A range of mission applications have been considered for particle-bed reactor systems. These applications seek to capitalize on the various attractive attributes of this technology, including high specific impulse and high thrust-to-weight ratio. At least four classes of missions have been specified:
+ Strategic defense interception;
+ Long-range ballistic missile;
+ Space launch vehicle;
+ Piloted Lunar and planetary missions;
In each of these cases, a number of specific applications have been characterized for PBR-based systems. Although PBR rockets do appear capable of performing many of these missions, a number of questions remain:
+ Does the PBR system provide a unique capability compared to other approaches?
+ Does this capability, regardless of its technological mode of implementation, have sufficiently high priority to warrant implementation?
In general, it would seem that all of the identified applications fail one or both of these tests. In many cases, notably those related to strategic defense and offense, the PBR-based system does not provide unique capabilities, or does not uniquely provide these capabilities. In other cases, such as space launch vehicles and piloted space exploration, significant questions remain as to whether these applications are of sufficiently high national priority to warrant significant development expenditures.
1 - Strategic Defense Interceptor
One of the initial applications investigated for the SNTP reactor was:(1)
"... as a potential far-term candidate for ground-based early intercept of ballistic missiles."
The performance of the PBR rocket for this mission is dependent on the thrust-to-weight ratio of the reactor, as well as the burn-time of the PBR stage. For a given T/W ratio, there is an optimum burn time that will maximize the range of the interceptor. Longer burn times require greater propellant loads, which increases the overall mass of the interceptor and thus reduces its acceleration and thus range. As can be seen in Figure IV-28, nuclear rocket systems with a T/W ratio of 3.0, which is typical of NERVA-class reactors, have negligible potential for executing this mission. For systems with a reactor T/W ratio of 30.0, which is typical of SNTP PBR designs, the optimum burn time to maximize range is approximately 100 seconds. A reactor with a T/W ratio of 60.0 exhibits a similar range-maximizing burn time, although range performance decays much more rapidly with increasing burn time than is the case with a system with a T/W ratio of 30.0. It should be noted that the 60.0 T/W ratio is beyond the range of values generally associated with PBR systems, and should be regarded as a limiting ideal case.
Several potential mission modes are possible for such an interceptor:
Defense of the continental United States from a single site, providing nation-wide coverage from the single site permitted under the ABM Treaty;
Early engagement of multiple warhead ballistic missiles prior to full deployment of independently targeted reentry vehicles and penetration aids.
Early interactive discrimination of ballistic missiles prior to full deployment of reentry vehicles and penetration aids.
Late-midcourse interactive discrimination of ballistic missiles reentry vehicles and penetration aids.
However, analysis of these applications does not suggest that PBR-based systems would provide particularly attractive capabilities. In the first case, other less technologically challenging means are available. The PBR system offers very marginal capabilities in the second case, and the two interactive discrimination cases face other technological and physical challenges which are not resolved by the use of a PBR booster.
i - Single-Site Continental Defense
Small PBR-powered anti-missile interceptors (Figure IV-29) with burn-out velocities in excess of 7 km/sec are capable of placing kinetic kill vehicles on ICBM-like trajectories. This would permit intercepts at ranges in excess of 5,000 km, approximately 18 minutes after launch. Such interceptors could clearly be capable of supporting the defense of the entire continental United States against ICBM attack.
There are two issues, however, that raise questions about the utility of this application:
Existing interceptor concepts, using solid propellant motor boosters, have similar capabilities.
The presumed political attractiveness of this capability -- an ABM Treaty compliant national defense -- is contingent on a reading of the ABM Treaty that is not universally accepted.
SDI work on ground-based interceptors has been based on the approach used in the Homing Overlay Experiment (HOE) that successfully intercepted a warhead in 1984. This was followed by the Lockheed-built Exo-atmospheric Reentry-vehicle Interception System (ERIS), which incorporated a much smaller and lighter kill vehicle. It was initially thought that Lockheed's work on ERIS would naturally translate into responsibility for production of the operational version of this system. But competitive development of the operational Ground Based Interceptor Experiment (GBI-X), a smaller and more sophisticated version of the ERIS, was awarded to Hughes, Martin Marietta and Rockwell in mid-1990, with one of them to receive the final contract.(2) The relative progress among these three generations of interceptors is indicated by the mass of the kill vehicle, which dropped from near 1,200 kilograms with Homing Overlay, to less than 200 kilograms with ERIS, to about 25 kilograms with GBI-X. This interceptor will have a range sufficient to cover the entire continental United States, depending on the types of sensors that are used in its support.
As signed in 1972, the ABM Treaty permitted the two parties to deploy 100 ABM launchers at each of two sites, one at the national capital area, and the other an ICBM missile field, separated by no less than 1300 kilometers, and they could construct no more than 15 interceptor launchers at agreed ABM test ranges. In 1974, a Protocol to the Treaty reduced the number of permitted sites to one. The United States chose to deploy a system at the Grand Forks ICBM field in North Dakota, and the Soviet Union chose to defend Moscow. The 1974 Protocol does, however, permit the parties to switch their deployment areas, provided notice is provided during the five-year review conference. But this right to relocate may be exercised only once.
Thus the United States would be within its rights under the ABM Treaty if it deployed 100 interceptors at Grand Forks, and given proper prior notification, also within its rights to instead deploy 100 interceptors in the vicinity of Washington, D.C.
The question of whether the long range and potentially nationwide coverage of these interceptors would affect their status under the Treaty is unclear. Article I of the 1972 Treaty obligates the Parties "not to deploy ABM systems for a defense of the territory of its country and not to provide a base for such a defense, and not to deploy ABM systems for the defense of an individual region except as provided for in Article III of this Treaty."
Writing in 1974, John Rhinelander (the chief legal advisor to the ABM Treaty negotiating team), argued that:
"The prohibition on a "base" for a nationwide defense systems ... would appear to limit the capability of ABM missiles ... to those that could not provide a thin defense covering substantially the whole of the territory of the U.S or U.S.S.R. Thus, a "thin" nationwide defense is prohibited, even in the unlikely event that technological advances made feasible thin coverage from the two ABM areas which each side is permitted..."
During the negotiations, the American side considered proposing an explicit limit on the maximum range of interceptors that could be deployed under the Treaty, and decided not to do so. The logic of this decision was that the intent of the Treaty was to ban the deployment of systems that could effectively defend the entire country, and that the limit of 100 interceptors would preclude such an effective defense, regardless of the extent of their coverage.
On the other hand, while the possibility of a range limit was the subject of internal U.S. Government discussions, the decision against proposing such a limit means that this understanding of the meaning of territorial defense was not the subject of bilateral discussions. Thus, while this might have been the position of the American side, there is no basis for believing that this was in fact also the Soviet position.
The requirement that the two permitted sites be separated by no less than 1300 kilometers contained in Agreed Statement C of the Treaty also provides some insight into the Article I distinction between "ABM systems for the defense of an individual region" which are permitted, and a "defense of the territory" which is prohibited. The American intent was to preclude Soviet deployment of additional ABM sites in the western Soviet Union, where they could protect urban areas as well as missile fields. The western borders are generally about 1000 kilometers from Moscow; thus the 1300 kilometer separation requires a deployment east of Moscow, in areas of much lower population density.
The 1300 kilometer separation is also greatly in excess of the range of interceptors existing when the Treaty was signed. Both the American Spartan and Soviet Galosh interceptors had ranges of a few hundred kilometers. Thus the Treaty stipulated that two isolated regions could be defended, but that the two initially permitted sites could not have areas of overlapping coverage, which might be construed as a base for the defense of the entire country.
The issue of what would constitute a "base" for a territorial defense is also not clear. At the time of the negotiations the principal concern of the American side was large phased-array radars, regarded as the long-lead-time item in providing a basis for a territorial defense. It was felt that interceptors could be deployed in a matter of months, whereas radars could take several years to construct.
This might lead one to the conclusion that only ABM radars, and not interceptors, are the subject of the Article I prohibition on providing a base for territorial defense. But the actual text of the Treaty does not make such a distinction.
Absent a review of the negotiating record it is not possible to state definitively what the agreement was between the two sides in 1972. Based on the public record, the ratification record and subsequent practice do appear to provide explicit guidance on this subject. At a minimum, given the ambiguity of the public record, it cannot be automatically assumed that the Soviets would concur with the assertion that the deployment of 100 interceptors at a single site capable of defending the entire United States would be Treaty consistent.
Given these observations, that there are other more conventional means of implementing continental defense from a single site, and that it is unclear that such defenses would be consistent with the ABM Treaty, it is difficult to identify uniquely attractive attributes for this application of the PBR interceptor.
ii - Post-Boost Phase Intercept
There are a variety of measures that the offense can take to reduce the effectiveness of a defense, short of directly attacking the defense itself. These passive countermeasures seek to reduce the effectiveness of the defensive sensors and weapons. Taken together, they can substantially increase the overall cost of the defense at relatively low cost to the offense, thus reducing or eliminating any favorable marginal cost advantage for the defense.
ICBM warheads spend most of their time in flight in a ballistic trajectory in the vacuum of space. This mid-course period lasts about 25 minutes. During this time the offense can deploy a number of countermeasures against defensive sensors, such as balloons, chaff and corner reflectors. Although these simple devices burn up as soon as they reenter the atmosphere, the brief time remaining to the defense in the terminal phase after reentry is insufficient for an effective defense.
One of the most effective countermeasures to defensive sensors during the mid-course is a decoy which would be an aluminized mylar balloon (similar in construction to the Echo I satellite or a weather balloon) that would be carried by the missiles post-boost vehicle (MIRV bus) and inflated as the actual reentry vehicles are dispensed. Each balloon would weigh only a few kilograms, with a diameter of several feet. If the balloons are made to look like warheads, they would serve as simulation or traffic decoys, increasing the number targets that the defensive sensors must track by several orders of magnitude. Typical numbers that are discussed suggest that each attacking warhead could be accompanied by as many as 10 to 20 balloon decoys.
Balloons can also serve as anti-simulation decoys, in which each of the actual warheads is also enclosed in a balloon. Although the light-weight balloons would cool more quickly than heavy warheads when exposed to the cold of space, the addition of small heaters to the empty balloons could prevent heat-sensitive infra-red sensors from discriminating which balloons actually contained warheads.
Although balloons are the most stressing passive countermeasure for defensive sensors, there are additional means available to the offense. Chaff, consisting of fine metallic fibers or strips of foil, can confuse anti-missile radars in the same way that they have confused air-defense radars since World War II. Aerosols and mists of appropriate composition can similarly confuse passive infra-red sensors. Both radars and imaging laser sensors can be confused by corner reflectors.
While development and production of such countermeasures is a complicated process, there is no doubt that the Soviets have mastered these techniques. Thus even a small attack by Soviet missiles, such as is contemplated in scenarios involving accidental or unauthorized launches, would be attended by such countermeasures. And there is little reason to doubt that any other country capable of building long range missiles tipped with nuclear warheads would also be capable of providing these systems with countermeasures that would severely complicate the task of the defense.
The difficulty of dealing with these mid-course countermeasures is the principal reason for the interest in boost-phase and post-boost phase defenses. However, defense in these phases requires space-basing of sensors and weapons, leaving the defense more vulnerable to direct attack. The defense is faced with a tradeoff between improving mid-course decoy discrimination and enhancing the survivability of its space-based components.
Existing missiles take about 1 minute to reach the altitudes at which the defense's space-based weapons become effective, and so the boost phase of liquid-fuel missiles includes about 4 minutes in space, and solid-fuel missiles about 2 minutes in space. Changing from liquid to solid fuel thus cuts in half this time available for the defense to attack missiles in their boost phase.
However, further reductions in the length of the boost-phase are possible, using so-called fast-burn boosters. These missiles achieve the velocity needed for intercontinental ranges while at very low altitudes, too low for most space-based weapons to be effective. Such fast-burn boosters require special high acceleration solid rocket motors and special coatings to protect against atmospheric friction. But the technology required to build fast-burn boosters is very similar to that required for ground-based ABM interceptor rockets, and both the United States and the Soviet Union could build such boosters by the end of this century. These special features could result in a reduction of 50% in the number of warheads that a missile of a given size could carry. From the standpoint of the defense, this would represent a "virtual" attrition of the offensive force. But a substantial investment by the defense would have been negated at relatively low cost to the offense. The alternatives facing the defense would be either to move to more exotic boost-phase weapons, such as gamma ray lasers, that could penetrate further into the atmosphere, or to concentrate on the later phases of the defense.
In addition to these countermeasures, the offense can also take a number of active measures to reduce the effectiveness of the defense. These include actions that will interfere with sensors and communications links, as well as the physical destruction of space-based components of the system.
Discussions of the survivability of a space-based defense usually presume that the direct attack on the defense occurs simultaneously with the launching of large number of offensive missiles. In a sense, this scenario is similar to that of the counterforce first strike that has dominated the strategic debate for over a decade. However, attacks on space-based defenses may also resemble the extended attrition campaigns typical of anti-submarine warfare. In a time of crisis, the offense could act over a period of days or weeks to degrade the capabilities of the defense. This gradual though inexorable reduction in defensive capabilities could provide significant bargaining leverage during the crisis, without the risk of extensive collateral damage to populations on Earth.
The most direct means of degrading the defense would be to attack and destroy its space-based components. One of the most frequently mentioned means of doing so would be the space mine, a concept popularized by Richard Garwin. A space mine would consist of a small satellite, carrying a nuclear or conventional explosive. On command from the ground, the space mine would move to within lethal distance of its target, and detonate. There is ample precedent for such an approach in the Soviet intelligence trawlers that shadow American carrier task groups at sea. Although the defense could proclaim protective keep-put zones around its space-based components, indicating that if the space mine came within these zones it would be subject to destruction. Although it would be technically possible to do this, the escalatory impact of this action might prove politically unappealing in a time of crisis.
All of these considerations pose a major quandary for the defense planner. On the one hand, boost phase intercept is the most highly leveraged phase of defense, permitting a single interceptor to potentially destroy up to a dozen warheads and hundreds of decoys, and to do so in an environment essentially uncluttered by offensive decoys. On the other hand, gaining access to the boost-phase is generally taken to require space-basing of defense assets, which renders them highly vulnerable to offense countermeasures, notably defense suppression.
This quandary could be resolved if there were some means of gaining access to the boost phase other than by basing intercept systems in space. Forward deployment of interceptors, either at sea or on aircraft, is one possible avenue, but this raises even greater concerns about vulnerability to defense suppression.
Although there has been neither public discussion nor any indication of SDIO analysis, it is interesting to consider whether the PBR technology may offer a resolution of this quandary. Attempting to gain access to the boost and post-boost phases of an ICBM's trajectory would require an interceptor with a significantly higher acceleration than the roughly 7 km/sec provided by the previously discussed interceptor. Placing a larger PBR stage atop MX Peacekeeper solid stages could provide terminal interceptor velocities in excess of 11 km/sec (Figure IV-30). Such a vehicle was discussed in the SNTP Environmental Impact Statement, which noted:(3)
"... studies show that other propulsion systems cannot equal the potential performance of PBR engines even when scaled to very large vehicle sizes. For example, to achieve results similar to a second stage PBR propulsion system would require a conventional propulsion system (solids of liquids) that would weigh 3 - 5 times more than PBR systems. The relatively low weight of the PBR propulsion system allows the design of a rocket that is approximately the size of the current Peacekeeper vehicle..."
Such a vehicle would permit an interceptor to fly much "hotter" depressed trajectories, relative to the minimum energy ICBM trajectories (Figure IV-31). As a result, intercepts could take place 13 to 14 minutes after the ICBM was launched.
In the case of a large liquid-fueled missile, such as the SS-18 (Figure IV-32), which has a five minute boost phase and ten minute post-boost phase during which warheads and decoys are deployed, this would permit access only to the last minute or two of the post boost phase, providing negligible leverage compared with space-based boost-phase intercepts.
In the case of a solid-fueled missile, such as the SS-24 or SS-25 (Figure IV-33), which has a three minute boost phase and five minute post-boost phase, this would permit access neither to the boost nor post boost phase.
These calculations are based on simultaneous launch of the interceptor and the ICBM. Additional delays of seconds to a few minutes might result from lags in the warning system, and from launch authorization procedures. Such delays could largely negate even the limited ability of PBR interceptors to gain access to the latter post-boost phase of liquid-fueled missile trajectories.
Thus the PBR interceptor would not seem to provide a viable alternative to space-based systems for engagement of ballistic missiles during the early phases of flight.
iii - Nuclear Directed Energy Interactive Discrimination
The attraction of post-boost phase defense is that it offers the prospect of discriminating warheads from decoys while they are being deployed. Highly capable imaging radars or imaging lasers (LIDARs) would be used to actually take pictures of the post-boost vehicles while the balloon were being inflated. Warheads that were covered in balloons could be targeted for attack, while empty balloons could be ignored. Discrimination using imaging sensors would require computer data processing capabilities far beyond those available today, and it represents the most demanding computational requirement of the SDI. This task would be further complicated if the entire process of inflating decoy balloons were to take place inside a single very large balloon, or were to be obscured by chaff and aerosols.
Unfortunately, most of these discrimination techniques have proven difficult to implement in practice.
The difficulties posed by boost-phase interception, and the challenges of using active imaging sensors and passive infrared sensors for decoy discrimination, led to a focus in the mid-1980s on using interactive sensors. This technique relies on producing observable changes in decoys that would differ from changes produced in warheads.
Interactive Discrimination of real targets from decoys by using sensors (such as a neutral particle beam generator or an X-Ray Laser) that interact with a target or decoy, with targets changing in a fashion that is observably different from the changes in the decoy.
The Excalibur nuclear-pumped X-Ray laser was initially identified as a candidate for boost phase and post-boost phase interceptions of ballistic missiles. The interceptor would have to be forward deployed in a ground-based or mobile sea-based mode (probably on submarines to enhance survivability). Interceptor rockets would be launched on warning of attack (the "pop-up" mode). Each nuclear device would have a number of laser rods (perhaps twenty), that would be pumped by the explosion of the device. These could be aimed separately for decoy discrimination, or at the same target for booster interception.
The military and strategic utility of this system was recognized as being highly contingent of the scope and character of the Soviet threat. A space-based deployment mode for this system would render it very vulnerable to defense suppression attacks using anti-satellite weapons. Identifiable Soviet countermeasures, such as booster hardening and the use of solid fuel and fast burn boosters, could substantially reduce the effectiveness of this system deployed in a pop-up mode, within the timeframe for the deployment of such a system.
However, the difficulty of getting the X-Ray laser device sufficiently high above the atmosphere to be able to shoot at missiles during their boost phase led to a reorientation of the mission applications for this program. Increasing emphasis was given to anti-satellite applications, with a secondary missions including midcourse decoy discrimination.
In late 1984 the Excalibur effort started to increase its focus on mid-course decoy discrimination applications. The effects of the laser on its target, such as trajectory changes and physical disruption, could provide a means for discriminating between decoys and reentry vehicles. Taps by low-power lasers could produce vibrations in empty balloons that would differ from the vibrations induced in balloons that concealed warheads. X-ray lasers could deflect the trajectories of decoys to a greater extent than they would deflect the trajectories of warheads.
The discrimination mission also led to increased interest in other types of third generation weapons, such as high power microwave devices, visible lasers, directed electromagnetic pulse, charged particle beams, and kinetic energy weapons. This last concepts would be used to disperse a high velocity aerosol for interactive discrimination of midcourse decoys.
This application could demonstrate significant synergisms for applications with kinetic energy weapons, by significantly relaxing the number of targets that must be attacked, with associated reductions in sensor and battle management requirements. The brightness and pointing accuracy performance requirements for decoy discrimination are less demanding than for target destruction.
Although there has been no explicit discussion of the use of PBR-boosted interceptors in conjunction with nuclear-pumped interactive discrimination systems, it is interesting to consider whether the higher specific impulse and longer-duration accelerations of the PBR system might enhance the effectiveness of such systems.
The payloads of several hundred kilograms posited for the simplest PBR interceptors are not-inconsistent with projected mass values for third generation nuclear devices, although the associated pointing systems required for some of these concepts would be considerably more massive. These could be accommodated on larger interceptor designs.
Returning to the trajectory curves previously considered for other interceptor missions, it would appear that PBR-interceptors launched from the continental United States would not obtain direct line-of-sight access to ICBMs launched from Russia until 6 or 7 minutes after launch. This would clearly be too late to effect boost-phase intercept of either liquid or solid fueled missiles. However, it would permit access to the later part of the post-boost phase for solid fuel missiles, and it would provide access to much of the 10 minute post-boost phase for older liquid fueled missiles. This could provide an opportunity for interactive discrimination, given that the target complex of reentry vehicles, decoys and post-boost vehicles would still be relatively compact, permitting a directed energy beam with relatively wide dispersion to illuminate the entire threat cloud, rather than attempting to interrogate each individual object. This would significantly relax pointing and tracking requirements, resulting in a greatly reduced mass for this component of the system, as well as permitting the use of less sophisticated wider dispersion directed energy sources. However, the engagement ranges are of the order of 5,000 km, which would significantly complicate operations.
In any event, the utility of this system is primarily contingent on the efficiency of the discrimination mechanism itself, which is in doubt, rather than the boost system. Thus it is unclear whether the PBR would provide interesting or unique capabilities for early discrimination.
iv - Late-Midcourse Interactive Discrimination
The large mass and high power requirements of the Neutral Particle Beam (NPB) technologies rendered the NPB an unattractive candidate for the boost-phase interception mission.
However the NPB was regarded as a more attractive candidate for use in mid-course interactive discrimination. A neutral particle beam will penetrate through several centimeters of aluminum before it interacts with aluminum atoms to produce secondary radiation. The NPB could thus produce observable secondary radiation when it strikes a warhead which is much thicker than this depth, but it would pass through the very thin skin of a balloon decoy without interacting and producing observable secondary radiation. It was suggested that relatively simple sensors carried on interceptors could home in on the warheads stimulated secondary radiation.
Particle accelerator performance requirements for decoy discrimination are less demanding than for target destruction. A 10 ton NPB would be required for this mission, in contrast to the 75 ton device require for boost-phase interceptions. Current gradients of about 10 MeV per meter are anticipated.
This application could demonstrate significant synergisms for applications with kinetic energy weapons, by significantly relaxing the number of targets that must be attacked, with associated reductions in sensor and battle management requirements.
Major issues for this application include detector system capabilities, and the complicating effects of natural background radiation. There are very real questions as to whether these problems can be resolved.
The baseline booster for this mission is based on three MX Peacekeeper missiles strapped together. The very large diameter of this booster poses significant basing problems, as well as uncertainties concerning staging performance. A PBR based system, using the MX Peacekeeper stages previously discussed for post-boost phase intercept and interactive discrimination applications, could be used for this mission as well. A 10 ton NPB discrimination package could be lofted on a 2,000 km trajectory, permitting interrogation and discrimination of threat cloud objects prior to negation by ground-based interceptors.
However, as with the other discrimination concepts previously mentioned, the primary technical obstacle is with the effectiveness of the discrimination mechanism itself, and thus it does not appear that application of the PBR to these missions provides qualitatively novel capabilities.
v - Summary
Based on the foregoing, it does not appear that the particle bed reactor provides qualitatively unique strategic defense capabilities. Many of the missions that have been identified for PBR-propelled rockets are capable of being supported by conventional propulsion systems. Most of these applications face moderate to severe technical or operational challenges other than propulsion. Thus it is not surprising the Strategic Defense Initiative Organization determined in 1991 to suspend further support of this technology.
2 - Ballistic Missile
The Defense Department's Office of Strategic and Nuclear Forces has examined the use of the SNTP technology for a Light Strategic Missile, which could be deployed as a future replacement for the current force of Minuteman 3 intercontinental ballistic missiles.(4) Although this is not an application that has been studied by the SNTP program office itself, this application is illustrative of the range of uses to which the PBR system can be applied.
Current strategic missiles range from the intermediate range Pershing 2, which was dismantled under the INF Treaty, to the large 10-warhead MX Peacekeeper (Figure IV-34). The development of the single-warhead Small ICBM has been curtailed, in the wake of the end of the Cold War. These missiles all utilize solid propellants with specific impulses of about 280 seconds.
Particle bed reactors could deliver specific impulses two to four times greater, depending on the propellant used. Although a liquid hydrogen propellant missile could deliver a payload of up to 16% of its initial mass over intercontinental ranges, such a vehicle would be extremely large, given the low density of liquid hydrogen (Figure IV-35). Ammonia propellants have about eight times the density of liquid hydrogen. Ammonia has the additional advantage of ease of handling, in contrast to the notoriously fickle handling characteristics of liquid hydrogen. Given the difficulties of loading liquid hydrogen propellant, it is unclear that this propellant could be used in military systems that would require high alert rates, or the potential to be quickly brought to a high level of alert.
The specific power consumption for ammonia propellant, at a specific impulse of 475 seconds, it about three times less than that of hydrogen. The thrust-to-weight ratio of an ammonia engine is about twice that of a comparable hydrogen engine.
Use of higher density ammonia propellants would provide a throw weight of about 7% of launch weight. This compares with throw weights of 3% - 4% for current ballistic missiles. The actual throw-weight realized by the LSM depends on altitude at which the PBR engine is ignited. Higher altitude ignition may be desired in order to minimize environmental contamination concerns. Although this would be of marginal relevance during nuclear combat, it could have a significant impact on community interface acceptability during peacetime development and testing. Initiation of reactor operations above 185 kilometers would impose a roughly 30% penalty on missile throw-weight.
Ammonia-fueled Light Strategic Missiles could be considerably less massive than existing missiles with comparable throw-weights (Figure IV-36). The LSM-A, which is somewhat lighter than the Pershing 2, would have a throw-weight comparable to the Small ICBM, while the LSM-D, which is comparable in size to the Small ICBM, would have a performance similar to the Minuteman 3. These small missiles could be deployed in a mobile mode, enhancing their survivability.
However, these throw-weight characteristics are extremely sensitive to nuclear propulsion system performance, and are based on an optimistic assumption of a reactor thrust-to-weight ratio of 40 : 1. This is significantly higher than the 30 : 1 thrust-to-weight ratio assumed for most other PBR applications.
The end of the Cold War has substantially diminished political interest in new ICBMs, as witnessed by the declining fortunes of the Small ICBM. In the absence of a resurgence of interest in strategic force modernization, it is unlikely that the Light Strategic Missile applications would provide the basis for further development of PBR propulsion systems.
3 - Launch Vehicle
The results of the 1983 Defense Technologies Study, which formed the basis for the Strategic Defense Initiative program, clearly indicated a requirement to significantly upgrade currently programmed space transportation systems and provide additional space logistics capabilities. Even with programmed improvements, the Space Transportation System (STS) was viewed at the time as clearly unable to satisfy anticipated SDI requirements in a cost effective manner. Thus there was established a requirement to develop a heavy-lift launch vehicle for placing platforms of over one hundred thousand kilograms in mass into near earth orbit. In addition, capabilities to service, on orbit, a variety of space assets and to transfer satellites from low orbits to high orbits and to return them or move them from one orbit to another were also identified.
These requirements resulted in the initiation of the Advanced Launch System program, to develop a family of heavy-lift launch vehicles for SDI and other national requirements.
These requirements also led to a series of mission analysis studies comparing the performance of nuclear thermal, nuclear electric, and chemical propulsion, by the Air Force Astronautics Lab, Martin Marietta, Rockwell, Westinghouse, INEL and EG&G.(5) These studies considered a variety of mission applications, including a reusable orbital transfer vehicle and a disposable upper stage for a heavy launch vehicle (either shuttle-derived or Titan 4). The NERVA-derived nuclear engine selected for study had a thrust of 65-kN, slightly lower than the 72-kN Gamma engine previously designed by Los Alamos National Laboratory in 1985. Based on Nuclear Furnace test results, this engine used Uranium Carbide fuel elements operating at 3200 K for up to ten hours, yielding a 970 second specific impulse.
Subsequent studies focused on applications of PBR configurations. Advantages of the SNTP engine could include a specific impulse over twice that of existing rocket motors (1000 seconds versus 450), coupled with thrust-to-weight ratios approaching those of chemical engines. An upper stage using the SNTP motor could provide two to three times the payload to geostationary orbit provided by current chemical upper stages.
i - Orbital Transfer Vehicle
Babcock & Wilcox has also conducted an evaluation of a small Particle Bed Reactor applied to SDI low Earth orbit Orbital Maneuvering Vehicle servicing, which would be used to support multiple round-trips from 520 km initial orbit to 3000 km altitude constellation of space-based sensors and weapons.(6) The 9.5 MWt reactor, whose design is based prior company design studies for a 30 KWe closed cycle electrical supply system, produces over 2.5 kN of thrust with a 2100 K outlet temperature, and could transfer a payload of 4500 Kg from LEO to Lunar Orbit (Figure IV-37).
A study of the applications of the Particle Bed Reactor for orbital transfer vehicle applications was conducted for the Air Force Rocket Propulsion Laboratory by Brookhaven National Laboratory, Babcock and Wilcox, Garrett, and Grumman (Figure IV-38).(7) Although this study focused on military requirements for transferring payloads from LEO to GEO, many of the technical issues raised are applicable to Particle Bed Reactors generally.
ii - Launch Vehicle Upper Stage
One of the most prominently mentioned applications for SNTP systems is as an upper stage for military and other space launch vehicles. According to SNTP Program Manager LTC Gary Bleeker:(8)
"Our ultimate objective is to use a nuclear rocket engine to dramatically reduce cost, increase reliability and operability associated with routine access to space... The DOD's interest in nuclear propulsion is to apply its potential for safe reliable low cost access to space. In application to DOD missions we envision its use either as an upper stage or in a wide variety of potential launch vehicles..."
According to Bleeker, SNTP provides a significant:(9)
"... difference in capabilities and access to space for a given booster. Depending upon the configuration, increases in payload uplift range from about 70% increase in one case to a factor of five in other cases. In so much as a nuclear engine is similar to the conventional engine, we believe that these engines will likely cost about the same as similarly sized conventional engines."
The 1991 report of the Synthesis Group observed:(10)
"Newer concepts, such as the compact particle bed reactor, offer potential for high power density reactor cores which could lead to substantially higher integrated thrust-to-weight ratios. A high thrust-to-weight ratio engine would be particularly attractive for a second generation upper stage of an advanced heavy lift launch vehicle."
A range of applications for PBR launch vehicle upper stages have been suggested. Many of these utilize variants of the Centaur upper stage, which uses cryogenic liquid hydrogen and liquid oxygen propellants (Figures IV-39, IV-40).(11) Recent upper stage studies have focused on a higher thrust, 333 kN engine, which has the same specific impulse and thrust-to-weight ratio of prior engine concepts. The new designs also feature a restart capability, as well as improved provisions for managing post-burn decay heat. The design philosophy places greater emphasis on greater performance margin for improved safety and reliability.
In general these applications assume that the PBR engine would not be used at altitudes below about 100 km, reducing the potential for atmospheric contamination from fission products. However, these upper stages would begin operations while at suborbital velocities, raising the possibility that engine or guidance failures could result in the impact of the reactor on the Earth's surface.
MX Peacekeeper configurations previously discussed for SDI interceptor applications can also be used for space launch vehicle applications (Figure IV-41). While the present MX Peacekeeper ICBM, using three solid propellant stages, can place 1.4 tons of payload into low Earth orbit, a PBR upper stage and the MX first stage could place over three times this mass into orbit. Use of the first two stages of the MX with a PBR upper stage would increase the payload capability to over four times that of the baseline MX.
Atlas derivatives also demonstrate improvements, payload capability with the use of PBR upper stages, though less dramatic than in the case of the MX Peacekeeper (Figure IV-42). Relatively straight-forward modifications of the Atlas IIAS, incorporating the Centaur variants previously discussed, result in improvements of up to 50% in low Earth orbit
payload capabilities. More radical extensions of Atlas technology could provide payload capabilities double to triple that of the baseline Atlas.(12)
Titan launch vehicles could nearly double their payload capability to either low Earth orbit or to geostationary orbit through the use of PBR upper stages (Figure IV-43). Launch vehicles consisting of Titan 4 Solid Rocket Motor Upgrade strapon boosters, with large diameter PBR core stages (Figure IV-44), could double or triple that low Earth orbit payload of the baseline Titan 4.(13) The addition of a PBR upper stage could place more than 30 tons in geostationary orbit, over six times the capacity of the Titan 4. Even higher performance levels are possible with higher thrust PBR engines (Figure IV-45).(14)
National Launch System variants using PBR engines could reduce the number of types of vehicles that would have to be developed (Figure IV-46). Application of a PBR second stage to the NLS-3 first stage would provide payload capabilities similar to those projected for the NLS-1, while use of a twin-engine PBR second stage on an NLS-21 first stage would provide a payload capacity equivalent to that of the NLS-1.
Advanced Launch System configurations are no longer under active consideration, given the greatly reduced space logistics requirements of current SDI deployment plans, but earlier studies suggested that use of a core stage with PBR engines could more than double the payload capacity of the ALS booster (Figure IV-47).
iii - Launch Vehicle Main Stage
A cluster of six high-thrust PBR engines could be used in a single-stage to orbit launch vehicle which would not use chemical propellant boosters (Figure IV-45). This would require significant improvements in reactor and vehicle performance. Absent such improvements, AIAA Director of Science and Technology Policy Jerry Grey notes:(15)
"The physics and mechanics of nuclear rocketry make ground launch infeasible."
This proposal harkens back to the earliest concepts for nuclear missiles in the late 1940s, and some nuclear powered launch vehicle concepts advanced in the early 1960s. However, this application would require resolution of safety and environmental issues associated with ground-level operations of nuclear rocket engines that were not present when the earlier concepts were proposed. It is unclear that such operations would be tolerated by today's standards.
iv - Summary
Although the claims for launch performance gains are impressive, it is unclear whether these improvements would be realized in practice, or whether future launch needs would justify the significant development effort that would be required.
Most of these calculations have not taken into account the requirements to shield payloads and other vehicle components from nuclear radiation from the reactors operation, or to shield cryogenic propellants from the thermal flux from the operating engines. Resolution of these issues could significantly degrade the performance levels discussed to date.(16)
Probably of greater importance are questions concerning the utility of these enhanced payload capabilities. In the absence of a compelling advantage to using a nuclear engine, it is unlikely that the development cost, as well as the community interface problems posed by safety and environmental concerns, would be justified.
The primary mission for PBR orbital transfer vehicle configurations derived from the servicing requirements of large SDI space-based weapons constellations. Given recent developments in SDI architectures away from massive space-based systems, such requirements have largely disappeared. Indeed, the recent cancellation of the NASA Orbital Transfer Vehicle program calls into question the viability of OTV-like applications for the foreseeable future.
In general, the PBR upper stages have the effect of increasing the payload capacity of each launch vehicle up to the level currently offered by the next largest launcher in the American stable of boosters. Thus the MX Peacekeeper with a PBR upper stage would have the payload capacity of the present Delta, the PBR modified Atlas would have a performance equal to that of a Titan 4, and so on. Given the pre-existence of the Delta and Titan 4, there is no apparent need to replicate their current payload capabilities with PBR augmented versions of other smaller launch vehicles.
Highly modified versions of the Titan booster with PBR upper stages could provide payloads to geostationary orbits significantly in excess of current capabilities. While at present the Titan 4 with a Centaur upper stage can place 5 tons into geostationary orbit, replacing the Centaur with a PBR upper stage could double this payload mass (Figure IV-43). Use of Titan strap-on solid motors with a PBR core stage and upper stage could place over six times this mass, 31 tons, into geostationary orbit (Figure IV-44). Although such massive payloads might emerge at some point in the next century, there is no prospect that these capabilities would be utilized in the next decade or two.
Civil applications for large geostationary payloads, such as the Earth Observing System Geostationary Platforms, will not be launched until the early years of the next century, and under current plans can be accommodated on the Titan 4 Centaur. Long-range NASA plans have identified a handful number of larger payloads that might enter development in the first decade or two of the 21st century. But given the prevailing disenchantment with large scientific satellites, it is far whether these ideas will reach fruition. They are certainly unlikely to be realized on a time scale that would justify significant development of PBR engines in the near term.
Commercial applications for large geostationary payloads are extremely difficult to identify. Although large multi-user communications platforms were the subject of extensive study in the late 1970s and early 1980s, such concepts have been out of favor more recently. A surprise-free projection of commercial launch requirements over the next two decades would include primarily payloads of the Atlas-Ariane 4 class, with perhaps a few satellites of the Titan-Ariane 5 class.
Military requirements are the most difficult to anticipate, but again there is little apparent requirement for very large geostationary payloads, absent an unanticipated resurgence of interest in deployment of massive space-based anti-missile systems, such as were briefly in vogue in the mid-1980s. The largest current geostationary military payload is the Milstar communications satellite, which can be launched by a Titan 4 with a Centaur upper stage. The reorientation of the mission of this system from strategic to tactical applications is unlikely to lead to increases in the mass of follow-on spacecraft, and may result in significant reductions in spacecraft mass, given reduced survivability requirements. Each succeeding generation of large geostationary signals intelligence satellites has typically been about twice as massive as its predecessor, with the upcoming system requiring the full payload capacity of the Titan 4 with a Centaur upper stage. While the follow-on system that might be developed in the early 21st century could require a larger booster than is presently available, the end of the Cold War must cast doubt on whether such a large and expensive system could be justified. At a minimum, it is unlikely that this requirement alone could rationalize development of a PBR engine.
Although the addition of a PBR upper stage could eliminate the need for one of the three members of the NLS family of vehicles, the philosophy of the NLS program is minimizing the incremental costs of developing each of the vehicles by maximizing system commonality. The additional development and operational costs of the novel PBR upper stage element contravenes this philosophy.
Thus, on balance, it is difficult to identify a compelling launch vehicle mission application that would justify development of particle bed reactor engines.
4 - Space Exploration Initiative
Since it is presently funded exclusively by the Air Force, the SNTP effort is addressed solely to Air Force requirements. However, increased NASA participation in the program may result in consideration of civil space applications as well. The higher specific impulse offered by the PBR relative to the NERVA baseline offers increasing advantage when less favorable planetary alignments are used, and can enable faster missions that are not practical with NERVA. According to SNTP Program Manager LTC Gary Bleeker:(17)
"The same engine has almost immediate application to the nation's Space Exploration Initiative. It could also... reduce the cost of returning to the Moon, and enable safer and faster trips to Mars and other planets. It can also allow us to perform space exploration missions simply not previously possible. Successful technology development could result in a compact reactor capable of operating for several hundred -- perhaps a thousand -- seconds. This is sufficient for DOD, and probably NASA for their exploration. Manned Mars trips will require longer operating times, perhaps several thousand seconds."
In anticipation of such applications, the Idaho National Engineering Laboratory has conducted a preliminary analysis of the ground test facilities required to support development of a PBR rocket system with a thrust of 333 kN, operating at a chamber pressure of 3.447 MPa and a temperature of 3200oK. The engine was assumed to have a regeneratively cooled nozzle section with an 8:1 expansion ratio.(18)
i - Mission Analyses
Grumman has evaluated the applicability of the PBR to the Lunar Exploration Vehicle (Figure IV-48).(19) In contrast to the NASA baseline LEV which has an initial mass of 182 tons and requires aerobraking for Earth return, the PBR system would have an initial mass in LEO of 113 tons, and use propulsive braking for Earth return. In addition, the NASA baseline chemical + aerobrake system would leave significant LEV hardware elements at the Moon, while the nuclear alternative would permit returning the LEV to LEO. This analysis also suggested that Mars Transfer Vehicle initial mass in LEO would be about 20% less for a PBR vehicle with a specific impulse of 975 seconds, versus a NERVA-derived vehicle with a specific impulse of 925 seconds.
Another study by Grumman focused on reductions in flight time for piloted Mars missions (Figure IV-48).(20) This analysis concluded:
"... the projected engine T/W (10 : 1 with crew shielding) enables short opposition class mission durations, with round trip transit times on the order of 240 days, at initial masses of under 1000 tonnes for a mission centered about the 2016 Earth-Mars opposition. This represents a 60 day improvement over a mission with a low T/W nuclear thermal rocket. Conjunction class missions also benefit substantially, with one-way transit times as short as 60 days feasible."
Boeing Defense and Space Group considered the utility of PBR engines for a range of interplanetary missions (Figure IV-49).(21) Of particular interest is the evaluation of piloted expeditions to the moons of Jupiter using PBR engines, utilizing a Dash/Flyby mission mode, in which a Crew Excursion Vehicle dashes ahead to conduct a brief landing mission, and then rejoins the main spacecraft as it flies by the planet. Combining this mission mode with a PBR engine would enable a crew of eight to conduct this mission with a 4 to 6 year flight time.
An analysis by Babcock & Wilcox compared a 5,000 MWt PBR engine with a specific impulse of 950 seconds and a thrust-to-weight ratio of 25 : 1 with a NERVA-derived engine with a specific impulse 900 seconds and a thrust-to-weight ratio of 6.25 : 1. The study concluded that for conjunction-class split-sprint missions:(22)
"A Particle Bed Reactor-based nuclear thermal rocket was found to offer a 38% to 52% total mass savings compared with a NERVA-based nuclear thermal rocket... This advantage is primarily due to the higher thrust-to-weight ratio of the Particle Bed Reactor nuclear rocket."
An analysis conducted at NASA's Marshall Space Flight Center reached similar conclusions.(23) The study concluded:
"... the particle bed reactor engine with its high thrust-to-weight ratio (20) and high specific impulse (950 to 1050 seconds) will offer distinct advantages over the larger and heavier NERVA type engines..."
While the PBR was found to permit reductions of up to 200 tons in initial mass in LEO, this additional performance capability could also be used to reduce flight times:
"This time reduction can be quite significant, amounting to 15% to 20% of the total trip time, depending on the trajectory chosen."
INEL has conducted a mission analysis comparing the performance of nuclear thermal, nuclear electric, and chemical propulsion, in conjunction with the Air Force Astronautics Lab, Martin Marietta, Rockwell, Westinghouse, and EG&G.(24) This analysis considered a variety of mission applications, including a Lunar Transfer Vehicle, and a Mars Mission.
ii - Outstanding Issues
Despite the potential technical promise of the particle bed reactor, several major reservations remain. Adaptation of the Air Force SNTP system to SEI applications will require several modifications and improvements. Although the PBR has clear advantages over NERVA for those applications requiring high thrust to weight ratios and short operating times, the additional requirements of SEI applications may reduce or negate the attractiveness of this technology.
+ There are a number of areas of high technical risk in the PBR technology, and there is only limited technical data to support analysis of the actual level of risk involved;
+ Engine configurations appropriate for some applications, such as short burn time strategic defense interceptors, would require substantial modification to support other applications, such as piloted interplanetary missions. Piloted Mars missions will require run times of 3,000 to 5,000 seconds, rather than the 100 to 1,000 seconds presently contemplated for Air Force SNTP applications. Such applications would probably require operating at less-than-maximum temperatures, to maintain adequate performance and safety margin in the event of anomalous engine operations. This will require additional testing, and may result in alternate technical solutions, including alternate fuel geometries and performance characteristics;
+ Although present SNTP run times are consistent with those required for piloted Lunar missions, rating the system to improve reliability to support piloted operations will be required to support piloted Lunar missions, which will add to the mass of the engine;
+ Additional reactor radiation shielding is required for crew safety.
+ Resolution of these issues could substantially compromise the claimed advantages over conventional reactor designs -- in particular, the operating temperature of a PBR system configured for piloted interplanetary missions may only be about 100 K greater than that of an advanced NERVA-derived engine, resulting in a gain in specific impulse of only about 20 seconds.
+ While the high thrust-to-weight ratios anticipated for PBR engines could have a significant impact on the performance of missions, such as the anti-missile interceptor, where engine weight is a significant fraction of overall vehicle weight, this is likely to be much less significant for other applications, such as piloted interplanetary missions, where the engine is only a very small fraction of total vehicle mass.
Given these difficulties, some at NASA have been reluctant to embrace the PBR for SEI. John Clark, Deputy Manager for Systems at the NASA Lewis Research Center, concluded that PBR advocates are:(25)
"... trying to sell their program as the nuclear propulsion system that would do both (the Air Force and NASA's) jobs. We don't see that."
The application of PBR engines to the Space Exploration Initiative poses several questions:
Does particle bed technology offer significant performance improvements relative to more mature NERVA-derived solid core configurations?
Are these advantages sufficiently clearly established to warrant selection of the PBR as the preferred propulsion option at this time?
On balance, both of these questions would appear to elicit negative responses at this time.
While proponents of the PBR suggest that it may offer significant gains in performance relative to other concepts, these claims are based on assumptions that are difficult to validate either analytically or on the basis of the very limited test data presently available. While some of these claims may be validated by subsequent experiments, much work remains to be done.
Given the very low level of budgetary support for efforts directly linked to the Space Exploration Initiative (a few tens of millions of dollars annually), it would seem premature to expend many tens of millions of dollars annually on the development of PBR engines on the basis of their potential contribution to SEI missions. In the likely absence of equally robust funding for other propulsion alternatives, such as aerobraking, the PBR engine could become by default the propulsion system of choice for SEI.
The future of the particle bed reactor appeared was inextricably linked to the future of the Space Exploration Initiative. In the absence of significant increases in SEI funding, did not receive sufficient budgetary support to permit extensive hardware fabrication and testing.
1. United States Air Force Systems Command Phillips Laboratory, "Phillips Laboratory Announces Program in Space Propulsion," Office of Public Affairs release 92-02, 13 January 1992.
2. "Hughes, Martin and Rockwell selected for GBI program," SDI Monitor, 31 August 1990, page 197-198.
3. Department of the Air Force, Space Nuclear Thermal Propulsion (SNTP) Program -Final Environmental Impact Statement, 19 September 1991, partially declassified 11 March 1992, page 1.1-1.
4. Broad, William, "Pentagon Considering Reactors for Missiles," The New York Times, 20 August 1991, pages C1, C9.
5. Ramsthaler, Jack, and Baker, David, "Comparison of a Direct Thrust Nuclear Engine, a Nuclear Electric Engine, and a Chemical Engine for Future Space Missions," chapter 27 in Space Nuclear Power Systems 1988, (Orbit Book Company, Malabar, FL 1989), page 171-183.
6. Malloy, John, and Potekhen, Dick, "A Conceptual Study of the Use of a Particle Bed Reactor Nuclear Propulsion Module for the Orbital Maneuvering Vehicle," 24th Intersociety Energy Conversion Engineering Conference, 1989, volume 5, paper IEEE 899423.
7. Powell, J.R., et al, "Nuclear Propulsion Systems for Orbit Transfer based on the Particle Bed Reactor," chapter 28 in Space Nuclear Power Systems 1988, (Orbit Book Company, Malabar, FL, 1989), pages 185-198.
8. United States Air Force, "Scoping Meeting on the Environmental Impact Statement for the Space Nuclear Thermal Propulsion Program," 7-9 April 1992.
9. ibid.
10. Stafford, Thomas, America at the Threshold, (Arlington, VA, Synthesis Group, 1991) page 66.
11. "Space Technology," Military Space, 8 April 1991, page 8.
12. Henderson, Breck, "New Thermal Propulsion Gains to Speed Rocket Production," Aviation Week & Space Technology, 20 January 1992, pages 20-21.
13. Elmer-Dewitt, Philip, "Star Wars Does it Again," Time 15 April 1991, page 36.
14. Asker, Jim, "Particle Bed Reactor Central to SDI Nuclear Rocket Idea," Aviation Week & Space Technology, 8 April 1991, pages 18-19.
15. Kiernan, Vincent, and Lawler, Andrew, "SDIO's Timberwind Could Imperil Nuclear Space Programs," Space News, 8-14 April 1991, pages 3, 21.
16. Asker, Jim, "Particle Bed Reactor Central to SDI Nuclear Rocket Idea," Aviation Week & Space Technology, 8 April 1991, pages 18-19.
17. United States Air Force, "Scoping Meeting on the Environmental Impact Statement for the Space Nuclear Thermal Propulsion Program," 7-9 April 1992.
18. Whitbeck, Judson, and Olsen, Tim, "Preliminary Study of Facility Options for Ground Testing of a Nuclear Thermal Propulsion Engine," Idaho National Engineering Laboratory Informal Report, EGG-NPO-9548, June 1991.
19. Ludewig, H., "Particle Bed Reactor," NASA Nuclear Propulsion Workshop, NASA Lewis Research Center, Cleveland OH, July 10-12, 1990.
20. Venetoklis, P., et al, "Fast Missions to Mars with a Particle Bed Reactor Propulsion System," AIAA/NASA/OAI Conference on Advanced SEI Propulsion Technologies, Cleveland, Ohio, 4-6 September 1991, paper AIAA 91-3404.
21. Donahue, Benjamin, "Nuclear Thermal Propulsion Vehicle Design for the Mars Flyby with Surface Exploration Mission," AIAA/SAE/ASME 27th Joint Propulsion Conference, Sacramento, CA, 24-26 June 1991, paper AIAA-91-2561.
22. Walton, Lewis, and Malloy, John, "Nuclear Propulsion Tradeoffs for Manned Mars Missions,"Proceedings of the Eighth Symposium on Space Nuclear Power Systems, Albuquerque, NM, 6-10 January 1991, pages 603-606, AIP paper CONF-91-0116.
23. Emrich, W., and Young, A., "Nuclear Propulsion System Options for Mars Missions," AIAA Space Programs and Technologies Conference, Huntsville, AL, 24-27 March 1992, paper AIAA 92-1496.
24. Ramsthaler, Jack, and Baker, David, "Comparison of a Direct Thrust Nuclear Engine, a Nuclear Electric Engine, and a Chemical Engine for Future Space Missions," chapter 27 in Space Nuclear Power Systems 1988, (Orbit Book Company, Malabar, FL 1989), page 171-183.
25. Garbre, Andrew, "A Nuclear Rocket for Mars?" The Idaho Statesman, 20 April 1992, page 1.