Innovation leads the way to attractive inertial fusion energy reactors–Prometheus-L and Prometheus-H

Fusion Engineering and Design 25 (1994) 125-143

ELSEVIER

Fusion Engineer!ng and Design

Innovation leads the way to attractive inertial fusion energy reactors Prometheus-L and Prometheus-H Lester M. Waganer McDonnell Douglas Aerospace, P.O. Box 516. St. Louis. MO 63166, USA

Abstract Two conceptual inertial fusion energy (IFE) reactor design studies employing innovative system concepts have been completed for the US Department of Energy. These concepts enable power plants with inertially confined plasmas to be economically competitive with other energy sources and provide safety and environmental advantages. A K r F driver employs 960 electric discharge laser amplifiers to enhance driver reliability and target illumination with a loss of one or more amplifiers. A non-linear laser architecture uses Raman accumulator cells to combine and enhance the beam quality and stimulated Brillouin scattering cells for beam compression. Optical delay switchyards maximize the utilization of beam energy to provide proper beam pulse forms to the target. Grazing incidence metal mirrors are the final optical elements that employ a high rigidity SiC support structure and graded thickness aluminum reflective surface material to obtain a life-of-plant optical element with a direct line-of-sight to the target within 20 m of the target. Sixty of these laser beamlines symmetrically illuminate the direct-drive target. A performance and economic systems code determined the optimum laser beam energy as 4 MJ corresponding to a target gain of 124. When pulsed at 5.65 Hz, the fusion power is 2807 MW. To reduce the cost of traditional, lengthy, multiple-beam heavy ion drivers, a single-beam LINAC driver with storage rings was adopted. A charge state of two was used to shorten the length and cost of the driver. An ion energy of 4 GeV reduced the number of beams. The LINAC is rapidly pulsed 18 times. Pulses are contained in storage rings and combined to form 2 prepulse and 12 main beams. These are subdivided to illuminate the indirectly driven target from two sides. Triplet coil sets ballistically focus the beams on the outside of the blanket. Channel transport is proposed to deliver the beams in two 6 mm diameter channels to the target. A total beam energy of 7 MJ is delivered to the target to obtain a gain of 103 and fusion power of 1818 MW at 3.54 Hz. A common reactor design is used for the laser and heavy ion beam systems. Low activation SiC material is used for the first wall and blanket systems. The first wall is protected with a thin film of liquid lead that is evaporated by each microexplosion and recondensed between explosions, thus providing protection and vacuum pumping of target debris. A lithium oxide breeder is cooled with low pressure, high temperature helium that minimizes stored energy and improves system safety and activation. The plant Level of Safety Assurance is one, and waste disposal is class C or better. Double-walled steam generators maintain low tritium permeation to the environment. High-temperature helium and first wall lead coolants are used with a 42% efficient, advanced Rankine cycle to deliver a power output of 1000 MW for both plant designs. All systems were optimized to deliver the lowest cost of electricity.

1. Introduction In 1990, the D e p a r t m e n t o f E n e r g y ( D O E ) , Office o f E n e r g y Research, c o n t r a c t e d with an Elsevier Science S.A. SSDI 0920-3796(94)00050-H

industrial design t e a m to p e r f o r m two c o n c e p t u a l design studies o f inertial fusion energy ( I F E ) c o m m e r c i a l p o w e r p l a n t s [1]. T h e t e a m was led by M c D o n n e l l D o u g l a s A e r o s p a c e ( M D A ) a n d in-

126

L.M. Waganer [ Fusion Engineering and Design 25 (1994) 125-143

cluded Canadian Fusion Fuels Technology Project, Ebasco Services Inc., KMS Fusion Inc., SPAR Aerospace Ltd., TRW Space and Electronics Group, and the University of California at Lost Angeles. Dr. Mohamed Abdou provided consulting services to the team. The DOE formed a Target Working Group (TWG) to provide normalized, unclassified target data to the team. An Oversight Committee was appointed to advise and assist the team in conduct of the study. These studies were to advance the state-of-theart for IFE physics, technology, and engineering. The studies were also encouraged to seek innovative design approaches and cost-effective solutions while improving the safety and environmental impact aspects of the reactor designs. The program objectives were defined in the contractual statement of work (SOW). The main objectives are summarized below: adopt common ground rules for design development and comparison tasks; conduct parametric trade studies using developed system codes; develop conceptual designs for two IFE reactor power plants; estimate plant capital and operating costs; assess critical technical issues and define research and development requirements; compare two IFE reactor designs. The team was to choose the two most attractive IFE drivers to be used in the conceptual designs. The KrF laser and the heavy ion beam drivers were chosen to be examined. A systems code was developed to model completely the drivers, the reactor plant, and the balance of plant (BOP) elements. The modeling included physics, system performance, capital cost, and operational costs. The design components were developed, modeled, and then incorporated into the code. The code allowed system trade studies to be conducted at the plant level to determine the effect on the cost of electricity (COE). The systems code evaluated all design options on the basis of performance and economics, while the system designers and analysts assessed the more intangible factors. The final study task evaluated quantitatively the two Prometheus designs, based on (1) physics feasibility, (2) engi-

neering feasibility, (3) economics, (4) safety and environmental, and (5) research and development requirements. As stated above, the charge given to the industrial team was to advance the state-of-the-art in innovative ways to stimulate the fusion community and the DOE. One of the purposes of fusion reactor conceptual designs is to push the frontiers of an integrated concept within the technical community. New innovations have been formulated and proposed, and perhaps even demonstrated on a small scale. These design teams are to assess the viability of these new ideas in the integrated context of a complete reactor. Will it work as well as thought? Are there other systems that suffer or benefit? Does this impose an environmental impact? Are there sufficient resources? The STARFIRE Commercial Reactor Conceptual Design [2] pushed forward such frontiers, for example in the use of current drive. When these IFE Reactor Design Studies were begun, there was a general conception that inertial drivers were destined to be large and expensive. Only at very large power production sizes (i.e., greater than 5 GWe) would the power plants become economically competitive. This was the environment in which the team began to search for innovative concepts to confront those conceptions. There have been several prior studies which influenced the direction of the Prometheus designs. The encouraging results of the Heavy Ion Fusion Systems Assessment (HIFSA) [3] demonstrated the performance benefits of a heavy ion LINAC coupled to direct and indirect IFE targets. Even though the multiple beam driver used in that study used higher charge states to reduce the length of the driver, the driver system costs remained too high to be competitive with other alternatives. Thus solutions were sought to reduce significantly the cost of the driver system. Direct comparisons of the performance and economic results from prior studies were avoided generally because each study had specific and unique guidelines to be satisfied. Extrapolating and interpreting prior study findings in light of the present study tends to distort comparisons. The Prometheus studies [1] and the OSIRIS/SOMBRERO studies [4] were conducted in the same

127

L.M. Waganer /Fusion Engineering and Design 25 (1994) 125-143

time frame. They collaborated to adopt common design ground rules. Furthermore, economic guidelines for these current IFE reactor design studies were coordinated with the most current magnetic fusion energy ( M F E ) reactor design study, ARIES I [5] and subsequent ARIES designs, to enable detailed performance and economic comparisons. These normalizations would allow comparison at a basic level. Design philosophy, level of technical optimism, and technical basis yielded significantly different approaches to the same problem, even when comparing the most current IFE designs. Within the Prometheus study, parametric trades studies based on the COE were conducted on all major system options such as multiple beam versus single beam LINACs, driver energy level, direct vs. indirect targets, and other driver-specific options. Many other system options such as first wall and blanket approaches were evaluated as to technical risk, design margin, capital and operating costs, lifetime, safety, and environmental impact. As an example, the Prometheus study examined and evaluated prior IFE reactor wall protection concepts in groups of granular solid protection ( C A S C A D E [6]), thick liquid jet ( H Y L I F E [7], SENRI [8], and H Y L I F E II [9,10]), liquid film ( H I B A L L / I N P O R T [11], LIBRA [ 12,13]), and gas or magnetic protection (SOLASE

[14], SIRIUS [15]). For reasons detailed in Section 5, the liquid lead first wall and the Li20 solid breeder/SiC structure best satisfied the study guidelines and requirements for PULSAR.

2. Target selection There is a desire to believe that the target, the laser driver, and the reactor chamber can be designed and optimized separately. However, their designs and performance are intimately tied together. These studies were constrained in that the design and performance of the targets were not within the scope of these studies. All target performance data were provided to the design team by the TWG. Both direct-drive and indirect-drive targets were considered for the laser driver whereas only the indirect targets were available for the heavy ion beam driver. Fig. 1 illustrates the available performance data for the available targets. The gain for the laser indirectly-driven (ID) target is considerably less than that for the direct-drive targets (constant spot (SC) and zoomed spot (ZS)). The required number of beamlines is roughly the same for both direct and indirect drive targets with similar beam quality; hence, the overall system

Ion Beam with 6 mm Diameter Focal Spot For a Set of Ion Ranges

KrF Laser 200

200

- - ' ~ " 0.025 g/cm 2 0.05 g/cm 2 - - - ~ 0.1 g/cm 2 - - I I - - 0.2 g/cm 2

150

150 C=

(.9 100 I.--

1 ,

1oo

5o

50

0

2

4

6

Input Energy (MJ)

8

10

0

2

4

6

8

10

Input Energy (MJ)

Fig. 1. Comparison of baseline gain curves for KrF laser and heavy ion systems. Laser curves are for direct drive (zoomed and constant focal spot) and indirect drive targets. Heavy ion curves show the variation with ion range.

128

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125-143

performance and cost are indicative of the respective target gain throughout the laser input energy range. The zoomed spot direct-drive target also indicates a significant improvement in performance over the constant spot target. However the technological complexity of a laser focusing system required to zoom the focal spot in concert with the imploding target on nanosecond time scales was deemed to be too costly and complex to consider as a serious option. Thus the constant spot, direct drive target was chosen for the laser driver. Only performance data for the indirect-drive, heavy ion target were provided by the T W G although, in principle, an indirectly-driven target is possible. Typical data for the heavy ion target are shown in Fig. 1. A complete set of data for a range of focal spot sizes was provided. The 6 mm diameter was found to be a good compromise between improving performance at smaller spot sizes and the complexity of achieving the smaller spot sizes. The performance plotted in Fig. 1 illustrates the gains as a function of input energy for a set of ion range curves (from 0.025 to 0.2 g cm-2). These ion range values correspond to ion energy levels from 2.4 GeV to 12.4 GeV respectively. At the upper ion energy levels, the improvement in target gain becomes increasingly small at significant cost and complexity to increase the ion energy level. The final heavy ion design point was selected to be at 7.0 MJ delivered to the target using 4 GeV lead ions (ion range of 0.045 g cm-2).

3. Non-linear optic KrF laser driver K r F lasers do not easily meet the IFE target irradiation requirements. No KrF amplifier can be constructed to produce multi-megajoules per pulse. The standard, electron-beam pumped KrF amplifier pulsed power suffers significant losses in efficiency for pulse durations less than 200 ns. Thus, in order to meet the demanding IFE target irradiation requirements, multiple KrF amplifiers and pulse compression by a factor of about 100 are needed. Previous KrF laser IFE driver architectures have been based on a few very large laser

amplifiers to generate the multi-megajoule output energy and "angular multiplexing" to provide the I00 x pulse compression factor. Aside from being overly complex in a demanding reactor operating environment, this K r F IFE driver architecture does not fail gracefully--loss of a single large KrF amplifier would force the IFE reactor to shut down. Moreover, the complicated "angular multiplexing" optical structure required to accomplish the pulse compression mission is costly, must be constructed either in a vacuum or in a non-Raman active gas, is prone to optical damage, occupies a very large volume, and can be inflexible if different temporal laser pulse shapes are required. The requirement to have the desired laser performance while achieving power plant reliability with a much lower capital cost suggested investigation of non-traditional laser driver approaches. The traditional concept of K r F large-area amplifiers with multiplexed optics was considered, but the main emphasis centered on the use of smaller, more reliable electric discharge KrF laser amplifiers coupled with non-linear optical (NLO) techniques. These NLO laser architectures have been developed over the last decade by the aerospace industry for use in the Strategic Defense Initiative. The results of this technology were applied to achieve the desired performance with the required maintenance, reliability, and cost goals. The laser driver architecture chosen for the Prometheus-L design is based on moderate energy (approximately 6kJ) electric discharge excimer laser amplifier modules. These amplifiers are coupled with N L O systems for beam combination and pulse compression. This approach yields a safer and more reliable design with design flexibility and capabilities to meet the target demands for beam quality and illumination requirements. The Prometheus-L laser driver delivers 4.0 MJ of laser energy approximately at 248 nm. This energy is delivered with 60 beams symmetrically arranged to converge simultaneously on the direct drive target. The N L O system architecture incorporates optical delay lines to tailor each arriving beam to have a long (80 ns), low energy precursor beam followed by a shorter (6 ns), high energy main pulse. Beam profiles are tailored to be flat topped to illuminate fully the target from all 60 locations.

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125-143

GAS DUCT ~ OUTLET KrF

~

~

/ ~ ~uCOUSTICoR{TYP, s

~

KIFE?:T;;MLN:::M:::T

'- HELIUM COOLANT INLET "'--- KrFGASHEATEXCHANGER - HELIUM COOLANT OUTLET Fig 2. Electric discharge laser subsystem.

The beams interact with the target to compress and heat the target to fusion conditions. The primary pulse of laser energy is provided by 960 electric discharge laser power amplifiers (16 for each of 60 beamlines), each producing approximately 6 kJ in a 250 ns pulse. These amplifiers are driven from a common master oscillator through a series of beam splitters and pre-amplifiers. Fig. 2 illustrates a module of 16 power amplifiers combined for a single beamline. The KrF working gas is circulated through the lasing cavities and the waste heat is removed to maintain proper thermal conditions in the lasing cavity. This waste heat is used in the thermal conversion process to improve the effective driver efficiency. The relatively low quality beam outputs from the 960 excimer laser power amplifiers (ELPAs) are directed into 60 Raman accumulator cells (RAC) that combine the energy of the individual amplifiers for each beamline. The Raman accumulators are 1.8m square aperture by 5m long cavities filled with D~. The RAC approach is attractive because it converts the relatively low

129

quality, high power pump beams into a high quality, high power output beam. The output beam from each RAC is 81 kJ and 250 ns long. This is achieved at a high conversion efficiency of 88%. The next step in the Prometheus-L optical train is to compress the 250 ns long beams exiting the Raman accumulators into shorter beam pulses suitable for the primary and precursor interactions with the target (approximately 6 ns for the main pulse and approximately 80 ns for the prepulse). This is accomplished with the use of a stimulated Brillouin scattering (SBS) cell. The leading edge (10 ns) of the long duration pulse from the RAC is "chirped" or frequency shifted by an amount equal to the Brillouin shift in the S F 6 gas used as the SBS gain medium. The duration and modulation depth of the "chirped" signal permit control of the shape and pulse duration of the resulting compressed pulse. The entire 250 ns beam with the "chirped" 10 ns leading edge travels the 37.5 m length of the SBS cell and reflects off an internal mirror aligned normal to the beam. As the "chirped" leading edge and the remainder of the beam reflects back onto itself, the frequency- shifted photons in the leading edge experience high optical gain from the S F 6 gas pumped by the incoming beam. This results in a highly compressed beam, suitable for the main pulse. Although the SBS cell has some control over the pulse shape and energy content of the emerging beams, additional temporal tailoring of the beam is required. The residual energy in the unextracted portion of the original 250 ns pulse is used to generate a prepulse beam. An optical switchyard of delay lines inverts the sequence of the leading and trailing pulses from t.he SBS cell to form proper prepulse and main pulse durations and energy content while achieving high energy conversion efficiency in the SBS cell. An illustration of the entire optical train from each set of 16 power amplifiers through the Raman accumulator, SBS cell, and delay lines for one of the 60 beamlines is shown in Fig. 3. The 60 beams are directed from the SBS cells and delay lines through the reactor building wall, down shielded beamlines, to the reactor cavity as shown in Fig. 4. To protect upstream optics and

130

L.M. Waganer / Fusion Engineering and Design 25 (.1994) 125-I43

POLARIZER LONGPULSE S & DELAYLINE

--OUTPUT 8EAM TO TARGET

SHORTPULSE DELAY LINE

~

POCKELSCELL(TYP) 1~ TURNING MIRROR

SBS CELL

PLATEJ L/4 C:HIRPER RAMAN ACCUMLATORJ TYPICAL DISCHARGELASER-OUTPUTBEAM STOKES SEED BEAM FLOW

Fig. 3. Raman accumulator,SBS cell, and delay line. driver plant equipment, a neutron pinhole is provided upstream of the final turning mirror. A grazing incidence metal mirror (GIMM) is used at a distance of 20 m from the center of the reactor cavity. This long-lived component will be located in the direct line-of-sight of the target. The final focus mirror is not in the direct line-of-sight of the target but is in a high radiation zone. Neutron traps are provided to help minimize radiation down the beamlines. The GIMMs and the final focus mirrors are equipped with fast-acting actuators to align the beams on the targets. Tracking systems will provide information from the targets at the end of the injector and, if needed, within the reactor cavity. The environment of the GIMM is the harshest of all the laser optical elements in that it has direct line of sight to the target. Radiation effects (and component lifetime) coa~ld be mitigated by increasing the distance from the center of the reactor. The increased distance increases the difficulty of focusing on the targets at greater

distances and increases the volume of the building or enclosure housing the beamlines, Thus the GIMM designers were encouraged to locate the GIMM as close to the target as possible yet achieve a tong-lived component. The surface of the mirror was chosen to be either AI or Mg since these materials have superior reflectivities at grazing incidences of 85 ° at 248 nm for KrF beams. Neutron irradiation results in a number of detrimental effects to the mirror's optical performance. The main areas of concern are the decrease in the resistivity (and increase in the laser light absorptivity) and deformation of the mirror surface (defocusing of the laser beam). The increase in absorptivity is on the order of 0.5% and will saturate within a few days of operation. Thus the combined effects of resistivity due to point defects and transmutations and surface roughening by 14 MeV neutrons will 0nly degrade the reflectivity of the 85° incidence aluminum mirrors from 99.4% to 99% at the end of the mirror lifetime. Neutron-induced swelling is the principal damage

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125 143

BLANKET/FW VACUUM VE

PINHOLE OPENING IN SHIELDING

7

BULK SHIELDING WALL FINAL FOCUS MIRROR -j

REACTOR BUILDING WALL

/

20 °\

.C

2.83 ,, 1.33

~ 5.oo~

f. f ~ . i

.11.34 .

1.31 ° -

131

/

1.90 ° -

l---NEUTRON

....... 20.0 . . . . . .

TRAP

/

/

GRAZING INCIDENCE MIRROR

DIMENSIONS IN METERS

/

COLLIMATING MIRROR- -"~=== VACUUM WINDOW -" - J / TURNING MIRROR LASER BEAM FROM SBS CELL

Fig. 4. Final optics configuration

reactor building.

Fig. 5. Cross-sectional view of the G I M M .

mechanism with a strong dependence on type of material and impurity level. High-purity aluminum is the better choice for the surface material to minimize the swelling effects. Tapering the thickness of the A1 surface layer is estimated to allow the G I M M to be a life-of-plant component. A cross-sec-

tion of the G I M M is shown in Fig. 5. The structural support of the mirror is composed of a low- swelling SiC composite to form a high-rigidity support structure. The SiC structure is composed of cooling channels running beneath the mirror. A concrete support frame controls thermal deformations.

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125-143

132

volved the accelerator configuration (single versus multiple beam) and the technique for delivering the beams to the target (channel transport versus some form of direct ballistic focus). The design team selected a single-beam LINAC with storage rings for the accelerator configuration based on the clear economic advantage for this approach. Fig. 6 illustrates a capital cost advantage of $560M and 12% for the single beam over the multiple beam LINAC. Channel transport was selected on a more qualitative basis. Channel transport offers many potential engineering advantages over focusing the beams directly onto the target. Furthermore, analyses indicate that when the beams are partially stripped, there is more than enough beam current to support the formation of a self-pinching transport channel. The concern lies in the capability to maintain a stable 5.6 m long channel in the surrounding cavity environment and the repeatability of aiming this channel at the proper spot on the target. These are critical issues that can only be addressed through future theoretical and experimental work. The system studies led to the selection of a 7 MJ target incident energy design point for the

Considerable effort was directed toward identifying the physical processes governing the transmission of the optical beams through the reactor cavity. The predominant consistuent in the gas atmosphere within the reactor cavity is lead vapor. Trade studies were conducted to determine the optimum combination of wall protectant temperature, cavity radius, repetition rate, beam losses by various mechanisms, and vacuum pumping requirements. The studies showed that the lead protectant operating at its highest possible temperature (525°C) for compatibility with external stainless steel piping and a cavity radius of 0.5 m would result in a 3 mTorr lead vapor pressure. This vapor pressure is not anticipated to cause significant loss in beam energy content or quality during transit across the cavity.

4. Heavy ion beam driver The heavy ion driver must efficiently and costeffectively deliver the required energy to the target within a focal spot radius on the order of 6 mm diameter from two sides of the cavity. The two key design choices considered for this study in-

1.300

1500 () \ \

1.200 i

.~-.

--~--'MB' CdE -0SB COE ---0--- MB Capital Cost - SB Capital Cost

~

f l ) 13O0

69-

"5

j

uJ 1.100 "5

03

1100 ~ Q_

o

o

~.4,----~4)'

•

1.000

kL

9O0

0

$ >

rr

.4)--

0.900

.

41)

700

_

o

)-

0.800

. 2

. 3

. 4

.

.

.

. . 5 6 7 8 Energy to Target (M J)

500 9

10

Fig. 6. Comparison of projected COE and driver capital cost for multiple and single beam LINACs. Systems are all 4 GeV, + 2 Led with 3 mm radius focal spot.

L.M. Waganer /Fusion Engineering and Design 25 (1994) 125-143

133

Precursor Beam

LINAC

I

I

-j f

Main Beam 1Buncher EntranceI

~

Storage Ring ~

_ ~

I

.uoc e,f'qJl

1

n-

Main Beam Fig. 7. Beamline configuration--routing from the LINAC to the reactor.

Prometheus-H reactor. A 10% energy loss was budgeted for forming the beam transport channels and transmission to the target, so the LINAC is designed to output 7.8 MJ. Lead ions with a + 2 charge state were chosen for compatibility with the first wall protection scheme. Detailed studies of transport lattice scaling determined that the required 7.8 MJ could be provided with 4 GeV ions using only 18 beamlets. Low ion energies are desirable because they provide improved target performance; however, the number of beamlets is a concern at low energies because it can become quite large (greater than 50) for some lattice scaling choices. This is a

particular problem for the single beam LINAC because of core recycling losses and beam stability and scattering loss in the high current storage rings. The resulting beamline configuration at various points along the driver is illustrated in Fig. 7. The driver consists of an injector, a ramped gradient section, a fixed gradient section, a stack of storage rings, a buncher accelerator section, a drift compression section, and a final focus section. These are shown in plan view in Fig. 8. Switching sections will also be required to insert and extract the beams from the storage rings, but these were not specifically considered for this study.

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125-143

134

STORAGE RINGS (14)

FINAL 180 rn LONG 7 BEAMS/SIDE

AUXILIARY BUILDING 120 X 60

ADMINISTRATION BUILDING 80 X 40

Fig. 8. Heavy ion reactor plant site plan.

The overall length of the LINAC is 2.2 km. Transport lattice scaling was chosen so the mean beam radius of 9.4 cm remains constant over this entire distance. Additional constraints were imposed on the quadrupole axial packing fraction (less than 80%) and aspect ratio (beam radius/ quadrupole length less than 0.25). These lead to a common magnet radial build as indicated in Fig. 7; however, the magnet length and field strength must be adjusted to provide proper focusing. The quadrupole inner and outer radii were chosen as 1.65 and 3.3 times the mean beam radius, based on guidance provided by LBL. An additional 12 cm was added to the magnet's outer radius to determine the core inner radius. This provides space for cryogenic insulation, magnet dewar, insulator rings, accelerator structure, and vacuum access. Ions are injected into the LINAC at an energy of 6 MeV. The beamlet current is 14 A at this point and the pulse length is 15.5 ms. The initial voltage

gradient is low (approximately 40 kV m-~), but it rapidly increases to the design limit of 1 MV m - t over the 223 m ramped gradient section. The local voltage gradient scales with beam energy in this section are based on accepted limits. This section contains 300 quadrupoles, terminating at a beam energy of approximately 100 MeV where the limiting gradient is reached. The initial 70% quadrupole axial packing fraction is high (but less than the 80% design limit) because focusing is needed every 0.5 m at this energy (lattice half-period). The packing fraction decreases to 39% at the end of the ramped gradient section because the lattice halfperiod grows to 1.33 m. The fixed gradient section (1 MV m -~) continues from the 100 MeV point to the final energy of 4 GeV. This section is 1.95 km long and contains 578 quadrupoles. The pulse length decreases from 1.64 ms to 85 ns in this section, so the beamlet current increases from 132A to 2.53kA. The

L.M. Waganer [ Fusion Engineering and Design 25 (1994) 125-143

quadruopole packing fraction drops to 19% at the end of the fixed gradient section because the lattice half-period has grown to 4.85 m. The quadrupole length also increases, from 0.52 to 0.91 m, to offset the increased beam stiffness. A flux swing of 1.5T was selected to reduce induction core losses (proportional to DB) since they are recycled 18 times per pulse for this design. The corresponding core thickness increases from 37 to 107 cm in the ramped gradient section owing to the increase in gradient. Thereafter, it decreases because the pulse length shortens as the beam energy increases. Driver efficiency was a concern for the single beam design, but the 1.5 T flux swing and low number of beamlets led to a projected efficiency of 20.6% for the Prometheus-L driver using Metglas loss curves suggested by LBL. The target pulse is generated by operating the L I N A C in a burst mode. The 18 beamlets are sequentially accelerated by cycling the induction cores on a 30 kHz timescale so the longest resident time within the storage rings is less than 1 ms. The beamlets are stacked vertically in 14 storage rings as indicated to provide a common bend radius and path length. Only 14 storage rings are needed because two storage rings collect three beamlets each to form a single prepulse for each side of the target. The 12 remaining rings each collect single beamlets that are used to form the main target pulse. This arrangement provides the recommended prepulse energy content of approximately 30%. Once all beamlets are collected and properly time sequenced, they are released from the storage rings and sent to the two buncher accelerators as indicated in Fig. 7. The prepulse buncher is 196 m long and induces a 4.9% velocity tilt on these beamlets. This causes the 256 ns long prepulse to compress to the required 30 ns over the 180 m drift distance to the target. The main pulse beamlets are sent through a shorter 85 m long buncher since they only require a 2.1% velocity tilt. This compresses their pulse length from 85 ns to the 7.3 ns required for target implosion over the same 180 m drift distance. The final time phasing and energy content of the prepulse and main beamlets are shown schematically in Fig. 9. A multiple-beam transport lattice is employed in the buncher sections, as indicated, to minimize core volume.

135

3ULSE

PREPULSE

I ~ I ~ " TOTARGET

7.3 ns ~ . ~

~I-- 30 ns - - ~

Fig. 9. Schematic diagram of a heavy ion beam prepulse and main pulse.

After the beams exit the buncher section, they are divided as indicated in Fig. 7, with six main and one prepulse beamlet directed to each side of the reactor cavity. The beam radius is allowed to increase in the buncher and drift sections to ease matching with the final focus magnets. The physical arrangement of the final focus magnets and beamlines is illustrated in Fig. 10. Triplet quadrupoles are used to focus the beams down on a point at the rear surface of the blanket. A lead vapor cell provides electrons that space-charge neutralize the beam at the exit of the last quadrupole. This permits the final 6 mm diameter focal spot to be attained. At the focal point, a lead vapor jet is provided that strips the beam ions to a high charge state. This is the mechanism for creating the transport channel. It also serves to isolate the reactor cavity pressure environment (approximately 100 mTorr) from that required in the beamlines (approximately 0.01 mTorr). The prepulse beamlet is located on the channel axis and arrives first at the lead gas jet as indicated in Fig. 10. The six main beamlets follow immediately and they arrive in parallel. Both the prepulse and the main pulse have a significant current margin for self-pinching (greater than five times), so channel formation is certainly feasible. The tendency of the surrounding plasma to generate a reverse current that might destroy the channel is a concern.

136

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125-143 ........

TARGET INJECTION SYSTEM OUADRUPOLEIN FINAL TRANSPORT HEAVY ION BEAM (7 PLACES)~

FIRST WALL PANEL

.%

REACTOR VACUUM

TRIPLET FINAL~ FOCUS MAGNETS

_ j"~cE~=~ / -

FINAL FOCUS-- J VACUUM ENCLOSURE

/

VACUUM P U M P - - ~ " ~ PORT SHIELD PLUG-BULK SHIELD W A L L - - /

7/

~ " // j -~\"

Fig. 10. Physical arrangement of final focus coils, plenum and shielding penetrations.

If transport channels are found to be viable, questions still remain concerning the ability to direct them at a target that is 5.6 m away with an accuracy of +_lmm. The initial path of the pinched beam channel will be defined by its net momentum. The single prepulse beams can be steered directly and can easily be pointed with the required accuracy. The direction of the main pulse, however, depends on the momentum inbalance between the six main beamlets. The questions involve whether momentum balances cause steering of the channel or whether the prepulse beam provides a focusing mechanism. These questions are a related part of the transport channel critical issue.

5. Reactor cavities

The need for innovative ideas also exists in the reactor design approach. The inertial fusion envi-

ronment has the harsh nuclear environment associated with magnetic fusion, including neutrons and high energy ions. In addition, the instantaneous nature of the inertial fusion reaction causes a host of new challenges. The heating of the first wall arrives first, with the prompt X-rays, gamma rays and neutrons, followed by the electrons and ions, and lastly the blast effects. Deeper in the blanket and shield, the instantaneous effects are attentuated to longer time scales. The thermal conversion of the fusion processes smooths out the power fluctuations until, at the heat exchangers and the thermal conversion system, the power is seen as a steady state process for repetition rates of 5 - 1 0 Hz. Safety and environmental impact played a significant role in the choice of materials and the design approaches employed. A top priority was the desire for inherent safety and minimum activation. This desire largely determined the material choices for the first wall, blanket, and shield. The first wall employs low-

L.M. Waganer ] Fusion Engineering and Design 25 (1994) 125-143 /

~'Ir~.

0.5 mm

Lead

Film

_

SupportStructure

~

137

_

~

Non-PorousFace

Fig. 11. First wall panel, cross-sectional view.

activation SiC composites which are also favored by the MFE designs of ARIES II and IV [16]. Both long-term and short-term activation is small, thus minimizing waste disposal problems and providing negligible decay heat. Li and LiPb were rejected for safety reasons in favor of a Pb wall protectant. Pb has toxicity and radioactivity concerns, but these were carefully estimated and minimized in the design. The blanket also uses SiC structure and reflector, with low-activation Li20 breeder and helium coolant. The tritium inventory in the breeder was minimized. Use of He at relatively low pressure, together with multiple containment barriers, makes blanket failures unlikely and the consequences benign. The shield material also was chosen to reduce activation. Instead of concrete, Prometheus uses an innovative, highly effective shield consisting of an aluminum structure, water coolant, and B4C, lead and SiC absorbers. Another major guiding principle was the incorporation of a sound engineering basis while proposing innovation. The required research and development and technical risk was minimized by adopting near-term technologies that can be extrapolated from existing data. This study's guidelines mandated a reactor design that could be operational in 2020-2030. SiC composites are commercially available today, although development will be required for use in a neutron radiation environment. Pb has been used as a coolant in the past, and technologies for using liquid metal as a coolant are well developed. Similarly, in the blanket, helium cooling is an established technology. The database for Li20 is being rapidly developed for the MFE fusion program. While not a driving force in the design, the relevance of Prometheus technology to MFE allows

an effective research and development program to be developed with minimum cost and time to completion. The research and development needs are bounded and predictable since the extrapolation from existing technologies is minimized. Cost penalties can be expected as compared with design concepts which are novel, or even radical; however, this was judged to be a reasonable strategy given the time schedule for fusion development. A cylindrical cavity with hemispherical ends was chosen to improve the maintainability of the reactor components. A wetted wall with a layer (0.5 mm) of liquid lead on the surface was chosen as the first wall protection scheme, shown in Fig. 11. The Prometheus design concept and configuration were chosen following a careful review of existing designs in both the IFE and MFE literature. A wetted-wall design was adopted with separate first wall and blanket as the best choice for the Prometheus reactors. Wetted walls have many potential engineering advantages, including good beamline accommodation, relaxed repetition rate limitations (compared with thick films), flexible engineering features, and low inventory and flow rate of the liquid film. The wall protection scheme chosen for Prometheus uses a thin liquid Pb film supplied from Pb coolant tubes through a porous structure of SiC composite material. The first wall coolant must have acceptable neutronic properties (either breed, multiply neutrons, or be transparent), such that the choices are limited to Li-bearing materials and neutron multipliers. Pb was selected for a number of reasons. Pb has a safety advantage over Li, good neutron multiplication, and chemical compatibility with SiC. Its thermophysical properties provide good operating temperature ranges. Its relatively high saturation temperature

138

L.M. Waganer t Fusion Enghwerhlg and Design 25 (1994) 125 143

leads to good conduction heat transfer into the coolant, its boiling point is not too high for materials temperature limits and compatibility, and the relatively high bulk coolant temperature gives good thermal conversion efficiency. Bi and BiPb were considered as alternative multipliers, but they have much higher radioactivity. Some of the outstanding disadvantages of Pb include high density and activation. The lead used in the target and the first wall coolant presents an activation concern, resulting in a considerable inventory when compared with the blanket materials. Bismuth, a common impurity in natural lead, produces Po -~"j in a high neutron environment. A detailed estimation of Po 2"~ production from the Prometheus first wall system was made from the D K R K F Cross-section Library and decay data in the Table of Radioactive Isotopes [17]. The bismuth concentration in the lead coolant is limited to 40 wppm to maintain the Biological Hazard Potential of Po-'"L significandy less than all other first wall system radionuclides. Production from bismuth is the dominate mechanism up to 1 year of irradiation, which can be controlled by limiting the bismuth concentration. After one year, production of Po -'~° from neutron interactions with natural lead begins to dominate. The calculated Po 2j" inventory due to both mechanisms is one order of magnitude lower than the ESECOM [18] Case 1 limiting case. Thus it is concluded that the Prometheus design can be classified as LSA = 1 with regard to Po 2m. Rectangular coolant channels for the liquid lead cool both the surface and bulk material. The composite SiC material is also graded in density from fully dense at the back face to 10% porosity at the front face. This porosity allows the lead to migrate from the coolant channels to the surface. A portion of the lead film is vaporized by the target explosion and is subsequently recondensed on the surface. A separate Pb film injector is used near the top of the cavity to establish and maintain a film in contact with the wall surface in that area. The lead coolant is introduced at the top of the cavity and it flows downward to the bottom of the cavity. The SiC first wall structure must be flexible enough to withstand cyclic loading from the blast,

Table I Predicted Nfetimeof SiC wall and blanket components Wall Porosity (%) Life for Prometheus-L (years) Life for Prometheus-H (years) Normal lifetimc (years)

Blanket

0 2.9 2.5

10 7.7 6.7 5

0 4.2 3.6

10 I 1.2 9.6 10

but strong enough to support itself and the interhal pressure of the film. SiC occupies a minimal first wall volume fraction (approximately 10%) to ensure good neutron multiplication. The thickness of SiC behind the liquid film is 5 mm to assist heat conduction into the coolant flow. The heat deposited in the first wall system is 1267 MWt which is over 40% of the generated thermal power. Removal of this heat requires a mass flow rate of 54 422 kg s -~ with an inlet bulk temperature of 375 C and an outlet bulk temperature of 525'C. Lead recondensation on the first wall surface is the primary vacuum pumping mechanism in the cavity. The surface temperature of the lead determines the vapor pressure. With a repetition rate of 5.65 Hz, it is calculated that the base pressure of the cavity will be 3 r e t o r t . The non-condensible gas load created by the He, unburned deuterium and tritium, and the hydrogen from the target capsule is pumped with cryogenic vacuum pumps and root blower backing pumps. One of the principal reasons for adoption of the SiC material was its resistance to radiation damage, namely swelling. The strong directional bonding and the mass difference between Si and C render the crystalline form of //-SIC exceptional radiation resistance characteristics. SiC/SiC composites manufactured with the CVI process will contain up to 10% porosity which helps mitigate the effects of swelling. A design limit of 5 % - 6 % volumetric swelling defines the lifetimes shown in Table 1 for a range of porosities. The first wall modules are attached to the front face of the blankets. The chosen blanket design builds on the existing data and design base developed by MFE. However, the decoupling of the first wall protection from tritium breeding allows more design freedom and enables development of

L.M. Waganer / Fusion Enghteering and Design 25 (1994) 125 143

FirstWall

Lead Coolant Tritium PurgeManifold

Channels

'

\

SiC ReflectorBlock

Lead Film \ /~ebble Bed Breeder~ \

L

J Inlet m unelP

/oolant,n,et ,,

Eo

I lllllII~I~-~,:--:; ....... {

~He

Purge

IIII|IIwr-.r

SiC Composite

: i IIIIII|IIJ|mLE "

.

r

IrlIIIIIIIIiI~~

II

-'-

~.

. Tritium/He

oufpu,

"CoolantOutlet

~6o cm-7~.L2o cm~'-- " ~ " OutletPlenum Structure TritiumOutputManifold Fig. 12. Schematic diagram of first wall and blanket conceptual design.

a safe, low-technology blanket system. The general blanket module design is shown in Fig. 12. The structural material is SiC composite, serving all the structural functions including coolant tubes, pebble bed container, outer blanket module wall, reflector region, and coolant plenum. Helium at 1.5 MPa is the coolant medium extracting heat from the blanket region. The tritium is bred in an Li20 pebble bed breeding zone. Low pressure helium gas removes the tritium gas for processing. The blanket system removes 1782MWt of power from the system. The inlet temperature of the helium is 400°C and the exit temperature is 650°C. There is an overall energy multiplication ratio of 1.14 within the first wall and the blanket. The blanket tritium breeding ratio is 1.2. The tritium inventory of the blanket is calculated to be 100 g. The peak and average neutron wall loads are equivalent to 6.5 and 4.3 M W m -2 respectively. The blanket is constructed in longitudinal modules. Penetrations are allowed for the entrance of the laser beams and the vacuum ports. Plenums are provided in the annular space outside the blanket region. Adequate space is allowed for remote maintenance of the blanket and first wall piping. A 0.5 m thick ferritic vacuum vessel is provided outside the blanket and plenum region to help contain the reactor vacuum conditions. The bulk shielding was placed close to the vac-

139

uum vessel with a l m space for maintenance access, support structure, and plumbing. The basic bulk shielding shape adopted was a simple right circular cylinder 20 m in diameter and 35 m high (inside dimensions). The adopted composite shield, composed of AI, SiC, B4C, Pb, and water, is 1.3 m thick. The figure also shows the shielding provided around the three vacuum plena. Similarly, Fig. 4 illustrates the 25 cm thick shielding provided around individual beamlines. At the end of the beamline, neutron trap and thicker shielding is provided to stop neutrons with direct lineof-sight to the cavity interior. The overall configuration of Prometheus is a low aspect ratio cylinder with hemispherical end caps. This configuration was selected for several reasons. (l) Maintenance of a cylinder is easier than a sphere. Maintenance paths were all straight vertical lines and the configuration allows independent removal of first wall panels and blanket modules. (2) A cylinder provides better control of film flow. Problems protecting the upper hemisphere can be reduced with higher aspect ratio, in which the distance from the blast to the upper end cap can be maximized. (3) A cylindrical configuration is more consistent with conventional plant layouts. After a review of the design requirements for the laser-driven and heavy ion-driven reactors, it was decided that a single reactor concept would accommodate both designs. The size of the reactor cavity would be tailored to the required fusion power, thus the heavy ion beam reactor would be slightly smaller in size.

6. Reactor integration As previously discussed, both the laser-driven and heavy ion-driven plant could be accommodated by a common design of the reactor chamber. However, the integration of the two types of beamlines resulted in significantly different reactor building designs. For the laser concept, an elevation view of both the laser driver and the reactor buildings is shown in Fig. 13. The reactor building is 86 m in diame-

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125-143

140

KrF DR,VER BLDG ANNULUS

i"

153 DIA ,,= --

.d!

/ i

II

I IIll IIIII

I ELECTRICDISCHARGE--i LASER SUBSYSTEM / /I

_s

ELEV +5.4

I II u

,,dI

•I

ELEV +44.2 --86 DIA REACTORBLOG

ELEV-27.311

I,F

--

j .1 O,A

I""

~RAMAN ACCUMULATOR

IIIl"~ ~-SRscELL IIII II "-~.'~%'."~.o';;~";''c"'""

~,...o.,c~o.~.~.~,,~

t-UJILEv -45.3

lrq~ll\

~

J

129 DIA

"I

Fig. 13. Prometheus-L reactor building provides space for shielded beamlines. The driver building surrounds the reactor building,

I

g

HEAVYION MAIN BEAMS (6) ON CONICAL ARRAY(8.54 °)

U

TARGET SYSTEM

REACTOR VACUUM VESSEL

FINAL FOCUSVACUUM PUMP

LAST QUADt IN FINAL TRANSPORT BEAMLINES

~

HEAVYION PRECURSOR BEAM (3 BEAMLETSIN CENTER BEAMLINE)

FINAL FOCUS VACUUM ENCLOSURE

TRIPLET --~ FINAL FOCUS MAGNETS

BEAMS TRANSITION TO VERTICAL STACK AT WALL

o','

59 m DIA REACTOR BUILDING

Fig. 14. Prometheus-H reactor cavity arrangement.

SOME ELEMENTS ROTATED FOR CLARITY

L.M. Waganer [ Fusion Engineering and Design 25 (1994) 125-143 Table 2 Major design parameters and features of the Prometheus plants Parameter

Prometheus-L (laser)

Prometheus-H (heavy ion)

Net electric power (MWe) Gross electric power (MWe) Driver power (MWe) Auxiliary power (MWe) Cavity pumping power (MWe) Total thermal cycle power (MWt) Blanket loop power (MWt) Wall protection loop power (MWt) Usable driver waste heat (MWt) Usable pumping waste heat (MWt) Thermal conversion efficiency (%) Recirculation power fraction (%) Net system efficiency (%) Fusion power (MW) Neutron power (MW) Surface heating power (MW) Fusion thermal power (MWt) Thermal power to shield (MWt)

972 1382 349 36 25 3264 1782 1267 193 22 42.3 30 31 2807 2027 780 3092 43

999 1189 137 28 25 2780 1597 1162 NA 21 42.7 16 36 2543 1818 725 2797 38

Cavity radius (m) Cavity height (m) First wall protection or coolant medium In/out temperature (°C) Breeder material Structural material, wall and blanket Blanket heat transfer medium In/out temperature (°C) Cavity pressure (mTorr Pb) Neutron wall load, peak/average (MW m -2) Energy multiplication factor Tritium breeding ratio

5.0 15.0 Liquid lead (375/525) Li20 pebbles SiC 1.5 MPa helium (400/650) 3.0 6.5]4.3 1.14

4.5 13.5 Liquid lead (375/525) LizO pebbles SiC 1.5 MPa helium (400]650) 100 7.1/4.7 1.14

1.20

1.20

Target illumination scheme Number of beams

Direct drive symmetric 60

Driver output energy (M J) Overall driver efficiency (%) Type and number of KrF amplifiers Beam combining technique Pulse compression technique

4.0 6.5 Electric discharge, 960 Raman accumulators Stimulated brillouin Scattering NA NA NA NA 100 124 497 5.65 79.4 72.0

Indirect drive, two-sided 18 in LINAC (12 main + 6 in 2 prepulses) 7.8 (7.0 to target) 20.6 NA NA NA

Ion accelerated Charge state Final energy (GeV) Type of accelerator Final beam transport efficiency (%) Target gain Target yield Repetition rate (pps) Plant availability (%) COE (mills/kWh, 19915)

Lead + 2 4.0 Single beam LINAC 90 103 719 3.54 80.8 62.6

141

142

L.M. Waganer / Fusion Engineering and Design 25 (1994) 125-143

ter, dictated by the length of the shielded beamlines. The annular driver building contains all the laser systems and surrounds the reactor building. The laser driver option uses 960 electric discharge lasers to provide a highly reliable power amplifier system. Sets of 16 laser amplifiers are grouped into modules which supply 60 beamlines equally spaced around the perimeter of the building. N L O laser elements provide the beam combining and compression functions to deliver high quality beams on the target. The N L O beamlines are arranged vertically to minimize building volume and distance to the reactor chamber while preserving geometry symmetry. The N L O elements of the RACs, the SBS cells, and the optical delay switchyard are shown. Beam pathlength-matching routes are shown. Only four of the 60 beamlines are shown in this view, along with some of the primary coolant piping. The 60 beams pass through an optical focus at a neutron pinhole to minimize neutron activation in the driver building. The final optical element is a final focus mirror to focus and turn the beams. A G I M M is the final optical element that lies in the direct line of sight of the center of the cavity. The 60 beam openings in the first wall range from 15 cm to 20 cm diameter. The integration of the heavy ion beam into the reactor building is less complicated than the laser beamlines. Fig. 10 illustrates one of the two sets of heavy ion beamlines that focus down on the backside of the blanket. This final focus system is displayed in the elevation view of the heavy ion reactor building as shown in Fig. 14. The main heavy ion pulse beams are arranged in an 8.54 ° conical array with the precursor beams on axis. After the channel transport is formed at the back of the blanket, the two beams proceed to the target at the center of the cavity. The channel transport concept is attractive because there are only two 2 cm diameter openings in the first wall and blanket which maximizes the wall and blanket area and shielding. Small antechambers are used to house the final focus beam elements and provide an area with increased shielding apart from the main heavy ion beam driver tunnel.

7. Summary Two inertial fusion energy power plants have been conceptually designed and analyzed. Innova-

TO Sh~Id

~ ~ Laser Syslem (Heavy Ion System)

I 7B2 M~V

MW

\ 2 2 MW In,~o, ~ , , ~ I ~ 349 MW t 136MW ~ ~ ( .... )11~ (28 MW)-- i Systemil T134MW{ 110 MW) T 972 MWe (999 MWe) Wasle Heal Nel EleCtric Power

Fig. 15. Overall plant power flows for Prometheus baseline designs.

tive concepts for the drivers and the reactor cavity have been integrated to yield a safe, environmentally attractive, and economical approach. Table 2 presents the high level design and performance parameters for both plants. The power flow diagrams for the two power plants are shown in Fig. 15. The higher efficiency of the heavy ion driver enables lower fusion power and recirculating power to achieve the same nominal net power output. Both designs have attractive features. The economics slightly favor the heavy ion case, but the levels of the unknowns are too large to make a discrimination at this point. Both represent significant advancements in the state-of-the-art in I F E reactor design.

References [1] L.M. Waganer et al., Inertial fusion energy reactor design study--Prometheus-L and Prometheus-H, Final Report, DOE/ER-54101, MDC920008, March 1992 ( Department of Energy, Washington, DC). [2] C.C. Baker, M.A. Abdou et al., STARFIRE--a commercial tokamak fusion power plant study, ANL/FPP-80I, September 1980. [3] The heavy ion fusion systems assessment (HIFSA), Fusion Technol. 13 (2) (1988). [4] W.R. Meier et al., OSIRIS and SOMBERO inertial fusion power plant designs, Final Report, WJSA-92-01, DOE[ER/54100-1, March 1992 (Department of Energy, Washington, DC). [5] F. Najmabadi, R.W. Conn et al., The ARIES-! tokamak reactor study, Final Report, UCLA-PPG-1323, 1991 (University of California at Los Angeles, Los Angeles, CA) Vols. I and II.

L.M. Waganer /Fusion Engineering and Design 25 (1994) 125-143 [6] J.H. Pitts, Development of the CASCADE inertial-confinement fusion reactor, Fusion Technol. 8 (I, Part 2B) (1985) 1198. [7] J.A. Blink et al., The high-yield lithium-injection fusionenergy (HYLIFE) reactor, UCRL-53559, LLNL, December 1985 (Lawrence Livermore National Laboratory). [8] C. Yamanaka et al., Concept and design of ICF reactor SENRI-I, Institute of Laser Engineering Report ILE8127 P, 1981 (Institute of Laser Engineering). [9] R.W. Moir, A review of inertial confinement fusion (ICF), ICF reactors, and the HYLIFE-II concept using liquid FLiBe, UCID-21748, 25 September 1989. [10] R.W. Moir, HYLIFE-II inertial confinement fusion reactor design, Fusion Technol. 19 (1991) 617-624. [I 1] B. Badger et al., HIBALL a conceptual heavy ion beam driven fusion reactor study, UWFDM-450, Fusion Engineering Program, 1981 (University of Wisconsin). [12] G.A. Moses et al., Overview of LIBRA light ion beam fusion conceptual design, Fusion Technol. 15 (2, Part 2A) (1989) 756.

143

[13] M.E. Sawan et al., Chamber design for the LIBRA light ion beam fusion reactor, Fusion Technol. 15 (2, Part 2A) (1989) 766. [14] R.W. Conn et al., SOLASE--A conceptual laser fusion reactor design, UWFDM-220, December 1977 (University of Wisconsin). [15] S.I. AbdeI-Khalik et al., SIRIUS-M: A symmetric illumination, inertially confined direct drive materials test facility, 2nd Int. Conf. on Fusion Reactor Materials, April 1986. [16] R.L. Miller and R.A. Krakowski, Options and optimizations for tokamak reactors: ARIES, Ninth Topical Meeting of the Technology of Fusion Energy, Oak Brooks, IL, October 7 11, 1990. [17] E. Browne and R. Firestone, Table of Radioactive Isotopes, Wiley, New York, 1986. [18] J.P. Holdren et al., Report of the senior committee on environmental, safety and economic aspects of magnetic fusion energy, UCLA-53766, 25 September 1989 (University of California at Los Angeles, Los Angeles, CA).