US2813070A - Method of sustaining a neutronic chain reacting system - Google Patents

Description

1957 E. FERMI ETAL 2,813,970

METHOD OF SUSTAINING A NEUTRONIC cum: REACTING SYSTEM Filed Nov. 28. 1945 7 Sheets-Sheet 1 Nov. 12, 1957 EfFERMl ETAL 2,813,070

METHOD OF sus'mmmc A NEUTRONIC CHAIN REACTING SYSTEM Filed Nov. 28, 1945 7 Sheets-Sheet 2 Nov. 12, 1957 E. FERMI ETAL' 2,813,970

METHOD OF sus'mmmc A NEUTRONIC cum; REACTING SYSTEM Filed Nov. 28, 1945 7 Shepts-Sheet s E. FERMI ET AL Nov. 12, 1957 METHOD OF SUSTAINING A NEUTRONIC CHAINREACTING SYSTEM Filed Nov. 28, 1945 7 Sheets-Sheet, 4

mu m HL Wiles 6'. Zfler-e airless es zmwm Nov. 12, 1957 E. FERMI EIAL METHOD OF SUSTAINING A NEUTRONIC CHAIN REACTING SYSTEM Filfid NOV. 28, 1945 7 Sheets-Sheet 5 Nov. 12, 1957 E. FERMI ETAL 298139976 METHOD OF sus'mmmc A NEUTRONIC CHAIN REACTING SYSTEM Filed Nov. 28, 1945 '7 Sheets-Sheet e F1E-E HMI Nov. 12, 1957 E. FERMI EI'AL METHOD OF SUSTAII INING A NEUTRONIC CHAIN REACTING SYSTEM Filed NOV. 28, 1945 7 Sheets-Sheet 7 eooooooooeooooeoooo 00.00000000000000000 oeooooooooooeooeooo ooooobvojo ooooooe oooooe-ooosoo ooooo ooecQeoooooeoa 0 000 oooooooooooomoaoooo ooo ooooeooooeofooo coo ooeooooooao coo coo oooaeoooooo can 000 oooeoooeooo ooo ooooeooeeoooaoo ooo oooo ooooooeoooeoeo oooo ooeooooseboooo aeooo ooooonooooooo ooooooe b 9000 000 ooooooboooooeoooooo oeooooooooooooooeoo a ooooeooeooooooooo oooooaaooeeooeo eooooob oooo oeoe awesome moo 09 90000068990 00000900900060 9 0000000000000 oooooaoucoeoooo aoooeoomeoeooo OQBQQAvOBOQQOQQ oooooooooooooe ODOOGQOQODOQBO ooooeoooo oe0oo ooueooeoooooodo odeoooeoooeoooe OGQQOOOOQOOQQOD oeooeooaoooooao ooeo eooooooqoo oeoeood oooe 0990900009009 0 0 0 9 0 O o O 0 0 G 0 O0 O0 O0 00 O0 00 0Q OGBSQOQOGQOOQGBOQOG ooooooooooooueouoov eouoooeoeaoooneoofi oooeooocooooeoeoeoo OQ'OOO'QOQOOQQQOOQO 000039 000000 000660 000000 OQQOOQ 099090 OOQGOO 000006 009900 0060.0 0.680 00 00 09 09. 00 .86 .00 09B United States Patent METHOD OF SUSTAINING A NEUTRONIC CHAIN REACTING SYSTEM Enrico Fermi, Santa Fe, N. Mex., and Miles C. Leverett, Oak Ridge, Tenn., assignorsto the United States of America as represented by the United States Atomic Energy Commission Application November 28, 1945, Serial No. 631,406

1 Claim. (Cl. 204-1542) natural uranium contains the isotopes 92 and 92 in the'ratio of approximately 139 to 1. Hereinafter in the specification and the claim the term uranium is to be understood as referring to uranium and its chemical compositions of normal isotopic content unless otherwise indicated by the context.

In a self-sustaining chain reaction with slow neutrons, 92 is converted by neutron capture to the isotope 92 The latter is converted by beta decay to 93 and this 93 in turn is converted by beta decay to the transuranic element 94 By slow neutron capture, 92 on the other hand, undergoes nuclear fission to release energy appearing as heat, gamma and beta radiation, together with the formation of fission fragments appearing as radioactive isotopes of elements of lower mass numbers and with the release of secondary neutrons.

The secondary neutrons thus produced by the fissioning of the 92 nuclei have a high average energy, and must be slowed to thermal energies by passing said neutrons through a material in which the neutrons are slowed by collisions. Such a material is known as a moderator.

Among the basic parts of a neutronic reactor system may be listed the following:

(1) A fissionable material such as uranium is properly dispersed in a moderator such as carbon or deuterium oxide.

' (2) In most cases, a neutron reflecting or scattering layer is provided around the active portion for returning at least a portion of escaping neutrons into the active portion/'- The reflecting layer can be an extension of the moderatorbeyond the active portion. 7

(3) In all but the lowest power output reactors, coolant channels or pipes are passed through, or in heat exchange relationship to the active portion, in order that a stable temperature can be maintained in the reactor.

(4) In most cases, loading and unloading mechanisms are provided to charge and remove uranium from the active portion.

(5) A shield is almost invariably provided around the reactor to minimize escape of radiations biologically harmful to operating personnel. Such shields may comprise layers of iron and hydrogenous material closely surround- 0 neutron reproduction 1n an average state of balance is provided.

ice

In the discussion to follow, the term neutronic reactor is used to mean the active portion and the reflector if used, and the term neutronic reactor system is used to mean the entire system including shields, charging and cooling systems, etc.

In considering the requirements for an operating neutronic reactor, the ratio of secondary neutrons produced by the fissions to the original number of primary neutrons initiating the fissions in a chain reacting system of infinite size using specific materials is called the reproduction or multiplication factor of the system and is denoted by the symbol K. If K can be made sufliciently greater than unity to create a net gain in neutrons, and the system made sufficiently large so that this gain is not entirely lost by leakage from the exterior surface of the system, then a selfsustaining chain reacting system can be built to produce power by nuclear fission of natural uranium. The neutron reproduction ratio (r) in a system of finite size differs from K by the leakage factor and operating poisoning factor, and it, too,- must be greater than unity to permit the 'neu. tron density to rise exponentially. Such rise will continue if not stabilized at a desired density corresponding to a desired power output.

During the interchange of neutrons in a system comprising bodies of uranium of any size dispersed in a neutron moderator, neutrons may be lost in four ways; by absorption in the uranium metal or compound, by absorption in the moderator, by absorption in impurities present in the system or added as a result of operation, and by leakage from the system.

Natural uranium, particularly by reason of its U con tent, has an especially strong absorbing power for neutrons when they have been slowed down to moderate energies, these energies being termed resonance energies. The absorption of neutrons in uranium at these energies is termed the uranium resonance absorption or capture. It is caused by the isotope U and does not result in fission but creates the relatively stable nucleus 94 It is not to be confused with absorption or capture of neutrons by impurities, referred to later. Neutron resonance absorption in uranium may take place either on the surface of the uranium bodies, in which case the absorption is known as surface resonance absorption, or it may take place further in the interior of the uranium body in which case the absorption is known as volume resonance absorption. Volume resonance absorption is due to the fact that some neutrons make collisions inside the uranium body and may thus arrive at resonance energies therein. After successfully reaching thermal velocities, a large percent of the neutrons are also subject to capture by U Without fission to produce 94 It is possible by proper physical arrangement of the materials in the moderator to control the amount of uranium resonance absorption. By the use of light elements such as graphite and beryllium for example, fewer collisions are required to slow the neutrons to thermal energies, thus decreasing the probability of a neutron being at a resonance energy as it enters a uranium atom. During the slowing process, however, neutrons are diffusing through the slowing medium over random paths and distances so that the uranium is not only exposed to thermal neutrons but also to neutrons of energies vary thesystem in the form of spaceduranium masses or bodies 3 of substantial size, either of metal, oxide, earbide, or combinations thereof. The uranium bodies can be in the form of layers, rods or cylinders, cubes or spheres, or approximate shapes, dispersed throughout :the'graphite, preferably in some geometric-pattern.

The term geometry is used to mean any pattern or arrangement wherein the uranium bodiesare distributed in the moderator with at least a roughly uniform spacing and are roughly uniform in size and'shape, ,or-aresystematic. in variations of size, shape or spacing-toproduce a volume pattern conforming to a generally symmetrical system. If the pattern is -a repeating or rather exactly regular one, the structuremay be conveniently des c ribed as a lattice. The uranium bodiescanbe'in the form of layers, rods, or cylinders, cubes or spheres, or approximate shapes, dispersed throughout the moderator. Optimum conditions are obtained with natural uranium by-using metal spheres. V

The. resonance losses in uraniumconstitute one ofthe critical factors in coordinating the. total losses permissible ina neutronicreactor. V V

The'thermal neutrons are also subject tocapture: by

thernoderating material. While. carbon. and-beryllium have very small. capture cross-sections for thermal neutrons, an appreciable fraction of thermal neutrons (about 10 percent of the neutrons present in the system under best conditions with graphite). is lost by cap turein the slowing material during diffusion therethrough. This means that when volume ratios are changed, the absorption in the moderator will also change. a

In addition to the above-.mentioned-losses, that are inherently a part of the nuclear chain reactionprocess, impurities present in both the moderator and the uranium add a very important neutron loss factor in the chain. Such impurities may be originally present, or be formed during operation. The effectiveness of variouselements as neutron absorbers varies tremendously. Certain elements such as boron, cadmium, Samarium, gadolinium, and some others, if present even in afew parts per million, could very likely prevent a self-sustaining .chain reaction from taking place. If impurities, solid, liquid for gaseous, and in elemental or. combined form,- are present in too great quantity, intheiuraniu mbodies or the-slowing material or'in, or by absorption from, the free spaces of the system, the self-sustaining chain freaction cannot be attained. f The amounts ofimpurities that may be' permitted in a system, vary witha'number of factors, such asthe specific geometryof the-. sys temuand the form in which the uraniumis nsed -that is, whether natural or containing more than. the natural Iamountof a fissionable'isotope, whether as metal. or oxide and-als o factors such as the weight ratiosbetweenthe uranium and the slowing down material, and .thetypeofmoderating material used- -for example, whether...deuterium, graphite or beryllium. Although all .of these considerations influence the actual. permissible amount. oft each impurity material, it has fortunately been .ftiundthatin general the effect of any given impurity .Or impurities can be correlated directly with the weight of the impurity present and with the K factor of the-s ystem, .so. that knowing the K factor for a" given geometry .andcornposition, the permissible amounts of particular impurities can be readily computed without taking individual account of the specific considerations namedabove. Diifer'ent'impurities are found to affect the reaction to Widely different extents; for example, relatively considerable quantities of elements such as hydrogen may bepresent, and, as previously suggested, the uranium may be inthe formof V a pa cu rsyst m h r predu t r on chang s occur in the atmospheric pressure. A sensitive barometer is thus obtained. In general, the inclusion of combined nitrogen is to be avoided.

The effect of impurities on the optimum reproduction factor K may be conveniently evaluated to a good approximation,- si rnply by -means;ofcertain constants known as danger coefiicientsiwhich'areiassignedto the various elements. I These danger coefiicients for the impurities are each multipliedbyjthe percent by weightof'the corre sponding impurity; and ,the" total'sum of these products gives a value'knbw'n as the total danger-sum. Thistotal danger sum is subtracted from the reproduction factor K as Ycalculated fof pure-materi'als and for-the specific geometry under-consideration.

a The danger coefficients are defined in terms of the ratio of the weight of impurity per unit mass of uranium and are based onthe cross-section for absorption of thermal neutrons'of the various elements; I These. values mayzbe obtained from; physics.- textbooks ion the subject :anditthe danger coeflicient computed by theafot'mul'a'.

wherein T1 represents the cross-section-for the impurity and Tu thecross-section for the'uranium, A1 theqatomic weight of the=irnpurity andAu the ,atomic weight :-for uranium. If the-gimpuritiesare in-thecarbon, theyxare computed as their percent of the weight. of the uranium o t e sys em:- T t e flsq i of impu s or n y in the system as placed .in operation can .be evaluated'in termsof the factor-- This fact is usedto insure :control of the reactor, .by deliberatelyiinserltingneutron absorbing materials to a greater or less extent inthe-reactor when. bu lt 1 a i crea e in n u o -de s y, w han! sequent stabilization of the neutron-density at a desired value, as.will be later explained.

The size of the. sy st em.for proper operation will vary depending upon the K factor; of the ,system,--and upon other thingssuch as, for example-,, th .type -'of;mode rator andpower of operation. If :the reproduction factor K is greater-than unity, the number'of neutronspresent will n e sew xp n y, p o i ed he s c r s mad sufliciently large. .If, on the contrary, the structure is small, with a largesurface-tovolumeratio,-there willbe a rate of loss of neutrons fromlthe" structureby leakage throughthe outer surfaces, which may overbalance the rate of neutron production inside thestructure sci-theta chain reaction will not be self-sustaining. For each value f the p u fa to g e er-tha ni y, ere is thus a minimum overallsize of a given active. structure known as the critical size, above which thetrate of;loss

O e ons-b d fiu on toe wa of he uct e nd leakage away from the-structuretis less -.than th e .ratetot production of neutrons vvitl1 in.-the-.:system,' thus imaging the-chain.reactionfself sustainingt Therate of difiusion o n u q aw y f o a ar t ct e i .whieh th y-ar being created through-gthe exterior surfaee ,thereofmay -be treated-gby mathematical analysis when the-value "of K and certain other constantsare'known. j I

In the case of a spherical structure employing uranium bodies. embeddedv in graphite-in the geometries disclosed he and th ut .a rna r flecto l the pllow ps formula gives the critical overall radius whenK is known:

Critical 'sphereof radius R ft., K''- 1=C/R where" C 'is a constant-that varies slightly: with? geometry and fornormaluranium-graphite lattices mayhave a value close to 7.4.

ora re ns r para el p p d sm me'r tneuas sphetical, the critical {size can be computed from the formula where a, b, and c are the lengths of the sides in feet. The critical size for a cylindrical structure is given by the formula, irrespective of the shape of the uranium bodies Cylinder heighthft. 0.59 RadiusRft. K 1 C h R However, when critical size is attained by definition, no rise in neutron density can be expected. It is therefore necessary to increase the size of the structure beyond the critical size but not to the extent that the period for doubling of the neutron density is too short, as will be explained later. A desirable reproduction ratio (r) for an operating structure at a stabilized power with all control absorbers removed is about 1.005. This will permit the neutron density to rise in a reasonable time and yet permit control by insertion of sufiicient neutron absorbers to bring the effective reproduction ratio to unity for continued operation at some predetermined neutron density. The size at which the reproduction ratio of 1.005 can be obtained may be computed from modifications of the above formulae for critical size when corrected for power of operation. For example, for spherical active structures the formula K r=7.4/R may be used to. find R when K is known and r is somewhat over unity. The same formula will, of course, give r for given structures for which K and Rare known.

It can be seen from the above discussion that if, after a neutronic reactor has been built to a fixed operating size, the K factor should change for any reason, critical size would change. The reproduction ratio r would then change in accordance with the difference between critical size and operating size. If K increases, critical size decreases, and r decreases. If K decreases critical size in creases, 1 decreases, and if K decreases sufficiently, critical size may coincide with or become less than operating size and the reactor will become inoperative unless increased to a new operating size where r is again greater than unity.

During the lifetime of operation of a neutronic reactor, K may change for several reasons which may be grouped under long term effects and short term effects.

During the neutronic reaction fission products, including such extremely absorptive elements as samarium and gadolinium, for example, may form from the fission of the uranium isotopes U and 94 Samarium, as well as some other elements formed in this manner, has very high neutron absorption characteristics and may increase the total parasitic neutron absorption in the reactor, with consequent decrease in the reproduction factor. However, the absorption of neutrons by the uranium isotope U leads to the production of 94 Plutonium is fissionable to an even greater degree than the isotope U so that formation of plutonium tends to increase the reproduction factor K, notwithstanding decrease in the U content of the uranium in the reactor due to destruction by fission. The increase in K over a long period of production of 94 may run as high as between 1 and 2 percent.

Still another factor tends to increase the reproduction factor in an operating reactor. Certain neutron absorbing impurities are present in graphite moderators, for example, sufiicient to reduce K by from 1 to 1 /2 percent. A substantial portion of this reduction is due to boron inherently present in the graphite. This boron is gradually transmuted, probably to lithium, during the operation of the reactor, and as lithium has a very much less neutron capture capability than boron, K will increase.

The long term effects can be summarized by stating that K will probably initially increase due to production of 94 and then decrease as the fission products build up.

In addition to the long term changes in K to be expected, there are shorter and even more significant changes in K that take place in reactors operating con- 1 6 tinuously at high neutron densities, as for example, U- graphite reactors operating at from 5,000 to 500,000 kilowatts power output.

The short term elfect of xenon formed as a radioactive decay element from iodine produced in the reactor during operation, becomes significant after 4 to 5 hours of continuous operation and changes K' by an amount equal to the equilibrium effect of the amount of xenon 135 produced and destroyed at the power of operation. Xenon 135 decays to barium, but is changed by neutron absorption to xenon 136. Xenon 135 has an extremely high neutron capture cross-section of about 2,500,000X10 cm. whereas the capture cross-sections of iodine, barium and xenon 136 are relatively small- Thus while the xenon 135 is present, due to the operation of the reactor, K is reduced by an amount related to the continuous power output in a graphitemoderated reactor as follows:

w kw.=0 .03 K 500,000 kW.=0.02 K 100,000 kW.=.0l3 K 10,000 kW.=.00l2 K The end result of this xenon 135 effect is that a reactor having an actual size capable of supporting a chain reaction at low, intermittent powers, may not be large enough to support a chain reaction at higher continuous power outputs. In consequence, the active portion of the reactor must be made large enough in the first place, if continuous high power operation is to be attained. I

While it is, of course, apparent that a decrease in K might cause a system fixed in size to become inoperative it is not so apparent that an increase in K would be undesirable. However an increase in K is undesirable for the reason that if K increases, r will also increase and the reactor may become dangerous to operate, if r reaches or exceeds 1.01.

In order that the significance of a reproduction ratio of 1.01 be more fully understood, the mechanism of fission will be discussed further. Not all of the fast neutrons originating in the uranium body, as the result of fission, leave the uranium immediately. With natural uranium about one percent are delayed neutrons. These delayed fast neutrons may appear at any time up to several minutes after the fission has occurred. Half these neutrons are emitted within six seconds and 0.9 within 45 seconds. The mean time of delayed emission is about 5 seconds. The neutron reproduction cycle is completed by 99 percent of the neutrons in about 0.0015 second, but if the reactor is operating with a reproduction ratio near unity, the extra 1 percent may make all the difference between an increase or a decrease in the activity of the reactor. The fact that the last neutron in the cycle is held back, as it were, imparts a slowness of response to the pile that would not be present if the fission neutrons were all emitted instantaneously.

For cases in which the reproduction ratio (r) dilfers from unity by appreciably less than 1 percent, the rise of neutron density, or more specifically the value N to which the number of neutrons has risen from an original value No, after a lapse of time of t seconds during and before which the pile has operated at a fixed value of r (N0 being the number of neutrons at the beginning of t, i. e., after disappearance of transient effects due to any preceding change in r) is given by N o wt where L.i a(1'-1) T In this formula or is the fraction of the neutrons that are delayed, i. e., 'a=.0067, and T is the mean time of delayed'emission of the delayed neutrons- 5 seconds.

asiao'ro that.is,'.-N/No=2.72 in 28.5 seconds. Hence doubling of -.theneutron density occurs about every 20 seconds and continues indefinitely. The above formula thus indicates the rate of rise for relatively lowvalues of r and shows how thereduction of t-he rate of thedelayed neutron effect :is particularly significant in the stated lower range of-J. values. ..Strictly' speaking, the given equation holds onlyfor'lthe. steady state, i. e., where r has been held constant for some time; an additional transient term must be included to obtain an accurate representation of the neutron density during the first few seconds after a sudden change of r. g,

If r were made exactly 1.01, a more detailed theory shows that the neutron density would be more than tripled each second. However, if the reproduction ratio'r is several percent greater than unity, so that the one percent delayed neutrons are unimportant compared with r1, the density increases at a much more rapid rate as given approximately'by r where l is 0.0015 second, the normal time to complete a cycle. Thus if r were tobe made 1.04, the neutron density would increase in 1.5 seconds by a factor of approximately 10 over its original level. However,- if r were 1.02 and 1.03, the factor by which the neutron density would be multiplied each second, would be ll00 and 700,000 respectively. It is thus apparent that too high a reproduction ratio in a practical system leadsto the necessity of inserting what may be considered as 'an'excessiveamount of controlling absorbers to reduce the effective reproduction ratio to unity. An exceedingly dangerous condition could exist if by accident these absorbers were suddenly completely removed, as the time required for reinserting the absorbing material might be too' long to prevent destruction of the system. As the same eventualdensity can be obtained with a reproduction ratio onlyslightly over unity, as with a higher ratio, only at a slower rate, the lower reproduction ratios which exceed unityby not substantially more than .01, an amount equalto the'percentage of the neutrons formed which are delayed neutrons are preferred in practice in the interest of safety.

It is therefore desirable to provide a reactor in which compensation for variation in the reproduction factor may be-rnade when necessary.

Asmassive'iron and concrete shields, which may be of the order of 20 feet in thickness customarily surround an operating reactor, any requirement that a reactor be increased in size beyond a size limited by the shielding structures, could not be met because of the immobility of said shields. Furthermore, if a decrease in size is to be made, the entire system must'be such as to be fully as 'efliciently operated after decrease as before.

It is the main objectof the present invention to provide a novel'n'eutronic reactor that may be initially constructed in such a manner as to allow for either an increase or decrease in thesize of the active portion.

Broadly stated, compensation for changes in the K factor with a consequent requirement for change in operating size can be accomplished in-accordance with the present invention by initially providing a moderator mass greater in size than would be required for a uranium-moderator structure of a given initial K factor, and then initially loading uranium into that mass only in an amount and with such distribution 't'o'provid'e areactor of'a predetermined operating reproduction ratio r. Then, if K should change,-"the-'amountiof uranium in the moderator mass can -beincreased or decreased to provide 'continuedoperatio'ri'with the predetermined r; 7

The amount 'of uraniumloadedintofa moderator mass can be changedbecause-any portion of the moderator extending beyond the uranium containing mass acts as a reflector, returning by the scattering action of the moderator nuclei a largenumber of the escaping neutrons. The use of a reflecting layer of moderator 2-feet thick, for example,'around a neutronic reactor can reduce the critical size of the reactorby several percent of the radius thereof.

It is not the'reduction in size of the reactor thatis important in this instance. The main consideration is that a reflecting layer one to two feet thick around a reactor is nearly as eflicient a reflector as a reflecting layer of infinite thickness. Asia close approximation, reflecting layers Zto 10 feet thick, for example, may be considered equivalent. In consequence, when the moderator mass is larger than that portion of the moderator containing uranium, the active portionfalways has an eflicient reflector, providing only that a moderator layer of sufficient thickness remains around the uranium bearing portion to give proper reflecting efficiency. I

As a result, a mass of moderator can be charged or loaded with varying-amounts of uranium to form a portion of operating size, asneeded, in all cases being surrounded by a reflectinglayer of substantially equal reflecting efficiency.

It is therefore-another object of the present invention to provide a neutronic reactor system of fixed dimensions wherein the size of thelactive portion supporting the chain reaction can be changed to suit circumstances without loss ofjefiiciency.

There .are .other advantages of a system wherein the active portion can be changed to suit conditions. As noted above, the active portions'of neutronic reactors can be in the overall shape of a cube, sphere or cylinder. Further, it has been stated that the uranium can be placed in the reactor in the form of spheres, rods or layers. Thus, it is desirable that provision be made that a moderator mass be provided such that active portions can be loaded, not only to varying overall sizes, but to varying overall shapes, with varying" types of uranium and with varying geometries of the uranium with respect to the moderator in the portion loaded.

It follows that another object of the present invention is to provide a neutronic reactor system in which uranium or other fissionable material can be loaded with'varying shapes of the loaded. portions,and with varying sizes and shapes of the uranium bodies themselves, as desired;

Other objects and advantages of the present invention may be more clearly understood by references to the following description and the attached drawings which illustrate, as an example, one form the invention may take. This example'is not to be taken as limiting, as other forms within' the scole of; the appended claim will be readily apparent to those skilled'in the art.

In the drawings:

Fig. 1 is a longitudinal view partly in section andpartly in elevation of an air cooled neutronic reactor system illustrating the present invention;

Fig. 2 is a cross-sectional view, partly in elevation, taken as indicated by the line 22 in Fig. 1;

Fig.3 is a plan view of the system shown in Figs. 1 and'2;

Fig. 4 is a longitudinal sectional view partly in elevation of a jacketed slug;

Fig. 5 is a longitudinal sectional'view, partly in elevation, .of'a horizontal'channel during a loading and unloading operation;

Fig. 6 is .a cross-sectional view taken as indicated by the line 66 in Fig. 5; V

Fig. 7 is. a longitudinal sectional view, partly in elevation," of one form of loading device; 7

Fig. 8 is a view partly in section and partly in elevation takenas'indicated bythe line 8-8 in Fig. 7;

Fig.9.is" a topjplan' view. of the loading device shown in Figs. 7 and'8 I Fig. is an end view of the loading device shown in Figs. 7, 8 and 9;

Fig. 11 is a view partly in section and partly in elevation of a friction drive used in the loading device;

Figs. 12 and 13 are diagrammatic longitudinal sections showing loadings of difierent sizes, the section lines being left off for purposes of clarity;

Figs. 14 and 15 are diagrammatic cross-sections taken as indicated by lines 1414 and 1515 in Figs. 12 and 13, respectively, section lines being omitted for clarity.

Fig. 16 is a diagrammatic cross-section of a cubically loaded neutronic reactor, section lines being omitted for clarity.

Referring to the drawings, the present invention is illustrated by reference to an air cooled graphite-uranium reactor, sometimes known as a pile. The operation and construction of such reactors are more fully described in Fermi et al. Patent 2,708,656, dated May 17, 1955.

Such a reactor broadly comprises a mass of graphite blocks closely piled or stacked into a cube 10 as shown in Figs. 1 and 2. This graphite cube may be, for example, 24-26 feet on a side and rest on a concrete foundation 11. The graphite cube 10 is pierced with horizontal air channels 12, of square cross-section, with one of the diagonals vertical. The channels may be readily made by grooving adjacent blocks (Fig. 6). The channels are .175 inches on a side and extend completely through the reactor, from an inlet face 14 to an outlet face 15. Two thousand channels may be provided, for example, and as will later be brought out, any unused channels can be plugged. Only a few of the channels are shown in the drawings for the sake of clarity.

Adjacent the inlet face 14 of the cube, the foundation is continued downwardly to form the floor of an inlet air duct 16 extending outwardly. The inlet air duct 16 is completed by concrete side walls 17 and top 19.

At some distance away from the graphite cube 10 the inlet duct is turned upwardly to terminate in an air filter 20, relatively close to the surface of the ground. A fan or blower 21, here illustrated as electrically driven, is installed on the floor of the inlet duct just below the air filter, access to the fan being conveniently obtained through duct door 22, behind the fan.

The concrete top 19 of the inlet air duct is continued upwardly as an inlet end shield 24, positioned parallel to but spaced away from inlet face 14 of the cube 10 to form an inlet chamber communicating with the air channels 12.

Above the inlet chamber 25 and the cube 10 the concrete is continued horizontally to form a top shield 26, and side shields 28 are built up from the foundation 11 to enclose cube 10. Shields 26 and 28 closely approach the top and side faces of the cube, to minimize air flow around the outside of the cube. A small amount of air circulation, however, may be desirable over the top and side faces to cool these faces.

At the outlet face 15, an outlet end shield 30 of concrete is provided. End shield 30 is parallel to and spaced from the outlet face 15 of the graphite cube to form an outlet chamber 31 communicating above with the base 32 of a stack 34 projecting upwardly and formed as a continuation of the concrete top side and outlet end shields. Thus the cube 10 is completely enclosed by concrete shields, with a duct system operating by virtue of pressure provided by fan 21 to conduct air from close to ground level through channels 12 into the stack and then into the atmosphere well above ground level at the top of the stack. The concrete shields may be from ten to twenty feet thick in accordance with the maximum desired operating power of the reactor and serve as shields to reduce escape of neutrons and gamma radiation.

As a neutronic reaction will take place when uranium bodies are properly spaced in a moderator mass of a certain finite size, the above described device can be made chain reacting by placing uranium bodies in the horizontal channels in such a manner and insuch an amount that a neutron reproduction ratio of slightly over unity is obtained, exclusive of all neutron losses within the reactor and from the exterior of the reactor. This reproduction ratio may be defined as the ratio of the number of neutrons gained by fission to the total number of neutrons lost by absorption in the uranium, absorption in the moderator, absorption by impurities in the reactor or added as a result of operation, and by leakage from the reactor for a reactor of finite size.

Using the graphite mass as the moderator to slow fast neutrons to energies where they again are able to create fission in 92 the device will have a reproduction ratio of unity when approximately 700 of the channels 12 in the graphite cube are each loaded with 68 aluminum jacketed uranium slugs 35 lying end to end, with a channel spacing of 7 inches measured center to center, and with the loaded channels roughly forming a cylinder (Figs. 14 and 15). Both graphite and uranium should be of highest possible purity. The diameter of the cylinder will be about 18 feet.

However, more than a unity reproduction ratio is required, as when the reproduction ratio is exactly unity no rise in neutron density will occur. Under such conditions the device will not develop high neutron densities or power in the form of heat. By loading additional channels, i. e., making the active portion greater than critical size, however, the reproduction ratio within the reactor can be brought above unity in order that arise in density can occur. Then this excess neutron reproduction can be absorbed by neutron absorbing materials deliberately inserted into the reactor in order to hold the reproduction ratio at an average value of unity after a desired power output has been obtained, as a result of the initial rise in density.

Consequently, in accordance with the amount of excess reproduction ratio desired, about 1,000 channels may be loaded with uranium slugs. The diameter of the cylinder will now be about 21 feet, leaving a minimum of 3 to 4 feet of graphite around the cylinder except at the ends. Most of the channels not loaded with uranium may be closed by inserting plugs, preferably of graphite, in such channels in order to conserve air. Some of the channels, however, in the peripheral portions of the cube may be left open for cooling of the graphite in those portions.

One preferred form of slug construction is shown in Fig. 4. Each uranium metal slug is 1.1 inches in diameter and 4 inches long covered with an aluminum jacket approximately 20 mils thick in good heat conductive relation to the uranium. The slugs weigh about 2 /2 pounds each.

In forming the slugs 35, the uranium portion 36 is machined to size, cleaned in trisodium phosphate and then washed in water. Aluminum or other non-fissionable metal jacket cans 37 are provided having an inside diameter somewhat larger than the uranium portion. This can 37, open at one end only, is slipped over the uranium after being cleaned in benzine and hot water. The can 37 with the uranium inside is then passed through a sizing die of 1.134 inches diameter. This die, being of smaller diameter than the 1.1 inch uranium portion plus the two 20 millimeter walls, draws the can in tight thermal con-tact with the uranium.

A cup-shaped cap 38 is then place-d base down inside the projecting portion of the can 37 and is seam welded to the can. The projecting portion is then cut off above the seam weld 40 and the remaining projecting portion including the weld, spun over the adjacent end of the slug. Thus each jacket completely encloses and seals the uranium, preventing air from corroding the uranium and, as Will be pointed out later, also preventing fission fragments created by nuclear fission at the surface of the uranium from entering the air stream.

' gamma radiation.

a reed Thelchahnels are loaded with uranium until the reproduction ratio; with neutron absorbersremoved, is about L005 'to'f1.00,6., This ineans'that for every two hundred neutrons, starting .in each neutron generation about two hundred and one. neutrons are produced in the reactor over and above all losses. Under these conditions andtaking into account the fact that about one percent of the neutrlonsof fission arev delayedin their emission for a mean timeof about seconds, the neutron density of the reactor will double every 8 to seconds. With some part of the neutron absorbers inserted but with the insertion of less than the amount of neutron absorbersrequired tornake the reproduction ratio unity, the rise is slower. Whenthe neutron absorbers are almostbut not entirelyinserted the doubling of the neutron density maytake severalhours. Then when a desired density has been reached,'the reproduction ratio can be-reduced to-unity so that-the desired density is continuously maintained bythe introduction of neutron absorbing material into the reactor.

The neutron absorbing material. is introduced into the reactor by'means. of a control rod- 41 as shown in Fig. 2. This control rod extendsinto' the graphite cube, sliding in a channel therein and'is operated from'outside of side-shield 28 as by rack and pinion 42. The rod is made from, or incorporates therein, an efiicient neutron absorber, such as for example, cadmium or boron. A sheet of cadmium riveted to a steel strip forms a satisfactory control rod. As the depth of insertion of the rod determines the amount of neutron absorbing material inside the reactor, the critical position of the rod is where the rate of neutron absorption by the rod balances the reproduction ratio at unity. Thus, by moving the rod outwardly from the critical position the neutron density in the reactor will rise. Moving the rod inwardly from the critical position causes the reproduction ratio to fall below unity, and the reaction stops. Thus the reaction is always under control, and as the rise in neutron density is exceptionally slow as the rod approaches the critical position; manual control is possible. Other and similar rods 141, shown schematically in Fig. 2 may be provided, if desired, for rapid progression into the reactor to stopthe reaction in case of failure of the control rod to stop the rise in neutron density for any reason. Such rods are termed safety rods.

During operation heat is released in the reactor in accordance with the neutron density therein. Most of the heat arises from the kinetic energy of the fission fragmentsand about 92 percent of the energy is released in the uranium. About 6 percent is released in the graphite due to neutron absorption therein and about 2 percent escapes from the reactor in the form of neutrons and Consequently, the reactor can only be operated at a power dependent upon heat removal to the point where a stable temperature obtains. Otherwise, the reactor will accumulate heat to the point that the device may be damaged. Since aluminum melts at 658 C., stable temperatures below this value should be used although with jackets of other non-fissionable metals, such as beryllium, the stable temperature may be increased, although if the temperature should rise too high the uranium bodies might be damaged even when using beryllium jackets as uranium of the type, used in neutronic reactors melts at about 1100 C.

A stable temperature is obtained in the device of the present invention by'passingatmospheric air through the reactor, and inthe specific example shown and described, the air is passed throughthe graphite channels and directly in contact with thev aluminum jackets of the slugs. Under these circumstances the reactor can be operated continuously'at250 kilowatts electrical equivalent of heat by passing 32,'OOOI;cubic feet per minute through the reactor with a maximum'temp'erature rise of the-slugs to about 100 C., andat500"kilowatts"continuously"with about 5 0,000 cubic feet per minute of air with a maximum metal temperature of 200"C. 'The output of the reactor can be stabilized at still higher powers by the use of larger fans if desired.

Having discussed generally the operation of the reactor, and the temperature stabilization thereof by air cooling at elevated powers and neutron densities, there is now de-, scribed one means and method by which the reactor can be loaded and, unloaded, in order that the neutron irradiateduranium-can be removed for further processing such as the recovery of 94 formed in the uranium, and fresh uranium inserted for subsequent operation of the reactor, and in any amount desired.

To accomplish loading of the slugs 35 into the various airchannels. 12, the concrete of the inlet end shield 24 is pierced with a plurality of loading apertures 45, as shown in Figs. 1 and 5, each aperture being aligned with the axis of slug positions in the air channels 12. Normally, during operation of the reactor, each aperture is closed by a removable lead plug extending through the shield 24 only. When it is desired to load a channel with new slugs, the lead plug in shield 24 for that channel alone is re-. moved, and acharging tube 47 inserted, extending through the inlet end shield 24, across the inlet chamber 25 and entering the corresponding air channel 12 as shown in- Fig. 5, The outerend of charging tube 47 is provided with a flanged nipple 49: shaped to engage a nipple recess 50 of a loading'mechanism indicated generally by numeral 51. It willbe noted that the charging tube is smaller than the air channel 12 and that air can pass through the channel being unloaded. The air'should circulate duringunloading, although it may be at reduced velocity.

Loadingrnechanism 5.1 .comprises a loading magazine 52, a loading plunger 54, and a plunger drive 55, as shown in Figs. 7 to 10 inclusive.

The loading mechanism 51 is mounted on an elevator platform 56 mounted to be raised and lowered in an elevator frame 57capable. of moving along the outside of inlet end shield 24.0n elevator tracks 58. Base 5,9.of the elevator frame is providedwith a platform 60,pro-. jecting outwardly on the same level as the top of a supply car 61travelling on supply car tracks 62. Supply car 61 is used-to bring a supply of slugs to the elevator for use in the loading mechanism 51.

The slu'gs' 35 when received at the elevator, are loaded into an inclined loading channel 64 in the loading magae zinc,- in side by side relationship and feed by gravity to the bottom thereof. a The bottom of loading channel 64 is a part-of a plunger bore 65 extending through the loading magazine ending in the nipple recess 50 cooperating with flanged nipple 49on charging tube 47 so that the V plunger bore 65 and the loading'bore in guide tube 47 partially irradiated and in consequence radioactive. The

thick walls of the-magazine then act as a shield for the radioactive slugs, and in this case a heavy cap 70 may closethe upperopening of the loading channel 64. Iron or lead. may be used for the body of the magazine. In addition, the use of thick metal in the magazine, particularly around the plunger bore- 65, reduces radiation that might pass through the'interconnected loading aperture 45 andguide tube 47 either from radioactive slugs therein'or from theirradiated slugs in the reactor when charging tube 47 is empty. 7

Tlieslugs 'arefed from magazine 52 by a reciprocating motion of plunger 54' operating in plunger bore 65. Plunger 54l'may be of iron to act as a shield when inserted into charging tube 47 and is supported outwardly by 13 plunger bearing 71 on the. opposite side of a plunger drive 55.

Plunger drive 55 in simplified form, may be a friction wheel 74 driven by motor 75 as shown in Figs. 7, 9 and 11 opposed by an idler wheel 76 pressed against plunger 54 by spring 77. Motor 75 is reversible and under control of the loading operator. Plunger 54 is sectional, having a threaded end 79 capable of making connection with additional plunger sections. Suflicient sections are provided to insert the plunger entirely through an air channel 12 when required.

In the initial loading of the graphite cube 10, loading is started with the more central air channels, and 68 slugs are placed in the guide tube 47 and pushed into each connected channel. Plunger 54 is operated to push the slugs into the channel until the outer end of the first slug is at the outlet face 15. The plunger is then withdrawn, leaving the outer end of the last slug about 16 inches from the inlet face 14, for purposes explained later.

Proceeding outwardly and preferably concentrically, additional channels are loaded, meanwhile checking the neutronic activity of the reactor. As the activity increases as the loading approaches a critical size, that is, the size where the reproduction ratio will be exactly unity, the approach to critical size can be predicted by extrapolation of observed neutron density values with respect to the volume of the cube loaded with uranium. The neutron density values can be obtained from an ionization chamber 8% (Fig. 2), for example, using any well known indicating circuit, or by measuring the radioactivity of indium, foils, for example, induced by neutron irradiation when inserted into the reactor.

As the critical size is approached, the control rod 41 is inserted deeply into the reactor to prevent a self-sustaining chain reaction; and loading is continued until the desired maximum reproduction ratio of, for example, from 1.005 to 1.006 is attained. This ratio can be checked by removal of the control rod and measuring the time taken by the reactor to double its neutron density. From this period, the reproduction ratio can be mathematically computed.

When the desired number of channels are loaded the active core of the reactor may contain from 34 to 50 tons of urnaium, and will be ready for operation. Graphite plugs for the unused air channels may be'loaded in a manner similar to that described for the uranium slugs.

It will be noted that on 4 sides of the graphite cube excess graphite will be present. On the fifth side, i. e., at the inlet face, graphite will also extend 16 inches beyond the uranium. On the remaining side, i. e., the outlet face, no graphite extends beyond the uranium. Thus, sides of the active portion (the uranium bearing portion) are surrounded by graphite. This graphite constitutes a reflector and reduces the amount of uranium required to reach critical size.

Graphite and other neutron moderators can be used as reflectors around the active portions of a ,neutronic reactor, as such materials scatter neutrons passing through them and thereby change the direction of the neutrons. Statistically, due to the scattering action, some of the neutrons that otherwise would be lost to the active portion are returned-to it, thus reducing exterior loss and thereby reducing critical size. In graphite-uranium reactors having roughly a spherical shape for the active portion there can be a reduction of 3 to 3 /2 feet in diameter of the active portion when it is surrounded by a layer of graphite from 3 to 5 feet thick. In the present instance the reduction in size is slightly less as only 5 sides of the active portion are surrounded with the reflecting layer.

It will be noted that more air channels are originally provided in the graphite cube than are required for the disposition of the uranium containing slugs. The manner in which the uranium is disposed in the graphite is known as the geometry of the system, and this geometry may be varied to suit conditions, for example, the system as described may be said to have rod geometry with cylindrical loading as the end to end relationship of the slugs forms, in effect, a long rod of uranium. Further, the use of full length rods in all channels concentrically arranged around the center of the active portion, gives that portion a generally cylindrical shape.

Lump geometry may also be used in the reactor, and is accomplished by separating the individual slugs by graphite rods, for example 4 inches long. In this case additional channels will be loaded to bring the amount of uranium almost to the same value as that used in the rod geometry, although the lump geometry is slightly more eflicient than rod geometry.

Furthermore, a sperical shape is somewhat more efficient for the active portion than a cylindrical shape, and can be approached with either rod or lump geometry by shortening the extent of uranium loading in the channels as the peripheral concentric layers are added, until an approximate sphere is formed. A suitable compromise is the use of one or more concentric rings of outer channels around a cylindrically loaded central portion, with only half the number of slugs in the outer channels and disposed with the ends of the slug row equally distant from the ends of the fully loaded rows. Other arrangements will be apparent to those skilled in the art, and can be attained in the structure described because of the provision of more channels than are to be used.

After the reactor is loaded the fan is started and the control rod is withdrawn until a rise in neutron density to a desired power output is attained. The control rod is then moved forward into the reactor until a neutron balance is obtained with the reproduction ratio at unity, thus maintaining the chain reaction at the desired operating power. Small variations from the unity reproduction ratio will occur during operation, due to temperature variations of the cooling air, and to change in barometric pressure and to minor variations in air pressure delivered by the fan. However, such variations are compensated by slight inward or outward corrective movements of the control rod, either by hand in response to indicated variations in neutron density, or automatically by direct linkage of the control rod to the output of the ionization chamber 80. However, such automatic control is no part of the present invention.

During operation of the reactor, the air passing through the reactor becomes radioactive due to the fact that itis subjected to intense neutron irradiation. investigation has proved that the only significant radioactivity present in the air after having passed through the operating reactor is that of Argon, having a minute half-life. At high power output, however, this activity may be present in the exhaust air to the point that it would be biologically dangerous to operating personnel unless highly diluted during its radioactive decay. For that reason the air passing through the operating reactor is not delivered to the atmosphere at ground level but is exhausted at a substantial distance above ground, such as for example from the top of a 200 foot stack, with the result that when and if any of the radioactive Argon reaches ground level it is so dispersed in and diluted by fresh atmospheric air that less than 0.1 Roentgen per day will be received by any persons on the ground, either close to or away from the stack. Thus, theair is passed only once through the reactor and does not acquire excessive radioactivity.

The sole presence of the above noted type of radioactivity, however, is predicated on the use of the jackets sealed around the uranium bodies. Aluminum is preferred for the jackets, as aluminum has a relatively low neutron capture capability and, consequently, can be used in substantial amounts in the reactor without absorbing or capturing suificient neutrons to prevent'a self-sustaining chain reaction from occurring. Aluminum also corrodes very slowly in hot air.

The jackets have two functions, both of which reduce radioactivity of the cooling air. The first is'to prevent oxidation of the uranium. While considerable oxide could be tolerated in the reactor itself if the uranium were to be used in unprotected condition, some of the oxide particles would 'be picked up by and exhausted in the cooling air. As these particles would be highly radioactive and relatively heavy, the proper dispersal thereof would be a difficult problem. a

In addition, if operation should be accomplished in the reactor with bare uranium, fission fragments from nuclear fissions occurring on the surface of the uranium would also be projected into the air stream and would be carried out by the air stream. These fragments. are exceptionally radioactive and could not safely be. dispersed into the atmosphere. When jackets are used, these fragments are stopped by the jackets and cannot enter the air stream. 7

Thus, the jackets prevent corrosion of the uranium and prevent fission fragments and corrosion products of uranium from entering the air stream. As fission fragments will pass through an extremely small hole, one method of monitoring the reactor for jacket failure, such as for example, a weld crack is to monitor the radioactivity of the stack gas. If the stack gas shows any substantially radioactivity other than that'of Argon then it is clear that a jacket failure has occurred. Such monitoring of the stack gas is normally a routine procedure of an operating air cooled reactor, but forms no part of the present invention.

After operation of the reactor for a sufficient length of time for an amount of 94 to be created sutficient for chemical separation, such as for example 100 days at 500 kilowatts, the reactor is shut down by inserting the control rodfully into the reactor. After about one half hours wait, during which all delayed neutron omission will have ceased and the more violent radioactivity subsided, the reactor may be unloaded.

The unloading may be accomplished in two ways, either by using the plunger to push the slugs out of the channels so that they fall by gravity out of the outlet face 15, or by using the plunger to insert new slugs in the channels, each slug so inserted pushing an irradiated slug out of the outlet face 15. In the first instance the graphite cube will be left empty after unloading. In the second instance the insertion of new slugs is continued until all or a predetermined part of the irradiated slugs are out of the reactor, having been replaced by fresh material. Thus, the reactor is left ready for the next run. Under ordinary circumstances the latter manner of unloading is preferred.

In either case the slugs drop by gravity from the outlet face into outlet chamber 31, falling on to two angularly disposed pad plates 90 positioned to intersect the falling slugs, in the bottom half of outlet chamber 31 as shown in Figs. 1 and 2. The two plates slant to a centrally disposed outlet pipe 91 extending downwardly through foundation 11 and provided with spaced valves 92 and 94. The slugs fall by gravity into pipe 91 above valve 92.

Outlet pipe 91 opens into a lower coffin chamber 93 that in turn connects with a tunnel 96 carrying car tracks 97 on which a cofiin car 99 may be moved by means of cable 100. Coffin car 99 supports a plurality of slug coffins 101 in position to be successively positioned beneath the lower opening of pipe 91.

Valves 92 and 94 are operated by means of rods 102 and 104, respectively, from behind a heavy lead shield 105, as shown in Fig. l. A crane 106 is used for placing cofiin caps 107 on each cofiin after it has been filled with irradiated slugs.

Before unloading is started, both valves 92 and 94 are closed, and the upper'portion of pipe 91 is filled with water from water inlet pipe 109. A proper water level is maintained above valve 92 by water outlet pipe 110. The air circulation is maintained, although it may be re- V r r 16 duced to about 25 percent of the operating value. Slugs are then pushed. out of'the reactor to fall on topad plates andthen. roll by gravity into the water in the upper part of outlet pipe 91., Y Y Y In' order that there be no material damage .tothe jacketing of the slugs, plates 90'are..prefer'ably padded with a soft material that does not deteriorate under neutron irradiation, and that will be able to withstand the slug impacts. A satisfactory pad has been. found to be 4 inch cotton duck on felt laid on wood backed by steel. I Combinations of various synthetic elastic materials have also been found satisfactory.

After a number of slugs have been collected above valve 92' sufficient to fill a coffin 101, unloading is stopped and valve 92 is opened, permitting the slugs. and the water to drop through thevalve and remain inthe space between valves 92 and. 94." Valve 92 is then closed, the water level re-established and unloading continued. In the meantime, valve 94 is opened permitting the slugs and water to fall into one of the cofi'ins 101. The car is then moved to register the opening of the next coflin with theend of the outlet pipe and the first coflin is capped. The procedure is continued until all of the irradiated slugs are in cofiins. These coffins may then be taken to a soaking pit (not shown) to remain until the radioactivity has decayed to a point where the slugs can be submitted to chemical removal of the products formed therein by irradiation. After days operation the aging period may be about 30 days. I

Re'movalof the irradiated slugs under the conditions specified is performed for two reasons. Firstly, the slugs are so highly radioactive that they cannot be safely approachedby personnel without adequate shielding being interposed, and, 'secondly,.for,some time after removal from the reactor this radioactivity is so intense that self absorption'of the emitted radiations causes self heating of the slugs. possibly. sufficient to melt the slugs if not cooled in some manner. By unloading during maintenance ,of'the air stream, by dropping-the slugs at once into water, and by keeping the slugs in water until the more violent radioactivity has subsided, melting is prevented, as the slugs are cooled as they boil the water in-which they are'immersed. The slugs are then stored or aged under water-until ready for chemical treatment, as for example, for thirty days.

It is to be noted that reactor operation at 250 to 500 kilowatts heat equivalent has been mentioned. These powers, however, are in'no way maxima as theoperating powers are dependent solely on the air supply available and the permissible maximum slug temperatures. Ordi narily slug temperatures of from 100 C. to slightly over 400 C. are permissible without'special treatment of the uranium bodies before jacketing. In case, however, it is desired to operate the'reactor with a maximum slug'temperature above about 430 C., then precautions should-be taken to remove occluded hydrogen from the uranium to less than .0002 percent in order to prevent swelling of the jackets by release of this hydrogen; At temperatures below 430 C. no swelling occurs from the release of occluded hydrogen, as uranium hydride is formed, preventing rise'of internal pressure.' Above 430 C., however, internal pressure from released hydrogen may, with unprocessed uranium, swell the jackets. The occluded hydrogen, however, is easily removed from the uranium bodies by heating them to a temperature of from 580 C. 'to'600 C. under continuous evacuation by a vacuum pump until equilibrium is reached. Theuranium bodies are then cooled 10 to 20 hours, preferably in an argon atmosphere, and then jacketed. When processed in this manner no swelling-will occur at any temperature. Thus, if desired, the reactor can be operated at powers of from 1000 kilowatts to 5000 kilowatts, when the proper amount of air is supplied, and the hottest slugs permitted to rise in temperature to from 400 C. to 500 C.

Having described the reactor with one type of loading,

17 changes in loading required by operating conditions can be made, as illustrated by the diagrams shown in Figs. 12 to 16.

In Fig. 12 and 14, a loading is shown, for example, where only minor amounts of intermittent power are required. For example, 800 tubes can be loaded providing an excess r of just enough to reach the required power, the power being lowered or the reaction stopped before the appearance of the xenon efiect. The solid black areas indicate uranium loading.

If, however, even higher powers are desired, the moderator blocks can, for example, be almost completely filled up, as shown in Figs. 13 and 15, as for example 1600 channels can be loaded to provide an excess r of about .01. This will permit, other conditions being consistent, a large continuous power output with some excess r available after xenon poisoning is taken care of, for control of minor variations by the control rod.

Thus by making the moderator blocks large, so that at least suflicient channels are provided to allow enough channels to be loaded for the maximum power contemplated and for maximum expected poisoning both short and long term, the reactor structure except for the loading, need not be changed.

Furthermore, the moderator blocks pierced with more channels than may be initially required, permits loading to be made in various patterns.

The loading in Figs. 12 and 14, and 13 and 15 is what is known as cylindrical loading, in that the active portion defines a cylinder. loading may be desirable, as shown by the diagram of Fig. 16, in cross section. The longitudinal section will be the same as shown in Fig. 13. It will be seen that the cubical loading can also be contracted or expanded in size to meet operating conditions. By progressively shortening peripheral rods an approximation of a spherical overall shape of the active portion can be obtained.

By so providing space for increasing or decreasing the size of the active portion, all presently known operating conditions requiring change in critical or operating size can be obtained. Thus, when the neutron reproduction ratio decreases due to impurities formed during operation of the reactor some plugs of graphite are removed and additional bodies of thermal neutron fissionable material are added to the passages thus provided.

However, cubical or parallelopiped While the theory of the nuclear chain fission mechanism in uranium set forth herein is based on the best presently known experimental evidence, the present invention is not to be bound thereby, as additional experimental data later discovered may modify the theory disclosed. Any such modification of theory, however, will in no way affect the results to be obtained in the practice of the invention herein described and claimed.

What is claimed is:

The method of sustaining a neutronic chain reacting system comprising operating a reactor containing a solid neutron moderator of graphite having a plurality of passages adapted to receive bodies of thermal neutron fissionable material, and containing sufiicient bodies of thermal neutron fissionable material in certain of said passages and solid moderator plugs of graphite in other of the passages to achieve a chain reaction having a neutron reproduction ratio greater than unity, removing some plugs of solid moderator from some of the other said passages, and adding more bodies of thermal neutron fissionable material, said additional bodies being added to the passages thus provided when the neutron reproduction ratio decreases due to impurities formed during operation of said reactor and said additional fissionable material being suflicient to maintain the neutron reproduction ratio at greater than unity.

References Cited in the file of this patent UNITED STATES PATENTS 2,708,656 Fermi et al May 17, 1955 FOREIGN PATENTS 114,150 Australia May 2, 1940 114,151 Australia May 3, 1940 861,390 France Oct. 28, 1940 233,011 Switzerland Oct. 2, 1944 OTHER REFERENCES Power, July 1940, pp. 56-59.

Smyth: Atomic Energy for Military Purposes, Aug. 1945.

Goodman: The Science and Engineering of Nuclear Power, vol. 1, page 275, Addison-Wesley (1947.)

Kelly et al.: Phy. Rev. 73 1135-9 (1948.)

Images