TRIUMF: Canada's national laboratory for particle and nuclear physics

Conceptual design of the TRIUMF thermal neutron facility Thorson, I. M.; Arrott, A. S. Jul 31, 1971

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
51833-TRI-71-03.pdf [ 31.65MB ]
Metadata
JSON: 51833-1.0302687.json
JSON-LD: 51833-1.0302687-ld.json
RDF/XML (Pretty): 51833-1.0302687-rdf.xml
RDF/JSON: 51833-1.0302687-rdf.json
Turtle: 51833-1.0302687-turtle.txt
N-Triples: 51833-1.0302687-rdf-ntriples.txt
Original Record: 51833-1.0302687-source.json
Full Text
51833-1.0302687-fulltext.txt
Citation
51833-1.0302687.ris

Full Text

TRIUMF CONCEPTUAL DESIGN OF THE TRIUMF THERMAL NEUTRON FACILITY I .M. Thorson and A.S. Arrott Simon Fraser University Department of Physics Burnaby 2, British Columbia Canada TRI-71-3 July 1971 ABSTRACT The report describes the conceptual design of the proposed thermal neutron facility (TNF) based on the neutron source produced by dumping the residual beam from the main TRIUMF proton beam line into a high atomic weight target. The maximum useful thermal neutron flux in a water-reflected, 120 em diam heavy-water-graphite moderator assembly surrounding a 15 em diam Pb-Bi target is ~0.9 x 10 13 cm- 2 sec-1 at 50 kW residual beam power; this figure is higher by a factor of ~2.5 for a natural uranium target. The neutron irradiation and beam tube facilities are described as well as facilities for carrying out pro-ton beam irradiations. Heat dissipation and residual activity problems are discussed and a preliminary cost estimate is given. The total estimated cost, without any contingency allowance, is $434,000, of which 60% is for shielding. i i C 0 N T E N T S 1. INTRODUCTION 1.1 Basis for the Facility 1.2 Purpose and Requirements 2. GENERAL DESCRIPTION Page 3 2.1 Vacuum Tank 3 2.2 Vacuum Window 4 2.3 Moderator Assembly 5 2.4 High Energy Neutron Facilities 6 2.5 Proton Irradiation Facilities 6 2 6 Thermal Neutron Beam Facilities 7 2.7 Neutron Irradiation Facilities 8 2.8 Neutron Production Target Assembly 10 2.9 Tabulation of Irradiation and Beam Tube Facilities 12 3. FACILITY CHARACTERISTICS AND SPECIFICATIONS 3.1 Proton Beam 3.2 Neutron Source 3.3 Neutron Flux Distribution 4. SHIELDING, ACTIVATION AND HEAT PRODUCTION 4.1 Operating Shield Requirements 4.2 Residual Activity and Radiation Fields 4.3 Residual Activity in Cooling Water 4.4 Heat Production and Dissipation 4.4.1 Heat Production 4.4.2 Target Heat Dissipation 5. FABRICATION AND COST ESTIMATE 5.1 Shielding Optimization 5.2 Actual Shielding Costs 5.3 Building Costs 5.4 Thermal Neutron Facility Core Costs 5.4.1 Vacuum Tank and Shielding Plugs 5.4.2 Vacuum System 5.4.3 Cooling System 5.4.4 Instrumentation and Control 5.4.5 Moderator-Reflector Assembly 5.4.6 Target Assemblies and Access Tubes 5.5 Facility Total Cost Acknowledgements References Appendix A. Appendix B. Figures Fast/Thermal Ratio Optimization Two-Component Shielding Optimization iii 14 14 14 17 26 26 29 33 37 37 40 44 44 45 47 48 50 50 51 51 52 53 55 56 58 61 64 II Ill IV v VI VII VIII IX X XI XII XIII XIV XV Ll ST OF TABLES Beam Tubes and Irradiation Positions in the Thermal Neutron Facility Neutron and Heat Production Results from Monte Carlo Calculations for 500 MeV Protons on 10 and 20 em diam Targets Energy Groups and Source Strengths for FORGOD Calculations Neutron Diffusion Parameters for FORGOD Calculations Flux Perturbations due to Structural Components in the Reference Design Assembly Cascade Neutron Leakage Power and Relaxation Lengths Operating Dose Rates and Shield Estimates for the Thermal Neutron Facility at 100 kW Beam Power, 500 MeV Saturated Residual Activity and Radiation Fields Induced by High Energy Reactions in the Thermal Neutron Facility at 100 kW Beam Power Saturated Residual Fields Inside the Steel Vacuum Tank from Cascade Neutron Reactions at 100 kW Beam Power Residual Activity Production Cross-Sections for High Energy Nucleons on Oxygen Induced Activity and Residual Fields in the Thermal Neutron Facility Coolant and Moderator Heat Distribution in the Thermal Neutron Facility at 100 kW Proton Beam Power Proton Beam and Uranium Target Parameters Shielding Costs for the Thermal Neutron Facility Thermal Neutron Facility Cost iv Page 13 15 18 19 22 27 28 30 32 35 36 38 43 46 54 l. (a) (b) (c) (d) 2. (a) (b) (c) (d) 3. (a) (b) 4. 5. LIST OF FIGURES Plan view of the thermal neutron facility (TNF) Lateral elevation section of the TNF Longitudinal elevation section of the TNF Partially cut-away isometric view of the TNF Split-section, plan view of the TNF core Lateral elevation section of the TNF core Longitudinal elevation section of the TNF core Partially cut-away isometric view of the TNF core Lateral section of the main target assembly Longitudinal section of the main target assembly Neutron yield from Pb and U as a function of incident proton energy Angular distribution of cascade neutrons from 500 MeV protons on Bi 6. (a) Evaporation neutron and heat production from 500 MeV protons on Bi (b) Radial distribution of heat from neutron-induced fission in natural uranium ]. Thermal neutron flux distribution for various moderator configurations 8. Thermal neutron flux as a function of moderator thickness for 020 and graphite 9. Thermal flux depression from substitution of graphite for 020 at the outside of the moderator assembly 10. Thermal flux dependence on H20 reflector thickness 11. Thermal neutron flux perturbations due to absorption 12. (a) Neutron energy group flux distributions in the TNF moderator assembly for a Pb-Bi target (b) Neutron energy group flux distributions in the TNF moderator assembly for a natural uranium target 13. Fast/thermal flux ratio for various moderator configurations 14. Neutron energy group flux/source factors for cascade induced source neutrons 15. Fast/thermal logarithmic flux gradient ratio in the TNF moderator assembly v 16. Gamma dose rate danger parameter for iron as a function of depth in the shield of the TNF core 17. Costs of spherical, two-component, iron-concrete shields as a function of outer shield radius 18. Optimized cylindrical shield costs as a function of proton sp i 11 rate vi 1 . INTRODUCTION 1.1 Basis for the Facility The . thermal neutron facility is based on the copious neutron source result-ing from high energy proton bombardment of thick, high atomic weight targets . The specific neutron source strength per 500 MeV proton incident on such targets is estimated to vary from 8 for lead or bismuth to ~32 for natural uranium. Thus for a 100 ~A proton beam current the total neutron source strength is in the range 5 x 101 5 sec-1 to 2 x 1016 sec-1, Approxi-mately 40% of the 32 neutrons produced in natural uranium come from the thermal neutron induced fission; thus the exact source strength is dependent on the detailed target and moderator assembly design. Most of the source neutrons are ••evaporat i on•• neutrons with an approximate 1 y Maxwellian kinetic energy distribution of mean energy ~3 MeV. These neutrons are moderated to thermal energies in low atomic weight materials surrounding the target source. The primary protons are stopped in the high atomic weight target. The shielding requirements for the facility are set by the average ~o.6 high energy 11 cascade11 neutrons per incident 500 MeV proton that emerge from a 6 in. diam, 11 in. long target; they have a strongly forward peaked angular distribution. 1.2 Purpose and Requirements The thermal neutron facility has four main functions: 1) to stop the residual proton beam downstream from the last meson production target in the main beam line; 2) to provide a facility for irradiating and activating specimens with high energy protons; 3) to provide a facility for irradiating and activating specimens with neutrons of all energies from thermal to the incident proton energy; and 4) to provide collimated neutron beams from the assembly core, primarily of thermal energies but also at epithermal and higher energies. - 2 -The first function is essential to the operation of the main TRIUMF beam line; thus it will be necessary to ensure a high reliability for all essen-tial components of the facility, and to make provision for easy replacement or substitution of components that require maintenance. The priorities given the other components of the facility will depend on the interests of the users and the requirements of the TRIUMF organization as a whole, and may evolve over the 1 ifetime of the facility. Thus the design of the original assembly should allow modifications to the core of the facility at reasonable cost, to reflect the probable changes of interest. No detailed consideration has been given to using an enriched, fast or thermal fission, sub-critical assembly to increase the neutron flux because of the substan-tial criticality and containment problems associated with such a device. - 3 -2. GENERAL DESCRIPTION The overall plan view of the facility is shown in Fig. la. The proton beam is transported from the meson target tunnel to the facility core through a 6ft wide trench. This trench, which allows access to the proton beam line from above after removal of the demountable shielding blocks, can be used for experimental arrangements which would involve small angle bending of the proton beam, probably on an intermittent basis. It will also provide access to the beam line and proton beam collimator and window, shown in Fig. 2c, for maintenance and modification. 2.1 Vacuum Tank The core of the neutron facility is contained in a thick-walled (~1-2 in.) steel tank, shown in an elevation section perpendicular to the proton beam direction in Fig. lb. The vacuum-tight vessel consists of two arms, one running vertically to the top of the shielding and the other at 30 deg to the horizontal to the swimming pool on the north side of the facility. In the region of the core and iron shield surrounding the core, the cross-sections of both legs of the vessel are rectangular. The vertical leg has inside dimensions 4'0'' along the beam line by 4'8" perpendicular to the beam line, to allow removal of the core moderator assembly. The vertical vessel section changes to circular of inside diameter 6'6" at 7ft above proton beam line and increases to 7'0" inside half-way along the circular leg, to suppress neutron streaming. The slanted leg of the vacuum vessel, going to the swimming pool, has an inside rectangular cross-section 4'0" along the beam direction and 3'0" perpendicular to the beam out to a distance 10 ft from the proton beam. At this point the cross-section changes to circular, of inside diameter 5'6",for the remainder of the distance to the swimming pool. Both the vertical and slanted legs will be filled with steel-jacketed shielding plugs with holes through them for access to the various irradiation and neutron beam source points. The vertical plug will be removable in two or more sections with the main crane. The plug for the slanted section can be fitted with compressed air lines and ports and external baffles on the bottom side to provide an air pad during insertion and removal. - 4 -The entire vessel will be evacuated when the facility is operated to elim-inate any radiolysis and noxious gas induced corrosion problems and to avoid airborne activity. The demountable seals for the system are all placed in easily accessible regions near the surface of the shielding where maintenance is convenient and radiation damage will not be a problem. All of the components inside the tank must be vacuum-tight but because the vacuum requirements are modest (~lo-2 Torr) fairly high outgassing and leak rates should be tolerable. The primary pumping point will be just under the proton beam 1 ine in the trench near the inlet to the facility core, as shown in Fig. lc. 2.2 Vacuum Window The vacuum in the proton beam line at the point of pumping will probably be adequate to satisfy vacuum requirements upstream from that point. Because this situation cannot be demonstrated rigorously for all operating circum-stances, and to provide back-up in the case of component failure, provision is made for a vacuum-tight window between the thermal neutron facility vacuum and the proton beam line vacuum at the edge of the facility core. The window, shown in the inset on Fig. 2c, is a simplified version of the type described by Penner. 1 The beam power density is limited to about 3.5 kW cm- 2 or less by heat dissipation in the neutron facility target. Thus the heat generated by the beam in each wall of the window is less than 1 W cm-2 and can be transferred from the thin window to the water-cooled jacket by natural convection in the 30 lb/in. 2 helium contained between the windows. Since cooling water does not come in contact with the windows, the corrosion limitations cited by Penner do not apply, and any of aluminum, titanium or stainless steel windows of thickness 35 mg cm- 2 can be used. The estimated maximum temperature difference between the windows and cooling jacket is ~200°C for the highest beam power considered, namely, 250 kW at 350 MeV. Basically, the separation between the proton beam line vacuum and the thermal neutron facility vacuum is at the wall of the trench adjacent to the facility core. However, to reduce residual activity in the trench, the vacuum window is placed at the end of a beam line stub which stops just - 5 -ahead of the main proton beam coli imator. This collimator, in a 12 in. diam port between the trench and the facility core, will normally absorb very little proton beam because the beam transport system downstream of the last meson production target will focus the beam to clear the collimator. Occasionally, at very high beam power, the limitation on peak beam power density may require expanding the beam and allowing some sig-nificant fraction (up to 25%) to be removed by the collimator. The collimator will have to be capable of dissipating the resulting heat. 2.3 Moderator Assembly The core of the facility consists of two basic parts, the moderator assem-bly suspended from the shielding plug in the vertical leg of the steel vacuum tank and the water-cooled neutron production target attached to the end of an 18 in. diam tube running through the slanted shielding plug to the swimming pool, designated ST-18 in Figs. 2b and 2d. The neutron pro-duction target is described in Sec. 2.8 below. The moderator assembly is basically a 4 1 611 diam, 3 1 1011 long cylindrical tank whose axis is coincident with the incident proton beam,as shown in Figs. 2a-2d. The out-side tank contains an inner tank and several re-entrant tubes and ports to provide access to the neutron beam source points and irradiation facilities. In particular, there is a 4 in. diam tube along the tank axis, for the proton beam, intersecting at the centre of the assembly a 6 in. wide by 14 in. long rectangular tube along a diameter, which takes the main target assembly. The inner tank is basically 46 in. diam by 38 in. long and contains standard reactor grade heavy water (0.2% H20) and graphite blocks in the outer 8 in. This graphite reduces the 020 inventory requirements at the cost of a small decrease in thermal flux. Heat will be removed from the 020 and graphite by natural convection and conduction through the inner tank wall to the H20 in the ~4 in. space between the inner and outer moderator tanks. The H20 has the combined functions of moderator assembly cooling and neutron reflection. The outer tank is cut off on the bottom side at a radius of ~19 in., as compared to the basic radius of 27 in.; the inner tank is cut off at 15 in. and the graphite layer is left out on this side because this volume is inaccessible for experimental purposes. This modification to the basic geometry does not affect the flux level - 6 -significantly at other points in the assembly. The target assembly is not centred longitudinally in the moderator tank assembly, as can be seen in Fig. 2c. The 14 in. long target space is 22 in. from the front (up-beam side) of moderator assembly and 10 in. from the back (down-beam side) of the moderator assembly. Again, the graphite has been left out of the assembly on the down-beam side, because the forward-directed, high energy cascade neutron component from the target would give poor signal-to-noise conditions for most thermal neutron experiments in this region. 2.4 High Energy Neutron Facilities For high energy neutron irradiations a 6 in. diam tube, shown in Fig. 2b as ST-6-C, intersects the proton beam centreline immediately outside the steel vacuum tank. This tube is extended for ~9 ft beyond the proton beam centre-line to allow mounting of experiments requiring low background conditions in a detector viewing the irradiation facility space through tight collima-tion. The section of 4 in. diam tube in the moderator tank between the target assembly and the outside of the tank can be flooded with heavy water when the cascade neutron facilities are not used. A horizontal, 2 in. diam tube, HT-2, extends from one side of the bulk shield to the other and intersects ST-6-C irradiation tube at the proton beam centreline. This tube can be used for high energy irradiation experiments requiring short irradiation and transit times. The 6 in. diam, horizontal beam tube, HB-C, from the facility core to the 6 ft wide down-beam trench is co-linear with the extended proton beam centreline. This beam tube can be used for beam experiments with high energy (cascade) neutrons mounted in the trench below the fairly massive removable shielding which they will require. 2.5 Proton Irradiation Facilities Two alternative sites for irradiation of specimens by the proton beam will be available in the facility. The primary site is immediately ahead of the main neutron production target, as shown in Fig. 3. The irradiation facility with its separate cooling circuit will be an integral part of the target assembly. The specimen volume will be 2 in. wide along the - 7 -direction of the proton beam and cover the whole 4 in. diameter of the incident proton beam. The detailed design of the proton irradiation target will have to satisfy the heat dissipation problem for each individual target material. Thus, because thermal conductivity varies widely, separate designs for each target assembly may be required. The proton irradiation targets will be inserted from the swimming pool through a water-lock into the ST-18-P and driven into position by the cool-ing water flow. They will be removed by reversing the cooling water flow or, if the transient conditions for reversing the flow are unacceptable, by means of a small retractable cable. When the proton irradiation facility is not in use, a hollow evacuated chamber will be inserted in the beam to minimize the beam power dissipation in the coolant and to direct the flow on the chamber walls for adequate heat removal. The alternative proton irradiation site is between the proton beam collimator in beam pipe HB-P and the facility core. Access from the swimming pool is through the tube ST-6-P shown in Fig. 2b. This position can be used where the thermal neutron flux at the main target may produce unwanted activity in the irradiation specimen. Other proton irradiation facilities for special purposes could be provided in the 6 ft wide trench along the beam 1 ine ahead of the thermal neutron target core. 2.6 Thermal Neutron Beam Facilities The six horizontal beam tubes, HB-1 through HB-6 on Figs. 1 and 2, extend from the facility core to the experimental areas outside the heavy concrete shielding and are intended primarily for thermal neutron beam experiments. All tubes are rectangular in cross-section (2 in. high by 8 in. wide, inside dimensions) from the vacuum tank wall through the 4ft thick iron shielding surrounding the vacuum tank. Outside the iron shielding the beam tube cross-section is circular at 12 in. inside diameter with the centreline of the 12 in. diam section offset vertically upwards by 2 in. from the centreline of the 2 in. x 8 in. section. The 12 in. diam section will be used for collimators to define the neutron - 8 -beam geometry and may be used for double monochromating of the neutron beam where very good signal-to-background ratios are required . The fixed shielding shown in Fig. 1 will give operating radiation fields through the shield in the vicinity of the neutron beam ports at 60 deg to incident proton beam direction of the order of 40 mrem h- 1 for a total beam power of 100 kW. The fields near the 120 deg ports will be approximately 1/2 this value. Supplementary shielding required around spectrometer apparatus at the end of the beam ports will reduce the field in the inhabited portions of the experimental area by one to two orders of magnitude. The four beam tubes, HB-3 to HB-6, at 60 deg and 120 deg to the proton beam direction 11see 11 into the 020 moderating assembly to a point 6 in. to 8 in. below the centreline of the main neutron production target, as shown on Figs. 2b and 2c. To reduce the cascade neutron flux, and fast neutron contamination generally in the beams, diametrically-opposite tubes view each other through 4 in. of 020 moderator and the re-entrant channel walls. The two beam tubes, HB-1 and HB-2, at 135 deg to the incident proton beam direction view the 020 moderator at a face even with the front face of the main target but offset 10 in. to 12 in. below the target centreline. The beam intensity in the 135 deg tubes will be approximately 65% that in the other tubes because of the additional offset, but the fast/thermal flux ratio will be decreased from ~o.4 to ~0.2. One other beam tube, SB-1, at 45 deg to the horizontal and 90 deg to the incident beam direction is also provided, as shown in Figs. 1 and 2. It will also be of 2 in. by 8 in. cross-section increasing to 12 in. diam outside the iron shield and wi 11 view the 020 moderator source at an offset to the target centreline of 8 in. to 10 in. It is intended for use where an extended neutron flight path may be required. 2.7 Neutron Irradiation Facilities Several re-entrant tubes provide access from the swimming pool to the core of the thermal neutron facility through the shielding plug in the slanted (30 deg) leg of the vacuum tank. These are intended primarily as irradia-tion facilities although they could be used for some purposes as beam tubes. Three re-entrant tubes to the facility core from the top vacuum - 9 -seal through the shielding plug in the vertical tank leg can be used either as irradiation positions or as vertical beam tubes. The vertical tubes in the assembly core can be seen in Figs. 2b, 2c and 2d. The vertical 2 in. diam tube, VT-2, passes through the moderator assembly and allows access to the outside of the main neutron production target. This source point will have the highest available fast and epithermal neutron flux in the assembly. The thermal flux will also be near the maxi-mum value at this point but could be suppressed in the emerging beam by a thin cadmium window. The 4 in. diam vertical tube, VT-4, into the 020 moderator will provide a beam or irradiation facility at a thermal flux level approximately two-thirds that at the target surface but with a fast-to-thermal flux ratio reduced from ~1.0 to 0. 15. VT-4 is upstream of the front face of the main target, thus avoiding a high cascade neutron flux. The 8 in. diam tube, VT-8, into the graphite assembly has a thermal neutron flux reduced by a factor ~3 compared to the 4 in. tube in the 020 region; the fast-to-thermal flux ratio is ~0.06. Six re-entrant tubes through the shielding plug in the slanted, 30 deg leg of the vacuum tank provide access from the swimming pool to the facility core. Two of these, the cascade neutron irradiation facility, ST-6-C, and the alternative proton irradiation facility, ST-6-P, have been described previously. The 18 in. diam tube, ST-18, carrying the main neutron pro-duction target and primary proton irradiation assembly ends at the edge of the 020 moderator; it also carries three re-entrant irradiation tubes, ST-18-2, -3 and -4' extending into the 020 moderator region from ST-18. They can be seen in Fig. 2 as 2 in., 3 in. and 4 in. diam tubes and wi 11 give thermal neutron fluxes near or at the maximum in the assembly, but with high fast-to-thermal flux ratios. When they are not in use they will be filled with graphite plugs to exclude all water except that needed for natural convection cooling. The 14 in. diam tube, ST-14, also stops at the edge of the 020 region but carries a 4 in. diam tube ST-14-4 extending into the 020 moderator. Both this tube and the 8 in. diam one, ST-8-B, extending into the graphite moderator at the up-beam side of the assembly, will have the lowest cascade neutron flux levels in the facility. A - 10 -second 8 in. diam tube into the graphite- ST-8-F on Fig. 2- will provide a facility for irradiations where cascade neutron contamination is not significant. Finally, a 2 in. diam tube, HT-1, passes through the 020 moderator 5 in. below the proton beam centreline and 5 in. ahead of the neutron production target front face. This channel will be fitted with an internal tube extending from one side of the shield to the other, with vacuum connections at the outside of the shield. The through tube must be removed before the moderator assembly is moved but need not be disturbed for withdrawal of any of the slanted tubes, including ST-18. The through tube is intended primarily for short-irradiation, short-transit-time experiments but could alternatively be used for tightly-collimated beam experiments. 2.8 Neutron Production Target Assembly The main target assembly, shown in Fig. 3, will be the most critical element in the thermal neutron facility because of the high heat dissipa-tion in the target and possible radiation damage in the assembly structure. The target head is connected to the end of the access tube ST-18 by a bolted all-metal seal. To change or service the target head, tube ST-18 is pulled back into the swimming pool after either lowering the water level in the pool below the inlet or installing a temporary water-lock and gate valve, depending on the radiation field problems. The target head can then be removed from ST-18 under water in the pool for servicing there or in the hot cells. The material for the main proton beam target has not yet been chosen. The choice is between the Pb-Bi eutectic or natural uranium. In both cases heat dissipation, while not appearing to present any insurmountable prob-lems, will require further study to establish a design that is sufficiently reliable. The Pb-Bi target would be operated in the molten state, relying on natural convection to transfer the heat deposited - about 3/4 of the incident beam power - to the outside of the target container. The container, basically a 6 in. diam, 11 in. long cylinder, will be cooled by light water with no cooling channels in the primary proton beam. - 11 -The alternative target material, natural uranium, has a much higher neutron production resulting from the primary proton bombardment, approximately 19 n/p as compared to 8 for Pb-Bi, both for 500 MeV protons, but also pro-duces an additional ~13 neutrons per incident proton from thermal neutron fission. Thus the total source strength is four times that for the Pb-Bi target. The absorption of thermal neutrons in the uranium reduces the factor by which the peak flux is increased from 4 to approximately 2.5; however, the region of peak flux is moved away from the target to a region of the moderator accessible to beam tubes and irradiation facilities. The fast-to-thermal flux ratios are comparable for both targets. The main disadvantage with the natural uranium target is the high heat production - approximately 6 times that for Pb-Bi for a given amount of proton beam power. This requires internal cooling of the target with a resulting increase in cooling water activity and radiolytic gas production. The total activity production in the target is also increased, by approxi-mately the factor of the neutron source strengths. The distribution of activity according to species is also changed substantially, but it is not clear that the hazards from fission products are significantly different from those due to the 21 0Po in the Pb-Bi target. The natural uranium target will either be in the form of a stack of 6 in. diam, 0.2 in. thick, 0.040 in. aluminum-clad uranium metal plates with 0.08 in. cooling water spaces between them or 0.3 in. diam, 0.020 in. aluminum-clad uranium metal spheres close-packed in a 6 in. diam, 11 in. long aluminum container. Both arrangements have a mean density similar to that of the Pb-Bi target. The fabrication of the aluminum-clad sphere for the close-packed target is not as well established as that for the plates, but the average heat flux across a metal-water interface is reduced by a factor 1.5 for the spheres as compared to the plates. Either design would be adequate for proton beam power of 100 kW or less. For higher beam power it would be necessary to decrease the plate thickness or increase the proton beam size, thus losing some beam power in the collimator ahead of the thermal neutron facility core. The cooling ·of all targets will be by light water in a closed circuit with cooling lines up the 18 in. diam target access tube, ST-18, to the swimming - 12 -pool and hence to an intermediate heat exchanger in the adjacent pump room. The volume in the target assembly not required for structure, cooling tubes, etc. will be filled with graphite or beryllium oxide to reduce thermal neutron absorption. 2.9 Tabulation of Irradiation and Beam Tube Facilities The neutron irradiation and beam tube facilities that have been incorporated into the design are shown in Table I. The approximate distance from the centre of the target, the inside dimensions of the irradiation or beam tube site and an estimate of the thermal, fast and cascade neutron flux, as dis-cussed in the next section, are listed. Most of the irradiation facilities have been incorporated in the design in response to specific requests or fairly clearly foreseen requirements. Specific requests have been made for only four or five of the neutron beam tubes. Additional beam facilities have been incorporated to cover a number of contingencies, such as uncertainties in the effective cascade neutron component as a function of angle between the incident proton beam and the thermal neutron beam. The slanted beam tube is intended for any long flight path requirements that might arise. The cost of installing beam tubes in the fixed shielding during initial construction is relatively trivial; it would be nearly impossible at a later stage. - 13 -TABLE Beam Tubes and Irradiation Positions in the Thermal Neutron Facility Approx1 Tube or Flux for 500 MeV, 100 11A Beam4 Facility Radial Thimble Size 10 12 cm- 2 sec- 1 Component Distance ins ide dimension Thermal Fast2 Cascade 3 em em Beam Tubes: HB-1, -2 29 5 X 20 6.0 1.0 0.007 (16.0) (3.0) HB-3, -4 20 5 X 20 9.0 3.5 0.003 HB-5, -6 0.006 SB-1 23 5 X 20 8.0 2.2 0.011 Hori zonta 1 (22. 0) (7.0) Thru-Tubes: HT-1 23 5 (d i am) 7.5 2.2 0.007 (21 .0) (7.0) HT-2 56 5 (d i am) 0-0.2 0.02 0.03 (0.07) Vertical Access Tubes: VT-2 12 5 (d i am) 10.0 10.0 0.24 ( 15. 0) (40.0) VT-4 29 10 (d i am) 6.5 1.0 0.007 (20.0) (3. 5) VT-8 54 20 (d i am) 2.0 0.12 0.003 Slanted (6.0) (0.4) Access Tubes: ST-18-2 14 5 (d i am) 10.0 7.0 0.15 (25. 0) (20.0) ST-18-3 18 7.5 (d i am) 10.0 4.5 0.04 (25. 0) (11.0) ST-18-4 21 10 (d i am) 9.0 3.0 0.14 (20.0) (9 .0) ST-14 51 35 (d i am) 2.0 0.15 0.002 (7.0) (0. 5) ST-14-4 36 10 (d i am) 5.0 0.6 0.004 ( 15. 0) (2. 0) ST-8-B 58 20 (d i am) 2.0 0.12 0.002 (5 .0) (0.5) ST-8-F 53 20 (d i am) 2.0 0.15 0.016 (5. 0) (0.6) ST-6-C 56 15 (d i am) 0-0.2 0.02 0.03 (0. 07) 1) From centre of the neutron production target (~8 em down-beam from front face). 2) Fast flux quoted here is estimated from evaporation neutron component still having energies greater than 1.46 eV. 3) Cascade flux is at energies ~10 MeV. 4) Unbracketed values apply to the Pb-Bi target or both the Pb-Bi and natural-U targets; bracketed values apply to natural-U target. - 14 -3. FACILITY CHARACTERISTICS AND SPECIFICATIONS 3. 1 Prot on Beam The emittance of the proton beam at the irradiation facility will depend on the configuration of meson production targets between the cyclotron and the facility, which may change from time to time. The maximum angular divergence of the proton beam at the target introduces no significant per-turbation on any characteristic of the proton or neutron irradiation facilities. The lateral distribution of the beam at various points along the beam line, and in particular at the target, is important, however. The two controlling considerations are the maximum tolerable specific heat deposition rate in the targets and container walls, and the spill from the wings of the distribution into the collimator in the proton beam port HB-P. The first requires a wide lateral distribution and the second a restricted lateral distribution. Both requirements should to some extent be satis-fied simultaneously by the emittance characteristic expected for the beam emerging from the collimator following the last meson production target, namely a Gaussian distribution with wings cut off sharply. The effects of collimator scattering on the lateral distribution at the target have, however, not been estimated in detail, and the beam characteristics have been estimated on the basis of a full Gaussian distribution. The proton · beam transport system should be capable of focusing the beam to a lateral distribution at the target, both horizontal and vertical, of standard deviation cr, in the range 1.0 em to 5.0 em for all possible meson target configurations and proton energies. The design of the beam transport system should be such that component failure any place in the system will not lead to a smaller lateral distribution at the target. 3.2 Neutron Source The effective neutron source strength when high energy protons are incident on cylindrical lead and uranium targets has been measured experimentally. 2 Fig. 4 shows the neutron yield as a function of proton energy for lead at two different target diameters and uranium at a single diameter from the lNG Proposa1. 3 Table II shows the results of Monte Carlo calculations 4 of evaporation and cascade neutron production from 10 and 20 em diam bismuth targets. The parameters in the calculations have been adjusted to give agreement with the experimental data cited above. - 15 -TABLE II Neutron and Heat Production Results from Monte Carlo Calculations for 500 MeV Protons on 10 and 20 em diam Bi Targets* Results for Results for Quantity Estimated 10 em diam 20 em diam Number evaporation neutrons 7.6 8.4 Escaping evaporation neutron energy, MeV 30.0 28.8 Inelastic y-ray production in target, MeV 5.3 9.0 Number cascade neutrons 0 . 66 0.50 Escaping cascade neutron energy, MeV 47.5 37.8 Heat production in target, MeV 370 376 *The calculations were done 4 for 40 em long targets of p = 9.75 g cm-3. The results should not change signifi-cantly for a 27.5 em long target of p = 10.25 g cm- 3 The results are normalized to one proton incident on the target. - 16 -Fig. 5 shows the average angular distribution of the cascade (En > 10 MeV) neutron component leaking from the 10 and 20 em bismuth targets as esti-mated by the Monte Carlo calculations. Both the particle current and the kinetic energy current are shown. The cascade neutrons will suffer further non-elastic collisions in the moderator, structure and shielding, increasing the total neutron production over that shown in Table I I. These neutrons are important in estimating the fast (1 MeV < En < 10 MeV) neutron flux distributions at various points in the assembly. Fig. 6a shows the spatial source distribution of the evaporation neutrons and heat along the length of the bismuth target. The radial distribution is essentially that of the incident proton beam. The neutron source strength and heat production are both increased substan-tially for a ~atural uranium target. Interpolation of the experimental data in Table VI I.A.3 of lNG Proposal 3 gives a value for the effective neutron source strength of 19.6 neutrons per 500 MeV proton incident on a solid, 10 em diam, depleted uranium target.* Fig. VI 1.3 of the lNG Proposal shows an estimated total heat production of ~65 MeV/neutron when 500 MeV protons are incident on uranium. Thus, allowing for some dilution by low mass elements in an actual target, the total heat production due to the inter-nuclear cascade collisions is ~1150 MeV per 500 MeV proton incident on uranium. The spatial source distributions of these neutrons and heat are nearly the same as those for bismuth in Fig. 6a, after scaling for the difference in density. A natural uranium target has a second major component of source neutrons and heat arising from thermal and epithermal neutron induced fission. The effective multiplication constant k for this sub-critical assembly is dependent on the detailed target and moderator assembly geometry. Based on the neutron diffusion calculations described below the k value for a 40 kg natural uranium target surrounded by a D20-graphite moderator, with *The residual 23Su content of the target used to obtain the data in the lNG Proposal is not reported nor is it stated what corrections, if any, were made for the thermal and epithermal neutron induced fission source. The errors involved in ignoring these possible effects are probably comparable with the overall accuracy of the data, ~10%. - 17 -aluminum structure, is approximately 0.4. Thus the total neutron source per incident 500 MeV proton in the assembly is N=N 0 1 1-k 1 = 19 1-0.4 = 32 where N = 19 is the external source, in this case from the inter-nuclear 0 cascade. The ~13 extra neutrons come from ~5.3 low energy neutron induced fission reactions, which deposit an additional ~1050 MeV in the target as heat. The spatial distribution of these sources is essentially that of the neutron capture. Fig. 6b shows the fission heat source distribution in a 9 em radius sphere of 70% natural uranium, 15% Al and 15% H20, normalized to the cascade source strength resulting from one 500 MeV proton incident on the target. The distribution for the actual cylindrical target will be very similar. The peaking in the thermal neutron induced component at the outside of the target complements the centrally-peaked source from the inter-nuclear cascade. 3.3 Neutron Flux Distribution The neutron flux distribution for various structure, coolant and moderator configurations has been estimated using a four-energy-group, one-spatial-dimension,neutron diffusion computer code called FORGOD. 5 All of the calculations reported here were estimated on the basis of spherical geometry, with the complete angular symmetry demanded by the code. The spherical target radius was usually taken to be 10 em for a target volume approximately equivalent to that of a 14 em diam by 27 em long cylinder. The neutron energy groups used are those shown in Table I I I as well as the neutron source strength in the energy groups, based on estimates made by S.A. Kushneriuk.6 The estimate includes an allowance for energy degrada-tion of the evaporation neutrons by inelastic scattering before they emerge from the Pb-Bi target. The macroscopic neutron absorption La' energy group removal Lr and transport Lt cross-sections required by the code are shown in Table IV. They were taken from various sources. 7- 9 Energy Group 1 2 3 4 - 18 -TABLE II I Energy Groups and Source Strengths for FORGOD Calculations Evaporation Neutron Source in Target for one 500 MeV Energy Range Proton Incident Pb-Bi Nat-U E > 1 MeV (fast) 5.7 15.1 1 MeV> E > 1.46 eV (intermediate) 2.3 3.8 1.46 eV > E > 0.127 eV (epithermal) 0 0 E<0.127eV (the rma 1) 0 0 - 19 -TABLE IV Neutron Diffusion Parameters for FORGOD Calculations Evaporation Slowing Neutron Down Cascade Multiplicity Absorption Removal Transport Neutron Cascade Energy Cross- Cross- Cross- Mean Free Neutron Material Group Section Section Section Path Collisions Ea E1' Et A. a M am-1 am- 1 am-1 am D20 1 0.0 0.5700E-Ol o. 1540 120.0 1.6 2 0.0 0.1320E-Ol 0.2530 120.0 0.0 3 O.IOOOE-04 0.7560E-Ol 0.2760 120.0 0.0 4 o.4oooE-o4 0.0 0.4060 120.0 0.0 H20 1 0.0 0.7800E-Ol 0.1120 125.0 1.6 2 0.0 0.6100E-Ol 0. 3770 125.0 0.0 3 0.4400E-02 0.6050 0.6400 125.0 0.0 4 0.1900E-Ol 0.0 2.080 125.0 0.0 Graphite 1 0.0 0.1360E-Ol 0.1310 76.0 1.6 2 0.0 0.4200E-02 0.3360 76.0 0.0 3 0.6000E-04 0.2490E-Ol 0.3620 76.0 0.0 4 0.2400E-03 0.0 0.3700 76.0 0.0 Pb 1 0.0 0.0 0.2000 17.5 5.0 2 0.0 0.0 0.3300 17.5 0.0 3 0.1400E-02 0.0 0.3600 17.5 0.0 4 0.5000E-02 0.0 0.2600 17.5 0.0 Bi 1 0.0 0.0 0.1700 20. 1 5.0 2 0.0 0.0 0.2600 20.1 0.0 3 0.2500E-03 0.0 0.2600 20 .I 0.0 4 o.8500E-03 0.0 0.2500 20 .I 0 .0 Fe 1 0.0 0.5100E-Ol 0.2540 19.2 2.8 2 o.46ooE-02 0.1600E-02 0.6900 19.2 0.0 3 0.5300E-Ol 0.1230E-Ol 0.8600 19.2 0.0 4 0.2300 0.0 0.8600 19.2 0.0 Zr 1 0.0 0.1300E-Ol 0.2000 25.0 3.5 2 0.5000E-03 0.5000E-03 0.3000 25.0 0.0 3 0.2000E-02 0.2400E-02 0.2700 25.0 0.0 4 0.7000E-02 0.0 0.2700 25.0 0.0 AI 1 0.0 0.1930E-Ol 0.1500 47.0 1.7 2 0.4000E-03 0.6000E-03 0.1050 47.0 0.0 3 0.2900E-02 0.2500E-02 0. 8400E-Ol 47.0 0.0 4 0.1250E-01 0.0 0 .9600E-01 47.0 0.0 Concrete 1 0.0 0.5000E-Ol 0.2200 56.0 1.6 2 0.0 0.8000E-02 0.2400 56.0 0.0 3 0.3200E-02 0.6700E-Ol 0.2700 56.0 0.0 4 O.llOOE-01 0.0 0.5300 56 . 0 0.0 Natural 1 0.4800E-02 0.5000E-Ol 0.2900 10.6 15.0 Uranium* 2 0.4000E-Ol 0.8000E-03 0.5000 10.6 0.0 3 0.9400E-Ol 0. 1600E-02 0.4800 10.6 0.0 4 0.3200 0.0 0.5300 10.6 0.0 *An additional parameter vEf., the product of the fission cross-section and the number of neutrons ~reduced per fission, is required for systems contain-ing natura11)uranium. T ;} values used for the energy groups are 1) 0.036 , 2) 0.017, 3 0.123 and 4 0.421. - 20 -Table IV also defines the evaporation neutron source in the various materials resulting from non-elastic collisions of the cascade neutrons emerging from the target. The cascade flux was assumed to have a combined exponential removal and l/r2 geometric dependence. The exponen-tial relaxation lengths Ac are those estimated previously for plane slab geometry, 10 and the evaporation neutron multiplicities M were derived from Monte Carlo intra-nuclear cascade results.ll,12 Fig. 7 shows the thermal neutron flux distribution for various moderator assemblies. In all cases the inner 10 em radius is the Pb-Bi target followed by various layers of moderator without any structural material. As estimated previously, 10 a large all-graphite assembly gives approximate-ly one-half the flux of a large heavy-water system. A light-water moderator system (not shown in Fig. 7), while giving a peak thermal neutron flux comparable to that of the graphite system, has a very small useful volume and very steep flux gradients. The smaller- 60 em outside radius-heavy-water assembly has a peak thermal flux of ~80% of the larger assembly; it falls off more quickly with increasing radius, falling below that of the large graphite assembly at r ~ 47 em. Replacing the 020 with graphite in the region r = 40 em to r = 60 em makes 1 ittle difference to the thermal flux in the region inside r = 30 em. Fig. 8 shows the dependence of thermal neutron flux on the outer radii of 020 and graphite moderator assemblies at two field points in the moderators. The values for a 60 em 020 moderator are approximately 70-80% of those for an infinite 020 system. The thermal flux in the graphite moderator, even in a large assembly, is less than that for the 020 moderator of ~so em outside radius for the field points shown. Fig. 9 shows the reduction in thermal flux at the two field points when the outside layer of a 60 em 020 moderator is replaced by graphite. The reduction is shown as a function of graphite layer thickness. Both assemblies have a 10 em thick H20 reflector outside the moderator. From Figs. 8 and 9 we see that a thermal neutron flux of ~75% of that in a large 020 system can be achieved in a 020-graphite-H20 assembly of 40 em, 60 em and 70 em outside radii, respectively. This has been chosen as the reference design for the thermal neutron facility moderator system. - 21 -Fig. 10 shows the dependence of the thermal neutron flux at three field points in the moderator on the H20 reflector thickness. The reflectivity of H20 is essentially saturated at a layer thickness of 5 em. The fraction of thermal neutrons leaking through the reflector, however, drops from ~40% at 5 em to ~10% at 10 em, both estimated for a Pb-Bi target. Thus if thermal neutron activation outside the reflector is an important considera-tion,a thick reflector may be required. Fig. 11 shows the effect of the basic structural components on the thermal flux distribution for the 30 em 020, 20 em graphite, 10 em H20 moderator system chosen for the reference design. The Zr shell represents a zirconium alloy target assembly structure and the H20 the target coolant. The aluminum moderator tank thicknesses do not allow for any internal structure such as bracing walls and beam port walls; the 2% (by volume) allowance indicates the flux depression effect from the additional struc-ture. None of the calculations estimates the flux depression due to the increased leakage from beam ports. In view of the fairly large leakage load on the proposed assembly, estimates of these flux depression effects will be required using more elaborate calculational techniques and, preferably, experimental measurements. Table V shows the percentage flux perturbation caused by the various com-ponents at three different radii in the reference design assembly. The energy groups are the same as those given in Table I I I. Other estimates (not shown) indicate that the flux depressions shown for 2% aluminum in the 020 moderator region would not be significantly increased by extending this structure to the graphite region. The substitution of a zirconium alloy for aluminum on an equal volume basis reduces the thermal flux perturbation by ~30%. Stainless steel structure produces thermal flux depressions greater than those shown in Table V for aluminum or zirconium by more than an order of magnitude. Fig. 12a shows the spatial flux dependence of the four neutron energy groups for the reference assembly with the basic structure and a Pb-Bi target. The neutron source consisted of the evaporation neutrons in the Pb-Bi target, as shown in Table I I I, and a cascade component produced by 1 ) 2) 3) At r = Target Assembly Moderator Tank 2% Al in At r = Target Assembly Moderator Tank 2% Al in - 22 -TABLE V Flux Perturbations due to Structural Components in the Reference Design Assembly Spherically Symmetric Assembly Equivalent to Reference Design !:1<111,% (fast) 17.5 em 0.9 em Zr +8.0 0.4 em H20 -1.5 0.3 em Al inside 020 +2.1 0.5 em Al outside graphite 0 020 region +1.0 35.5 em 0.9 em Zr +4.7 0.4 em H20 -1.1 0.3 em Al inside 020 +1.6 0.5 em Al outside graphite +0. 1 020 region +3.0 [ 10 em Pb-Bi Target 30 em 020 Moderator 20 em Graphite Moderator 10 em H20 Reflector 1':1<112, % M>::l' % (intermediate (epithermal) +2.8 +0.6 -4.5 -5.5 +0.8 +0.2 0 0 +0.4 0 +3.9 +3. 7 -3.4 -3.6 +1.1 +1.1 +0.1 0 +2.8 +2.8 At r = 50.5 em Target 0.9 em Zr 0 +3.3 +3.6 Assembly 0.4 em H2 0 -0.4 -2.6 -3 0 1 Mcxlerator 0.3 em Al Tank inside 020 +0.3 +1.0 +1.0 0.5 em Al outside graphite +1.7 +0.5 +0.3 2% Al in 020 region +1.1 +3.3 +3.3 l:l<ll'+' % ( therma 1) -50 1 -1.5 -2.6 -1.1 -8.7 -2.2 -2.3 -1.2 -2.2 -9.5 -1.5 -2.4 -1.0 -4.8 -8.4 - 23 -non-elastic collisions of the average 0.6 cascade neutrons per incident proton emerging from the target. These cascade neutrons, which produce evaporation energy neutrons in the moderator and shield with a relaxation length and multiplicity given in Table IV, have a very anisotropic distri-bution, as shown in Fig. 5. Thus the turn up in the fast and intermediate flux components in the reflector region near the shield wall will be dependent on the position within the assembly. The average cascade neutron current occurs at an angle of ~60 deg to the incident beam direction. Fig. 12b shows the flux distributions for the various neutron energy groups for a natural uranium target. The fluxes for all energy groups in the assembly are generally a factor of ~2.5 higher for the natural uranium target than for the Pb-Bi target, except for the thermal neutron flux in the uranium target. The thermal neutron flux depression due to capture in the uranium target does not extend very far into the moderator, however. Fig. 13 shows the ratio of the fast and intermediate group flux to the thermal flux as a function of r for the same moderator assemblies as in Fig. ]. This ratio approaches that of the large 020 moderator assembly in the 020 region of the reference assembly; even at the centre of the graphite region the spectrum in the reference assembly is softer by a factor 3 than for the all-graphite moderator. Fig. 13 is estimated on the basis of an angular integrated source of 0.7 cascade neutrons per 500 MeV proton incident on the target. To give an indication of the importance of this source on the lower energy flux groups, Fig. 14 shows the spatial dependence of the coefficients ai(r) when the group fluxes are written in the form where S is the total (integrated over 4TI) cascade source strength from a Pb-Bi target and ~i(r) is the energy group flux dependence for S = 0. At aS= 1, for example, the flux in the energy group arises equally from the target evaporation source and the cascade induced source. Because the estimates are made on the basis of isotropic angular distributions, the results can only be interpreted as outside limits on the energy-group fluxes for the actual anisotropic case. - 24 -For scattering experiments with thermal neutron beams the background noise level in the detectors is usually limited by the epithermal and fast neutron contamination in the beam. The question arises as to the optimum source point for the experimental beam tubes when both the thermal flux and fast-to-thermal flux ratio are functions of r, the distance from the centre of the assembly. The problem can be defined by choosing the signal-to-noise ratio M of the experiment to be performed. The condition to be satisfied is then (see Appendix A) 1; R-n F{r) d dr R-n T(r) = M + 4 2 where F(r) and T(r) are the fast and thermal flux at r, respectively. Fig. 15 shows this ratio for the reference assembly and Pb-Bi target with the basic structure and the average cascade source included in the estimate. The fast flux is taken as the sum of groups 1 and 2, i.e. all neutrons above 1.46 eV, as they are the most important background-producing component. For neutron elastic scattering experiments where large values of M (~10) usually apply, the optimum operation is at small r; for experi-ments at nearly equal signal and background, the optimum operating point is at r ~ JO am. The optimum working point for all experiments on this criterion is insider~ J? em, all based on a Pb-Bi target. The results are nearly the same for the natural uranium target. Four of the experimental beam tubes, HB-3 to HB-6 inclusive, intended primarily for diffraction experiments, at 60 deg and 120 deg to the incident proton beam direction view the 020 moderator at 15 to 21 em from the assembly centre. The two neutron beam tubes, HB-1 and HB-2, at 135 deg to the incident proton beam direction view a source at 26 to 32 em from the centre of the target, and thus are optimized for fairly low signal-to-background ratio experiments. The slanted neutron beam tube ST-1 views a source at 20 to 26 em from the assembly centreline and has an intermediate fast-to-thermal ratio. The cylindrical symmetry of the moderator assembly about the beam line makes it desirable to adopt horizontal axes for diffraction experiments. - 25 -It is anticipated that most spectrometers will use counters which move in a vertical plane about an axis perpendicular to a beam tube. The problems associated with this somewhat unconventional approach have been studied and found not to be serious, while at the same time some advantages have been noted. - 26 -4. SHIELDING, ACTIVATION AND HEAT PRODUCTION 4.1 Operating Shield Requirements The shielding requirements for the thermal neutron facility are set by the very high energy (cascade) neutrons resulting from proton collisions in the target. Because these neutrons have a very anisotropic angular distribu-tion, as seen in Fig. 5, the shielding requirements are significantly different at various angles from the incident beam direction. As well as the variation in particle angular distribution the average kinetic energy carried by the cascade neutrons is also peaked in the forward direction due to the correlation in outgoing angle and energy. The shield require-ments are essentially proportional to the energy carried by the cascade particles rather than their number. No detailed data are available for cascade neutrons from uranium; they are expected to be similar in angular distribution and spectra to those from bismuth. The effective relaxation length of the cascade neutron energy is also dependent on the angle relative to the incident beam direction because of the differences in spectra. The numbers in the box in Table VI show the estimated relaxation lengths when protons are incident on a plane slab of aluminum at various directions to the plane normal . 10 This variation in relaxation length was extrapolated linearly in cose to the backward angles and applied proportionally to the forward relaxation lengths of iron and heavy concrete to make the shielding estimates for the thermal neutron facility. Dose rate estimates were based on the cascade neutron kinetic energy emerging from the target at various angles, as shown in Fig. 5, the data in Table VI and the estimate that lo- 11 W cm- 2 corresponds to 2.5 mrem h-1.10 Table VI I shows the estimated dose rates at various points around the thermal neutron facility. They are estimated for a shielding assembly composed of an iron cube 13ft to a side, with its centre displaced down-beam 6 in. from the centre of the target. This displacement plus the asymmetry of the moderator assembly gives 5'2 11 of iron in the forward direction, 41 011 of iron in the lateral direction, and 3'611 of iron in the backward direction, with proportionally more iron in the corner directions. The remainder of the distance from the outside of the iron to the field - 27 -TABLE VI Cascade Neutron Leakage Power and Relaxation Lengths Cascade Angle to Neutron Incident Kinetic ;\AZ. Ape ;\HC Beam Energy deg MeV/str•p g cm- 2 g cm- 2 g cm- 2 0 15.5 127 151 140 30 11.8 125 149 138 60 5. 2 121 144 133 90 1.7 117 139 129 120 0.55 114 135 125 150 0.24 111 132 122 180 0.18 109 130 120 - 28 -TABLE VII Operating Dose Rates and Shield Estimates for Thermal Neutron Facility at 100 kW Beam Power, 500 MeV Angle from Total Heavy Proton Beam Distance Iron Concrete Fie 1 d Point)~ Direction from Target Attenuation Attenuation deg ft Trench d.b. 0 10 2.4 X lo-4 8.4 X 10-2 (lO,O,b) Trench u.b. 180 8 1.6 X 10-3 1.4 X 10-I (-8,0,b) SB-1 Exit 90 18 6.0 X 10-5 3.2 X 10-4 (0,-14,281) Side Shields 90 19 1.05 X 10-3 1.5 X 10-5 (0,±19,b) Side Shields 45 22 2.5 x lo- 5 1.5 X 10-5 (16,±15.5,b) HB-5, -6 Exit 60 18 5.7 X lo-4 J. 65 X lo-4 (9,±15.5,b) HB-3, -4 Exit 120 15 3.5 X 10-4 9.6 X lo-4 (-13,±13,b) HB-1 , -2 Exit 135 18 3.1 x lo-4 8.9 X 10-5 (-13,±13,b) Top Trench d.b. 60 18 5.7 X lo-4 1.65 X 10-4 (10,0,284) Top Trench ·u.b. 120 17 3.5 X lo-4 1.4 X lo-4 ( -8 ,o ,284) Above Target 90 20.5 1.04 X lo-3 3.6 X 10-6 (0 ,0 ,289) Target W.C. at S.P. 90 31 w.c . att '\Jl,3 X Jo-6 (0 ,27 ,284) Dose Rate mrem h-1 J. 3 X 10 5 1.5 X 104 3.0 2.0 0.2 40.0 22.0 0.7 40.0 2.5 0.45 70.0 *The bracketed quantities are the Cartesian co-ordinates in feet from the centre of the target assembly; the x-axis is along the direction of the proton beam, the y-axis is lateral to the beam in the horizontal, norther-ly direction, and z is the vertical co-ordinate. Other terms used mean: d.b. -down-beam from the target; u.b. - up-beam from the target; b- beam line elevation= 268.5 ft; W.C. -water column for main neutron production target; S. P. -swimming pool. - 29 -points 1 is ted in Table VII was assumed filled with heavy concrete (p = 3. 8 g em- 3) in making the dose rate estimates. The only exception is the last point in the table for the junction of the swimming pool and the water column that affords access to the target assembly. For this point the attenuation by the water is greater than for an infinite medium because most of the elastic scattering effectively removes the neutrons into the heavy concrete plug. The effective removal cross-section for the water in the column was estimated from the total scattering cross-section for oxygen, discounted by 10 to 25% depending on neutron energy, to allow for the fraction of the neutrons remaining in the column after elastic scattering. The permanent shielding on either side of the trench up-beam from the thermal neutron target is 10 ft of heavy concrete and on the trench down-beam from the target, 12.5 ft. The 6ft wide trenches will normally be filled with heavy concrete block shielding, providing a top shield and supplementing the side and bottom shielding. Local hot spots from beam spilled by .the beam transport system or experimental assemblies in the up-beam trench can be shielded by iron or other high density shielding. One safety consideration for the shielding along the beam transport system is the dose-rate outside the fixed shielding when the entire beam is spilled in an accident condition at a point protected only by the general shielding. If a 100 kW beam is spilled at such a point in the up-beam trench, the maximum dose rate at any point in the adjacent experimental areas protected only by the 10ft of fixed heavy concrete shielding will be ~2 rem per minute. 4.2 Residual Activity and Radiation Fields The most highly activated component in the thermal neutron facility will be the target. Column 2 of Table VIII shows the activity induced in the Pb-Bi target by high energy nucleon reactions after a long irradiation at 100 kW beam power. The estimate is based on Barbier and Cooper•s tabulationl3 and an allowance of a factor 2 for inter-nuclear cascading in the target. These data are, in turn, based on Rudstam•s empirical recipe 14 Decay Time 1 hour 1 day 1 week 1 month - 30 -TABLE VIII Saturated Residual Activity and Radiation Fields Induced by High Energy Reactions in the Thermal Neutron Facility at 100 kW Beam Power Pb-Bi Target Zr Target Structure Al Moderator Tank y-activity Field @ 1 m y-activity Fie 1 d @ 1 m y-activity Field @ 1 m kCi rem h- 1 kCi rem h- 1 kCi rem h- 1 70 2430 4.7 1860 4.2 220 40 1520 3.3 1420 2.5 100 20 1020 2.2 1060 1.6 40 8 500 1.6 840 1.5 39 - 31 -and apply to the spallation activity only. There is an additional high energy induced component, from high excitation fission. This reaction, which occurs in ~10% of the non-elastic collisions for bismuth, should not charige the total activity significantly but will affect the atomic number and mass distribution of the residual species. Insufficient data exist at present to define the complete distribution in detail, but it will be concentrated in the regions immediately below the target mass number and at low mass numbers for the spallation reactions, and will have a rather broad peak in the region of one-half the target mass due to fission. The total activity in the natural uranium target will be greater by a factor somewhat larger than the ratio of the neutron source strengths, i.e. a factor ~4. Most of the products will come from fission, both high energy and thermal. One potentially troublesome species in the Pb-Bi target, namely 210 Po, will not be present to an appreciable extent in the natural uranium, however. Because some of the active species will be volatile the integrity of the target containment system must be high. The target is doubly contained when viewed from any other point in the facility, first at the target vessel and secondly by the closed target coolant system. There is a third activity containment stage to any inhabited area of the TRIUMF facility provided by the vacuum systems and/or the water column and swimming pool. Column 3 of Table VI I I shows the residual radiation field at 1 m distance from the 47 kg Pb-Bi target for various decay times. The estimate includes the effect of self-shielding by the Pb-Bi. The induced activities in the zirconium target structure and the aluminum moderator tank are also shown in Table VI I I. The estimate for the zirconium assumes a total path length of 1.5 em to the incident protons and 1.0 em to the outgoing cascade neutrons. The aluminum estimate assumes a path length of 1.5 em to the outgoing cascade neutrons and 1.25% of the incident beam is scraped off by the proton inlet tube. The residual field estimates make no allowance for self-shielding in either case. Table IX shows the residual field from activity produced in the steel vacuum tank by the cascade neutrons. This source is concentrated in the - 32 -TABLE IX Saturated Residual Fields Inside the Steel Vacuum Tank from Cascade Neutron Reactions at 100 kW Beam Power Residual Field Decay Time Danger Parameter'~ at Target Position mrem h- 1 @ 106 cm-2 sec-1 rem h- 1 1 hour 61 1000 1 day 53 850 1 week 42 670 1 month 21 340 * The danger parameter is defined as the radiation field that would exist at any point inside an infinite body of the material containing the uniformly-distributed source for which the values apply. It depends only on the source strength per unit volume and the transport properties of the radiation in the medium under consideration. The radiation fields outside thick bodies can be obtained by multiplying the danger parameter by the fraction of the 4TI solid angle subtend by the source object at the field point . - 33 -forward direction from the target so that the residual field at the target position, the field point considered in estimating Table IX, is somewhat higher than for other points inside the vacuum tank. Fig. 16 shows the danger parameter for the cascade neutron induced activity as a function of depth in heavy concrete shielding along a line at 90 deg to the incident beam direction. The estimate is based on Barbier and Cooper's data 13 for iron for a decay period of 1 day. Thermal neutron capture will also produce residual activity in various components. The 210po activity from neutron capture by the 209Bi in the target will reach equilibrium at a value of ~o.4 kCi for a beam power of 100 kW. It will not contribute significantly to the residual radiation field. Thermal capture by 94zr in the zirconium will produce a double a-decay source of 0.5 kCi each for 95zr and 95Nb giving a saturated residual radiation field of 430 rem h- 1 at 1 m distance,assuming a total container mass of 11 kg and 100 kW beam power. Thermal neutrons leaking out of the moderator assembly will be captured primarily in the H20 reflector; assuming that 10% of all neutrons produced in a Pb-Bi target leak into the steel tank, the saturated 59 Fe activity will reach a level of ~40 Ci and give a residual field inside the tank of ~70 rem h- 1 • If the steel contains 50 ppm Co, the residual field from 6°Co at saturation is the same as that from 59Fe. Manganese activity will 1 ikely give much higher residual fields but will decay fairly quickly. As long as the 60co concentration in the steel next to the thermal neutron moderator tank is less than 0.1%, the residual fields will be dominated by the cascade neutron spallation activity. The residual fields in the beam line trench upstream from the target will depend on the spill rate from the beam line. The danger parameter in iron from cascade neutrons leaking back through the iron and heavy concrete shielding is estimated to be 1-2 mrem h- 1 after one day decay. The residual field from cascade neutrons leaking into the more heavily shielded trench down-beam from the target is ~3-4 mrem h- 1 after one day. 4.3 Residual Activity in Cooling Water The target and proton irradiation facility will be cooled by water. The incident protons and leaking cascade neutrons will produce activity by - 34 -high energy reactions with the oxygen. Table X shows the principal activities produced in the target cooling system, their half-lives and pro-duction cross-sections per oxygen atom. The reaction energy thresholds are also listed for the simple reactions. The thresholds for the more complicated reactions are generally higher. The cross-section for lSo and 16N production are the estimated effective values for a fairly slowly varying spectrum immediately above the reaction threshold, as found in BNL-325, Vol. I Supp. #2. 7 The tritium production cross-section is an experimental value 15 at 650 MeV, and the other cross-sections are those measured in the GeV region. 16 They are probably correct within a factor 2 for the protons incident on the target; they probably are an over-estimate for the leaking cascade neutrons. Assuming an incident proton path length in the target cooling water of 2.5 em and an average cascade neutron track length of 1 .0 em in the target cooling channel, the equilibrium source strengths for 100 kW of beam power are shown in column 2 of Table XI. The evaporation neutrons will contribute significantly to the activity production for the lowest threshold reaction, namely 16N production. The results of the Monte Carlo calculation4 indicate that approximately 0.4 evaporation neutrons per incident proton will leak out of the 15 em diam Pb-Bi target with energy greater than ~11 MeV. Thus the effective source strength for the 16o(n,p)l6N reaction is 0.4 + 0.6 = 1.0 neutrons per incident proton. Only the 0.6 cascade neutron per incident proton contributes to the production of the other activities in Table X. The third column of Table XI shows they-radiation field at 1 m from the unshielded sources of column 2. The residual fields will not allow access to the cooling system when the facility is operating at power levels above approximately 100 W, and the llc activity will require a decay period of 3 or 4 h to afford free access following an operating period at 100 kW. The 16N activity will set the operating shielding requirements for the cooling system piping and heat exchangers. The 7Be activity can be removed by filtering. 17 The tritium will build up fairly slowly, 8 Ci/year at 100 kW beam power, but would require dilution by a factor 104 for a once-per-year turnover of an assumed 200 i total coolant inventory to maintain drinking water standards. Residual Species 3H 7 Be uc 13N 1s0 16N - 35 -TABLE X Residual Activity Production Cross-Sections for High Energy Nucleons on Oxygen Reaction Energy Production Reaction Threshold Cross-Section MeV mb lGo(p,t) 40 16o(n,t) 160(p,5p5n) 10 16o(n,4p6n) 160(p,3p3n) 12 16o(n,2p4n) 16o(p,2p2n) 6 16o(n,p3n) 16o(n,2n) 18 15 16o(p,pn) 16o (n ,p) 11 40 Half-Life 12.26 y 53 d 20.5 min 10 min 124 sec 7.35 sec Active Species 3H 7 Be 11c 13N 1s0 16N - 36 -TABLE XI Induced Activity and Residual Fields in the Thermal Neutron Facility Coolant and Moderator Target Coolant D20 Moderator H20 Reflector Saturated Radiation Saturated Radiation Saturated Radiation Activity Field@ 1 m Activity Field@ 1 m Activity Field @ 1 m Ci rem h-1 Ci rem h-1 c i rem h-1 140 'V0 920 'V0 100 'V0 35 1.0 230 6.6 24 0.7 42 23 280 150 30 16 21 12 140 80 115 8 52 29 340 200 37 21 45 160 1400 5000 150 530 - 37 -The fourth and fifth columns of Table XI give an estimate of the various activities in and residual fields from the 330 t inventory of 02 0 moderator. The estimate is based on 47% of the cascade neutrons being removed by non-elastic collisions in the moderato~ and the fraction of collisions resulting in each species was found by taking the ratio of the production cross-sections from Table X to the total non-elastic cross-section for oxygen, 270mb. The sixth and seventh columns are estimates for the 10 em thick H20 reflector-moderator coolant at the outside of the moderator tank. In all cases they are approximately 11% of the 020 moderator estimate. One other water circuit comes in contact with the assembly core and will contain some activity; these are the water columns which provide access to the central target and various irradiation facilities from the swimming pool. Most of the activity production in this water will be in the column to the central target and high flux irradiation tubes which extends down to the graphite-0 20 interface. Assuming that no graphite is used above the target assembly or in the irradiation tubes and that these are filled with water, the activity level will be comparable to that in the H20 reflector circuit. If graphite plugs are used in the water column and irradiation tubes to reduce activity and neutron capture in the moderator, the activity level will be reduced by a factor 2 to 5, depending on the final details of the assembly. Tritium is also produced in the 02 0 moderator by thermal neutron capture in the deuterium. The equilibrium 3H activity from this mechanism is ~3.2 kCi at 100 kW beam power, compared to the 0.9 kCi due to oxygen spallation reactions. 4.4 Heat Production and Dissipation 4.4.1 Heat Production Table XI I shows the distribution of the heat production when 100 kW of beam power is incident on both Pb-Bi and natural uranium targets by 500 MeV protons. The total heat dissipation in the case of Pb-Bi is some 3% less than the beam power because of binding energy differences between ~e initial and final nuclei and nucleons involved in all of the inter-nuclear cascade Component Target Moderator Reflector Shield - 38 -TABLE XII Heat Distribution in Thermal Neutron Facility at 100 kW Proton Beam Power Heat Production Heat Component kW Pb-Bi Target Nat U Target Ionization and recoil 75 440 Evaporation neutrons 5.6 21.0 Cascade neutrons 2.7 2.7 Capture y-rays 2.5 14.5 Moderator total 11 38 Capture y-rays 2.3 6.0 Cascade neutrons 1.1 1.1 Reflector total 4 7 Cascade neutrons 5.2 5.2 Capture y- rays 1.8 4.5 Shield total 7 10 TOTAL HEAT 97 495 - 39 -reactions. The two most important components of this energy difference are the separation energies of the secondary, and higher order, charged particles that stop by ionization and the difference in binding energy for neutrons in Pb or Bi nuclei as compared to hydrogen, a major absorber of the thermalized evaporation neutrons. The estimates in Table XI I are based on the reference design of the target and moderator assembly with a 40 em outside radius 020 moderator region, a 60 em outside radius graphite moderator region separated from a 10 em thick H20 reflector region by a 0.5 em aluminum wall. If the reflector thick-ness is decreased from 10 em to 5 em, the thermal neutron fraction leaking out of the reflector increases from ~0.6 neutrons per incident proton to ~2.6 neutrons per proton for a Pb-Bi target. This change would increase the heat load in the shield by ~2 kW with a partially compensating reduc-tion in the reflector heat load. The heat load in the shield is contributed primarily (~60-75%) by the cascade neutrons; thus it has a very anisotropic distribution. The cascade neutron power flowing into the shield in the forward direction from the target has a peak intensity of 1.1 W cm- 2 . This heat is to be absorbed in cooling coils at the inside of the vacuum tank. Based on a plane-slab model with exponential deposition of the cascade neutron energy as heat, the maximum temperature in a monolithic iron shield would be ~30°C above that at the tank wall. The temperature gradient required to drive 1.1 W cm-2 through concrete, of thermal conductivity k = 0.009 W cm- 1oc- 1 is 122°C cm-1. Thus in the forward direction it will be necessary to 11 thermally bond 11 the iron shield to the vacuum tank with something of higher conductivity than concrete. At 90 deg to the incident beam direc-tion the energy current into the shield is ~0.15 W cm- 2 at 100 kW beam power, and the thermal conductivity of the concrete grout should be adequate to transmit the heat to the vacuum tank wall. The heat deposited in the moderator assembly will be transmitted by natural convection and conduction through the aluminum tank wall to the H20 reflector. The reflector will be cooled by circulating the reflector water to the heat exchanger. - 40 -4.4.2 Target Heat Dissipation Two heat dissipation problems occur for the Pb-Bi target. One is the transfer of the heat from the outside of the zirconium alloy target con-tainer to the H20 coolant. The best mechanism to employ is nucleate boiling at the surface. The limiting heat flux for this mechanism is 125 W cm- 2 for a water-cooled surface based on 2/3 of the critical heat flux in a natural convection system. The peak heat production rate at the front of the target for 100 kW beam power (see Fig. 6) is 4.6 kW cm- 1. Thus assuming no lengthwise circulation of the molten Pb-Bi eutectic in the target, the average heat flux around the circumference of the 14 em diam target is 105 W cm- 2 . This would only leave a 20% allowance for non-uniformity in the circumferential heat flux for a natural convection cool-ing system at this power level. For lower proton beam energies the heat dissipation will be more strongly peaked at the front of the target. The cooling system for the target will be pumped H20 giving much higher flow velocities than can be achieved by a natural convection system. This is expected to allow heat dissipation rates significantly higher than for natural convection systems, but careful attention will be required to the flow patterns in the detailed design of the target assembly. The other critical point is the mechanism used to transfer heat from the Pb-Bi bulk to the container wall, namely, natural convection. This will be tested experimentally by mocking up a target assembly with electrical heaters. Dissipation of the heat generated in the zirconium alloy walls of the proton irradiation facility and target by the proton beam is quite easily handled by the face cooling. For a 3 mm wall the peak heat production from a 100 kW beam of Gaussian radial profile of standard deviation 2 em is 62 W cm- 2 • At the front face of the target, however, an additional 292 W cm- 2 of heat is generated in the first em of the Pb-Bi eutectic. The conductivity of the eutectic is 0.14 W cm- 1 °K- 1; thus a temperature gradient of 450°C cm-1 would approximately double the heat flux through the face centre, to 125 W cm- 2 • The actual gradient depends on the competition between convection and con-duction in the eutectic and will require more analysis. - 41 -The temperature gradient required to drive a heat flux of 125 W cm- 2 in zirconium is approximately 1000°C cm- 1 • Thus the temperature drop across a 3 mm wall would be approximately 300°C. The estimated temperature difference between the boiling film and the surface for nucleate boiling is ~50°C for a heat flux of 125 W cm-2 ; the heat flux is proportional to this temperature difference to the third or higher power over the nucleate boiling range. No expansion or contraction problems on melting or solidifying of the Pb-Bi eutectic are expected from a structural point of view. The eutectic (mp = 125°C) has only a very small volume change at the transitio~ and the target chamber will have an attached void, probably in the form of a tube entering at the top to accommodate the hydrogen and helium gas evolved from protons, deuterons, tritons and a-particles stopping in the target. At 100 kW beam power the gas production is approximately lt per year at STP. This void will also accommodate target expansion when it is in the 1 iquid state. Two forms of the natural uranium target are still being considered. One is a stack of 14 em diam, 0.6 em thick aluminum-clad plates with 0.2 em spacing for cooling water; the other is a 15 em diam aluminum container filled with 0.8 em diam aluminum-clad uranium metal spheres. The void fraction for close-packed, uniform-sized spheres is 0.26; this space is filled with pumped cooling water. As we see below the sphere target has superior heat dissipation characteristics and is preferred if a satis-factory and reliable thermal bond can be achieved between the cladding and uranium. The most important disadvantage of the uranium target is the more strin-gent limitation on the peak beam power. In the molten Pb-Bi target the convection circulation will provide a flattening effect on the peak temperatures; for the solid uranium targets the peak temperatures and heat fluxes follow the peak power proportionately. Thus the lateral profile of the proton beam entering the target must be controlled somewhat more closely. The basic proton beam profile will be essentially Gaussian for some con-figurations of the meson production targets upstream from the thermal - 42 -neutron facility. For other configurations the wings of the distribution will probably be removed in upstream collimators. For a Gaussian density distribution the peak power density at the centre (r = 0) of the beam is given by h(O) = H 2TI cr 2 where H is the total power in the beam and a is the standard deviation of the distribution. The heat production rate from the inter-nuclear cascade produced by 500 MeV protons on uranium metal is estimated to be 26.3% cm- 1 of the incident beam power. Thus, the maximum heat source in the uranium for a Gaussian proton beam density distribution is given by h = 0.263 H 2TI cr 2 The second heat component, namely thermal and epithermal fission, is essentially independent of the incident proton beam profile but contributes a significant amount to the total heat. The heat removal capabilities of the targets are expected to be limited by the heat flux that can be transferred across a metal-water interface. Table XI I I shows the uranium target and incident proton beam parameters for four different target designs and three different beam power levels, based on a maximum heat flux limit of 150 W cm-2. We see that both the disc and sphere targets are capable of dissipating the heat from a 50 kW beam with-out dumping a substantial fraction into upstream collimators. However, only the small sphere target is capable of accepting a 250 kW beam at the collimation chosen for the reference design, namely 10 em diam. A second potential problem with the internally-cooled natural uranium target is that of radiolytic gas production in the cooling water. A much larger fraction of the proton kinetic energy is deposited in the water at high linear energy transfer (LET) for the uranium target than for the Pb-Bi target. High LET radiation favours the production of hydrogen and oxygen gas; thus it may be necessary to install a catalytic recombination unit in the cooling circuit for a uranium target. - 43 -TABLE XIII Proton Beam and Uranium Target Parameters -p ~T Beam Power Target g cm- 3 oc 50 kW 150 kW 250 kW Stacked Plates 0.4 emU l cr = 1. 72 em 3. 22 em 4.55 em 0.2 em Al 9.9 50 Ns = 0.015 0.30 0.54 0.2 em H20 Stacked Plates 0.6 em U l cr = 2. 16 em 4.24 em 6.46 em 0.2 em Al 12. 1 75 Ns = 0.066 0.50 0.74 0.2 em H20 Close-packed Spheres 0.5 em diam U l cr = 1 . 01 em 1. 78 em 2.36 em 0.05 em Al cladding 9.2 106 Ns = rvlo-5 0.018 0.104 25% H20 Close-packed Spheres 0.] em diam U l cr = 1 . 23 em 2. 20 em 2.96 em 0 . 05 em Al cladding 10.4 128 Ns = 2xlo- 4 0.077 0.24 25% H20 cr is the standard deviation of the assumed Gaussian lateral density distri-bution of the incident proton beam for a heat flux limit at metal-water interfaces of 150 W cm- 2 • N5 is the fraction of the incident beam outside a radius of 5 em. ~T is the temperature difference between the centre and surface of the target element for a surface heat flux of 150 W cm- 2 • pis the average target density for the uranium, aluminum and water. - 44 -5. FABRICATION AND COST ESTIMATE 5.1 Shielding Optimization To shield a point source of radiation there is an economic advantage in using materials of superior attenuation properties irrespective of their cost, providing the cost is directly proportional to the quantity used. For most practical applications, where point geometry cannot be assumed at small distances, only materials whose per unit weight prices differ by less than two orders of magnitude need be considered in minimizing the shielding costs. For line source geometry the unit price range offering an economic incentive is smaller,and for plane geometry situations only the minimum cost materials satisfying all other requirements need be con-sidered. The foregoing assumes that the radiation attenuation is directly proportional to mass, an adequate approximation for most circumstances. In optimizing the shielding costs for the thermal neutron facility, iron shielding, heavy aggregate shielding, standard concrete shielding and space costs were considered. Appendix B describes the two-component optimization calculation. Fig . 17 shows the total and direct costs for idealized (complete spherical symmetry) two-component shields for parameters approximating those for iron-ilmenite concrete and iron-standard concrete systems. While the total cost of the heavy-concrete shield is somewhat less than for normal concrete, the more important difference is in the shield size. This is of great advantage if the size of the available area is fixed, as on the south side of the proposed thermal neutron facility, or if the effective source strength for an experiment is distance dependent. To achieve a shield of approximately 6 m outside radius using standard concrete, the cost would be substantially higher than for heavy concrete. The prices shown in Fig. 17 are estimated only. The heavy concrete unit price has no explicit component for form-work and reinforcing. This will probably amount to ~10% of the heavy concrete component cost for the thermal neutron facility shield, thus reducing the apparent total cost advantage of heavy concrete over light concrete in Fig. 17. The unit prices are probably not reliable to better than ~15%. The iron cost is based on 5c/lb for material in the form of locally-available billets, and an estimated 5c/lb for installation . - 45 -Part of the shield in the area of the thermal neutron facility serves to shield the beam line,which may have a distributed component. Fig. 18 shows the idealized shield (complete cylindrical symmetry) cost for normal and heavy concrete systems as a function of spill rate. Both have a trivial amount of iron around a 20 em diam beam pipe and can be taken as effectively fabricated from concrete only. We see that normal concrete is cheaper at the unit prices shown, but the difference is not large, at least for the total costs. The unit prices are appropriate to bulk shielding installations rather than block form, of course. The attenuation factor of the shield at a beam spill of 1.0 nA m-1 was assumed to be 0.3 x lo-4 for an estimated dose rate of ~2 mrem h- 1 outside the shield. A reduction in the total shielding cost can be obtained by using both heavy and standard concrete; in particular, a gradation of heavy to standard concrete with increasing radius would produce a minimum cost. We have not included these potential savings in our estimate. The solution to this optimization problem depends sensitively on the value put on the building space. 5.2 Actual Shielding Costs The estimated component shielding costs for the thermal neutron facility are shown in Table XIV. The iron shielding will be fabricated from extruded billets, flame-cut where ~ecessary to fit the steel vacuum tank, and beam tubes. A small amount of welding may be necessary to achieve adequate thermal contact between the billets and the steel vacuum tank in the forward direction. The ~10% void fraction in the iron will be filled with heavy aggregate grout, either as the iron billets are stacked or following stacking, whichever is cheaper. The iron shielding below the vacuum tank, and probably up to the horizontal beam tubes, will be installed before the vacuum tank. The vacuum tank with beam tubes will then be set in position and tested for vacuum integrity, after which the remaining iron shielding will be installed. The average iron installation cost is expected to be well within the 5¢/lb allowance used in Table XIV. The heavy concrete costs are based on a study by G.E. Crippen & Associates Ltd. IS Their study estimated the · cost of the bulk shielding for different - 46 -TABLE XIV Shielding Costs for the Thermal Neutron Facility Unit Price Unit Price Fabrication & Component Material Installation Amount Cost k$ Iron shielding 5¢/lb 5¢/lb 106 1 b 100 ------------------ ------- ------- -------Beam Line I sides 11 men i te concrete $98/yd 3 $28/yd 3 342 yd3 43. 1 Central Core Ilmenite concrete $98/yd 3 $28/yd 3 515 yd3 64.9 Cascade Neutron Trench Ilmenite concrete $98/yd 3 $28/yd 3 245 yd3 30.9 Water Column Standard concrete $30/yd 3 $10/yd 3 60 yd3 2.4 Removab 1 e Blocks, Beam Line Trench 11 men i te concrete $98/yd 3 $62/yd 3 80 yd3 12.8 Removable Blocks, Case. Neutron Trench Ilmenite concrete $98/yd 3 $62/yd3 54 yd3 8.7 1296 yd3 162.8 ------------------ ------- ------- -------TOTAL 262.8 - 47 -experimental area configurations and compared the cost of bulk shielding in the form of heavy aggregate contained in normal concrete structural walls with that of self-contained heavy aggregate concrete. The experi-mental area configurations considered were: 1) Beam experiments in 8ft to 10ft diam wells 2) Open experimental areas at the beam line level, as shown in Fig. 3) Combinations of 1) and 2) for the south and north halves of the facility, respectively While savings of ~$10,000 might be achieved by the well design (1, above) as compared to the open experimental areas (2, above), the flexibility of the larger experimental areas was judged so much greater that this was adopted as the reference design. The potential saving in using a normal concrete structure filled with loose, saturated heavy aggregate shielding was estimated to be less than $10,000 as compared to self-contained heavy aggregate concrete, based on the open experimental area configuration like that in Fig. 1. This poten-tial saving was not judged sufficient to compensate for the possible settlement and waterproofing problems which might arise with a loose aggregate design. The unit price of $98/yd 3 for ilmenite concrete has estimated components of $83 for material and $15 for plant equipment and mixing, based on $18/ton of ilmenite ore f.o.b. the Vancouver harbour. This ore price may be dependent on an orderl9 for at least 5000 tons; the thermal neutron facility wi 11 require approximately 3500 tons, and the rest can probably be used to advantage in the manufacture of removable shielding blocks for the meson and proton areas. The $28 placement costs are made up of $16 for actual placing and $12 for form-work and reinforcing, based on $2.50/ft 2 for these two components. The $160/yd 3 price of ilmenite concrete blocks assumes a size of at least 2 cubic yards per block, i.e. blocks of ~6.5 tons. 5.3 Building Structure Costs The other structural components required for the thermal neutron facility are the swimming pool and water-lock access to the hot cell region. Based - 48 -on Crippen's study of the structural alternatives, with interpolations for modifications to the original specifications, these components are estimated as: Swimming pool $ 4,500 (scaled by reduced perimeter) Support for swimming pool 9,000· (scaled by reduced height) Water-lock structure Water holding tank Structural wall to hot cell region Total 2,500 4,000 (scaled by reduced perimeter) 15,000 $35,000 The basis for the changes from Crippen's estimate are given by the comments except in the case of the water-lock where the figure was approximately doubled to allow for fittings which were not included originally. The structural wall does not explicitly include the floors or west wall in the hot cell region and is based on 1/4 of Crippen's estimate for the structural walls surrounding the entire thermal neutron facility. 5.4 Thermal Neutron Facility Core Costs 5.4.1 Vacuum Tank and Shielding Plugs The steel vacuum tank will not be removable from the facility after it is vacuum tested and installed. It will be embedded in the grouted iron and heavy concrete shielding, and any subsequent repairs or modifications will have to be done in situ. This indicates a need for a fairly conservative design of this component. The tank consists of two legs. One is vertical of rectangular cross-section of 4 1 011 x 4 1 811 for a length of 8'611 changing to a circular cross-section of 6'611 for 7 ft and expanding to 7'011 diam for the final 7 ft length to the top of the shield. The slant leg is rectangular 3 ft X 4 ft at the facility core, changing to circular of 5'611 diamat 7'611 from the vertical leg wall. The vacuum tank legs will closed and sealed at the outside of the shield thus providing easy access for any possible maintenance. Provision will also be made for making a vacuum seal in the vertical leg at the point where the tank cross-section changes from rectangular to circular, in case this is ever required. be - 49 -The vacuum tank is breached in several places for beam tubes. These tubes will be welded to the main tank and also embedded in the shielding, making it necessary to effect any repairs or modifications in situ . Normally the beam tubes will also form part of the facility vacuum system, with the seals being made at or near the outside of the shielding. As insurance against possible vacuum problems with these tubes, the tank walls will be machine faced and tapped to take bolted, all-metal vacuum seals over these ports if they should ever be required. Water-cooling coils for removing heat from the tank wall and surrounding shielding will be accessible from the inside of the tank. Based on 2 in. thick walls for all flat surfaces and 1 in. thick walls for round surfaces the total weight of the tank will be approximately 50,800 lbs. Assuming a price of 60¢/lb for an installed system, the total price of the vacuum tank is $30,500. Both the vertical and slanted legs require removable shielding plugs. The cost of the heavy concrete and iron shielding in these spaces has been allowed for in the shielding estimates above; thus only the container tanks for these plugs and additional fittings will be costed here. The slanted plug will be built in the form of a steel tank with tubes running length-wise to provide access from the swimming pool to the facility core. Some reinforcing will be necessary for flexural strength when it is handled on installation or removal. Based on steel tank and tube wall thicknesses of 1/2 in., and a fabricated steel price of 60¢/lb, the cost of the shielding plug for the vertical legs of the vacuum tank will be $6,000 and for the slanted leg $14,000. The 1/2 in. wall thickness is conservative and should allow enough reduc-tion to provide the necessary reinforcing. The slanted plug will have embedded air-lines on the bottom side of the plug with holes to the surface at regular intervals. These will be pressurized for installation and removal of the plug. For normal operation these lines are evacuated. The beam tubes from the vacuum tank to the outside of the shield will be a small part of the vacuum system and are included in the costs quoted above. - 50 -5.4.2 Vacuum System The vacuum required in the tank is not high, lo- 2 Torr or less being adequate. A system consisting of a two-stage mechanical pump and a Roots pump similar to one now in use at TRIUMF would have sufficient capacity, ~200 CFM at pressures above 2 x 10- 3 Torr. These pumps cost $4,000. Allowing $2,000 for installation and piping, the total cost of the vacuum system would be $6,000. The vacuum pumps will be situated in the trench above the proton beam line with access from above for servicing. This will reduce the piping run and provide the best vacuum in the vicinity of the proton beam and target, where the requirement is most stringent. 5.4.3 Cooling System There are six separately-identifiable cooling circuits, namely: 1) the main neutron production target, 2) the moderator-reflector, 3) the vacuum tank and shield, 4) the swimming pool, 5) proton irradiation facilities, and 6) the vacuum window and collimator in the proton beam line. The cooling lines in all cases are taken back to the pump room. For the main target and proton irradiation facility circuits the access is through the water columns to the swimming pool and then to the pump room. For the moderator-reflector and vacuum tank-shield circuit the lines run up the inside of the vertical leg of the vacuum vessel, through the vessel wall and a trench to the pump room. Access for installation and servicing of the horizontal runs will be from above by removing shielding blocks needed for protection from the 16N activity fields during operation . The heat load into the swimming pool is at most a few kilowatts. This heat can probably be removed through the piping leads to the other circuits such as the main target. If for any reason the temperature distribution from this arrangement is unsatisfactory, the heat can be removed by a primary circuit cooling coil. The other circuits will be taken to a common primary side of a heat exchanger. The secondary side of the heat exchanger will be cooled with raw water from the cooling tower or low-conductivity water from the heat exchanger in the services annex. To avoid spreading activity in the event of a leak, the primary side of the heat exchanger should operate at a lower pressure than the secondary side. - 51 -Based on recent experience with a similar system built for the central region model cyclotron, the pump (shelf price $700, 200 gpm capacity), heat exchanger ($2,000, 250 kW), installation ($800), and piping and manual valves, etc. ($6,000) will cost $9,500. The only circuits which may require automatic flow control are the main target and the two proton irradiation facility circuits at a cost of $1,500 per valve, for a total of $4,500. A stainless steel dump tank to take the entire coolant inventory of ~400 gallons is estimated to cost $2,000. Water treatment for the coolant can probably be done on a batch basis on facilities existing elsewhere in TRIUMF. Thus the total cost of the cooling system is estimated to be $16,000. 5.4.4 Instrumentation and Control The only control exercised in the thermal neutron facility, other than beam intensity and distribution which is exercised elsewhere, will be on the coolant flow to the target. It will be necessary to monitor some pressure, temperature and radiation level sensors by the central control and safety system. The following list would be desirable, at least at start-up of the facility, but may be cut back for normal operation: Vacuum system pressure Coolant system pressure Temperature Radiation fields 5 points 10 20 15 50 points At $70 per point the cost of this system, exclusive of sensor elements, is $3,500. As a rough allowance for sensors, double the cost to $7,000 for the whole system, installed. 5.4.5 Moderator-Reflector Assembly The moderator-reflector tank shown in Fig. 2 central moderator compartment of diameter 46 volume of ~1000 L If this were all filled $50,000. The outer 8 in. of this tank will is made in. and with 020, be fi 11 ed of aluminum. The length 38 in. has a the cost would be with uncanned graphite, leaving a 020 volume of approximately 330 ~ at a cost of $16,500 - 52 -and a graphite volume of 670 ~at a cost of $7,000, assuming a price of $3/lb. The cost of material and fabrication of the tank has been estimated at $10,000. 20 Thus the total cost of the moderator assembly is $33,500. Essentially no external components of the moderator system are required because heat is removed through the reflector. If it is necessary to dump 020, the shipping containers can be used. To cover the cost of a 020 filler tube and stand pipe for setting the pressure level in the 020 tank, we shall allow $500, bringing the total cost of the moderator system to $34,000. Others 21 are actively considering the replacement of 020 by un-canned graphite at the outside of reactor moderator tanks. While not all of the known possible problems have been explored thoroughly yet, it is unlikely that insurmountable problems will arise in our application because of the rather modest flux level and relatively limited lifetime requirements. If the graphite cannot be used in the moderator tank, the cost of the moderator system, as shown in Fig. 2, would essentially double to $65,000; this would probably be reduced by reducing the size of the moderator assembly. 5.4.6 Target Assemblies and Access Tubes The Pb-Bi target will be contained in a zirconium alloy vessel. To avoid problems with vacuum-tight joints between dissimilar metals, the external cooling jacket and proton irradiation facility will be made of aluminum. The uranium target cqntainer and cladding will be fabricated entirely from aluminum. The target head will be bolted to the 18 in. diam aluminum water column for disassembly under water after withdrawal of the water column into the swimming pool. The highly-activated head can then be passed to the hot cell through the water-lock for servicing. The cost of the main target has not been estimated in detail, as development work will be required to prove the heat dissipation capabilities of the design. Allowing only for detailed design and construction, but not development, the assumed cost of the target with full cooling and a dummy proton irradi-ation facility holder will be taken to be $10,000. - 53 -The vacuum window and beam collimator in the proton beam line have not been designed in detail. The assumed cost of $7,500 should cover all cooling circuit costs as well as the specific components. The aluminum tubes providing water column access to the various irradiation facilities from the swimming pool will cost ~$4,000 at $5/lb of fabricated aluminum and another $1,000 for vacuum seals, for a total of $5,000. Thus, the total cost of the targets and access tubes will be $22,500, exclusive of development costs. 5.5 Facility Total Cost The total cost of the thermal neutron facility is tabulated in Table XV. The probable errors on the cost estimates are different for the core costs and the shielding costs. The shielding cost estimate is expected to be accurate to ~10%, and for the core, say 20%. Thus, the total cost is probably accurate to 15% in terms of 1970 dollars. - 54 -TABLE XV Thermal Neutron Facility Cost Target assemblies and access tubes Moderator-reflector system Instrumentation and control Cooling system Vacuum pumping system Vacuum tank and shielding plugs Swimming pool and internal wall structure Shielding (from Table XIV) Total cost $ 22,500 34,000 7,000 16,000 6,000 50,500 35,000 171 ,000 262,800 $433,800 - 55 -ACKNOWLEDGEMENTS The design presented here is one of several considered in the past few years. We wish to acknowledge the many people and groups who have helped sort through these possibilities. First of all we thank our colleagues at TRIUMF who have, with patience and understanding, put up with periodic changes fn approach while at the same time putting down some of the wilder ideas. We are particularly indebted to Dr. D.W. Hone of Royal Roads Military College for his guidance in the early stages. Dr. B.D. Pate of Simon Fraser University did much to initiate this project. Dr. R.M. Pearce of the University of Victoria has been instrumental in keeping our aims realistic. We have benefited by interaction with groups at Chalk River Nuclear Laboratory, at Gulf General Atomics, at the Nuclear Engineering Department of the University of Michigan, and at the Argonne National Laboratory. More than a score of workers in these groups have generously shared the wisdom they have gained from long experience with reactors. We also wish to acknowledge the more direct assistance of several persons to this report. From the Chalk River Nuclear Laboratory we have obtained the advice of Mr. E. LeSurf on the problems of the Pb-Bi target and container materials and of Messrs. E.C.W. Perryman and I .A.W. Morrison on the feasi-bility of using graphite in a D20-aluminum moderator tank; we are grateful to Messrs. N. Smith and E. Critoph of CRNL for supplying the neutron diffusion code FORGOD and several preliminary calculations and to Dr. J.S. Fraser of CRNL for the inter-nuclear cascade calculations for the bismuth target. The firm of Dilworth, Secord, Meagher and Associates, Limited, consulting engineers, has been most helpful. We acknowledge in particular the advice of Mr. B.C. Stonehill and many design notes on several aspects of the facility written by Mr. W.A. Grundman and Dr. W.J. Wiesehahn; the drawings for Figs. 1 and 2 were done in the DSMA drafting office. A cost study of the shielding structure alternatives was supervised by Mr. D. Kwong of G.E. Crippen & Associates, Ltd. Finally we should like to thank Mrs. R. Hotell and Miss A. Strathdee for their patient typing and retyping of the several drafts of this report. - 56 -REFERENCES 1. S. Penner, "Handling High Power Electron Beams", IEEE Trans. Nucl. Sci. NS-14, 908 (1967) 2. J.S. Fraser, R.E. Green, J.W. Hilborn, J.C.D. Milton of CRNL and W.A. Gibson, E.E. Cross, A. Zucker of ORNL, "Neutron Production in Thick Targets Bombarded by High Energy Protons", Phys. Can. 1.!... (2), 17 (1965) _and J.C.D. Milton and J.S. Fraser, "A Monte Carlo Calculation of Neutron Production in Heavy Element Targets in the Range 0.3 to 1.0 BeV'', AECL-2259 (1965) 3. G.A. Bartholomew and P.R. Tunnicliffe, editors, "The AECL Study for an Intense Neutron Generator", AECL-2600 (1966) 4. J.S. Fraser, private communication (1968) 5. N. Smith and E. Critoph, private communication (1969) 6. S.A. Kushneriuk, J.M. Kennedy and P.M. Attree, private communication ( 1964) 7. H.J. Hughes and R.B. Schwartz, "Neutron Cross-section Compilation", BNL-325, 2nd edition (1958) and J.R. Stehn et aZ., Vol. 1, Supp #2 (1964) M.D. Goldberg et aZ., Vol. I lA, Supp #2 (1966) M.D. Goldberg et aZ., Vol. liB, Supp #2 (1966) M.D. Goldberg et al., Vol. IIC, Supp #2 (1966) J.R. Stehn et al., Vol. Ill, Supp #2 (1965) All of BNL-325, 2nd edition 8. H. Goldstein et al., "The Slowing Down of Neutrons in Hydrogenous Media - Status of Theory and Experiment", Geneva Conference on the Peaceful Uses of Atomic Energy, Proceedings, Vol. 18, 379 (1958) 9. Reactor Physics Constants, ANL-5800, 2nd edition (1963) 10. I .M. Thorson, "Shielding and Activation in a 500 MeV H- Cyclotron Facility", TRI-68-4 (1968) 11. R.G. Alsmiller, R. Leimdorfer and J. Barish, "Analytical Representa-tion of Non-Elastic Cross-Sections and Particle-Emission Spectra from Nucleon-Nucleus Collisions in the Energy Range 25 to 400 MeV", ORNL-4046 (1967) 12. H.W. Bertini, "Low-energy Intranuclear Cascade Calculations", Phys. Rev . .!l.!_, 1801 (1963) with erratum Phys. Rev. 138, AB2 (1965) and "Results from Low-Energy Intranuclear-Cascade Calculations, ORNL-TM-1225 (1965) - 57 -13. M. Barbier and A. Cooper, 11 Estimates of Induced Radioactivity in Acce lerators 11 , CERN 65-34 (1965) 14. G. Rudstam, 11Systematics of Spallation Yields 11 , Z. Naturforsch. 21A, 1027 (1966) 15. L.A. Currie, 11Tritium Production by 6 BeV Protons••, Phys. Rev. 114, 878 (1959) 16. P.A. Benioff, 11Nuclear Reactions of Low-Z Elements with 5.7 BeV Protons 11 , Phys. Rev . ..!.!1' 316 (1960) 17. D.O. Busick, ••7 Be Build-up in a Large Water Beam Dump at the Stanford Linear Accelerator Center11 , SLAC-PUB-521 (1968) 18. G.E. Crippen & Associates Ltd., North Vancouver, B.C., private communication (1970) 19. Houston Aggregate Co. Ltd., Jacques Cartier, P.Q., private communica-tion ( 1970) 20. Dilworth, Secord, Meagher and Associates Limited, Vancouver, B.C., private communication (1969-70) 21. E.C.W. Perryman, private communication (1970) - 58 -APPENDIX A FAST/THERMAL RATIO OPTIMIZATION For thermal neutron beam experiments the sensitivity-limiting condition is often the achievable signal-to-background ratio. The source of the back-ground, in turn, is often the fast neutron contamination in the beam. Inasmuch as a) the background is proportional to the fast neutron flux at the source point of the neutron beam, and b) the signal is proportional to the thermal flux at the same point, then the optimum choice for the source point is defined by the spatial distribution of the fast and thermal flux and the signal-to-noise ratio of the particular experiment considered. Consider the truncated Taylor expansions of the thermal flux T(r) = T - TP 0 and the fast flux F(r) = F - ~r 0 at a point r in a spherically symmetric system. The signal (or informa-tion) S is found by taking the difference between the combined signal and background, S + B, and the background measured separately. Thus S = (S + B) - B and on the assumptions a) and b) above S = [~T(r) + BF(r)- BF(r))t where t is the time of the measurement and ~ and 8 are constants depending on the details of the experiment. The absolute error on Sis given by and the fractional error, to be minimized, is - 59 -~ /~(T0 - TP) + 2B(F0 - ~p) E= -= I S ~(T - TP)t2 0 and this wi 11 have an extremum when dE 1 [ T [ o(T0 - TP) + 2(3(F0 - ~pJ)t -= 0 = dP d ~(T0 - TP) T - TP 0 and at the reference point P _dE _ _ --.,...1 _1_ [-T-( ~_T_0_+_2_f3_F_0_J_t dP t2~T0 T0 I or T(~T0 + 2BF0)2 To ~T + 2(3~ =------~1 2(~T0 + 2BF0)2 For a signal-to-background ratio S/B =Mat the reference point P B T ~T = Mf3F and -= ~ o o ~ MF 0 and the extremum condition is or or .L T(~T0 + 2~T0/M) 2 To 1 [T:J t [ M : 2 J t M + 2 M 1 (3~ = -+-2 ~T ~T + 2(3~ = 2~0/M)t 2(~0 + ~· = + 2(3~ 2[~T0 ]![(M+2)/MJ)! - 60 -or M+ 2 - M/2 ~ = M MTFO <I> (M + 4)T -= and Po 2T0 Thus d~ ( R.n F(r)} = M ; 4 • ! l tn T(r)) and this extremum can be shown to be the condition for minimum fractional error on the measured signal. Thus we see that even at low signal-to-noise ratios, the fraction rate of change in the fast flux as a function of r must be at least twice that of the thermal component at the optimum source point. - 61 -APPENDIX B TWO-COMPONENT SHIELDING OPTIMIZATION 1. Point Source To simplify the analysis we shall fix the source-point to field-point distance. This is reasonable for field points near the outside shield surface and has the simplifying effect of fixing the attenuation factor for the shield. For a two-component, point-source shield the direct cost P8 is then given by where r 0 is the radius of the source; in the case of the moderator-reflector assembly, for example, r 0 = 0.7 m), r 1 is the radius of inner shield of material costing c1 per unit volume, and r 2 is the radius of the outer shield of material costing C2 per unit volume. There is an additional cost Pa of the space in the building. This is taken to be proportional to the floor area covered; thus and the total cost P is For a total attenuation factor A= e-N, where AI and A2 are the exponential radiation relaxation lengths in shield materials 1 and 2, respectively. After writing r 2 in terms of r 1 from the equation for total attenuation, - 62 -the total cost is given by p = A3r1 3 + A2r12 + A1r1 + Ao where A3 4rr (c1 + c2 (8 3 - 1)] =-3 A2 4rr (3s2 rac2 + c:)) =--y-A1 4rr (3a8(aC2 + ~) =--y-Now the minimum cost is given by dP dr =0 1 or 3A 3r 12 + 2A2r 1 + A1 = 0 2. Line Source = -2A2 + /4(A2 2 - JA3A1J 6A 3 for minimum cost. For a uniformly-distributed line source the total cost per unit length of line is given by where all parameters have essentially the same meaning as in the point source case above. Again on the assumption of exponential attenuation, the outer radius r 2 is written in terms of the attenuation exponent and the transition radius r 1 , giving - 63 -where a. = NA. 2 + A.2ro/A.l s = 1 - >..2/A.l A2 = 1r[c1 + C2(S2 v] Al = S(1Ta.C2 + 2Ca) A -0 - 1T(a.2c2 - ro 2c1J + 2a.Ca and the minimum is given by dP O dr1 = or For the beam line the source radius r 0 would be that of the beam tube plus the tolerance distance to the inside shield surface, say 15-20 em for most cases. b I . 0 I ~ 'PROTON REAM SECTION A-A Fig. 1 (a) A plan view shows the basic components of the thermal neutron facility, with most of the thermal neutr n beam tubes dotted in. A E.L.Z64'-0" .... -~ ~ \ ' . -.:J, . . I~ . , . c:.l . , , c I HAl N VAGUUH TANK •,_ <I" -' -= .. 7'-0"+ I __ ....st:._~-----4---I RON .SHIELDING 6'-6"+ .J c HEAVY CONCRETE SHIELDING ~ ; • ....-t- '"\ ~... ' ' ~ ~-.<a ' " ..... '... ... HT-1 /// ' / / ~ SECTION 8-B 5 IHHING 'POOL (HzO) Al--/ 4ft X 4ft WATERLOCK m HOT CELLS ~ . • : . .-..q . ·, ···~- •• ,·.r . ~ ' "~ '(' ,' '· . ,. ' ·. ·~ ~~ .' :· .. ~ · Y/ / Fig. 1 (b) An elevation section through the centre of the TNF perpendicular to the proton beam shows the basic outline of the vacuum tank and some of the neutron beam tubes imbedded in the iron and heavy-concrete shielding. HEAYV CO NCRET"E: SHIELDINGt TRENCH~ HB-c ST-6 -C (EXTENSION) ' .. \ ' .. ~ < • • . ·~-~ 6. • ',~ ~ h) .. ~ ". ". ""' t: .. .~~ . ...., __ ...,_ __ ,.. ' .. .. \ ~· ' •" . .(:·,,. TRENCH~ -~ I .:· v ;pROTON BEA H ;.~ 1 TUBE ~- VAGUUfvt 'PUHPING LINE SECTION C-C Fig. 1 (c) An elevation section along the proton beam 1 ine of the TNF shows the proton beam tube HB-P, the ci1 scade neutron beam tube HB-C and the trenches for gaining access to these two facilities. The trenches are filled with demountable shielding (not shown on the figure) when the TNF is operational. -- 2'-0" SWIHH lNG 'POOL ( Hz.O) VIEW D MAIN VAC..UUH TANK HB · Z. HB-P HEAVY CONCRETE SHIELDING Fig. 1 (d) A partially eut-aw · . I e dtng they wt11 require when the -faci1 ity i~r~~~~a~7~~a~ ~ne and cascade neutron tre ~ches without the b1'ock shi 1 . ay tsome~rtc view of the TNT shows the . HB-1 L . I 0 0 I I .. ~ SB-1 ST-G-C EXTENSION 1------- --- ----"PROTON BEAM SECTION A-A MODERATOR TANK HB-e I HB-4 _j I IRON SHIELDING:. H 'B-6. E Fig. 2(a) A split-section plan view of the core of the thermal neutron facility shows the outline of the moderatlor tanks and some of the neutron beam tubes. liJEAH TUBE SB· I s.T-e.-c EJ'.TENSION 7' - 0 ' .. _,l:~D,_. _____ _ e.·- e.- 1.0 SECTION B-B MAIM VACUUM TANK SLANTED SHIE.l.DING "PLUG Fig. 2(b) An elevation section of the TNF core perpendicular to the proton beam 1 ine shows the basic target and moderator arrangement and most of the neutron beam tube and irradiation facilities. SECTION D-D 'PROTON IRRADIATION VACILITY HAl f\l NE:UTRON "PRODUCfJON TARGE:T ·-~ .. HB-G SECTION C -C VT- Z "PROTON BEAM COLLIMATOR ~~! 1..-.-.J~~======• I IRON HEAY'( C.ONCR~ ' 1 . . .......:!~ r r , <r ~ . -i~ . ~ I . ' ~ ,~. ~ HB-P _He C60LEO VACUUI-1 WINDOW t:NLARGED DETA'L SCALE.~ 1'- O" ~ tt PF?OTO N_ BEAM ---r------t- VACU UH PUMPING L..l ~E Fig . 2(c) An elevation section of the TNF core along the incident proton beam 1 ine shows the vertical beam tubes and the vacuum line from the tank to the up-beam trench. ST-&-C t:.XTENSION HzO ~EFLECTOR TANK --...-- DzO -GRAPHITE VIEW E N.T.5 HT-1 HODEJ?ATOR TANK HB-4 B-Z. Fig. 2(d) This isometric view of the vacuum tank in the vicinity of the TNF core has most of the shielding cut away for a clear view of the beam tubes a~d irradiation facilities. ST- 14-TO SWIMMING POOL • B ST-18-P ST-18 I I I ~~~~~~~~777777~~~~~ ~----,1 BOLTED, ALL-METAL VACUUM SEAL II ST-18-3--J.! II II II II II II II II II II II 1.!:_-----_-_Jj ALUMINUM TARGET -COOLANT CONTAINER ALUMINUM- CLAD URANIUM METAL SPHERES I ~ MODERATOR TANK I -DOTTED OUTLINE I I I I I I I I rl!z:z::l_=_z!l~ ALUMINUM TARGET c:&Jbo'U~~~ ,/ -COOLANT CONTAINER H2 0 COOLANT CHANNEL -~s MOLTEN Pb-Bi IN or A ZIRCALLOY CONTAINER SECTION A-A Fig. 3(a) A-A is a section through the centre of the main target assembly perpendicular to the incident proton beam direction showing both the aluminum-clad uranium metal spheres target and the molten Pb-Bi target. ST-18-P HB-P I I I l~:s:s:t t.: __ , A----~ 1 TO SWIMMING POOL TARGET COOLANT OUTLET INLET ST-18 -~-----J ~------r- ST-18-2 1 (BEHIND) PROTON BEAM ---------CENTRELINE -------., I L.:-I PROTON IRRADIATION j' FACILITY A~--I I t--- OUTLINE OF PRODUCTION TARGET I MODERATOR TANI I ----------~ SECTION B-8 Fig. 3(b) B-B is a section of the main target assembly along the proton beam centreline showing the neutron production target and the proton irradiation facility. 60 z 0 b 50 ~ a.. ~ z Ill c 0 z C/) 40 z 30 0 ~ ~ :::::) Ill z "-0 20 ~ Ill m ~ :::::) z 10 NEUTRONS PER INCIDENT PROTON ON Pb AND U 0.5 1.0 PROTON ENERGY ( GeV) e U TARGET, DIMENSION (DIAMETER X LENGTH) 10.2 em X 61 em 1.5 Pb TARGET, DIMENSION (DIAMETER X LENGTH) 20.4 em X 61 em 10.2 em X 61 em 2 .0 Fig. 4 The experimentally determined 2 • 3 neutron yield from protons incident on Pb and U is shown as function of proton energy. CASCADE NEUTRON ANGULAR DISTRIBUTION FOR 500 MeV PROTONS ON A Bi TARGET AVERAGE OF M.G. RESULTS FOR 10 em AND 20 em DIAMETER TARGETS e CASCADE NEUTRONS (R.H.S. ) 0 CASCADE NEUTRONS KINETIC ENERGY (L.H.S.) z ~ 0 -0.1 1£ ._ z 1&1 AVERAGE NEUTRON c ~ . z ct AVERAGE 0 ct a: l&J NEUTRON K. E. .... Cl) a: 1&1 • Q. z 0 a: .... ;:::) 1&1 z -.01 1&1 • c ct u en ct u O.IL---------L---------~--------~--------~--------------~ 1.0 0 -1.0 cos e Fig. 5 The figure shows angular distributions of cascade neutron particle and kinetic energy currents emerging from a cylindrical bismuth target when 500 MeV protons are incident along the centreline. z ~ ~ a.. 0.6 ~ z ~ 0 z -'e u ~ u ~ c 0 f z 0.3 0 cr ~ ~ I.&J z z 0 0 ~ 3 a.. ~ I.&J 0.1 EVAPORATION NEUTRON AND HEAT PRODUCTION VS DEPTH IN TARGET Bi TARGET, AVERAGE OF M.C. CALCULATIONS FOR IOcm AND 20 em DIAMETER TARGETS 500 MeV PROTONS INCIDENT • HEAT PRODUCTION - R.H.S. o EVAPORATED NEUTRON PRODUCTION- L.H.S. z 24 ~ cr a.. ~ z I.&J 20 e u z cr I.&J a.. I 16 E u > ., ~ z 12 z 0 t= Cl) ~ I.&J 8 c ~ I.&J %: ~ I.&J 4 ~ ~ ~------~------~------~~----~~----~~----~~----~~----~~----~0 40 80 120 160 200 240 280 320 360 0 X g cm- 2 Fig. 6(a) Evaporation neutron and heat production is shown as a function of depth in a bismuth target for 500 MeV protons incident 0.6 :z().5 0 1-0 a:: a.. > ;;o.4 0 0 10 HEAT FROM NEUTRON INDUCED FISSION vs RADIUS NATURAL URANIUM TARGET, 9 em RADIUS SPHERICAL GEOMETRY TOTAL THERMAL ~0.3~-------------------­CI (.) z a:: w0.2 a.. If) ·~ (.) > G) 0.1 ~ 0 0 TARGET 0.7 NAT. U, 0.15 AI I 0.15 H20 2 4 6 r,CM 8 EPI-THERMAL AND FAST 10 12 Fig. 6(b) Heat production from thermal and epithermal neutron induced fission in a natural uranium, 9 em radius spherical target is shown as a function of radius in the target; the values are normalized to one 500 MeV proton incident on the target. 14 t-z L&J a u ~ a= L&J a.. N •e u X ~ ...J LL ...J <( ~ a= L&J I t--3 THERMAL NEUTRON FLUX DISTRIBUTION SPHERICAL GEOMETRY- VARIOUS MODERATORS 10 em Pb- Bi TARGET 10 em H20 REFLECTOR NO STRUCTURE 0 140 em 0 20 MODERATORS AS SHOWN • 110 em c 11 50 em DzO • 30 em 0 2 0- 20 em C 10 L---------------~---------------L--------------~------~ 0 20 40 60 r em Fig. 7 The thermal neutron flux distributions are shown for various moderator assemblies surrounding a 10 em radius, spherical Pb-Bi target. All moderator assemblies are surrounded by a 10 em H20 reflector. Moderator dimensions shown are spherical shell thicknesses . 3.0 -1&.1 > ~ ...J 1&.1 ~ 2. -)( ~ ...J lL ...J <I ~ ~ 1&.1 :I: t-1.0 0 THERMAL FLUX AT r = 17.5 em AND 35.5 em VS OUTER MODERATOR RADIUS D20 AND GRAPHITE MODERATOR 10 em Pb-Bi TARGET, 10 em H20 REFLECTOR, r•l7.5em r • 35.5 em -------+-}r • 17.5em --- _._ -- GRAPHITE ----- -~- r • 35.5 em --..... ----------------..... -50 100 150 OUTER MODERATOR RADIUS, CM Fig. 8 Thermal neutron flux is shown as a funct ion of outer moderator radius at two field points i n the D20 and graphite moderator regions. 200 -LLI > ~ ct 1.0 ...J LLI a: -X :) ...J l&.. ...J ct ~ a: LLI I ~ 0.9 THERMAL FLUX DEPRESSION FROM GRAPHITE SUBSTITUTION FOR D2 0 AT OUTSIDE OF 60 em RADIUS MODERATOR ASSEMBLY r • 35.5 em 10 20 GRAPHITE THICKNESS- t, em Fig. 9 The thermal flux depression from replacement of 020 by graphite is shown for two field points of radius r as a function of graphite thickness; the basic assembly has a 60 em outer moderator radius, followed by a 10 em thick H2 0 she 11. 2.0 -LLI > t-c:t ...J LL1 1.8 a: -X ~ ...J l&.. ~ 1.6 ~ a: LLI ::r t-1.4 1.2 THERMAL FLUX DEPENDENCE ON H20 REFLECTOR THICKNESS AT FIELD POINT RADII r x,__---------------x--x/ / r • 50.5 em I X X r • 35.5 em r• 17.5 em I. 0 III::...-------------.....1.-----------...L....-----' 0 5 10 H20 REFLECTOR THICKNESS- d, em Fig . 10 The thermal neutron flux dependence on the H20 reflector thickness is shown for three field points in the reference design moderator assembly. z 0 r-0 IE 162 ~ ~ 0 0 It) r-z lU 0 0 z a: lU Q. N I E u X ~ ...J "- 10 0 THERMAL NEUTRON FLUX VS RADIUS D20- GRAPHITE MODERATOR SPHERICAL GEOMETRY 0 NO STRUCTURE OR COOLANT • 0.9 em Zr~ 0.4 em Hj) 0.3 em AI AT TARGET- MODERATOR BOUNDARY 0.5 em AI AT MODERATOR-REFLECTOR BOUNDARY X STRUCTURE AS ABOVE PLUS 2~. AI IN D20 MODERATOR Pb-Bi I TARGET 10 D2 0 MODERATOR 20 30 I 40 r, em GRAPHITE MODERATOR 50 60 Fig. 11 The estimated thermal neutron flux distributions for three different structural configurations demonstrate the effects of neutron absorption in the target and moderator assemblies. 70 z ~ ~ a. ~ ~ ENERGY GROUP NEUTRON FLUX VS RADIUS Pb- Bi TARGET SPHERICAL GEOMETRY ~ 105------..... ~ t-z LAJ a u z a: ~ N 'e u X :;:) ..J "-a. 5 ffi NEUTRON ENERGY GROUPS • E > I MeV (FAST) ll I MeV> E > 1.46 eV (INTERMEDIATE) e 1.46 eV > E > 0.127 eV lEPI-THERMAL) o E < 0.127 eV (THERMAL) 'Pb- Bi I TARGET I DzO MODERATOR ~ 0 .9 em Zr, 0.4 em H 0 0 .3em AI 0 10 20 30 r em GRAPHITE MODERATOR 40 50 Fig. 12(a) The flux distributions for neutrons in four energy groups are shown for the reference moderator assembly and the Pb-Bi target. The estimate is based on spherical geometry with structure as shown above the radius ordinate at the bottom of the figure. t-z I&J ~ 012~ 70 ~ <( I&J m z ~ 0 a: a. <( ~ 0 0 t-<( X :;:) ..J "-a. ~ 0 a: C) z ~ 0 a: Q. 10-t .....---=E;.;...;;N=ER;...;..G.;;;...Y.;.._G;;;...;.R...;..;O;;...;U;;..;..P~N...;..;E~U;;....T.;....;.R...;..;O;;..;,N...;..._F;....;L~U;;..;,X...;.......,;v~s-R;...;..A;......;.;D~I;..;;;;.U...;;;.S __ __,. NATURAL URANIUM TARGET SPHERICAL GEOMETRY > 10-2 • 2 0 0 10 t-z &&.1 Q 0 !: a: &&.ll<f3 Q. N '2 0 0 NEUTRON ENERGY GROUPS .& E > I MeV (FAST) 6. I MeV > E > 1.46 eV (INT.) • 1.46eV > E > .127 eV (EPI-THERMAL) o E ( .127 eV (THERMAL) 0.7U II 0.998 ~0 0.15H20 0.002 ~0 0.15AI~ 0.5AI 0.5~0 20 r,cM 40 0.85 GRAPH. II H20 0.15 ~0 ~ 0 .5AI 0.5~0 60 Fig . l2(b) The flux d istributions for neutrons in four energy groups are shown for the reference moderator assembly and the natural uranium target. Except for the targets, the details of assemblies used in making the estimates in Figs . l2(a) and l2(b) are identical. C( ::l. 0 2 -0 Fe 0 FAST I THERMAL FLUX RATIO VS RADIUS 10 em H20 REFLECTOR, 10 em Pb-Bi TARGET NO STRUCTURE MODERATORS AS SHOWN 0 140 em 020 e II 0 em GRAPHITE !:. 50 em DzO • 30 em 0 20 - 20 em GRAPHITE 20 40 60 r, em Fig. 13 The fast-to-thermal flux ratios are shown for the same assemblies as in Fig. ]. The fast component includes all neutrons with energies greater than 1 .46 eV. .... z UJ I.) lL lL UJ 0 I.) ~r---------------------------------------------------~~ 0.1 0 CASCADE NEUTRON SOURCE COEFFICIENTS REFERENCE ASSEMBLY NO STRUCTURE o THERMAL e EPI- THERMAL 6 INTERMEDIATE • FAST 20 r em Fig. 14 40 60 The coefficients o.i are shown as a function of radius in the reference moderator assem-bly, where the ith group flux is factored into the form ~i(r,s) = ~i(r.o}(l+o.i(r)s) and s is the angle integrated cascade neutron component, assumed for purposes of the estimate to be isotropically distribu~d. The value of s in the actual assembly with a Pb-Bi target varies from 2.2 in the forward direction to 0.06 in the backward with angle integrated value of 0.6. 7 6 5 4 3 2 0 RATIO FAST/THERMAL LOGARITHMIC FLUX GRADIENTS R= d(lnF{r))/d{lnT{r)) IN dr dr 20 REFERENCE ASSEMBLY Pb- Bi TARGET M (M+4)/2· 0 2.0 I 2.5 2 3.0 5 4.5 10 7.0 40 r,cm Fig. 15 60 The ratio of the fast/thermal logarithmic flux gradient is shown as a function of radius in the reference assembly with a Pb-Bi target . The contributions to the energy group fluxes by the cascade neutrons have been taken to be that at the angle integrated average. The table on the figure shows the optimum values for various experimental signal-to-noise ratios, M. -I s= DANGER PARAMETER FOR IRON IN HEAVY CONCRETE VS RADIUS AT 90° TO INCIDENT BEAM MODERATOR REFLECTOR 50 IRRADIATION PERIOD 5000 DAYS DECAY PERIOD I DAY MODERATOR-REFLECTOR OUT TO r = 70 em 100 kW BEAM POWER HEAVY CONCRETE 100 150 200 r em Fig. 16 The danger parameter (defined at the bottom of Table IX, page 32) for iron is shown as a function of depth in heavy concrete shielding outside thermal neutron facility core . The field point is at 90 deg to the incident proton beam of 100 kW power. An irradiation time of 5000 days and a decay time of one day was used in the estimate. 400r---------------------------------------------------------~ SPHERICAL SHIELD COST VS OUTSIDE RADIUS FOR ATTENUATION FACTOR= 10-8 300 200 10 UNIT PRICE A,CM IRON IOt/LB 18.9 ILMENITE CONCRETE $115/YD3 34.2 STANDARD CONCRETE $40/YD1 50.0 BUILDING AREA $50/FT 2 o DIRECT SHIELD COST IRON-ILMENITE CONCRETE • TOTAL COST IRON-ILMENITE CONCRETE A DIRECT SHIELD COST IRON-STANDARD CONCRETE • TOTAL COST IRON-STANDARD CONCRETE 0~------~------~------~~------~------~------~~----~ 4 5 6 7 8 9 10 II Fig. 17 OUTSIDE SHIELD RADIUS m The spherical shield cost for iron-standard concrete and iron-heavy concrete shields of attenuation 10- 8 for cascade neutrons is shown as a function of outer shield radius. Both total costs, with a component for floor area, and the direct shield costs are shown. ~ "" ..... "" ~~ ~ * ~ 20r-----------------------------------------------------~ OPTIMIZED IRON-CONCRETE CYLINDRICAL SHIELD COSTS ~ COST 10TALCOST em 18 STANDARD CONCRETE ILMENITE CONCRETE IRON $401 YD.1 $1151 YO~ lOt/ La $50 I FT2 50.0 34.2 18.9 ILMENITE CONCRETE AREA STANDARD CONCRETE 16 14 ILMENITE ~CRETE 12 SHIELD COST 10 STANDARD CONCRETE 8 6 4 2 OL-------~----------------L---------------~----------------~ 0 1.0 10 100 LOG SPILL RATE nAim Fig. 18 The optimized cylindrical shield costs are shown as a function of uniformly distributed proton spill rate along a 1 ine for iron-standard concrete and iron-heavy concrete shields. The amount of iron is very small and the shields are essentially all concrete. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.51833.1-0302687/manifest

Comment

Related Items