With minor exceptions, all design effort has been concerned with components and layout within the reactor cell. A layout was made of the drain tanks with interconnecting piping for all three salt circuits. Figure 9.1 is a plan view of the general arrangement. This preliminary layout was made to get an idea of the magnitude of the "tank farm" problem, and it appears that a satisfactory solution to this question will not be overly difficult.

Again, in order to get a preliminary evaluation of the nature of the communication problem between reactor plant and chemical processing plant, one method of effecting this interconnection was developed. Basically this plan involves a batch transfer between plants, the batches being moved between the two plants by gas pressure. Transfer to and from the chemical plant is in control of the reactor plant operator. Although the system of communication between reactor and chemical plant has not been worked out in any detail, it appears that the communication scheme can be done with certainty and safety.

We believe that the optimization of the reactor is now accurate enough to warrant serious detailing of the design. A basic concept of fuel cell, blanket, plenum chambers, and reflector thickness for a reactor with an average power density of approximately 40 kW/liter has reached a firm status. It is this module which will receive detailed attention. For comparison purposes a module of equal power but having twice the average power density was calculated. Table 9.1 shows the comparison between these two reactors regarding performance and economy.

The blanket and fuel heat exchangers are also considered to be in a sufficiently firm design to justify detailing of these components. Both heat exchangers have undergone significant changes in physical arrangement and will be discussed in some detail later.

The arrangement of components in the reactor cell as previously reported posed some very difficult stress problems. A new arrangement of these components appears to have removed or greatly lessened these problems. Analyses are currently in progress which will indicate whether or not further changes are required.

Operational experience with the MSRE has indicated clearly that the handling of fission gases from a molten-salt reactor requires very careful design. The gas system for the breeder reactor is a most important part of the plant, and a start has been made in the design of this system. A flowsheet which seems to describe this system has been developed. Also, analytical studies of gas stripping efficiencies and parameters which determine or define the nature of the gas sparging system have been initiated.

Associated with the gas handling problem is the matter of draining the reactor in the event of a pump stoppage. Modifications in the primary heat exchanger indicated by the sparge gas handling and core drainage influenced us to put an overflow line from the sump tank of the pump into a fuel dump tank. These changes in the primary heat exchanger have been rather extensive, although the basic heat exchanger remains very much as the original concept.

Since the reference design is based on the modular reactor concept, the flowsheet of Fig. 9.2 is for one module. The number of the various components in this module is shown. The steam system is shown very sketchily for the 1000-MWe plant, and an indication of a steam header fed by the four modules is shown.

It will be noted that the maximum coolant-salt pressure is somewhat higher than has been reported previously. This pressure occurs at the discharge of the coolant pump, and, of the 260 psi at this point, 110 psi comes from a gas overpressure at the suction side of the coolant pump.

We believe it to be imperative that, in the event of a failure in any part of the salt systems, the pressures should be such that blanket salt would flow into fuel salt, and coolant salt would flow into blanket and fuel salt, depending on the location of the failure. This made it necessary to put a high bias pressure on the coolant salt. The pressures shown on the flow diagram represent design-point pressures and take into account static heads and dynamic pressure drops in all cases.

In previous flowsheets a separate coolant pump circulated salt through the reheater. We now bypass the discharge of the main coolant pump through a salt-flow regulating valve to the reheater. This valve is little more than a variable orifice, and we believe the use of such a valve is justified.

Because of the relatively high melting point of the blanket salt, it is desirable to connect the coolant circuits of the fuel heat exchanger and the blanket heat exchanger in series. Coolant salt leaving the fuel heat exchanger at a temperature of 1111°F enters the blanket heat exchanger, where it picks up 14°F to attain the design-point temperature of 1125°F. The coolant-salt line, carrying 37.5 ft3/sec, is a large pipe (~20 in. in diameter) and therefore rigidly connects these two heat exchangers. Since the blanket heat exchanger is located high in the reactor cell while the fuel heat exchanger must be located below the bottom of the reactor vessel, there is a shift in the center of gravity of the coupled heat exchangers when they are filled with salt.

In the original layout of the reactor cell which had the reactor, fuel heat exchanger, and blanket heat exchanger in a triangular array, a twisting moment resulted about the axis of the fuel connection between the reactor and the fuel heat exchanger. Also, the method of suspension mounting of the heat exchangers from constant load cells could not provide accommodation for the differential thermal expansion of the system under all conditions.

A new layout of the reactor cell shown in Fig. 9.3 was made. In this configuration the three components, reactor, fuel heat exchanger, and blanket heat exchanger, are mounted in line. The coolant lines connecting the fuel and the blanket exchanger are made concentric, and the shortest possible spacing between exchangers is used. The connection from fuel heat exchanger to reactor is as short as possible, using concentric piping for this as we have been doing previously.

The connection from the blanket heat exchanger to the reactor vessel has been changed. Instead of the short concentric pipe used previously, the inlet and outlet lines from reactor vessel to blanket exchanger are run independently. These lines are purposely made long and flexible by running them through four 90° bends. Thus we intend to decouple the interconnection between the reactor and the blanket heat exchanger by use of flexible piping.

The three components, reactor, fuel exchanger, and blanket exchanger, are rigidly interconnected then by only one link for each; the reactor connects by the concentric fuel line to the fuel exchanger, which connects by one concentric coolant line to the blanket exchanger. At design point these connecting lines are very near the ambient cell temperature of 1000°F. The fuel line is actually at 1000°F, and the coolant line is very near 850°F. With the three cell components in line, a different plan of support is now proposed. Let us point out that this arrangement has not yet been analyzed for stresses, so we cannot be sure it is satisfactory in its present conception.

The plan now involves mounting a stainless steel bed plate in the reactor cell in such a way as to be tied at one end but free to expand with cell temperature. This bed plate is located horizontally below the bottom of the reactor vessel and is mounted through necessary insulated supports to the load-bearing concrete of the cell. The reactor vessel and both heat exchangers are mounted rigidly to the plate as shown in Fig. 9.4. We believe that this general arrangement will work and that any overstressing found by analysis of the system can be corrected without departing drastically from the concept.

Vertical growth of all components can take place independently, and the only accommodation needed is a low-temperature sealing bellows at each pump-drive penetration through the cell membrane. Any control-rod drive mechanism would involve a similar bellows seal.

The treatment of the reactor cell has not received any attention, but it is certain that some form of radiation shield must cover the structural concrete and this shield must have cooling. Inside the radiation shield, between it and the furnace of the cell, there must be some thermal insulation.

The reactor is shown in the elevation drawing of Fig. 9.5. The reactor vessel has a cylindrical body 12 ft in diameter and approximately 17 ft high with dished heads on top and bottom. A heavy supporting ring essentially the diameter of the core carries legs which weld to the bed plate and provide the mechanical mounting for the vessel. From the center of the bottom head a concentric fuel line communicates with the fuel heat exchanger. This fuel line has an outside diameter of 24 in. with a concentric return line 16 in. in outside diameter. Fuel flows in these lines at a rate of approximately 18 fps.

Inside the reactor vessel are dished heads forming plenum chambers for distribution of the fuel cells which rise from these heads. These plenum chambers are removable from the reactor vessel, being held in place by a flanged mounting ring equipped with clamps. This flange has not been completely designed, but we are considering the use of flat graphite gaskets to prevent bonding of the flange surfaces under the high temperature and rather high loading pressure obtained at design conditions. The inner head communicates with the inner concentric line through a slip joint which permits movement of the head relative to the pipe. Such movement results from thermal expansion and must be accommodated. In addition, the slip joint allows removal of the core when the flange seal is broken.

The core of the reactor is made up of 336 cylindrical graphite fuel cells mounted as close together as tolerances permit to form essentially a cylindrical array approximately 8.3 ft in diameter. The graphite cells are extruded cylinders with center holes 1 1/2 in. in diameter, surrounded at 120° angles with three 7/8-in.-diam holes. At the top of each cell there is a graphite cap machined to provide a smooth communication between the four holes of the cell. Figure 9.6 shows the arrangement of one of these cells.

At the bottom of a cell, the graphite cylinder is joined to a transition nipple. This joining is done by a combination of a short threaded engagement for mechanical attachment and a rather longer brazed joint for leak-tightness. The connection to this nipple is a matter of development, and work is proceeding to establish the best practice for effecting this connection. The lower end of the nipple is threaded to screw into tapped holes in the upper curved head.

From smaller holes in the inner head, concentric with the upper one, smaller nipples extend through the upper head. These nipples are also attached to the inner head by a pipe thread, each one being screwed into place before attaching the fuel cells above.

As a fuel cell is put into place, the 1 1/2-in. center hole is slipped over the inner nipple. This engagement is a slip joint and allows for screwing the cell into place, and in operation it permits the necessary differential expansion between the cell and the inner nipple. After all cells are in place, the entire core can be leak-tested, the graphite-to-metal transition joints having previously been tested individually. When the assembled core has been tested, it can be lowered into the reactor vessel and the flanged joint can be clamped.

Just inside the reactor vessel there is a graphite reflector 6 in. thick. This reflector is only on the vertical walls; there is none below or above the core. It is made of rectangular-shaped pieces with mortised joints to form a self-supporting wall resting on a ledge at the bottom of the vessel. The volume between the reflector and the vessel provides a flow passage for the blanket salt as it discharges through holes in the bottom of the distribution ring located near the top of the vessel.

We are considering the use of 4-in.-diam graphite balls in the region between the core and the reflector. These balls have holes in them so that graphite occupies 40% of the volume in this 1 1/2-ft-thick region. A retainer of perforated metal prevents the balls, which float in the blanket salt, from crowding into the top head of the reactor vessel. Instead, by floating up against this retainer, a measure of support is afforded the core structure, the floating balls tending to exert a considerable force on the fuel cells.

With this core configuration a two-fluid reactor is achieved. Fuel salt enters the outer concentric pipe at the bottom of the reactor. It flows out to the outer radius of the core and in between the upper inner dished heads, feeding salt into the nipples emerging from the upper head. Fuel passes up through the three 7/8-in. holes of each fuel cell and down through the center 1 1/2 in. hole through the inner pipe nipple into the collecting plenum. From there it flows through the slip joint into the inner concentric fuel line to the pump suction at the center of the fuel heat exchanger.

The blanket salt flows into a distributor at the top of the reactor, and, emerging through holes in the bottom of this header, it flows behind the graphite reflector to the bottom of the reactor. There it flows up through the balls in the blanket annulus and out of the suction line near the top of the reactor.

Some of the blanket salt fills the interstices between the fuel nipples and the fuel cells. Circulation of the blanket salt is governed by thermal convection. Heat generation and temperature calculations are in progress to determine whether this flow is adequate.

The center location for a fuel cell is left vacant. This is for a control rod. In Fig. 9.6 this control rod is, represented as a hollow graphite cylinder 5 in. in outside diameter, 4 in. in inside diameter, and equipped with a gas inlet at the top. By controlling the gas pressure, blanket salt can be positioned at any height within this cylinder, and this blanket volume controls the reactivity of the core. Calculations indicate a change of 1.8% in reactivity between the full and empty condition of this cylinder.

The reactor has been optimized as far as nuclear performance is concerned. There is no very sensitive parameter, and rather large changes in all dimensions can be tolerated without changing the characteristics materially. Table 9.1 gives the calculated performance of this reactor. This core volume is 503 ft3, with a fuel volume in the core of 83 ft3. This gives an average power density in the core of 39 kW/liter.

For comparison, the performance of a reactor having a power density of 78 kW/liter is given in Table 9.1. This reactor has half the volume in the core and would give somewhat better performance. At present we do not know just what the limit is on graphite radiation tolerance. The smaller reactor, whatever the graphite limit, would be expected to have approximately half the core life of the larger one, which is used as the reference design.

Some general comments can be made about the reactor as it is currently conceived. As we stated above, the nuclear performance is so insensitive to parameters that the final reactor design is likely to be close to the present concept. We would prefer to have the core removable in modules rather than having to remove it in one piece. This is particularly desirable if we have to maintain the lower power density and if it becomes desirable to make reactor modules larger than the 556-MWt size with which we are concerned at present. We do not yet have what we consider an acceptable way of making these core sub-modules, but work is continuing to try to develop one.

The thickness of the heads has not received rigorous attention. At present the top or load-bearing head is 1 in. thick. This is thicker than the load seems to require but is used to give an adequate thickness for tapping for the nipples. The analysis of all stresses in these heads is being made but has not yet been completed. The same is true for the pressure drops.

The top of the reactor as shown has a welded joint which can be ground off for removal and re-welded when a new core has been installed. Probably this access joint will be moved up outside at least one layer of shielding in the final core design. In this case some of the shielding blocks would be supported by the top of the reactor vessel. The closure point for the reactor would, by this means, be more accessible.

The heat transfer work which had been done for the Breeder Reactor study was reevaluated. This was made necessary by the results obtained from the MSRE. It was found that the salt thermal conductivity was less by about a factor of 3 than the value used in the original heat transfer calculations. This difference resulted in degradation of the overall heat transfer coefficient, necessitating an increase in heat transfer surface of the heat exchanger of approximately 20%. A new heat exchanger design with a larger number of tubes and a slight change in the baffle arrangement was developed. The parameters of this new heat exchanger are given in Table 9.2.

While the new heat exchanger has a higher fuel holdup by reason of the additional tubes, a more favorable arrangement of the header and other reductions in parasitic volumes compensated for this larger number of tubes, so that the net fuel inventory in the heat exchanger-pump complex did not change significantly from that required in the first heat exchanger used in the reactor study.

A significant change in the heat exchanger design concerns the method of flanging the heat exchanger into the coolant circuit piping. The coolant flows up around the heat exchanger and is discharged symmetrically across the outer bank of fuel tubes. In this arrangement the toroidal coolant header is replaced by a jacket around the tubes. Figure 9.7 shows the arrangement of this new heat exchanger.

After traversing the outside of the fuel tubes in a countercurrent flow pattern, the coolant salt leaves the heat exchanger through a central discharge pipe which connects to the center of the concentric coolant-salt line through a slip joint. The entire heat exchanger can be removed from the circuit by opening the large flange, cutting the large fuel pipe communicating with the reactor, and disengaging the center fuel pipe by slipping it out of the slip joint provided for this purpose. The drain and overflow lines and auxiliary pump lines must also be cut to remove the exchanger.

Removing and replacing a primary heat exchanger is a major repair. However, with the design as shown, the difficulties appear to be less formidable than if both coolant lines had to be cut and welded to perform the task. Adequate remote cutting, aligning, welding, and inspecting equipment and methods must be demonstrated for this application.

Because of the expansion of the fuel tubes, the toroidal header at the bottom of the exchanger will move relative to the vessel enclosing the exchanger. For this reason and also to permit removal of the exchanger, a drain line cannot be connected to this header. Instead, a dip line for draining fuel extends into this header from the top of the exchanger. This line connects through a freeze valve to the fuel drain tanks for the reactor. When the drain valve is thawed and the drain tank is vented, the fuel salt in the heat exchanger flows through this line into the drain tank. The gas pressure in the pump bowl is sufficient to effect a complete drain of the fuel salt from the system.

In case of a pump stoppage with the attendant necessity to drain the reactor, the following sequence takes place. Upon pump stoppage the reactor automatically drains into the pump sump tank and through the 5-in. overflow line; the salt goes to the drain tank. A gas flow into the reactor inlet plenum is furnished by the gas separator lines from the pump sump cover gas. This automatic drainage takes place in the order of a few seconds. However, the concentric lines to the reactor and the head space above the fuel tubes in the heat exchanger remain full of salt and will overheat if not cooled.

When the pump stops, the pressure in the lower header of the heat exchanger drops to the static pressure of the salt, about 14 psi. If now the gas valves in the line are opened, the stored gas will flow into the bottom of the heat exchanger, displacing the fuel into the pump sump from which it overflows into the drain tank. By proper sizing of this tank, enough fuel can be displaced from the heat exchanger to bring the level at equilibrium below the top tube sheet of the exchanger. The lines and volumes above the tube sheets will be empty of salt, and no overheating will result.

We have calculated the cooling required for fuel in the tubes of the heat exchanger under the stagnant conditions. Because of the geometry of the coolant salt system, there is convective flow of coolant salt even when no coolant pump is running. This convective flow of coolant is adequate to remove afterheat from the stagnant fuel in the tubes.

To provide cooling but to avoid overcooling requires some modification of the steam circuit of a module. We have not studied the modifications in sufficient detail to justify a discussion at this time.

The large sump tank on top of the heat exchanger is not intended to hold the drained fuel from the reactor. Because of the undesirability of installing the afterheat cooling system in the reactor cell and also in order to be able to drain the lines to the reactor, an overflow pipe was installed from the sump tank to a separate dump tank. This tank is equipped with a cooling system adequate for afterheat removal.

The sump tank on top of the heat exchanger appears to be larger than necessary inasmuch as excess fuel flows through the 5-in. drain line to the drain tank. The size of the tank is dictated by the fact that it must hold the reactor volume (83 ft3) upon starting the reactor. Fuel cannot be forced by gas pressure out of the drain tank fast enough to keep the fuel pump primed on startup. The fuel must be displaced into the sump tank and held there by gas overpressure in the drain tank until the reactor is filled by the action of the pump.

Not shown in the drawing of the heat exchanger is the gas separator and associated apparatus for handling the sparge gas for the fuel system. The plans for handling this gas are being worked out, and drawings of the equipment are not yet ready. The criteria and flowsheet are discussed in Sect. 4.

The blanket heat exchanger is shown in Fig. 9.8. This heat exchanger is similar to the fuel heat exchanger but considerably smaller and less complicated. There is no need for the gas sparge system, and there is no afterheat problem that has to be provided for here. In addition, the coolant flow is less complicated since counterflow is unnecessary. The temperature difference is modest between blanket salt and coolant, and the total rise in temperature is only 11°F as the coolant passes through. Table 9.3 gives the characteristics of the blanket exchanger.

In addition to optimization studies described above, work on MSBR reactor physics included (1) a series of cell calculations performed to examine the sensitivity of the MSBR cross sections and reactivity to various changes in cell structure and composition, and (2) several two-dimensional calculations of the entire reactor. All of the calculations were based upon the most accurate description of the system that was available as of January 1, 1967. The graphite-moderated portion of the core was 10 ft in length and 8 ft in diameter, contained ~0.2 mole % 233U, 27 mole % 232Th, and had a fuel volume fraction of 16.48% and a fertile volume fraction of 5.85%. The reactor had a 1.75-ft-thick radial blanket consisting of 60% fertile salt and 40% graphite, surrounded by a graphite reflector 6 in. thick. The top axial blanket was 1.5 ft thick and contained 60% fertile salt and 40% graphite. The bottom axial blanket was 1 ft thick and contained 3.18% Hastelloy N, 16.48% fuel salt, and 80.34% fertile salt. A number of structural details below the lower blanket were included in the two-dimensional calculations.

Figure 9.9 shows the geometry of a cell in the graphite-moderated core. Graphite dowels of appropriate size are located at the six corners of the cell to yield the desired fertile salt volume fraction. The cell calculations were performed with the code TONG and involved varying (1) cell diameter, (2) fuel distribution (i.e., fuel separation distance, s), (3) 233U concentration, (4) 232Th concentration, (5) fuel volume fraction, and (6) fertile volume fraction. Each of these parameters was varied separately while holding the others constant. Table 9.4 and Fig. 9.10 show the effect on reactivity of varying the parameters. The variations are shown relative to a reference cell which had a diameter (flat to flat) of 3.156 in. and a fuel separation distance, s, of 1/8 in.

Table 9.5 and Fig. 9.11 show information about the flux distribution for case 3. Table 9.5 shows the ratio of the average flux in the fuel to the cell average flux, the ratio of the average flux in the graphite to the cell average flux, and the ratio of the average flux in the fertile salt to the cell average flux for the epithermal and fast flux ranges. Figure 9.11 shows the thermal flux distribution in the cell. The results in Fig. 9.11 are based upon an annular approximation to the cell of Fig. 9.9.

The gas handling system for the MSBR serves several functions in the operation of the plant. These include: supplying helium for purging and pressurization of the gas spaces; rapidly removing 135Xe from the salt; transporting, removing, and storing the radioactive fission product gases; and processing the helium for recycle into the gas supply or disposal to the atmosphere. A scheme for accomplishing these functions is described below.

Preliminary studies of the migration of 135Xe to the graphite in molten-salt breeder reactors indicated that it would be possible to reduce the 135Xe poison fraction to an acceptable value of ½% if a gas stripping system could be devised which would process the entire reactor fuel-salt inventory in about 30 sec. Since the reactor system salt circuit time is about 9 sec, it is required that the entire inventory be processed every 3 to 4 passes around the circuit.

The required xenon processing time could be extended by processing the fuel for removal of the precursor, 135I. However, the advantages of 135I processing are limited by the fact that 20% of the total 135Xe produced in the fissioning of 233U appears directly, and very high processing rates would be required to obtain even modest gains. For example, if the entire fuel inventory were processed for 135I removal once every hour, the xenon removal process rate would only be extended to about 80 sec, and a very high 135I processing rate would increase the xenon processing time to only about 110 sec. Therefore, processing for iodine is not, at this time, being considered as a method of removing 135Xe.

The method proposed for removal of 135Xe involves stripping of xenon-enriched helium bubbles from the fuel stream by means of a gas separator located at the heat exchanger outlet. The bubbles will be generated by injecting helium into the salt at the fuel pump suction, and transfer of xenon from the salt to the gas will be effected during passage through the heat exchanger.

The flowsheet for the MSBR off-gas system is given in Fig. 9.16. Helium is injected into the suction of the fuel salt pump and is removed by centrifugal separation at the heat exchanger outlet. The liquid-gas mixture from the separator is then fed into a cyclone separator where the entrained liquid is removed and returned to the pump bowl. The gas from the pump shaft purge and the instrument lines is combined with that from the cyclone separator and fed into the 48hr xenon holdup system. A small portion of the flow from this system is fed into the long-term xenon holdup and then through the noble-gas separator to a recycle system for supplying clean helium to the pump shaft purge and the instrument gas lines. The remainder of the gas from the 48-hr xenon holdup system is fed into the gas injector system at the fuel pump suction. It is the salt-powered gas injector that serves as the prime mover for the high-gas-flow recycle system. Some of the design criteria of the critical components are described below.

1. The gas addition rate shall be sufficient to produce about 1% void volume in the salt at the pump suction.

2. The location and manner of gas injection shall provide a balanced distribution of bubbles in the liquid entering the pump impeller. This consideration affects the bubble distribution in the heat exchanger and the bubble size, as well as the pump dynamic balance.

3. There shall be operator control over the gas addition rate into the gas injector.

4. The supply for the gas injector shall be taken from the outlet line from the 48-hr xenon holdup helium recirculation system. There shall be a backflow preventer arrangement at this gas supply point to prevent salt from getting back into the charcoal beds as a result of a sudden pressure transient in either the salt or gas system.

5. The method of injecting bubbles into the pump suction shall be a fuel-salt-powered jet pump taking its salt supply from the heat exchanger discharge.

The suction pressure for the salt-powered jet pump shall be less than 10 psia at a flow rate of 6.5 scfm of helium.

1. The bubble separator shall be installed within the outer annulus of the fuel-salt line from the heat exchanger to the reactor.

2. The separator, including the entrance region, shall be kept as short as is reasonably possible, and the separator shall be as close to the heat exchanger as good design permits. The purpose of this restriction is to permit removal of the separator along with the heat exchanger and at the same time permit locating the heat exchanger close to the reactor.

3. The design of the system shall include assurance that the separator will not admit gas into the salt line to the reactor except when the fuel pump stops. Check valves or liquid submersion of the discharge will not be used to prevent backflow, since the reactor drain scheme will use the separator outlet as a gas source.

4. The bubble fraction of the salt as it enters the reactor shall be less than 0.1% during steady-state operation.

5. The fuel line outside the gas separator shall be free of protrusions all the way to the heat exchanger outlet. Equipment for cutting and welding the 24-in.-diam fuel line will occupy this space during some maintenance operations.

6. The gas outlet from the bubble separator and the gas line to the pump bowl shall stay within the salt line if possible. This will reduce the number of penetrations through the pipe wall.

7. A modified gas cyclone separator shall be an integral part of the bubble separator discharge stream. The liquid would be discharged to the pump bowl, and the gaseous discharge would go to the xenon holdup system.

The first stage of the high-flow 48-hr xenon holdup is a gas volume holdup whose several functions are listed below.

1. The first stage shall be a cooled volume for the decay of the very short-lived gaseous fission products. A reentrant tube system similar to the one in the fuel drain tank of the MSRE will be investigated for providing the cooling.

2. There shall be a provision for a final demisting of the gas as it comes from the cyclone separator.

3. Since there will be a total of more than a kilogram of fission products produced each day in the reactor complex, and since a sizable fraction is involved in the short-lived gaseous products, a method of collecting, cooling, and conveying to a disposal area must be included. The collecting device should pass only gas which has resided in the volume holdup for at least the specified minimum transit time.

4. The volume holdup system shall be sized to reduce significantly the heat load and particulate loading rate from short-half-life fission products in the first stage of the charcoal bed, which is immediately downstream.

Many of the remaining components in the off-gas handling system are reasonably well understood, and, while they must be designed and ultimately tested, it is believed that they will offer no critical problems. Some of these are listed below:

1. High-gas-flow charcoal bed design.

2. Biological charcoal bed design (low flow).

3. Noble-gas separator and disposal system. This includes 85Kr and tritium removal from the helium which is to be returned to the recycle system.

4. Helium compressor for recycle of clean helium.

5. Gas sampling system for purity control and for surveillance of the chemical condition of the salt.

6. Instrumentation and controls for operating the gas system.