable cross sections is more important than statistical errors in the data. 112 Neutron Spectroscopy Use of Silicon Surface-Barrier Counters in Fast-Neutron Detection and Spectroscopy (T. A. Love, R. B. Murray, H. A. Todd, J. J. Manning) The possible use of a neutron-sensitive semiconductor detector for neutron spectroscopy was discussed previously.7'8 Basically, the spec~- trometer consists of a thin layer of Li°F (~150 g/cmz) between two silicon- gold surface-barrier counters, each of which is seated in a fluorothene sheet. Neutrons are detected by observing the & + T pair resulting from the Li®(n,a)T reaction. The details of the counter construction were given previously,8 But some modifications have since been introduced. A schemétic diagram of the current design is shown in Fig. 10.9. The separation between the two silicon counters is now effected by a 0.002-in.-thick ring of Teflon, and "ANP Semiann. Prog. Rep. April 30, 1960, ORNL-2942, p. 117. 8ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 128. UNCLASSIFIED ORNL-LR-DWG 57789 SILICON: Bmm x {4 mm x 0,381 mm GOLD: 50 ug/ém® Li®F: 150 pug/cmt 25 ALUMINUM CAN TOP DIODE ASSEMBLY 245 ALUMINUM MOUNT 0002 in. THICK —a== TEFLON SPACER FLUOROTHENE DISK /7 TEFLON TERMINAL AQUADAG 0004 -in~dia ' GOLD WIRE /4/ . //'q SILVER v 5 PAINT S GOLD: 50 ug/cm o 1 APPROXIMATELY { INCH BOTTOM DIODE ASSEMBLY Fig. 10.9. Schematic Diagram of Sandwich-Type Silicon-Gold Surface- Barrier Counter. : 113 a small wire is used to make the electrical contact to the gold surface. The wire is cemented to the gold with a tiny drop of Aquadag and then grounded. A reverse bias of 100 v is applied through a resistor of ~500 000 ohms, and the signal is fed through a coupling capacitor to a low-noise high-gain preamplifier. A block diagram of the electronic system for the sandwich counter is shown in Fig. 10.10. The signal from each detector is fed into a pre- amplifier, which, in turn, feeds a DD-2 double delay line amplifier with a O.7-psec clipping time. The output of the amplifier is fed into a dis- criminator circuit biased to reject all pulses which represent less than 1l.5-Mev energy deposited in the diode. UNCLASSIFIED ORNL -LR-DWG 57795 DD- * )l PREAMP » e DISCRIMINATOR AMPLIFIER 500K iK _ HANN . /LA F 1100 v MIXER DD-2 SIGNAL | MULTIC EL 5 AU CIRCUIT AMPLIFIER ANALYZER AN —=Si COINCIDENCE { |{GATE . S00K 1K CIRCUIT . pl; PREAMP * 0b-2 DISCRIMINATOR 1:| AMPLIFIER Fig. 10.10. Block Diagram of Electronic System for Sandwich-Type Silicon-Gold Surface-Barrier Counter. The discriminator setting of 1.5 Mev was chosen on the basis of the response of a single detector exposed to thermal neutrons with no dis- criminator setting. For incident thermal neutrons, the triton and alpha peaks are clearly defined, as shown in Fig. 10.1l, and most of the counts below 1.5 Mev are attributable to gamma rays. About 5% of the pulses be- low this energy are from reaction products which have lost several hundred kev of energy. Thus a discriminator setting of 1.5 Mev essentially re- Jjects all the gamma-ray background with little loss of the counts due to the reaction products. A gamma-ray background with the discriminator 114 UNGCLASSIFIED ORNL-LR-DWG 57794 2400 600 ! | ALPHA PEAK ¥ 45 % OF TOTAL ABOVE CHANNEL 25 GAMMA BACKGROUND, 2000 ¥~ NOISE, ETC 500 fl,m ITONS 1600 400 T T = o = o 8 ¢ 2 S 1200 300 S 2 » c c - 800 % 200 ALPHAS 1.5 Mev 1 400 ' 100 & 9 0 900, v %‘e "=;= 0 0 10 20 30 40 50 60 CHANNEL NUMBER Fig. 10.,11. Pulse-Height Response of Single Surface-Barrier Detector Exposed to Thermal Neutrons (No Discrimination). setting of 1.5 Mev has the same effect as noise; that is, the peak from monoenergetic neutrons will be broadened. The loss of counts as a result of the reaction products will be much greater for higher energy neutrons, perhaps as much as 30% for 10- to 15- Mev incident neutrons, owing to the forward peaking of the reaction prod- ucts, Calculations are under way for determining the extent of the loss as a function of energy. The coincidence circuit of Fig. 10.10 requires that the pulses occur within 107 sec of each other, and the output bf this circuit is used to gate the multichannel analyzer. Thus, with this circuit, Lis(n,a)T events in which the alpha particle and triton deposit at least 1.5 Mev in sepa- rate counters are recorded, but those events that deposit less than this amount in either counter are eliminagted. The output of the preamplifiers is also. fed into a mixer circuit. The output of the mixer circuit is amplified by a third DD-2 amplifier 115 whose output is recorded in the multichannel analyzer. The input of each preamplifier also has provisions for accepting a pulse from a multichannel pulse generator. This arrangement allows each amplifier and preamplifier to be checked separately for linearity and the analyzer to be checked for both linearity and zero position. The pulse generator also allows a secondary check for stability. Experimental Data and Discussion. The response of the semiconductor spectfometer t0 monoenergetic neutrons up to energies of 3.5 Mev and also to 14.7-Mev neutrons was reported previously.7’8 Further studies of the response to monoenergetic neutrons in the energy region from 3.5 to 8.2 Mev and to groups of monoenergetic neutrons from the Be®(d,n)Bl? reaction have now been made. Background effects from (n, charged particle) reactions appeared at all neutron energies above the threshold energy (approximately 2 Mev) of Si(n, charged particle) reactions. Pulses from fhese charged particles were recorded only in the neutron pulse-height spectrum for incident neutrons above about 5 Mev, since the reactions are endothermic. 1In order to determine the background, a pair of silicon diodes without Li®F was constructed, and pulse-height spectra for diodes with and without Li®F were compared for monoenergetic neutrons at several energies. Later modi- fications now allow gains to be adjusted more accurately for matching both sets of diodes by use of a pulse generator. A comparison of the foreground and background for 5.73-Mev incident neutrons from the D(d,n)He? reaction is shown in Fig. 10.12. The peak at a pulse height corresponding to neutrons of 2.19-Mev energy is due to carbon contamination of the target and comes from the Clz(d,n)N13 reaction. Figure 10.13 shows a foreground-background comparison for neutrons of 8.13 Mev. | The pulse-height spectrum for groups of neutrons from the Be?(d,n)BLO reaction (deuteron energy of 1.85 Mev) is shown in Fig. 10.1l4a. This is to be compared with the spectrum obtained for the same reaction by the time-of-flight method® (see Fig. 10.14b). 1In Fig. 10.14 the numbers in Private communication from J. H. Neiler et al., ORNL Physics Division. 116 UNCLASSIFIED ORNL-LR-DWG 57792 900 T i ® FOREGROUND | © BACKGROUND GAINS NOT NORMALIZED 800 THERMALS ’ + 4.6 Mev C'z(dn} ni3 i 2.19- Mev NEUTRONS D (d,n)H2 + 4.6 Mev 5.73- Mev NEUTRONS + 4.6 Mev 700 fi H 600 4 ® @ 500 < o £z O @ ¢ c [ ] S 400 ® p *lo 300 3 ¢ .: P J 9 200 A . o 3 I . 100 - / » \. : b L ||'/ ® /';E o‘d ® \*.0” ':O .ou .-.. .. ' O &, l 0 o 90,9 0 20 30 40 50 60 70 80 CHANNEL NUMBER Fig. 10.12. Pulse-Height Response of Sandwich-Type Surface-Barrier Counter Exposed to 5.73-Mev Neutrons. parentheses are the values reported for neutrons from this reaction. As expected, the resolution of the sandwich counters at the lower energies is not good enough to resolve the 0.96-Mev group from the 1.33-Mev group, but at the higher energies it is quite good. The data in Fig. 10.14 represent measured pulse-height spectra, and the energy dependence of the neutron detection efficiency has not been taken into account in either case. The data shown in Fig. 10.15 demonstrate that the device gives a puise height which is linear over the range examined (0 to 14.7 Mev). 117 UNCLASSIFIED 6 ORNL— LR—DWG 57790 GAINS NOT NORMALIZED 4.62-Mev NEUTRONS A C12(d, 7) N‘.3 1 counts /channel # 8.13-Mev NEUTRONS D(d,n)He3 BACKGROUND 0 20 40 60 80 100 ‘ CHANNEL NUMBER Fig. 10.13. 'Pulse-Height Response of Sandwich-Type Ssurface-Barrier Counter Exposed to 8.13-Mev Neutrons. 118 l_l |_l UNCLASSIFIED . ORNL—LR-DWG 57793 103 £, =1.85 Mev Lab. ANGLE =0 deg 4q 10 THERMALS _L S NUMBERS IN PARENTHESES ARE REPORTED ENERGIES UNCLASSIFIED 2—Q1—077—-NS—11—13—-55—1 100 p.h.u. = 75 musec | ( ) {5.39) {0.986) 0.95 (G.Og) 4.00) 103 counts /channel {4.39) 440 (2.55) 2.'|f0 counts /microcoulomb Beg(d,n)s‘o—fd =1.85 Mev LAB ANGLE 20°—5-meter FLIGHT PATH 10 20 30 40 50 &0 70 80 300 400 500 600 700 800 900 1000 100 CHANNEL NUMBER PULSE HEIGHT UNITS Fig. 10.14. (a) Pulse-Height Response of Sandwich-Type Surface-Barrier Counter to Neutrons from the Be?(d,n)B'® Reaction. (b) Pulse-Height Response of a Time-of-Flight Spectrometer to Neutrons from the Be®(d,n)B'0 Reaction. UNCLASSIFIED ORNL-LR-DWG 57791 140 | - T(d’ fl) JES— - i 120 - T, n) — //////, 18.3 100 i /'fh—E!e9 (@, 7) 13.36 Mev o | | | = » 5 80 A B (0 ) 12. 42 Mev E ,//, | | ? ////'\\Bé(mnMQ25Mw | é 80 : .‘/ | /////"“-cm(d,n)7.77 Mev 40 ] . ' /*(j / THERMAL 4.6 Mev 20 — 0 | 0 2 a4 6 8 10 12 14 6 18 20 NEUTRON ENERGY (Mev)+ 4.6 Mev Fig. 10.15. Pulse-Height Response of Sandwich-Type Surface-Barrier Counter as a Function of Neutron Energy Plus 4.6 Mev. Conclusion. From the data of the figures represented here and those published previously, it is concluded that the full width at half maximum is relatively constant, of the order of 300 kev, regardless of incident neutron energy. The presence of gamma rays will produce a broadening of the neutron peak, but measurements have been made in gamma-ray fields up to approximately 200 r/hr without appreciably altering the response. Further, the device produces & pulse height that is a linear function of incident neutron enefgy. The background data for monoenergetic neutrons with energies of 4, 5.73, 7.40, 8.13, 13.7, and 14.7 Mev have served as a basis for an esti- mate of the background to be expected when the counters are used to measure a fission spectrum. Calculations predict that the foreground-to- background ratio will be about 5 to 1. A preliminary experiment was per- formed with neutrons from the Tower Shielding Reactor IT to get an esti- mate of the foreground and background, and the results indicated that the ‘background was approximately 20% of the foreground, which is in agreement with calculation. A measurement of the reactor spectrum will be made in the near future. 120 Further Developments. In an attempt to improve the counting effi- ciency of these counters, the effects of several variations in the counter design are being investigated. Specifically, a sandwich counter with an area approximately four times the area of the counters used in the above experiments is now under construction. In addition, the feasibility of substituting 1i® metal for Li®F as the neutron-sensitive material is being studied. Favorable results in these experiments will permit the use of about 20 to 50 times as much Li® as is presently used, which, assuming normal incidence to the plane of the Li® layer, will increase the counting efficiency of the counters from about 107 to about 10-° for neutrons in the Mév region and will result in a greater foreground-to-background ratio for a given spectrum. ©Since the foreground-to-background ratio is deter- mined by the relative number of higher energy neutrons present, this in- crease in Li® will allow a much "harder" spectrum to be examined. A Fast-Pulse Integrator for Dosimetry P. T. Perdue The integrator which was used for dosimetry at the Lid Tank Facility for several years is composed of an eight-channel discriminator whose out- put is fed to the first four binaries of two modified Q-1010 scalers. The 5-psec resolving time of this integrator was considered sufficient when used with the A-1A amplifier, since the resolving time of the amplifier was of this order, but with the advent of the DD-2 amplifier and its known 2.6-usec dead time it became desirable to decrease the resolving time of the integrator to within the same interval so that counting loss corrections could be calculated. Development of such an integrator10 was thus initiated by Lid Tank Facility personnel. The new integrator, which has a total resolution of 2.4 usec, was placed in operation early this year. It con- sists of two sections, each having four discriminator stages coupled to- gether as a true binary séaie of 16. The output pulses from both sections are fed to a single decade scaler. 10p, T, Perdue et al., Fast-Pulse Integrator for Dosimetry, ORNL re- port (to be published). 121 Increasing the speed in the integration process necessitated the development of several additional ideas. For example, to minimize the counting losses in the binary stages under the various "off-on" combina- tions, the slope of the input pulse from the DD-2 is being used to deter- mine the time interval between successive discriminator responses. A special pulse shaper is being used between binary stages to maintain a maximum pulse width of 0.2 usec at the input of the next stage. The width of the output pulse from the integrator sections to the decade scaler has been reduced to 0.3 upsec through the design of a shaper circuit, giving a resolution of approximately 1.0 usec in the scaler. Cathode interface problems, experienced in the Schmidt-trigger discriminators of each of the stages, have been eliminated through the use of a 6922 double triode, where both halves of the tube are in steady-state operation within the grid base. Operation of the instrument with various fast-neutron detectors has indicated that the statistical spread in the count rate from the integrator output is virtually the same as the maximum spread on the total number of pulses fed into the integrator. 122 11. BASIC SHIELDING STUDIES Energy and Angular Distributions of Neutrons Emerging from Planar Surfaces of Diffusing Media V. V. Verbinski The initiation of a program for obtaining basic information on the energy and angular distributions of neutrons in the electron-volt region throughout various shielding materials was reported previously, along with the results of a preliminary experiment for investigating the angular distribution of neutrons emerging from poisoned hydrogenous slabs.l As originally conceived, the over-all purpose of the program was to determine the applicability of several techniques for calculating neutron energies and angular distributions within high-performance materials, specifically LiH. The evaluation of the calculational techniques was to be aided by experimental measurements made with a proposed neutron chopper.2 Insofar as possible, both the calculated and measured data were to be obtained as energy spectra of neutrons at several points inside the shield, energy spectra of neutrons leaking from the surface of the shield, and angular distributions of the leakage neutrons. Recently, however, budgetary support for construction of the neutron chopper facility was withdrawn. In order to obtain other experimental data which could be compared with the calculations, plans were immediately initiated to perform a series of measurements at the Linear Accelerator Facility of General Atomic, San Diego, in early May 1961. -Three calculational methods are being utilized: +the NDA NIOBE (direct numerical integration of the Boltzmann equation) code, the moments method, and the ORNL O5R code. Where possible, the source energy and LiH geometry are similar so that direct comparisons of calculations can be made. The NIOBE code is set up to handle only spherical geometry, the moments method only infinite geometry, and O5R code any geometry. Placing a source at 1ANP Semiann. Prog. Rep. Oct. 31, 1960, ORNL-3029, p. 179. 2Ibid., p. 142. 123 a reasonable distance from the shield will allow comparisons of the spec- tral and angular distributions of leakage neutrons from slabs of LiH cal- culated with the NIOBE and O5R codes and will also allow comparisons of all three methdds for internal spectra at several depths in very thick slabs. Thus far, energy spectra of neutrons from a point source in an infi- nite medium of LiH have been calculated by the moments method for distances from the source of 11.5, 23, and 34 g/cm? and for four source energies | of 16.3, 1.48, 0.12, and 0.01 Mev. These preliminary spectra, normalized at 1 volt in Fig. 11.1, show that low-energy equilibrium is reached at all thicknesses for all input energies employed. This equilibrium spectrum is in agreement with earlier Monte Carlo calculations made at NDA. One re- sult of the Monte Carlo calculation, labeled "Fission,"” is for a fission input spectrum and a source distance of 26 cm. It can be seen that the spectral shape of this curve is practically identical with the curve for 1.48-Mev neutrons at 23 g/cm?. Some NIOBE calculations? of the angular distributions of neutrons leaking from LiH are also available, although they were not made specifi- cally for this program. Here, of course, a spherical geometry was used, but it is not expected that this will significantly affect a comparison with slab-geometry experiments. The angular distributions of leakage cur- rent, J(u), calculated by the NIOBE code for 5.51-, 1.58-, and 0.037-ev neutrons have the form 1 + Ay, where p = cos8@ and 6 is the angle from perpendicular; that is, J(u) < 1 + 3,25u for 5.51-ev neutrons J(u) 1l + 2.64u for 1.58-ev neutrons J(u) o« 1l + 1,02y for 0.037-ev neutrons 3Herbert Goldstein, Nuclear Development Corporation of America, private communication. 124 UNCLASSIFIED 2-01-058—-0-592 SOURCE ENERGY 16.3 Mev 1.48 Mey 104 SOURCE DISTANCE =11.5 g/cm? 0.12 Mev 0.01Mev ____________ FISSION (MONTE CARLO) 103 16.3 Mev - 1.48 Mev N -9- Wy Ll > = < | o 7 FISSION |,/ 102 16.3 Mev 1.48 Mev 10! 349/cm2 0.12 Mev 100 0.0t 0.4 1 10 100 1000 10,000 100,000 NEUTRON ENERGY (ev) Fig. 11.1. Moments Method Calculations of Neutron Spectra Times Energy at Various Distances from Monoenergetic Point Sources in an Infinite Medium of LiH. Spectrum at each depth normalized to 1.48-Mev curve at 1 ev. "Fission" curve from earlier Monte Carlo calculations. 125 The NIOBE calculations are to be compared with Bulk-Shielding Fa- cility measurements of the angular distributions of neutrons leaking from a 4-in.-thick LiH slab. The experimental arrangement was the same as that described previously,1 where neutrons from the Bulk Shielding Reac- tor were incident on the shielding sample after traveling through an air- filled collimator. On the opposite side of the shield sample was posi- tioned an air chamber from which air-filled tubes extended at various angles from the normal up to 66 deg. Detector foils were placed at the far end of each tube. These measurements gave the‘following results: I+ J(u) <1 + (2.00 * 0.25)u for 4.95-ev neutrons J(p) <1 + (1.75 + 0,25)u for l.44-ev neutrons + 1.4 J(u) o<1 + (0.9_0.4 Ju for subcadmium neutrons In the subcadmium measurements, instabilities were encountered in the counting circuitry, and the measurements are therefore being repeated. The linear dependence of J(p) on p is to be expected when the flux varies linearly near the surface of the LiH, and the flattening of J(u) with lower energies is due to a decrease in extrapolation length with decrease in energy. In contrast, the angular distribution in a non- moderating shield with high l/v absorption peaks strongly in the forward direction as the neutron energy decreases. Therefore, the slowing down in a hydrogenous material is seen to be of major importance in deter- mining the angular distribution of leakage neutrons. As mentioned above, measurements of leakage spectra and of spectra within LiH slabs were to have been conducted with a neutron.chopper fa- cility planned for the BSF. 1In conjunctiqn with this effort, a program was developed for the IBM-7090 computer to calcplate the fransmission function and burst shape of the mechanical chopper. In addition, several calculational techniques were developed for taking into account edge leak- age, divergence of the neutron beam, and off-axial effects of the outer 126 slits. Also, numerical techniques were evolved which resulted in large economies in computing time. A report is in preparation outlining these endeavors. The series of measurements to be conducted at General Atomic, San Diego, will begin in early May 1961. A preliminary measurement conducted at GA had adequately demonstrated the feasibility of the series of planned measurements. Although the Linac spectrum is not identical to a reactor spectrum, the NDA calculations in Fig. 1l.1 indicate that the results are not very sensitive to the value of source energy for the high-energy more penetrating neutrons. Experimental Verification of a Geometrical Shielding Transformation L. Jung J. M. Miller F. J. Muckenthaler J. Grant The Lid Tank Shielding Facility is currently being utilized to verify a geometrical transformation concept that the axial dose rate from a large source plate in a hOngeneous medium can be inferred from the dose rates from a small disk source. Expressed mathematically, 2 D(z,a) =D (%,—Ji-) + D z? + E;-, 21+ cee T vn v (1) a? a + D zZ< + y — where a = radius of source plate, 7z = distance from source plate along a line perpendicular to center of source plate, = I§ integer (a/A/f = radius of the small disk source). 127 During this investigation the source plate size will be varied by in- serting cadmium irises of sevéral different'diameters between the source plate and the incident beam of neutrons from the ORNL Graphite Reactor. Cadmium was selected for this purpose because it provides less scattering of thermal neutrons near the edge of the iris, thus giving less error in describing the source area. The use of cadmium gives rise to inaccuracies in the fast-neutron and gamma-ray dose rate measurements in the media- be- cause of cadmium capture gamma rays, however, and verification of the theory must therefore depend only on the thermal-neutron flux measurements. Thus far, only the measurements with a 7-in.-diam iris have been completed. 128 12. APPLIED SHIELDING Preanalysis of Pratt & Whitney Divided-Shield Experiment at TSF S. K. Penny, D. K. Trubey The preanalysis for the Pratt & Whitney divided-shield experiment at the Tower Shielding Facility has been completed. This preanalysis, | which predicts the neutron and gamma-ray dose rates to be measured in the experiment, was performed to serve three purposes. First, the calcu- lations were intended to guide the experimentalists in setting up the measurements; second, a comparison of the calculated and experimental re- sults will indicate how well the theoretical methods at hand can be used to predict measurements; and third, the disagreements will lead to im- provements in both the theoretical methods and the experimental methods so that they can be more directly correlated. The Pratt & Whitney divided-shield design consists of a highly asym- metric reactor shield surrounding the Tower Shielding Reactor II that is separated from the TSF compartmentalized cylindrical crew shield by ap- proximately 64 ft. The reactor shield, a sketch of which is shown in the following section of this chapter, incorporates two advanced shielding ma- terials, lithium hydride as the outer neutron shield and depleted uranium as the gamma-ray shadow shield. The design dimensions for the reactor shield are given in detail in Table 12.1, and those for the crew compart- ment are given in Table 12.2. For a complex shield such as the Pratt & Whitney divided shield, every conceivable source of radiation must be considered and carefully weighed before a decision can be made as to its importance. Also, every theoretical method considered must be examined as to its applicability, avallability of calculaticnal methods, and practicality before it is used in the calculations. Thus, calculations for a complex shield system rep- resent a fairly large undertaking, and the final results contain many nuances of the above considerations. TFor reasons of convenience and sim- plicity, the housing for the reactor chambers, etc., positioned above the reactor core was ignored in the calculations, and the reactor shield was 129 Table 12.1. Gamma-Ray Shield Thicknesses and Outer Radii of Pratt & Whitney Reactor Shield Nominal dimensions: inner spherical radius of shield 19.364 in. thickness of inner stainless steel 0.125 in. can thickness of LiH "cooling layer"” 6.430 in, annulus thickness of gamma-ray shield stain- 0.375 in. less steel support can thickness of forward steel bulkhead’ 0.25 in. thickness of rear steel bulkhead 1.00 in. v Nominal Gamma-Ray Outer Radius v Nominal Gamma-Ray Quter Radius Polar Angle Shield Thickness® of Shield® Polar Angle Shield Thickness® of ShieldP (deg) (in.) (in.) (deg) (in.) (in.) 0 3.68 54 .27 100 0 39.15 5 3.65 54,05 105 0 38.75 10 3.55 53.770 110 0 38.35 15 3.38 53.05 115 0 37.88 20 3.19 52.45 120 0 37.55 25 2.85 51.50 125 0 37.15 30 2.50 50. 56 130 0 36.75 35 2.09 49,75 135 0 36.15 40 1.60 48,95 140 0 35.65 45 1.04 47.95 145 0 35.00 50 0.68 47.25 150 0 34.45 55 0.35 46.31 155 0 33.56 60 0.05 45,55 160 0 32.95 65 0 b4ér .65 165 0 32.28 70 0 43.85 170 0 31.85 75 0 42,85 175 0 31.50 80 0 42,05 180 0 31.45 85 0 41.16 90 0 40.35 95 0 39.65 thick stainless steel support can. 130 aDimensions do not include 0.125-in.-thick outer stainless steel can or 0.375-in.- bThese dimensions are nominal to the inner surface of the outer stainless steel can and include allowance for 0.25 in. of foam plastic to take up gaps and allow for expansion with temperature. Table 12.2. Dimensions of Cylindrical Crew Compartment Inside diameter = 36.0 in. Inside length = 72.0 in. Material Region Thickness (in.) b Sides and front 0.20 H;0 Sides and front 8.00 Pb Rear 3.27 H-0 Rear 18.00 considered to be azimuthally symmetrical about the reactor—crew compart- ment axis. Radiation Sources The radiation sources which could conceivably be important are (1) fission neutrons from the core and from the uranium shadow shield, (2) prompt-fission and fission-product decay gamma rays from the core and shadow shield, (3) capture and inelastic scattering gamma rays produced in the reactor and reactor shield, (4) capture and inelastic scattering gamma rays produced in air, and (5) capture and inelastic scattering gamma rays produced in the crew compartment. The fission neutrons in the shadow shield were neglected because this source was quite small. | A1l inelastic scattering gamma rays were also neglected because no suit- able method was available for treating this source. In addition, the capture gamma rays in the ailr and in the crew compartment were neglected because no adequate calculational method was available for computing the distributions of thermal and epithermal neutrons in these regions. These sources are expected to be small since the low-energy neutrons are markedly suppressed by lithium hydride. Attenuation and Transport Processes The attenuation and transport processes used in the calculation can be divided into three categories: (1) the radiation emergenf from the reactor—reactor shield system, hereinafter called the leakage spectrum, (2) the subsequent scattering of this radiation in air, and (3) the at- tenuation by the crew compartment shield of radiation which is both un- scattered and scattered in air. The calculations for each category will be discussed separately, and the assumptions and approximations that were involved in each case will be given. Leakage Spectrum. The reactor shield leakage spectra for both neu- trons and gamma rays were calculated on the basis of "moments method" datal for lithium hydride. The data were used as a point-to-point at- tenuation kernel from a source point in the reactor—reactor shield system to a point in air. No attenuation or buildup in air was included in this 131 part of the calculation. The assumption that the energy spectrum at a point in air can be characterized for both gamma rays and neutrons by data for lithium hydride is Jjustified on the following bases. Monte Carlo calculations indicate that the emergent energy spectrum for gamma rays 3 is characteristic of the final material if incident on infinite slabs?s that material is at least one mean free path thick. These conditions are satisfied in the reactor shield, since lithium hydride is the final ma- terial (outer layer) and its thickness is relatively large. The magnitude of the gamma-ray energy spectrum depends, of course, on the actual total' number of mean free paths thickness. In the case of neutrons, the fact that fast-neutron removal theory4 works fairly well makes possible the use of the emergent energy spectrum characteristic of the last hydrogenous material (that is, lithium hydride), since the neutron slowing down is due primarily to hydrogen. Here the magnitude of the energy spectrum depends on the total equivalent thickness of the last hydrogenous material, where the equivalent thickness is calculated on a hydrogen density basis. Cor- rections for norhydrogenous materials are effected through the use of re- moval cross sections. Since the "moments method" data were obtained for an infinite medium, their use in penetration calculations for finite media overestimates the energy spectrum, particularly at the lower energies. Furthermore, using the method of a point-to-point kernel leaves the angular distribution of the radiation completely open to question. In this calculation it was as- sumed that all the radiation is moving along the line of sight from the 1H. Goldstein and J. E. Wilkins, Jr., Calculations of the Penetration of Gamma Rays, NY0-3075, June 24, 1954; D. D. Babb, J. N. Keller, and E. McCray, Curve Fits of Gamma-Ray Differential-Energy Spectra, NARF-59-36T (MR-N-251), Nov. 1, 1959; H. Goldstein, Some Recent Calculations on the Penetration of Fission Neutrons in LiH, NDA-42, Aug. 7, 1957. 21,, A. Bowman and D. K. Trubey, Stratified Slab Gamma Ray Dose-Rate Buildup Factors for Lead and Water Shields, ORNL CF-58-1-41, Jan. 16, 1958. | 3H. Goldstein, Fundamental Aspects of Reactor Shielding, p. 223, Addison-Wesley, Reading, Mass., 1959. “Tbid, p. 326. 132 source point to the field point. The actual case is that radiation emerges from all points on the shield surface with an angular distribution which depends on where the surface point is located. The errors involved in this assumption can be qualitatively evaluated. The neutron source for the leakage calculation is the number of fis- sion neutrons produced per unit volume per unit reactor power with epi- thermal fissions neglected. The gamma-ray source is the number of gamma rays produced per unit volume per unit reactor power of a given energy arising from neutron absorptions, which include fission and thermal-neutron radiative capture. In ordef to obtain these sources, the thermal-neutron flux throughout the reactor—reactor shield system must be known. The flux slope in the core and immediately outside the core was obtained by calcu- lations with the GM-GNU reactor code’ for the case of a spherical reactor surrounded by 20 cm of water. It was extended by joining it with the measured BSR thermal-neutron flux at distances greater than 10 em from the reactor boundary. It was then normalized to a total power of 1 watt. Since the medium surrounding the TSR-II in the Pratt & Whitney design is lithium hydride instead of water, it was necessary to convert the flux for water to that for lithium hydride. This can be done approximately by assuming that most of the slowing down of neutrons to thermal energies is effected by hydrogen and that the effect on the slowing down by non- hydrogenous materials is taken into account by the use of removal cross ~ sections. If leakage is neglected, the number of neutrons arriving at thermal energies by hydrogen collisions in corresponding volume elements of the two materials is equal under these assumptions. The corresponding volume elements are found through the transforma- tion from medium 1 to medium 2 by PrT1 = P22 SGNU-II, A Multigroup One-Dimensional Diffusion Program, General Motors Corp. Report 101, Nov. 12, 1957, 133 and thus Pi1 AV, = dV; -5 > P2 where pi = hydrogen density in medium i, r, = spacial coordinate in medium i, dVi = differential volume element in medium i. The expression for the number of neutrons slowing down to thermal energies (and consequently being absorbed) in a differential volume element in a pure hydrogen slowing-down medium is 1) (2 ( o r 2, 'r e © ’ Egl)¢(l)(r1) dv, = e ° Z£2)¢(2)(r2) dva , where Zii) = macroscopic removal cross section for nonhydrogenous ma- terials in medium i, Zéi) = macroscopic thermal-neutron absorption cross section in ' medium i, _¢(i)(ri) = thermal-neutron flux in medium i at rj’ so that, finally, - (1) 3 ¢(2) n = E—?;)— (0_2) ¢(l)(I‘1) exp (Ei,l) *Zi.z)pl/pz) ry . P2 &Y P1 _ For the sake of simplicity, it was assumed that the thermal-neutron flux‘had only the radial dependence that was calculated with the reactor code (and extended by BSR data) and modified to correspond to lithium hydride, although the reactor—reactor shield system is actually asymmetri- cal in shape., This would probably not be too much in error except close to the outer boundary were it not for the presence of the shadow shield. 134 Since this assumption ignores the presénce of the shadow shield, however, and consequently the strong depression of the flux it causes, the number of gamma rays born in the shadow-shield region is overestimated. (This was borne out by further calculations with the GM-GNU code performed by W. E. Kinney of ORNL after the major part of this preanalysis was com- pleted.) However, even with this overestimate the contribution from the shadow shield is small. In calculating the thermal-neutron flux, various regions of the TSR- IT were homogenized for the sake of convenience and also to-save time in the calculation of the gamma-ray leakage spectrum. Four regions were used: the central control plate housing, the core annulus, the lead layer, -and the remainder, which was mostly aluminum and water. The thermal- neutron flux was not averaged over these regions. Errors may be intro- duced by this homogenization, but they are probably small. By far the largest contribution to the gamma-ray leakage spectrum was that of the capture gamma rays born in the last homogenized region of the TSR-IT, The results of these calculations were used to predict dose rates in air and in the crew compartment from radiation unscattered in air, that is, the direct-beam dose rates. Direct-beam dose rates in air 50 ft from the reactor are given in Tables 12.3 and 12.4 and those in the crew com- partment are given in Table 12.5. The results were also used to obtain an equivalent point source to be used in the air-scattering calculations. The equivalent point-source angular distribution is found by calculating the number current spectrum on the surface of a very large sphere at a given angle and multiplying the results by the radius squared. Air Scattering. The angular and energy distributions of neutrons which are emitted from a monodirectional and monoenergetic point source and subsequently scattered in air were estimated by using data from the Convair D-35 Monte Carlo code.® This code incorporates inelastic as well as elastic scattering, and the distances from the source point to the field point range from 10 to 100 ft. It also estimates the dose rate in M. B. Wells, Monte Carlo Calculations of Fast Neutron Scattering in Air, NARF-60-8T, Vols. 1 and 2 (FZK-9-147), May 13, 1960. 135 air from the scattered neutrons; hence at this point the estimate of the dose rate 1n alr has been completed. Table 12.3. Calculated Neutron Dose Rates for a Point in Air 50 ft from the Reactor Neutron Dose Rates (rep/w-hr) 6 a (deg) Direct Scattered Total 0 5.91 x 107° 2.53 x 10™7 2.59 x 1077 5 6.17 X 107° 2.53 x 107 2.59 x 1077 15 8.29 x 107° 2.56 x 10~7 2.64 % 10°7 30 2.00 x 1078 2.66 x 10°7 2.86 x 107 60 1,17 x 1077 3.19 x 10-7 4.36 x 107 90 5.68 x 10™7 4.55 x 10~7 1.02 x 10- 120 1.60 x 10~ 7.34 % 10-7 2.33 x 10-° 150 5.03 x 1076 1.22 x 10~ 6.25 % 1076 180 1.12 x 10°° 1.70 x 10-° 1.29 x 10~? % = polar angle, the polar axis being the axis of metry through the reactor and crew compartment. sym- The positive direction of the polar axis is from the reactor to the crew compartment. - Table 12.4. Calculated Gamma-Ray Dose Rates for a Point in Air 50 ft from the Reactor Gamma-Ray Dose Rates (r/w-hr) 6 : a, (deg) Direct Scattered Total 0 1.42 x 107° 9.36 x 10-° 1.08 X 10°° 5 3.60 x 1076 9.48 x 1076 1.31 X 1077 15 1.26 x 10-° 1.01 x 10~° 2.27 x 107° 30 5,32 x 10°° 1.25 x 1077 6.57 x 107 60 2.91 x 10°% 2.%44 x 10-° 3,15 x 1074 90 3,91 x 1074 3,40 x 107° 4,25 x 1074 120 5.17 X 1074 4.37 x 1075 5.61 X 10~% 150 6.38 x 1074 5,28 x 1072 6.91 x 10~% 180 6.98 x 10™% 5.70 x 1077 7.55 X 10~% 136 %Defined in Table 12.3. Table 12.5. Calculated Neutron and Gamma-Ray Dose Rates for a Position Inside the Crew Compartment Neutron Dose Gamma-Ray Dose Rate Rate (rep/w-hr) (r/w-hr) Scattered radiation 1.55 x 107° 7.9 % 107° Direct beam 9.93 x 10-11 1.05 x 1078 Total 1.65 x 107° 1.84 x 1078 Monte Carlo calculations yielding energy and angular distributions and dose rates for gamma rays scattered in air are also available,7 but because the energy and angular distribution data, which fill four large volumes, are not in a form that can be readily used with computers only the dose-rate data could be used directly. For this preanalysis it was necessary to obtain the energy and angular distributions of gamma rays emitted from a point source by using the approximation of single scatter- ing with no attenuation on either leg in air. This is probably not ter- ribly wrong,8 except for the region of the front of the crew compartment, where it is certainly an underestimate. Air-scattered neutron and gamma-ray dose rates calculated by these procedures are given in Tables 12.3 and 12.4, respectively, for a point in air 50 ft from the reactor. Crew Compartment Attenuation. The neutron attenuation in the crew compartment was calculated by use of the Monte Carlo ABCD Code.® This code assumes the geometry of a cylindrical crew compartment, takes into account inelastic as well as elastic scattering, and assumes that there is a magnetic tape containing the angular and energy distribution of 7R. E. Lynch et al., A Monte Carlo Calculation of Air Scattered Gamma Rays, ORNL-2292, Vols, 1-5, Sept. 10, 1958. 8D. K. Trubey, The Single-Scattering Approximation to the Solution of the Gamma-Ray Air-Scattering Problem, ORNL-2998, Jan. 20, 196l. ?Herbert Steinberg, Monte Carlo Code for Penetration of Crew Com- partment-IT, TRG-211-3-FR. ' 137 neutrons scattered in air (that is, it uses the output from the Convair D-35 code referred to above). This angular and energy distribution is assumed to be uniformly distributed on the outside surface of the crew compartment in such a way that if the thicknesses on the crew compart- ment approach zero, the dose rate in the crew compartment approaches that in air. The code only allows one medium, making it necessary to replace the lead on the rear of the crew compartment with an equivalent water thickness, prescribed as follows: 10 ZPb tHZO = th —;; XH20 = 0.58 th 5 where t = thickness, Zib = removal cross section of lead = 0.116 cm™ 1, KH o= relaxation length of water = 10 cmn. 2 The small lead thickness on the side was ignored. The attenuation of gamma rays by the crew compartment was computed with the Monte Carlo TRIGR-P Codell for slabs. As mentioned in the pre- ceding discussion, the energies and angles of incidence upon the slabs representing thicknesses on the crew compartment were determined by the single-scattering approximation. The results were integrated 6ver the proper range of angles of incidence, and the dose rates were combined by weighting with twice the fractional area of the particular surface of the crew compartment., This gives the corfect limit as the crew éompartment thicknesses approach zero. However, the weighting should have been done at the time of integration over the angles of incidence and it should have been with the fractional projected areas onto planes normal to the in- cident directions. The neutron and gamma-ray dose rates inside the crew compartment after scattering in air are given in Table 12.5. 107, vy, Blosser, ORNL, private commnication. llMonte Carlo Calculations of Gamme Ray Penetration, TRG 125-FR-II (WADC 59-771). 138 Future Calculations As pointed out in the preceding steps, several improvements are needed for calculations of this type. For example, the moments-method type of calculation of the leakage spectrum should be replaced By some finite- medium calculation which also predicts angular distributions and can handle multiregion shields. It is doubtful that this can be done in the near future, however. If would also be desirable to be able to predict the thermal-neutron flux throughout the entire reactor—reactor shield system, as well as in air and in the crew compartment. In addition, some method for handling inelastic scattering gamma rays should be devised. This 1is possible, but it would have to be done in conjunction with the neutron slowing-down problem. Other aids would be the availability of air- scattered gamma-ray data in a usable form, which would probably be best done by recalculation, and the availability of crercompartment penetra- tion codes which take into account the crew-compartment geometry and allow for multiregion shields. The latter problem is not insurmountable. Mathematical Description of Calculation Consider a point isotropic source in a finite medium which is sur- rounded by a vacuum. This source gives rise to radiation leaking from all points on the surface of the medium. If we approximate the flux, not the current, on the surface by that in an infinite medium, then we are overestimating the flux. If we further assume that the radiation leaving the surface is directed along the line of sight from the source to the point on the surface, then the flux is overestimated in that di- rection and of course completely neglected in other directions. There- fore, under these assumptions, if we are interested in the flux at some point in the vacuum, the error in the estimated flux depends on hdw im- portant the contribution is from other parts of the surface which have radiation directed toward the point in the vacuum. For the system in question, the resulting leakage spectrum at the point directly to the rear of the reactor shield is certainly overestimated. The trend of the error is nebulous for the point directly to the front of the reactor shield, since the obvious errors tend to compensate. 139 This analysis also holds for the leékage spectrum headed toward the detector, if the surrcunding medium is air, but the contribution to the flux from radiation leaving the shield surface and subsequently scatter- ing in air must be considered. Here again, the error is nebulous, since it depends on both\the importance of the angular distribution (including the magnitude) of the radiation leaving the surface and the probability of scattering in air and reaching the field point. The estimate of the flux in air in accordance with the assumptions resulting from gamma rays produced in the reactor—reactor shield volume and not scattering in air is given by ~ ¢(r) : : _¢7(E:2) = zl'{; fdl” F:“F E Za(_r_)fyl(Eo) dEq X Z X [zmrz?M (Ef u(Eo,27) dz’; Eo):, ) 0 and the equivalent point source of gamma rays leaking from the reactor shield is §7(8,0) = = far ¢(x) T s [N (B ao x Z X [szM (Ef w(Eo, 27) at’; Eo)] , 0 , where ¢7(E,D) = flux of gamma rays of energy E per unit energy range at the spacial position D in air, n ~ — = o ~ il number of gamma rays per sec per unit energy range per steradian emitted from the equivalent point source in the direction given by the unit vector g, r = spacial position of a point in the reactor—reactor shield volume, 140 dr = differential volume element at r, = thermal-neutron flux at r, = thermal-neutron macroscopic absorption cross sec- tion (including fission) of the ith nuclear species at r, ' 71(E0) = number of gamma rays emitted with energy Eg per unit energy range resulting from a thermal-neutron absorption, 12 I = distance from the source at r to the shield surface measured along the direction of the unit vector D<:f£ in the case of the flux in air and along the unit vector { in the case of the equivalent point source, 1’ = distance from the point socurce to another point within the shield measured in the same manner as I, IJ(EO) l/) gamma-ray macroscopic absorption cross section at energy Eg and at the spacial position determined ; by 17, .l; W(Bo, 17) a1’ I total number of mean free paths at energy Eg for the distance I, 4r1?M(E,X;BEg) = 4m1° times the flux of gamma rays of energy E given by the moments-method solution for a point isotropic source emitting gamma rays of energy Eg at a dis- tance X in mean free paths at energy Eg in an in- finite medium of lithium hydride. These same quantities for neutrons are . r . (,bn(E’.]_).) = E de—I—D—_-—lz sz(_I;) [47TZ§N(E, le)] X b fLiH (17) X exp f (Z -5 (1) B )dl/]; ’ [ * t LiH sn(E,-'fi) = Z]:T—T: fd; ¢(r)vip(r) [4miZN(E,1))] X ’ LiH / -p(l’) / o[£ (47500 £22) ], 128, Troubetzkoy and H. Goldstein, A Compilation of Information on Gamma—Ray Spectra Resultlng from Thermal Neutron Capture, ORNL-2904, May 17, 1960. 141 where l f dl/ D(Z/)) € o PLin v = number of fission neutrons per fission, l I Zf(r) = macroscopic thermal-neutron fission cross section at r, Prif = hydrogen density in lithium hydride, 0(1’) = hydrogen density at spacial position determined by 1/, ZLlH = macroscopic removal cross section for lithium in t lithium hydride, Zr(l’) macroscopic removal cross section for nonhydrogenous material at spacial position determined by 17, 4vl: times the flux of neutrons of energy E given by the moments-method solution for a point isotropic fission source at a given distance ! 1n an infinite medium of lithium hydride. © 2 4WleN(E,le) The reactor—reactor shield is composed of spherical or, at worst, azimuthally symmetrical layers of materials. The problem of computing the lengths along the directions E<:f£ or 5 across regions is done.as follows. The radius of the layers is a function of the angle 6 that the radius makes with the axis of symmetry; i.e., R = radius vector to the intersection of a line in the direction § (from the source point) and the shield layer in guestion, R=10Q+r, R(8) cos 6 = 1 cos V¥ + Z, R2(6) = 12 + r2 + 21(r - Q), @ = angle between symmetry axis and the direction fi, = angle between symmetry axis and the direction fi, V Z = projection of r on the symmetry axis, l=—(£-§) i\/(£-5)2+R2—r2 = 0. If the layer is spherical, the problem is trivial. For the layers which are not spherical, an iteration technique is employed. Tables of R versus cos 6 and R versus R cos 6 are generated. An initial guess of the value of cos 6 1s made and hence an initial R is obtained. A value of I is then computed with this estimate of R by the quadratic formula, 142 4) and an estimate of R cos € is made with the resulting value of [ by Rcos @ =171cos ¥V +72 . Hence a new R is obtained for the next iteration. The test for convergence is I, i+ 1 - 7 < some small number, i where i denotes the order of iteration. When this is satisfied, Zi+l is accepted. This scheme works very well as long as R varies slowly compared with 1, which is the situation for practically all cases. Fortunately, the cases where the scheme does not work well are few in number and also un- important. The neutron flux spectrum and dose rates on the shield sur- face are presented in Tables 12.6 and 12.7, ahd equivalent point-source spectra are presented in Tables 12.8 and 12.9. The dose rate in air at spacial position D from radiation leaving the reactof—reactor shield system and scattering in air is given by D_(D) = fae faf s(E,0) T(E,8-D,D) |, K3 P I » Oy S 2 i dose rate in air due to a monodirectional monoenergetic point source, 0p! P = @ Il equivalent point source spectrum, o Il unit vector in the direction of D. The dose rate in the crew compartment due to radiation leaving the reactor— reactor shield system, scattering in air, and penetrating the crew compart- ment is Dg(D) = Sfam St s(,8) p(g,8,0) |, 143 Table 12.6. Neutron Leakage Spectrum at Shield Surface Neutron Flux (27 neutrons/w.sec-cm?.Mev) 8 (deg)® E =1.1 Mev E = 2.7 Mev E = 4 Mev E = 6 Mev E = & Mev E = 10.9 Mev 0 6.96 X 1072 3.11 X 1072 2.48 x 1072 2.02 x 1072 1.45 x 1072 7.79 x 10~ 20 1.21 X 1071 5.37 X 1072 4.29 x 1072 3,42 x 1072 2.43 x 1072 1.27 x 10~2 30 2.27 x 10°Y 9.85 x 1072 7.90 x 10-2 6.18 x 10”% 4.31 x 1072 2,20 x 102 40 4,90 x 107 2,10 x 107t 1.67 x 107} 1.28 x 107! 8.84 x 1072 4.38 x 1072 50 9.78 x 1071 4.14 x 107!' 3.27 x 107! 2.46 x 107! 1.65x 1071 7.98 x 1072 60 1.87 x 10° 7.81 x 10°Y 6.11 x 107Y 4.49 x 107! 2,94 x 10! 1.37 x 107! 70 4,24 x 100 1.7 x 10° 1.33 x 10° 9.46 x 107t 5,91 x 107* 2.60 x 107% 80 9,78 x 10° 3.92 x 10° 2.88 x 109 1.93 x 10° 1.13 x 10° 4.60 x 1071 90 1.87 x 10t 7.37 x 10° 5.28 x 109 3.36 x 10° 1.88 x 10° 7.4 x 107t 100 3.44 x 10t 1.34 x 10° 9.30 x 10° 5.62 x 10° 3.01 x 10° 1.07 x 10° 120 7.00 x 10% 2.67 x 10% 1.81 x 10t 1.02 x 10t 5.17 x 109 1.70 x 10° 140 1.63 x 102 6.13 x 101 3.95 x 10t 2.04 x 10t 9,78 x 10° 2.90 x 10° 160 6.12 X 102 2.26 X 102 1.33 x 102 6.06 x 10t 2.64 x 10t 6.62 x 10° 180 7.57 x 102 2.78 x 102 1.6l x 102 7.23 x 101 3,10 x 10! 7.62 x 10° %Defined in Table 12.3. Table 12.7. Neutron Dose Rates at Shield Surface e Dose Rate e Dose Rate (eg)? (rep/w-hr) (deg) (rep/w-hr) 0 6.92 x 107 80 6.63 X 1077 20 1.17 x 1078 90 1.20 x 10-% 30 2.10 x 1076 100 2.08 x 1074 40 4,36 x 1076 120 4.04 x 1074 50 8.29 x 1076 140 9.07 X 10-% 60 1.52 x 10°? 160 3.28 x 1073 70 3.17 x 10°° 180 4,02 x 1073 aDefined in Table 12.3. Table 12.8. Equivalent Point Source Neutron Spectra Neutron Current (27 neutrons/w:sec-Mev-Steradian) a (deg) E=1.1Mev E = 2.7 Mev E = 4 Mev B =6 Mev E = 8 Mev E = 10.9 Mev 0 1.43 x 103 6.40 x 102 5.11 X 102 4,14 x 102 2.97 x 102 1.60 x 102 5 1.42 x 103 6.36 x 102 5.08 x 102 4,12 x 10% 2.96 x 102 1.60 x 102 15 1.87 x 103 8.29 x 1072 6.61 x 102 5,31 x 102 3.78 x 102 2.00 x 102 30 4,10 x 103 1.79 x 103 1.43 x 103 1.13 x 102 7.90 x 102 4,07 x 102 60 2.85 x 10% 1.18 x 10% 9,23 x 10° 6.69 x 103 4.33 x 103 1.99 x 10? 90 2.12 x 107 8.29 x 104 5.90 x 10% 3.69 x 10% 2.04 x 10% 7.60 x 103 120 6.64 X 107 2,53 x 10° 1.70 x 10° 9.38 x 10% 4.75 x 104 1.53 x 10% 150 2.10 x 108 7.80 x 10° 4,75 x 10° 2.28 x 10° 1.04 x 107 2.82 x 104 180 4.90 x 10° 1.80 x 10° 1.04 x 10% 4,67 X 10° 1.99 x 10° 4.90 x 10% aDefined in Table 12.3. 144 Table 12.9. Equivalent Point Source Gamma-Ray Spectra Gamma-Ray Current (2 photons/w?sec-Mev-Steradian) 6 (deg)® - ' E = 0.5 Mev E=2.0Mev E=4.0Mev E =6.0Mev E = 8.0 Mev E = 10.0 Mev 0 3.25 x 106 1.01 x 106 2.94 x 10° 2.89 x 10° 2.25 x 10° 9.93 x 103 5 4.45 x 106 2.48 x 108 7.02 x 10° 8.13 x 10° 7.19 x 107 5,32 X 10% 15 2.30 x 107 7.23 x 108 2.24 x 10°® 2.65 x 106 2.88 x 108 1.73x 10° 30 1.52 x 108 2.79 x 107 8.44 x 106 9,54 x 108 1.18 x 107 7.47 x 103 60 1.78 x 10° 1.41 x 108 3.56 x 107 3.16 x 107 4.00 x 107 2.44 x 1068 a0 2.20 x 10° 2.13 x 108 5.11 x 107 4.28 x 107 5.50 x 107 3.35 x 106 120 3.23 x 10° 2.66 x 108 6.19 x 107 5.18 x 107 6.70 x 107 4.07 x 106 150 4.39 x 10° 3.17 x 108 7.03 x 107 5.72 x 107 7.39 x 107 4. b4l x 108 180 4.95 x 107 3.44 x 108 7.46 x 107 5.94 x 107 7.66 x 107 4.58 x 108 ®Pefined in Table 12.3. where P(E,fi,D) is the dose rate in the crew compartment due to radiation leaving a monodirectional monoenergetic point source, subsequently scatter- ing in air, and then penetrating the crew compartment at a distance D away from the source. For neutron dose calculations, values of P(E,fi,D) are obtalined directly by using the ABCD code and D-35 code for neutrons. For gamma rays it is computed on the basis of single scattering in air and the penetration of “infinite slabs using the TRIGR-P code. The procedure is as follows: P(E,,0) = .=~ [ap(5,8,0,8) + Ap(m,8,0,0) + ap(e,d,0,0)] total where P(E,,D,A) = dose rate behind an infinite slab with constituents and orientation the same as area A on the crew compartment from radiation leaving a monoenergetic monocdirectional point source, subsequently scattering in air, and then Penetrating the crew compartment which is at a distance D away from the source, Atotal = total crew compartment area, 145 AS = side crew compartment area, AR = rear crew compartment area, AF = front crew compartment area. Since the equivalent point source is azimuthally symmetrical, we use A A 1 2”- ~ P(E,0-D,D,4) = == [ d¢ P(E,q,D,4) =V 2T / 1 do A “mmemy J e L w9 B s (R X I[ES, cos (K,B,(b)] H[COS (K-)B}(b)] s where A A cos ¥ = (0 H(x) = {Oj x < 0. A better estimate of P(E,Q(@,D) would have been: 146 " or do ) S dp fo dp n — (E,¥+8) X P(E,8-D,0) = 2 2—1@,‘@ j 4\ DEPLETED URANIUM ORNL-LR-DWG 58223 MOVABLE WATER- FILLED PATCH AXIS OF _J LITHIUM HYDRIDE ALL DIMENSIONS ARE IN INCHES 4 0 4 8 12 e — e — INCHES Fig. 12.3. 152 Cross-Sectional View of Pratt & Whitney Reactor Shield. SYMMETRY - A\l 131.60 in. OUTSIDE UNCLASSIFIED 2—01—-056—15—P-490R2 e——37.05 in. OUTSIDE ‘—‘”-—7 73,68 in. OUTSIDE——————-‘iZOAS in. OUTSIDE —= 12.0in. (TYP) {10 10a 9 16 21 26 3 1242 in.—= frea— g G | t : ] z < 4 t < _~ 36.0in. DIA ~ D U D U rol oz K 8 et ‘{ 17.74in. Dm—T g 18.0 in. OD —0.188in. 1 0.188 in. Q.188 in. 12.36 in; 24.06in. (TYP} 7.96 in. O.IBBI‘Ln. 250in.% n N 0250, 22T Nosron %, ALL METAL THICKNESSES ARE 0.125in. (TYP) UNLESS SPECIFICALLY DESIGNATED OTHERWISE Fig. 12.4. Cross-Sectional View of TSF Compartmentalized Crew Com- partment. | UNCLASSIFIED 2-01-060-95 OAK RIDGE GRID NORTH —J <1 Q — 4 09 L) - | ¢ £ | //// // / — / . f ) " & = r /- . ) CENTER LINE OF TSF PAD REACTOR CENTER Fig. 12.5. Coordinate System Used for TSF Pratt & Whitney Ex- periment. ' ; 153 Table 12.14. Compositions of Crew Compartment Shielding Materials Compositicn Material Relative ateria Density Hydrogen Carbon Oxygen Boron (atoms/cm?®) (atoms/ecm®) (atoms/cm®) (wt %) _ x 1022 % 1022 X 1022 Water 1.00 6.69 3.34 Transformer oil: &6.7 wt % C, 0.87 6.6 3.78 12.7 wt % H Transformer oil and Borester '7: 0.92 6.2 3.35 0.45 2.9 50 wt % each Borated polyethylene: 3.8 wt % 0.9 7.8 3.94 3.0 B4C, 96.2 wt & (CH,), detector. When measurements are being made in the crew compartment, the distance from the reactor center to the rear surface of the crew shield cavity is 50 ft, with the crew shield centerline on the line determined by 8 = 0 and ¢ = 90 deg. The distance between the rear lead in the crew shield and the center of detection of the detector is designated by the symbol x, At times the reactor is rotated about its vertical centerline, in which case the counterclockwise angle through which it turns from its normal position (i.e., with the thickest point of the shield at 6 = O and ¢ = 90 deg) must be given. The symbol used for this angle is Q. Thus far, the experimental program is about one-half complete. There- fore, the description of the program which follows is divided into measure- ments already completed and p}anned measurements. Completed Measurements. Most of the measurements completed to date have been made in air around the reactor shield. At each point the measure- ments usually included the fast-neutron dose rate, the gamma-ray dose rate, and the thermal-neutron flux. The detector positions used for these measure- ments are given in Table 12.15. For each position a complete rotation of the reactor was made; i.e., O was varied from O to 360 deg. Samples of these measurements are shown in Figs. 12.6 and 12.7, which represent map- pings of the fast-neutron and gamma—ray dose rates in the horizontal plane 50 ft from the center of the reactor as a function of . 154 Table 12.15, Coordinate Positions for In-Air Measurements Around Pratt & Whitney Shield h, Reactor 0 ¢ (deg) Altitude (ft) 58.6 in. 90 100 65.1 in. 67.5 100 65.1 in. 112.5 100 50 ft 90 15 50 ft 90 50 50 ft 20 175 75 ft 45 125 75 ft 20 175 75 Tt 135 175 In addition to the in-air measurements at various distances from the reactor, several points were obtained with counters directly on the surface of the shield. Also, an altitude traverse from h = 6 ft to h = 175 ft was run for a detector position specified by p = 65 ft, € =0, and ¢ = 90 deg. For this run, the reactor was at o = O. Some of the dose-rate measurements inside the crew shield have also been completed. One altitude traverse was made with three detectors at the lateral positions shown in Fig. 12.8. For this traverse, the detector coordinate x was at 36 in. In all other cases, x was varied between 12.5 and 53.5 in., continuously for gamma-ray dose rates and thermal-neutron fluxes and in about &-in. steps for fast-neutron dose rates. Data of this type have been taken with the optimum crew compartment configuration using water, oil, and the borated oil mixture. Data for the borated oil mixture were also taken with all the front tanks (tanks 27, 28, 29, 30, 32, and 33) drained to determine the relative amount of radiation penetrating the forward compartments. | Measurements have also been made inside the crew shield as a function of the position of the water-filled patch. These were performed with the 155 optimum crew shield configuration, except that water rather than borated olil was used in the tanks. UNCLASSIFIED 2- 04 - 056 -XXX-832 ¥ -1 TISSUE o FAST-NEUTRON DOSE RATE (ergs-g w05 180 200 220 240 260 280 300 320 340 O 20 40 60 80 {00 20 140 160 180 a, REACTOR HORIZONTAL ROTATION ANGLE (deg) Fig. 12.6. Fast-Neutron Dose Rates 50 ft from Center of Reactor as a Function of @, TSR-II in Pratt & Whitney shield. ' 156 L » GAMMA-RAY DOSE RATE (ergs-q}'lssu E-hr"-m" ) Planned Measurements. Additional in-air mapping measurements are planned, primarily traverses in the angle ¢ with p = 25 ft and with various UNCLASSIFIED 2-01-056 -XXX -833 Tom 1072 wn ~N 10-3 104 180 200 220 240 260 280 300 320 340 O - 20 40 60 80 100 120 140 160 180 a, REACTOR HORIZONTAL ROTATION ANGLE (deg) Fig. 12.7. Gamma-Ray Dose Rates 50 ft from Center of Reactor as a Function of &, TSR-II in Pratt & Whitney shield. 157 values of 6., Collimated dose-rate measurements will also be made as a function of reactor rotation, with the collimator looking only at the shield surface, to determine scattered and direct components of the in-air measurements, and some foil measurements directly on the reactor shield surface will probably be / HOLDERS FOR FND, ASD, OR BF, made. CAVITY CENTERLINE POSITION C UNCLASSIFIED 2-01-056-XXX~-834 HOLDER FOR . PARAFFIN-COVERED \ BF3 COUNTER TRAVERSING MECHANISM MAXIMUM OF i-in. Pb SIDE SHIELDING INSIDE CAVITY Fig., 12.8. Counter Positions Inside 36-in.-diam Cavity in Crew Com- partment (Looking Toward 158 Reactor). ~d » 3 Gamma-ray and neutron spectral measurements are desired, the most urgent being the neutron spectrum at the rear of the reactor shield. If time permits, both neutron and gamma-ray spectra will be taken at several positions. TR -~ Reports previously issued in this series are as follows: 160 ORNL-528 ORNL-629 ORNL-768 ORNL~858 ORNL-919 ANP-60 ANP-65 ORNL-1154 ORNL-1170 ORNL=-1227 ORNL~1294 ORNL-1375 ORNL-1439 ORNL-1515 ORNL-1556 ORNL-1609 ORNL-1649 ORNL-1692 ORNL-1729 ORNL-1771 ORNL-1816 ORNL-1864 ORNL-1896 ORNL-1947 ORNL-2012 ORNL-2061 ORNL-2106 ORNL-215% ORNL-2221 ORNL-2274 ORNL-2340 ORNL-2387 ORNL-2440 ORNL-2517 ORNL-2599 ORNL-2711 ORNL-2840 ORNL-2942 ORNL-302%9 Period Period Period Period Period Period Period Period Period Period Period Period Period Period Period Period Period Period Period Perioad Period Period Perioad Period Period Period Period Period Period Period Period Period Period . Period Perioad Period Period Period Period S Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending Ending November 30, 1949 February 28, 1950 May 31, 1950 August 31, 1950 December 10, 1950 March 10, 1951 June 10, 1951 September 10, 1951 December 10, 1951 March 10, 1952 June 10, 1952 September 10, 1952 December 10, 1952 March 10, 1953 June 10, 1953 September 10, 1953 December 10, 1953 March 10, 1954 June 10, 1954 September 10, 1954 December 10, 1954 March 10, 1955 June 10, 1955 September 10, 1955 December 10, 1955 March 10, 1956 June 10, 1956 September 10, 1956 December 31, 1956 March 31, 1957 June 30, 1957 September 30, 1957 December 31, 1957 March 31, 1958 September 30, 1958 March 31, 1959 October 31, 1959 April 30, 1960 October 31, 1960 L1 b ) O R -gowumduwH ?flF:?fl?fl§1$1}+§2§1;:Fip*?-Fzgizzgzfiaz:;iz:w g QrrdorEmEHOn g ORNL-~3144 C-84 — Reactors—Special Features of Aircraft Reactors M-3679 (24th ed.) INTERNAL DISTRIBUTION M. Adamson 44, F. C. Maienschein W. Allen 45. W. D. Manly E. Beall 46. A. J. Miller S. Billington 47. K. Z. Morgan F. Blankenship 48. F. J. Muckenthaler P. Blizard 49, M. L. Nelson L. Boch 50. P. Patriarca BE. Boyd 51, S. K. Penny B. Briggs 52. A. M. Perry D. Callihan 53. P. M. Reyling E. Center (K-25) 54, H. W. Savage A, Charpie 55, A. W. Savolainen E. Clausing 56. E. D. Shipley S. Cockreham 57. 0. Sisman B. Cottrell 58. M. J. Skinner L. Culler 59. G. M. Slaughter A. Douglas 60. A. H. Snell L. Fowler 61. I. Spiewak P. Fraas 62. B. J. Sturm H. Frye 63. C. D. Susano J. Gray 64. J. A. Swartout L. Greenstreet 65. D. B. Trauger R. Grimes 66. D. K. Trubey Guth 67. G. M., Watson 0. Harms 68. A. M. Weinberg Hikido 69. J. C. White R. Hill 70. E. P. Wigner (consultant) . E. Hinkle 71, C. E. Winters E. Hoffman 72-78. Laboratory Records Department W. Hoffman 79. Laboratory Records, ORNL R. C. Hollaender 80-83. ORNL — Y-12 Technical Library B. Holland Document Reference Section Tnouye 84—86. Central Research Library H. Jordan W. Keilholtz P. Keim J. Keyes G. Lafyatis R. Laing S. Livingston N. Lyon G. MacPherson E. Maerker - 161 87. 88-91. 92-93. 9. 95-97. 98—99. 100. 101-102. 103-104%. 105. 106—108. 109. 110. 111-112 113. 114, 115-120. 121. 122. 123. 124~125. 126. 127. 128. 129. 130. 131. 132. 133-134., 135. 136. 137. 138. 139. 140-147. 148-150. 151. 152. 153. 154~155. 156. 157. 158. 159. 160. 161. 162. 163. 164 . 162 EXTERNAL DISTRIBUTION AiResearch Manufacturing Company Air Force Ballistic Missile Division ATPR, Boeing, Seattle AFPR, Douglas, Long Beach AFPR, Douglas, Santa Monica AFPR, Lockheed, Marietta AFPR, North American, Downey Air Force Special Weapons Center Air Research and Development Command (RDRRA) Air Technical Intelligence Center ANP Project Office, Convair, Fort Worth Albuquerque Operations Office Argonne National Laboratory Army Ballistic Missile Agency Army Rocket and Guided Missile Agency Assistant Secretary of the Air Force, R&D Atomic Energy Commission, Washington Atomics International Battelle Memorial Institute Brookhaven National Laboratory Bureau of Naval Weapons Bureau of Naval Weapons General Representative BUWEPSREP, Aerojet~General, Azusa BUWEPSREP, Chance Vought, Dallas BUWEPSREP, Convair, San Diego BUWEPSREP, Grumman Aircraft, Bethpage BUWEPSREP, Martin, Baltimore Bureau of Yards and Docks Chicago Operations Office Chicago Patent Group Defense Atomic Support Agency, Washington Director of Naval Intelligence DuPont Company, Aiken Engineer Research and Development Laboratories General Electric Company (ANPD) General Electric Company, Richland General Nuclear Engineering Corporation. Hartford Aircraft Reactors Area Office Idaho Test Division (LAROO) Knolls Atomic Power Laboratory Lockland Aircraft Reactors Operations Office Los Alamos Scientific Iaboratory Marquardt Aircraft Company Martin Company National Aeronautics and Space Administration, National Aeronautics and Space Administration, National Bureau of Standards Naval Air Development Center Naval Air Material Center Cleveland Washington b A P o 165. 166. 167. 168. 169. 170. 171. 172. 173-174. 175~178. 179. 180. 181—-182. 183, 184, 185. 186. 187. 188. 189-190. 191-193 194~205. 206—~230. 231. Naval Air Turbine Test Station Naval Research laboratory New York Operations Office Nuclear Metals, Inc. Oak Ridge Operations Office Office of Naval Research Office of the Chief of Naval Operations Patent Branch, Washington Phillips Petroleum Company (NRTS) Pratt & Whitney Aircraft Division Public Health Service Sandia Corporation School of Aviation Medicine Sylvania-Corning Nuclear Corporation Technical Research Group USAF Headquarters USAF Project RAND U. S. Naval Postgraduate School U. S. Naval Radiological Defense ILaboratory University of California, Livermore Westinghouse Bettis Atomic Power Laboratory Wright Air Development Division Technical Information Service Extension Division of Research and Development, AEC, ORO AL e TR A ihe S 163 SESTRICIED DRI . This document contains Repssssinsretpeimrpemsegesinsd in the Atomic Energy Act of 1954, 1Its t disclosure of its contents in any manner to an unauthor i ‘ wed. ' 3