The figures given in Table 11.2 indicate that orbiting of nuclear thermionic power plants, even in the megawatt range, is a problem which can be solved by modern launch vehicles such as the Saturn-5.* It is obvious from Table 11.2 that the main weight in plants of lower output is in the shielding.

*We recall that the Saturn-5 rocket carries an artificial satellite payload of 109 tons into orbit, whereas a l000-kW plant, according to the data of Table 11.2, would weigh approximately 15 tons.

Outside of the Soviet Union, designs for medium output power (20 - 200 kW) have been investigated, for example, in West Germany, with their ITR (Incore Thermionic Reactor) system [17, 21, 22]. The main purpose of the ITR system is to supply electric power for satellites in synchronous orbit which relay television programs directly to ordinary household antennas. The electric power required on board the satellite for this application is in the range of 1- 100 kW per TV program, depending on the area of signal reception on the earth.

The ITR system (Fig. 11.10) consists of an inner core which generates electric power and an outer driver core which provides the required critical size of the reactor. The required excess reactivity for two years of operation is 11%.

In the minimum size (200 kW of electrical power), the inner core contains 19 TFEs (thermionic fuel elements), each of which is surrounded by a moderator of zirconium hydride or yttrium hydride. The cuter core has 414 rods containing both fuel arid moderator.

The electrical power of a thermionic reactor can be increased with minimum alterations to the system by replacing some of the rods of the outer core with thermionic fuel elements. Thus, in the 30-kW version, the weight of the entire system is equal to 2,000 kg,

but in the 50-kW version it increases to only 2,500 kg. Radial beryllium reflectors in the form of rotating drums can be seen in Fig. 11.10. A neutron absorber, which provides coarse regulation of the reactivity, is located on part of the drum surface.

The main characteristics of the ITR system are presented in Table 11.3. Part of the thermionic fuel element - the diode converter with connections - is shown schematically in Fig. 11.11. The fuel (UO₂) in the form of pellets is located in channels 3.6 mm in diameter inside a molybdenum emitter (20 mm in diameter). The emitters are insulated from each other by ceramic rings, which simultaneously fix the interelectrode gap at 0.15 mm. Isolated channels are provided in the TFE to supply cesium to the diodes and to bleed off the gaseous fission products. Good equalization of the heat flux in the TFE is achieved by appropriately distributing the fuel and moderator concentration along the axis of the reactor. That is, there is less fuel in the middle region than at the ends. Special anti-diffusion layers are provided to reduce hydrogen loss from the hydride during operation. Some current-voltage characteristics of a single TFE diode, taken during electrical heating, are shown in Fig. 11.12.

Investigations are being carried out in the United States mainly on three designs [23, 24]: 10 kW, to supply power for the on-board equipment of automatic space stations; 40 kW, to supply power for a manned orbital space observatory; and 120 kW, to supply power for the auxiliary electric propulsion on unmanned satellites.

The 120-kW thermionic reactor has a neutron spectrum close to that of fast breeder reactors, while the smaller reactors have a moderator (ZrH).

Despite this appreciable difference it was possible to standardize the TFE’s considerably. The 120-kW thermionic reactor has 162 TFEs; the 10-kW reactor has 19 TFEs; and the 60-kW reactor has 60 TFEs. In the two latter reactors, the TFE region is surrounded by a booster core as in the ITR reactor. Each TFE has six diode converters, connected in series; the length of each diode is 7.36 cm. The fuel composition

consists of 90% UC and 10% ZrC. The emitter is tungsten, the collector is Nb, and the insulation is A1₂0₃.

The thermionic reactor for an orbital observatory (12 - 15 men) is being designed in two versions. In one version it is connected to the main station by a boom 60 m long, and in another version by a cable 3.2 km long. Electrical power is obtained through the cable by three-phase current at a voltage of 4.8 kV.
A great reduction in shielding weight is achieved in the cable version because of the separation of the reactor from the station.
A great deal of attention is being devoted to the reliability of the system, which is designed for continuous operation over a period of five years. Breakdown of individual diodes, radiator tubes and other components is compensated for by excess electric power at the beginning of operation.

In the United States, after testing TFEs for an extended period, full-scale ground tests of the thermionic reactors being designed is planned to begin during the period 1977—1978.

Some general characteristics of a thermionic reactor with a rating of 4 MW are presented in Table 11.5. This was designed by the General Electric Company to supply power for an electric propulsion engine.

There are plans to use thermionic converters not only in space but also undersea, for example, the French design of an autonomous nuclear power plant (20 kW) for operation on the ocean floor [25].

The thermal flux to TIC anodes at 700 - 800° C can be used by thermoelectric (semiconductor) generators (TEGs) [26] if water cooling is available. Thus, a compact plant for direct (non-mechanical) conversion of heat to electric power with an efficiency up to 30% can potentially be achieved.

The operating principle of a heat pipe can be easily explained by using Fig. 11.13 [27]. A heat pipe is a closed, evacuated vessel containing a small amount of working fluid. This fluid is evaporated in the hot zone, where heat is supplied; the vapor moves along the pipe and is condensed at the cold end, from which heat is removed. Return flow of the material is accomplished through capillaries built into the wall of the pipe, i.e., the wall is a kind of wick for the working fluid.* In a heat pipe, each atom (or molecule) which moves from the hot end to the cold end carries with it a very large amount of heat, equal to the evaporation energy of the liquid--this explains the enormous heat fluxes in heat pipes. The operation of heat pipes is independent of gravity, and consequently, they can be used in space nuclear power stations.

Two temperature ranges are of interest for TICs: 1400-2000° C to supply energy to the cathode and 500-900° C to dissipate heat from the anode. The development of heat pipes for anode temperatures is essentially completed [28]. Sodium is usually employed as the working fluid, and the walls can be made of stainless steel, nickel, or Nb-Zr alloy. There are reports of these pipes being tested for a period of 16,000 hours at 600° C [28]. A suitable working fluid for high-temperature pipes is Li, and for the highest temperatures is Ag. There are reports [28, 29] of a successful operation of heat pipes using lithium with walls of tantalum (including an yttrium additive to bind the hydrogen) for a period of 1000 hours at 1600° C, and one using silver with walls of tungsten at 1900° C.

Calculations [30] show that up to 5 kW can be transported at 800° C (with Na) and up to 500 kW can be transported at 1600° C (with Li) through a heat pipe 2 cm² in diameter and 1 m long. Heat flux densities along the pipe, up to 15 kW/cm at 1500° C, have already been achieved experimentally with lithium [29]. The heat flux densities supplied to the pipe in the radial direction reach 400 W/cm², which is already quite adequate to dissipate heat from a fuel element [7].

A thermionic reactor with heat pipes is being investigated at the Euratom Laboratory in Ispra, Italy for an output of 50 kW and 100 kW [31]. The core is a parallelepiped; heat pipes emerge through all the lateral sides. Each pipe operates its own diode converter. This unusual design was selected so that the load would be uniformly distributed among a large number of adjacent heat pipes in case of breakdown of any one heat pipe. To avoid high-temperature electrical insulation, the energy to the pipes operating at 1600° C is supplied by radiation from fuel elements having a temperature of 1900-2000° C. Heat transfer from the anode to the radiator is also accomplished by heat pipes.

An American design is described in [7], in which the cathode heat pipes exchange heat with the reactor pipes through a layer of high-temperature electric insulation (BeO) located behind the shielding. The converter is designed for a cathode temperature of 1800° K and an output electrical power of 100 kW at a voltage of 400 V. Versions were considered at the same time in which power varied from 10 kW to 10 MW, while cathode temperatures varied from 1400 to 2000° K. The weights obtained (in kg per 1 kW of electric power) are presented in Fig. 11.14 for two versions of shielding

An intriguing possibility is the use of the reactor fuel itself as the cathode material. Thus, numerous early experiments were carried out to study the emission characteristics of fuel material in Cs vapor [3]. Satisfactory results were achieved for UC with a ZrC additive. However, even in this case, the rate of evaporation was too high and the mechanical strength was insufficient. The rate of evaporation of uranium carbide (T_m.p. = 2400° C) at T = 2200° K is 5x10^-8 g/(cm².sec) [33], whereas the rate of evaporation of tungsten at this same temperature is almost five orders of magnitude lower. It is obvious that a converter with a suitable operation life cannot use cathodes for which the rate of evaporation exceeds 10^-9 - 10^-10 g/(cm².sec). Moreover, fuel structures become extremely brittle from operating at high temperatures--which sharply reduces reliability.

At the present time the main thrust of cathode development is with clad cathodes, i.e., the fuel is placed inside a cylindrical tube whose surface is the thermionic cathode. The highest thermionic emission is achieved with cesiated monocrystals, but oriented polycrystals of tungsten and rhenium (with a vacuum work function of approximately 5 eV) are also efficient cathodes [34]. The latter are prepared by decomposition at the cathode surface of volatile compounds

(e.g. WCl₆ or WF₆) or by electroplating. Various types of refractory alloys, in particular the W + Mo system and rhenium alloys, are also being considered as cathode materials [33, 35].

It is also necessary to provide sufficient purity in the materials used; impurities have in many cases been the cause of reduction in power output during extended tests. This degradation can be caused by the migration of impurities to the surface with a consequent change in the work function, and also by accumulation of impurities at the grain boundaries. The latter causes microcracks, which break the fuel and fission fragment containment and increase the thermal and electrical resistance [36].

The cathode and other high temperature parts should be carefully degassed, because dissolved gases can not only alter the electrode work function, but, in the gap, can also participate in transport reactions like the well-known water cycle (see §11, Chapter 2). Transfer of cathode material to the anode by transport reactions can be the cause of gradual degradation and even short-circuiting of the converter.

There are also difficulties in the selection of the anode material. From the viewpoint of conversion efficiency, the anode work function should be as low as possible and stable. Rather low work functions have been achieved in the laboratory. For example, a work function of 1.1 eV has been obtained with ZrC at low temperatures (T_A » 450° C). However, during the operation of a TIC, the anode surface is covered with evaporated cathode material. Therefore, it is difficult to achieve a work function below 1.5-1.6 eV with extended operation.

Molybdenum, nickel, or niobium are chosen in most designs as the anode material. From the viewpoint of work function, Mo is preferred; under TIC conditions, it has a work function 0.2-0.15 eV less than the others. Nickel is easier to machine and joins better to other materials. From extensive testing with 135 laboratory devices [11], one recommendation is that a rhenium anode, which acquires a work function of 1.47 eV, be used in combination with a rhenium emitter.

Insulation and structural materials. The insulation materials used in reactor TICs should provide reliable electric insulation between electrodes and also between the anode and the cooling system. These materials should have high mechanical strength and have good radiation and corrosion resistance at high temperatures in cesium vapor. The interelectrode insulation should also be vacuum tight and should combine well in a metal—ceramic seal. Under TIC conditions, Nb-A1₂0₃ junctions perform well, i.e., if these materials are carefully cleaned of impurities [36, 37].

materials is proceeding in two main directions: the development of insulation coatings and the manufacture of parts for electrical insulation [38], A1₂O₃ and BeO being used in the initial components. Numerous investigations have shown that these ceramics are rather stable under thermionic reactor conditions, at least up to 1100-1200° K. Cermet systems based on BeO and A1₂O₃ with additives of refractory metals (for example, Mo and Nb) are very promising. They retain the good insulation properties of their ceramic base and at the same time have new important qualities: better plasticity and thermal conductivity. By varying the percentage content of the metal additive, thermally matched cermet compounds can be developed. Also, multilayer insulators are being investigated.

In-core tests [39, 40] have shown that swelling of the insulating materials is caused by the flux of fast neutrons which have an energy greater than 0.1 MeV. Fast neutrons knock the atom out of crystal lattice positions, and the resulting mechanical stresses lead to microcracks along the grain boundaries. Special experiments have shown that ceramics with minimum grain dimensions and minimum impurity content (especially silicon) swell less and retain their vacuum density better. Such parts manufactured from A1₂0₃ with grain dimensions about 3-5 m m tolerate fast neutron fluxes (E_n > 0.1 MeV) up to 6x10²¹ neutrons/cm² at T » 1000° C. In this case, volume swelling does not exceed 3%, the modulus of elasticity decreased by only 30%, and the vacuum density, good thermal conductivity and high insulating properties are retained. Even some increase of resistance during irradiation was observed with BeO [41].

However, swelling even by 3% is not always permissible because it can lead to large mechanical stresses and deformations. For this reason, insulators based on yttrium oxide [39] are preferred because they do not undergo appreciable swelling in fast neutron fluxes up to 5x10²¹ neutrons/cm². However, it should be noted that Y₂O₃ has poorer insulating properties, lower thermal conductivity, and less mechanical strength compared to A1₂O₃. Yttrium oxide has a linear expansion coefficient close to that of niobium, which permits good vacuum tight and thermally matched Y₂O₃-Nb junctions.

The most acceptable structural materials for a neutron-environment are Nb, Mo, and stainless steel [33]. The use of Ta is undesirable because of its tendency to enter into numerous chemical reactions [36].

Materials also change their structure and mechanical properties under the effects of neutron radiation. Monocrystals of molybdenum and tungsten, irradiated in a BR-5 reactor with an integrated flux of 1.4x10²² (including 1.4x10²¹ fast neutrons with E_n > 1 MeV), were described in [42]. Calculation of the nuclear reactions showed that 0.2% Re is accumulated under these conditions in tungsten and 0.1% Nb and Zr is accumulated in molybdenum. Also, other significant changes in structure occur: single vacancies gather in a cluster and the crystal becomes divided into individual grains. On the boundaries between these grains gather dislocations and impurities. High-temperature annealing restores the initial mechanical and electric properties, but not completely.

Let us consider some of these problems. The fuel in [45] was sintered uranium dioxide UO₂ or sintered uranium carbide with zirconium carbide (90% UC + 10% ZrC).@ The uranium was enriched to 93% U²³⁵. The fuel was placed in a tungsten, vapor deposited (WF₆) cladding 1 mm thick. The anode was molybdenum. The geometric dimensions are - cathode diameter 1.64 cm, cathode area 13 cm², and interelectrode spacing 0.2 mm. The volume for fission product retention was 20 cm³. Heaters were provided in the system to vary the temperature of the cesium reservoir and the anode; thermocouples were provided to control the temperature at different points; and also potentiometric leads were attached to give the current-voltage characteristics. The maximum temperature inside the fuel element was not measured, but was estimated to be about 2200° C. At the surface, the temperature was 1650 - 1800° C.

The above tests showed that converters with U0₂ have a number of advantages over those with UC. The converter parameters with U0₂ fuel at T_C = 1730° C hardly varied over a period of 5000 hours; there was more of a tendency toward electrical power increase than toward a decrease. The authors [45] link this increase of efficiency to the diffusion of oxygen through the tungsten fuel cladding, which leads to a favorable change in the cathode and anode work functions. A similar phenomenon was observed in [46], where the output parameters were strongly dependent on anode temperature. The latter result was possibly related to the formation of cesium oxide on the anode.

*According to the data of [43, 44], a plasma diode with rhenium electrodes operated stably under laboratory conditions for a period of 20,700 hours at a cathode temperature of 2000° K with an output of P = 25.6 W/cm² and V_L = 0.77 V.

@ZrC was added to increase the fuel melting point.

Although the results with UO₂ are promising, one must bear in mind that uranium dioxide has one great disadvantage over uranium carbide--its rather low thermal conductivity. Thus, there are large temperature drops inside the UO2 fuel elements [47]. In a number of cases, cathodes with U0₂ have deformed more severely than those with UC [48].

With UC fuel, the initial electric power over a period of 2,500 hours decreased by approximately 20%, but efficiency decreased even more, which indicates an increase of thermal losses in the converter. The decrease of electrical power in [45, 48] is explained by the diffusion of uranium through the fuel element cladding which leads to a decrease of the vacuum work function of the cathode and of total efficiency. The opinion was advanced that the diffusion of uranium was enhanced by the lack of carbon compared to a stoichiometric composition. Therefore, the ratio of carbon to uranium concentration (C/U) was increased to 1.03-1.05. This, however led to an increase of carbon diffusion through the emitter cladding [48]. Measures were subsequently taken to reduce the effects of carbon diffusion through the emitter (a special structure of the tungsten cladding) and also increase the solubility of carbon in the collector material. Thus, stable results were achieved with a ratio of C/U = 1.01.

Containment in the TIC of inert gases and other elements produced in the fuel element during the fission of uranium has been studied in many investigations [49-51]. The main attention has been devoted to the yield of inert gases. The dependence of the effective diffusion coefficient of Kr and Xe on the temperature for a U0₂-Mo cermet is presented in [49]. The diffusion theory for inert gas containment in the fuel predicts a very strong dependence of yield on the temperature and operating time of the fuel element, which has been well confirmed by experiment. However, the anticipated dependence on the U0₂ particle size has not been observed: degassing of particles 125 m m in diameter hardly differs from 22 m m in diameter.

to observations [49], molybdenum coating of UO₂ particles does not affect the escape rate of inert gases from the fuel. The application of a molybdenum shell around the individual U0₂ particles (grains) also does not help, because the shells undergo considerable structural change and lose their gas tightness after 1000 hours of operation in the reactor at T = 2000° K [50].

As already noted, release of gases inside gas tight fuel elements may lead to swelling of the cladding and, in the final analysis, to short-circuiting of the cathode to the anode. Thus, an initial gap of 0.175 mm in [52] was short-circuited after 3596 hours (T_C = 1600° C) at a calculated pressure inside the fuel element of approximately 6 atm. A similar fuel element design in which the gases could escape into the cesium container operated normally, except that the heat dissipation from the emitter increased somewhat with time (approximately 10%) because of heat transfer through the gases.

Numerous reactor tests carried out during the past few years have confirmed that diodes with cathodes of Mo (coated with W) and with anodes of Mo may operate stably in the core for an extended period [53-55].

The world’s first complete thermionic reactor--the Topaz I installation--was tested in the USSR in 1970 [56, 57]. A structural diagram of Topaz-I is shown in Fig. 11.15.

High uniformity of energy liberation along the radius (the variation factor did not exceed 1.05) was achieved by design measures and by shaping the fuel and moderator (zirconium hydride). Rotary beryllium cylinders, with absorption linings of boron carbide, arranged on a radial reflector were used to regulate the reactor. The use of these cylinders made it possible to eliminate distortion of the neutron field during the

The current-voltage characteristics of the entire thermionic reactor are shown in Fig. 11.16 a and b; they are naturally dependent on the thermal output of the reactor. The electric power generated varied in the main section from 3.6 to 7.2 kW and in the auxiliary section from 1.2 to 2.7 kW.

Topaz-I successfully completed its assigned operating life--l000 hours. A Topaz-II model has had a total operating life of approximately 6000 hours [57].

Selection of a specific isotope as a heat source is dependent on a number of factors: the specific power generated by the isotope (watts per unit weight), the half-life, cost, etc. With all characteristics being equal, preference is given to isotopes which have only a - and b -activity, because better radiation shielding can be provided in this case and the assembly of the generator is easier [65]. When using a -active isotopes, removal of the liberated helium is required. Some data of the most promising isotopes, according to [58, 59], are presented in Table 11.6.

Consider now as an example the isotope generator with an output of several tens of watts (Fig. 11.17) developed by Belgian and West German companies [60]. The heat source is actinium, which changes to lead after a number of radioactive decays. The radioactive material is placed in an airtight fuel capsule of W-Re alloy, which has good thermal contact with the TIC emitter at one side of the capsule. The other sides have thermal insulation consisting of layers of metal foil separated by vacuum spaces [61]. The heat travels from the anode to the radiator which is connected to the anode by a heat pipe. The fuel is Ac₂O₃ particles, 500 m m in size, covered by a layer of 25 m m thick ThO₂ and enclosed in a tungsten matrix. This dispersion system; 1) easily passes helium (with conversion of one atom of Ac to Pb, 5 atoms of He are produced), which then emerges to the outside through the walls and does not interfere with the operation of the TIC; 2) combines with the oxygen obtained from the Ac₂O₃ during the decay of Ac; 3) does not swell and 4) at the same time, provides sufficiently high heat conduction (» 22% of the heat conduction of metallic tungsten) that there are no large temperature drops in the fuel. The thermal

power density in the matrix is 45 W/cm³ compared to 182 W/cm³ in pure Ac.

Polycrystalline oriented [110] tungsten, rhenium, or a W-Re alloy can be used for the cathode. The interelectrode spacing has been optimized in the design. The optimum for polycrystalline tungsten is d = 0.08 mm. It is larger for electrodes with higher vacuum work functions, which of course is desirable from the viewpoint of device reliability. The anode is made of molybdenum and has a minimum work function of approximately 1.48 eV at T_A = 405° C. All the materials used for the generator were checked for chemical compatibility, and gas tightness was tested over a period of 5000 hours at a cesium pressure of 20 torr.

The total thermal power of the device is 250 W. Heat losses were 95 W (through the thermal insulation, 13 W; from the non-operating part of the emitter, 17 W; and from the electron cooling of the cathode, 65 W). The output characteristics of the diode at T_C = 1527, 1577, and 1627° C are shown in Fig. 11.18. The output electric power for 1527° C is equal to 23 W with a converter efficiency of approximately 15% and total efficiency on the order of 9% (considering all heat losses).

The American SNAP—13 generator [62], which yields P = 12.5 W with h = 7% at T_C = 1375° C and P = 20 W with P = 8.6% at T_C = 1404° C, has similar parameters. The SNAP—13 generator tolerated all dynamic loads and was tested over a period of 13,000 hours during electric heating.

In 1966, investigation began in the United States of a standard isotope module with an electrical power of 100 W. Connecting such modules together, electrical power units up to 2 kW of electric power can be obtained. The module is designed to operate for one year (with Cm²⁴⁴) or 90 days (with Po²¹⁰). Compared to a thermoelectric generator giving similar power, TICs would have a number of significant advantages [61] for such a module:

Large isotope TICs to supply 2 to 10 kW of electrical power for electric propulsion engines and on-board equipment of unmanned spacecraft to distant planets--Saturn, Uranus and Neptune--have also been studied recently [63].

The main elements of a 5-kW generator (52 kW thermal) assembled prior to launch are shown in Fig. ll.19a. The shielding (LiH) and the flow-through water cooling system are also shown. The isotope capsule consists of pipes with curium oxide (Cm₂²⁴⁴O₃), between which the cathode and auxiliary heat pipes pass (for heat dissipation to the water container at launch). There are a heat shield, parachutes and a shock absorber, as well as a large special radiator for heat dissipation, to prevent dispersion of the isotope into the atmosphere in case of emergency reentry.

After injection into a free hyperbolic trajectory, the shielding system is separated and the generator is shifted a distance of 6 m to reduce radiation to the instrument compartment. The temperature increases and the TIC begins to operate. Heat is removed from the converter diode by the anode heat pipes, which form a radiator for heat rejection by radiation. The cathode temperature is 1900° K, anode temperature is 1000° K, and the interelectrode distance is 0.2 mm.

The low launch weight of the electric power plant, 142 kg (the weight of the separated shielding is 304 kg), and the long operation life, 72,000 hours, are noteworthy. The fuel load is 1.5 megacuries.

Besides isotope TICs, which operate in the ordinary arc mode, converters are being designed with a very small interelectrode distance which operate without compensation of electron space charge. Cesium vapor is introduced only to optimize the electrode work functions. The main advantage of these TICs is the possibility of operating at lower temperatures (1400° K or lower), which increases the reliability of the entire system, despite the decrease of the inter-electrode gap.

As an example, we cite some characteristics of the Isomite converter [64]. A sectional view of one of these converter meddles is shown in Fig. 11.20. The electric

power of the module may vary from 100 m W to several W, depending on the geometric dimensions. The Cs pressure in the gap is determined by the anode temperature, but electron scattering from Cs atoms can be disregarded because of the small gap (0.025 mm) and calculations can be carried out by formulas for the vacuum mode (see Chapter 2).

Special cathodes and anodes based on iridium, as well as oxidized tantalum and tungsten, are being developed for diode converters. At cathode operating temperatures, the work function is within the range 1.7-2.3 eV, and at anode temperatures, the collector work function is 1.34-1.54 eV. The Ta-C-Cs system was tested successfully over an extended period; it is suggested that the surface oxygen losses from the electrodes are completely compensated for by diffusion from the bulk tantalum. It is suggested that curium cermet with a heat release density of 19.1 W/cm³ or plutonium with a heat release of 3.5 W/cm³ be used as the fuel in the Isomite.

The individual elements are series-parallel connected to increase reliability. Thus, a total power of 50-200 W and the required voltage to convert direct current to alternating current are achieved. The design operation life of the TIC is 5-10 years.

The authors [64] feel that, based on existing technology, the Isomite can be developed for a series of space power sources with an efficiency of 7% and a specific power of 10 W/kg, with prospects for an increase of efficiency to 10% and of specific power to 50 W/kg.

It is important to increase the efficiency of central station electric power plants to conserve fuel and to reduce thermal waste. To raise efficiency, it is necessary to use a heat engine operating in the temperature range above the steam-turbine cycle (i.e., above 550-600° C).

Compared to MHD (see §1, Chapter 1), a TIC has a thermodynamic cycle closer to ideal and has a better temperature match with the steam turbine. However, it is necessary to have a larger total electrode surface for the TIC, and consequently, it is necessary to appropriately combine very many diode converters. TICs may also be used in nuclear power plants by increasing the coolant temperature, and can be used with fossil fuels, given the development of coatings resistant to combustion products [66]. The calculated technical and economic promise of electric power plants using thermionic converters is very intriguing [67].