D O E / E T / 1 5 4 2 0 — T l

DE03 0C4388

CESIUM-PLASMA-CONDUCTIVITY ENHANCEMENT IN THE

ADVANCED THERMIONIC ENERGY CONVERTER

Final Report

by

Constantine N. Manikopoulos, Pr inc ipal Invest iga tor Department of E lec t r i ca l Engineering

College of Engineering Rutgers, The State Univers i ty o f New Jersey

Piscataway, NJ 08854

NOTICE

PORTIONS OF THIS REPORT ARE ILLEGIBLE. It has been reproduced from the best available copy to permit the broadest possible avail-ability.

Sponsored by

U.S. Department of Energy Chicago Operations Office 9800 South Cass Avenue Argonne, IL 60439

Contract Identification Number DE-AC02-79ET 15420

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency Thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

DISCLAIMER Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

TABLE OF CONTENTS

Abstract iii

1. Introduction 1

2. Resonance Photoabsorption in a Cesium Discharge 3

a. Theoretical Background 3

b. Experimental Results in Absorption of Resonance Radiation 11

3. Ionization Processes in a Cesium Discharge in the

Presence of Resonance Radiation 13

a. Theory 13

b. Experimental Method 23

c. Experimental Results 25

4. Cesium Plasma Sustenance by Application of

Microwave Power 25

a. Experimental Measurements 25

b. Theoretical Considerations 32 References 40

ii

ABSTRACT

Two methods of plasma conductivity enhancement in a cesium vapor thermionic

energy converter have been studied. The first involves resonance photoabsorption

of several cesium lines and the second utilizes cesium plasma sustenance by

application of microwave power.

An extensive study of ionization processes in a cesium discharge in the

presence of resonance ionization was made. Calculations were made of expected

percentage excitation levels for several cesium resonance transitions for

different values of neutral density and temperature as well as incident radiation

power levels. The results of some of these computations were tabulated. Several

ionization schemes were considered. A number of cesium transitions were investigated

in the range of 799 to 870 nanometers for four different cesium reservoir tempera­

tures, 467, 511, 550 and 591 K. The related absorption coefficients of the radia­

tion lines in the plasma were deduced and tabulated. The resulting plasma

conductivity increase was recorded and the associated ionization enhancement was

deduced.

A microwave cavity was built where the emitter and collector of a simple

thermionic converter made up two of the cavity walls and resonant microwave power

was externally applied. The I-V characteristics of the thermionic converter

were studied under several microwave power levels in the range of 0-2 watts.

Significant shifts to higher currents were observed as the microwave power levels

were raised. It was also found that the knee of the ignition point of the I-V

curves shifted to lower voltage as the microwave power increased and then dis­

appeared completely at the level of about 1 watt.

In conclusion, both methods show promise as auxiliary ionization mechanisms

for the thermionic energy converter, especially at low emitter temperatures.

i i i

1. INTRODUCTION

The thermionic energy coverter has already been successfully applied to power

systems in a variety of space vehicles. The primary requirements in such applica­

tions are high power to weight ratio and reliable performance. Recently, though,

the focus of utilization of the thermionic converter has shifted from space to

terrestrial applications where efficiency and cost are the dominant considerations.

In the search of cost reduction, the use of stainless steel or other alloys is

favored over refractory metals which are expensive in all facets of construction.

The penalty paid is that the emitter temperature now attainable is about 1600°K

which is considerably smaller than the range of 2200-2400°K utilized with refractory

metals. The associated collector temperature desired falls in the range of 800°K.

These lower temperatures are a better match to conventional power generators util­

izing the steam cycle, a fact which makes the thermionic converter quite attractive

in topping applications.

It is clear that we are demanding a strikingly large reduction of the emitter

temperature in comparison to the space generation converters. The direct result

of this emitter temperature reduction is a considerable reduction in the conductivity

of the plasma. This is very undesirable in thermionic converters where high effi­

ciency and high power density are required. It is thus important to devise a high

efficiency ion generation scheme for the advanced converter in order to reduce the

plasma resistivity. This must be accomplished without disturbing the quality of

the performance of the emitter and the collector electrodes.

It is evident that some external source of energy must be directed into the

interelectrode spacing to produce cesium ions. The energy consumption of this

source can then be interpreted as an equivalent plasma arc voltage drop. The perform­

ance of the overall system of electrodes and plasma should result in considerable

reduction of losses in order to achieve a viable converter.

The arc-drop V. can be eliminated theoretically simply by reducing the inter­

electrode spacing d such that both collisional and space charge effects can be

neglected. In practice, however, this required value of d is much too small (mainly

because of space charge effect) to be a viable alternative. Following S. H. Lam [1]

we shall consider primarily thermionic energy converters which satisfy the follow­

ing practical constraints:

1. The spacing d is relatively wide, probably 0.5 mm or larger.

2. The output current density J is relatively high, probably of the order of 2

1 amp/cm or larger.

1

3.' The emitter temperature Tr is re lat ive ly low, probably in the range

1500°K-1800°K.

These practical constraints have the following immediate consequences. Items

(2) and (3) render the emitter impotent as a positive ion supplier; thus the con­

verter must operate with other posit ive ion mechanisms. Items (1) and (2) set a

lower l im i t for pd (m i l - t o r r ) ; in general we expect pd to be a number larger than

unity. Thus, the electrons must suffer col l is ions as they flow across the converter.

I f ions are produced by an auxi l iary process, an equivalent plasma maintenance

voltage drop V . must be introduced as a converter voltage loss [2]

I V V V* - auxi l iary power _ a a _ a d ~ output current I A"

where V is of the order of the ionization potent ia l , V., for the ignited tr iode and

is the order of the difference between V. and the ionizer work function (J)T for the

unignited t r iode. A = I / I is the current amplif ication factor. The adjusted a

barrier index would thus read:

\ = •c - d * <

In summary, advanced performance in converters can be achieved through the

significant reduction of 41 or the suppression of V.. But these accomplishments

must be achieved in such a way that a negligible voltage loss (V .) is introduced

by the auxiliary ion production.

A very direct and informative method of achieving expected efficiency increases

in the operation of a thermionic converter is by laser line pumping of the thermionic

cesium plasma. Let us consider an atom pumped by a laser from the lower cesium

states to some upper level. If in this process, the "bottleneck" in ionization is

crossed, it is expected that the excited cesium atom will preferentially be collision-

ally excited by electrons in the plasma to higher levels eventually becoming an ion.

In order to produce ions, direct photoionization will not be necessary; laser pump­

ing followed by step ionization in the plasma can accomplish this task. In principle,

such a laser line ionization process should be the most efficient in comparison to

all others, if for the moment we overlook the energy required to generate the laser

radiation. Furthermore, it is a "clean" method of studying the ionization require­

ments and effects in the thermionic plasma because it essentially decouples ioniza­

tion from the electron gas, i.e. no hot electrons are required for ionization.

2

Therefore, we can now increase the electron density without raising the electron

temperature. This is important in providing experimental verification to recent

theoretical predictions about the required level of the concentration in a converter

for significantly reducing the plasma arc drop losses.

A very intriguing, although admittedly very speculative, operating mode of a

thermionic converter could evolve in which a thermionic laser is combined with a

laser pumped region. In this design, part of the cesium plasma discharge would be

operating in the thermionic laser mode supplying laser radiation to another part

of the discharge where it would be used to enhance the ionization and thus efficiently

produce electrical power.

The cesium plasma resistivity in the thermionic converter has been the focus of

much attention in the past several years. Theoretical discussions and treatments

have been presented by several researchers in the field.

Laboratory experiments have been carried out in which we have achieved cesium

plasma sustenance or pumping by the application of either DC or pulsed microwave

power into a thermionic converter. External microwave power of high frequency and

high power are applied to the interelectrodes spacing. The electron temperature and

plasma density are thus raised while the plasma resistivity is lowered. By controlling

the input microwave power level and/or frequency, we may change the plasma density

and electron temperature. These changes can be achieved either in the DC or in

the pulsed mode.

There are several advantages in utilizing microwave power as an external source

of ionization, namely, simplicity of power coupling (electodeless), flexibility in

control (both input power level and frequency are easily adjustable), minimum dis­

turbance to the thermionic surfaces, and the possibility in some modes (particularly

TEM modes) to couple power into very narrow emitter collector separations. Some

preliminary work has been done in the area with encouraging first results [3,4].

2. RESONANCE PHOTOABSORPTION IN A CESIUM DISCHARGE

a. Theoretical Background

The amount of radiant energy that a gas, occupying a known volume at a known

vapor pressure, can absorb may be computed if the absorption coefficient of the

gas as a function of frequency is known and if the spectral distribution of radia­

tion incident on the sample is given. The observed profile or shape of an absorption

line of a monatomic gas is strongly influenced by a number of factors including,

for example, the neutral and plasma temperature, the frequency of collision between

3

the absorbing atom and atoms of the same or different kind, the frequency of collision

between the absorbing atom and positively charged ions or electrons, natural uncer­

tainty in the energy levels involved in the optical transition and instrumental

broadening of the absorption line.

If a continuum of radiation is passed through an absorption cell containing a

monatomic gas at a given vapor pressure, an attenuation in the intensity of the

beam will be observed in the region of a resonance line of the atom. The intensity

of the transmitted radiation will exhibit a frequency distribution similar to that

depicted in Figure 1. The absorption coefficient of the gas (k ) is defined by

the relation:

I^=I„c-V (1)

where Lis the thickness of the absorbing layer.

If L is measured in cm, then k has the dimensions of cm" . Using Eq. (1) in

conjunction with the measured transmitted intensity, k may be determined as a function

of frequency and the relationship can be plotted graphically as is shown in Figure 2.

An experimental determination of the absorption coefficient frequency contour

is of fundamental importance not only because it serves as a measure of the amount

of radiation that the sample absorbs, but also by virtue of the fact that the half-

width, Av, is inversely proportional to the lifetime of the individual excited state.

In addition, the qualitative shape of the contour gives an indication of the relative

importance of the various perturbations that the absorbing atoms are subjected to.

The most interesting, but complicated, form of line broadening is that caused

by collisions between the absorbing atom and other neutral or electrically charged

particles. A great deal of experimental and theoretical work has been entered into

in recent years in an effort to understand this phenomenon.

While it is customary to reduce an experimentally determined absorption coeffi­

cient (k ) to standard conditions, such a reduction is not always valid and is

essentially meaningless for pressure-broadened absorption lines of the type which

will be the subject of this discussion.

Once k has been determined for a given path length, L, the rate at which the

energy lying between \ and A + dx is removed from a light beam by a sample of dimen-

sions (1 x 1 X L)cm can be computed from:

^abs(x) = ^0^^) e-^L- (2)

4

X

o o UJ

5 z < a: h-

O >

z UJ

z 3

FIG.

1 0

FREQUENCY V IN CM*'

GENERALIZED ABSORPTION LINE SHOWING INTENSITY OF TRANSMITTED LIGHT AS A FUNCTION OF WAVE-NUMBER (FREQUENCY)

1 0 - i

FREQUENCY IN CM - ^

FIG. 2 GENERALIZED ABSORPTION COEFFICIENT-FREQUENCY CON 'GUR

5

providing that the value of k used in Eq. (2) was obtained experimentally by

measuring the transmittance of a gas sample which was under identical conditions of

temperature, pressure, and concentration of foreign gases as the gas for which the

calculation is made. This restriction is invoked for pressure-broadened lines but

is not generally necessary for atomic absorption lines. For example, within the

temperature and pressure range where the influence of interatomic forces are

negligible, the same k may be applied to gas samples under a variety of conditions

temperature, pressure, and concentration.

Generally speaking, at total vapor pressures of less than 1 atm, atomic absorp­

tion lines can be expected to be characteristically narrow and the wavelength

interval over which energy may be removed from a beam of radiation passing through

an atomic gas is usually of the order of a few angsrom units or less, for any given

atomic transition.

The resonance absorption lines (that is, the first number of the principal

series of absorption lines) of the alkali metals present an exception to the

generalization stated above since relatively long-range interatomic perturbations

come into play which result in significant variations in the quantized energy levels

of the absorbing atoms. As a result, the resonance absorption lines are broadened

to an unusually large extent. The manner in which the interatomic forces give rise

to line broadening can be illustrated with the help of Figure 3, in which the

potential energy of a system composed of two cesium atoms is plotted against the

interatomic separation.

When a cesium atom is not subject to the perturbing fields due to the neighbor­

ing atom, the energy difference E^-E, determines the wavelength of the radiation

which is absorbed by the atom. This is the case if the two atoms are separated by

an infinite distance. At this infinite separation the width of the absorption line

is determined by natural uncertainty and by the Doppler effect. The broadening due

to these two effects results in an increase in line width of the order of a hundredth

of an angstrom, and both of these broadening processes become negligible compared

to other broadening processes which occur at distances where the interatomic forces

become important.

The forces which occur between two neutral atoms of the same species which are o

in close proximity (i.e., at a separation of less than 300 A) are: (1) resonance

forces, (2) dispersion forces, and (3) exchange forces. The resonance and dispersion

forces occur at distances where the charge distribution of the two atoms do not

overlap. The exchange forces result from an overlapping of two charge distributions,

and give rise to the familiar valence or binding forces which result in a stable

6

POTENTIAL ENERGY

INTERATOMIC DISTANCE, r

FIG. 3 SCHEMATIC TRENDS OF POTENTIAL ENERGY CURVES DUE TO INTERATOMIC FORCES

diatomic molecule. It is important to emphasize, however, that exchange forces may

also be repulsive and, in fact, for two ground-state cesium atoms the probability

that a collision gives rise to repulsive forces is approximately three times as great

as the probability for attractive forces.

Expressions for resonance and dispersion forces can be obtained by expanding the

classical expression for the non-overlapping charge distributions and computing the

quantum mechanical interaction by means of perturbation theory. The resonance fores

are due fundamentally to the fact that the perturbing atom is essentially indis­

tinguishable from the absorbing atom and dominates the dispersion forces when the

probability of an absorbing atom colliding with an atom of its own kind is very much

greater than the probability for a collision with a foreign atom. The dominance of

the resonance forces over the dispersion forces is due to the fact that the resonance

forces give rise to a first-order perturbation which can be shown to be proportional _3

to r where r. is the internuclear separation. The dispersion forces arise from a

second order effect and are proportional to r~ . Both the dispersion and resonance

forces are especially large for the resonance lines of alkali atoms as can be seen

by examining the main features in the expressions for the interaction energies. _3

In addition to the r dependence, the principal features of the resonance

forces are that: (1) their strength is proportional to a dimensionless quantity, f,

which is known as the oscillator strength of the atomic transition, and (2) their

sum over all magnetic quantum numbers of the two states involved in the transition,

vanishes in the mean. (1) explains why the cesium resonance lines are more apt to o

be broadened than any other absorption lines such as, for example, the mercury 2537 A

line, since the oscillator strength for the combined cesium resonance lines is about o

0.98 while the oscillator strength for the mercury 2537 A line is about 0.03. It

also explains why the broadening of the higher members of the cesium absorption

lines, due to resonance, decreases rapidly in importance since the f number for

these higher members is very small with respect to unity. (2) merely means that a

resonance interaction may cause the atomic absorption line to broaden symmetrically

since an increase in the transition energy is just as likely to result as a decrease

in transition energy.

The dispersion forces are abnormally large for the alkali atoms since the

expression for the second order perturbation energy involves a sum of terms, each

of which has a denominator which depends on the difference in energy between the

ground level and the excited atomic levels. These differences are relatively small

for the alkali atoms and accordingly a large number of terms contribute significant

amounts to the dispersion energy. The dispersion forces for sodium and potassium

8

are important even at relatively large interatomic distances. The dispersion forces

for cesium can be expected to be even stronger than for sodium and potassium since

the separations between the pertinent atomic energy levels are smaller for the cesium

atom than for the latter two.

In Figure 3, line a , the energy difference Eo'^i s'fsrs to the sharply quantized

transition that occurs when the cesium atom is not influenced by a neighboring atom.

The lines connecting the energy level curves may be various lengths depending on the

distance of approach of a pertubing atom at the instant of the electronic transition.

Line b_represents the modification in transition energy that results at distances

where the resonance forces come into play; line c_ represents the transition energy

when resonance and dispersion forces are of the same order of magnitude. Transitions

may also occur from any point on the repulsive energy curve (dashed line). For

example: line d^which refers to a transition from the repulsive state to the por­

tion of the upper curve which corresponds to a region in which attractive dispersion

and resonance forces are important. The net effect of these various attractive and

repulsive forces on the lower and upper states of the atomic transition is a severe o o

broadening of the cesium absorption lines at 8521 A and 8943 A.

Line £ represents a transition from the stable ground state of the cesium

molecule to a stable upper state. It should be noted here that the upper potential

energy curve in Figure 3 has been greatly simplified. There are a total of eight

possible curves arising from a collision between an excited cesium atom and an atom o

in the ground state. Several molecular transitions which correspond to the 8521 A o

and 8943 A atomic transitions may result. Herzberg [5] has assigned the strong o o

7600 A and 9000 A molecular absorption bands to two of these transitions.

The molecular transitions can take on an important significance since they

provide an additional means to introducing energy into the vapor at higher pressures.

In the current studies it was observed that for pressures greater than about

1 mm Hg the contour of the atomic resonance absorption lines become assymetrically

broadened as the perturbations due to collisions become important. As the pressure

increases above a few mm Hg the atomic resonance lines become overlapped by diffuse

band-like absorption features which are undobutedly due to the CS2 molecule. The

net effect of the line broadening and molecular absorption is to cause absorption o o

in the 8521 A and 8943 A region to take place over a fairly broad wavelength region.

This is readily evident in the reproductions shown in Figure 4.

In addition to the diffuse absorption in the vicinity of the atomic resonance

lines, Cs« gives rise to many other absorption features. The most important of t o

these are a band system in the vicinity of 7600 A (Figure 5) and a number of 9

100% 1 INCH PATH t rsOTM « . , .wVtffti) .W

i'«fi»

^ /

J

Jl . . . , : : / , .

LAI .4^ .Vi,

1 7 6 . 7 A

;/j

f- • :^-—- — — r - - i t ;•» /./INCH

FIG. 4. ATOMIC RESONANCE AliSOKPTION AT 6321 t AS A FUNCTION OF PRESSURE

KX>%

PWEiSURE 4

tV • IT

r.

; ; 1 1 -

i

'• i , 1 ' •< I

^—7—K IMPURITY / - 0 1 %

i _

/

PRESSURt I 0 TORR

/ y^wA- i

f ' »

.ii

!|

k ]<..

ij;-

-RO iMPuRl" i ~ 0 I'l.

— / . .

- ^

^b' -* ?• & /c ' . ' - .'' A

FIG. 5. ch.oii:K *iou;cL'u.-u< .-.r.soKrno,^ AS A FI^ICCTIO;-. O,- i>';oS'.'ji;

10

diffuse absorption features on the long wavelength side of the 8943 A line. The o

band system at 7600 A shows up clearly in the spectra of cesium vapor at pressures

greater than 0.5 mm Hg. Several weak absorption bands have also been detected at o

other wavelengths but these are very weak in comparison with the 7600 A bands and

will not be considered at present. b. Experimental Results in Absorption of Resonance Radiation

A schematic of the experimental setup with which all the absorption measurement

were made is shown in Figure 6. Preliminary measurements were taken using a Jarrel-

Ash grating spectrograph to determine the effect of lamp filament temperature on

resonance absorption bandwidth. It was found that there was no measurable effect

on bandwidth over a filament temperature range of 1800-2250°K at constant cesium

pressure. This indicates that the observed absorption broadening is completely a

pressure effect over the pressure range .06 to 8 mm Hg and is not dependent on

radiation intensity within this filament temperature range.

Figures 4 and 5 are densitometer recordings taken of spectra obtained with

constant lamp filament temperature and cell environment temperature with the cesium

pressure varied from .06 mm Hg to 4 mm Hg. Figure 4 shows the broadening of the o o

8521 A absorption band. The 8943 A bandwidth was only slightly less than that at o o

8521 A. The half height bandwidth of the 8521 A band at the highest pressure in o

Figure 4 is approximately 175 A. Figure 5 shows the absorption growth of the

molecular band over the same pressure range. The data presented in these figures

give an excellent indication of the extent of increase of the absorption coefficient

with vapor pressure.

It is interesting to note the rapid increase of molecular absorption at the

higher pressures. It has been shown that the ratio of the neutral cesium molecule -3 -2

density to atomic density is between 10 to 10 . Therefore, if the absorption

cross section of the molecule is of the same order of magnitude as the atomic

absorption cross section, the sudden increase in molecular absorption could be

explained by a mechanism involving the production of cesium molecules in the o o

vapor. The absorption at the resonance lines at 8521 A and 8943 A is producing

excited cesium atoms which can combine to form the molecular ion. Cs* + Cs* + KE -> Cs2 + e (3)

One of the possible complex ion loss mechanisms in the plasma will be recombina­

tion of the molecular ion with thermal electrons. The product of this recombination

11

MONOCHROMATOR LENS

LIGHT^

PATH

PM . TUBE

POWER SUPPLY

THERMIONIC DIODE

HIGH POWER PULSE

GENERATOR X

^L. BOXCAR

INTEGRATOR

SIGNAL

GENERATOR

RECORDER

FIG. 6. SCHEMATIC DRAWING OF THE SPECTROSCOPIC APPARATUS

12

may be a neutral molecule. This process could then have the effect of increasing

the molecular to atomic ratio in the vapor. Absorption in the molecular band

can also cause excitation of the molecule resulting in dissociation in to an

excided atom and a ground state atom. It is further possible that the excited

cesium molecule produced by radiation absorption can produce ionic forms through

complex collision mechanisms with other excided species existing the the plasma.

Therefore, the molecular absorption may provide a further means of excitation and

ultimately ionization in cesium vapor plasmas.

The rubidium impurity which is observed in Figure 5 represents an impurity of

0.1 percent and, therefore, through the cesium pressure range indicated on the -5

figure, the rubidium constitutes a partial pressure range of 6 x 10 mm Hg to _3

4 X 10 mm Hg. It can be shown that for the pressure range of this work, there wi

be no overlap of the cesium molecular absorption bands with possible RbCs molecular

absorption bands.

3. IONIZATION PROCESSES IN A CESIUM DISCHARGE IN THE PRESENCE OF RESONANCE RADIATION

a. Theory

In a converter the resonance radiation is provided by the emitter where for o

gray body temperatures greater than approximately 1400°K, 8500 A is well within

the spectral range of the energy distribution. Figure 7 shows the total and

spectral emissivity of several refractory materials which are commonly used as

emitters. Notice that over the temperature range indicated, there can be almost

an increase of two in the total emissivity. Indicating that as the temperature

increases the radiating power will increase much more rapidly than its normal tem­

perature dependence would allow. In order to calculate the percentage of atoms in

the vapor which are excited to the resonance level by the incident radiation, the

power available to be absorbed must first be determined.

The energy density (U ) radiated by a black body emitter which is the energy

available for absorption is: 3 r . 1-1

(4) II _ 8Trhv

V ^3

U, = Sl§^e,

e x p f i ^ - l

<P (- ^ ) (5)

13

7 ?ft7T6e

.30

.28

>-1 -

>

(O CO

:E LJ

_J <

o K

.26

.24

.22

.20

.18

.16

.14

10 1400

TUNGSTEN

MOLYBDENUM

1600 1800 2000 2200

EMITTER TEMPERATURE,(°K)

.60

.55

50

45

40

35

30

.25

4 . CO ^taO**

> H-

> CO 00

2 UJ

_ ) < QC K CJ Ld CL CO

<< \x;

.20

.15

.10

FIG. 7. EMISSIVITY OF SOME REFRACTORIES

14

where

h = Planck's constant V = radiation frequency

c = speed of light k = Boltzmann's constant T = emitter temperature in degrees K

The power radiated (P ) is. V

^ = 2r c U dv (6)

V

Since the resonance absorption bandwidth, 6v = V2 - v, used in the integral evalua­tion is for complete absorption, the power radiated in this bandwidth is the power absorbed in this bandwidth, therefore.

p _ 2Trh£

abs ^2 ^^ v\ -^ dv (7) ^1

where e = spectral emissivity of emitter near the resonance wav elength. To,evaluate equation (7), Simpson's rule was used since 6v is small compared to v = «

Therefore, the equation to determine power absorbed as a function of emitter tempera­ture is:

2iThv „ ~hv.

the results are given in Figure 8. The percentage excitation is obtained by consider­ing an energy balance mechanism within the plasma. It will be assumed that the power absorbed in the plasma is equal to the power reradiated by the plasma.

P = P ^abs ^rad

p - ^ ^rad T

where W . = the energy radiated into the plasma and T = the radiation capture time or in other words the time in which the radiation is held in the plasma by successive absorption"and decay before it reaches the side walls of the container. If we

o

consider only resonance radiation from one level, say 8521 A, then W , is equal

15

TEMITTER ^ T POWER POWER AVAILABLE PERCENTAGE RADIATED FOR ABSORPTION EXCITATION

(.Olyu BANDWIDTH)

1400 °K .18 3.93 watts/cm^ .00196 v^atts/cm^ 2.19 7o

1600° K .21 7 8 2 watts/cm^ .00870 watts/cm^ 9 . 8 %

1800 °K .24 14.3 wat ts /cm^ .028wat ts /cm^ 31.2%

2 0 C 0 ' K .26 23 58 watts/cm^ .0727 watts/cm^ 81.4%,

2 2 0 0 ° K .28 37.37 wat ts /cm^ .156 watts/cm^ lOOVo"^

^ POWER AVAILABLE FOR EXCITATION WAS GREATER THAN AMOUNT NECESSARY FOR 100 % EXCITATION OF AVAILABLE ATOMS IF THIS WERE POSSIBLE.

Fig. 8. CALCULATED PERCENTAGE EXCITATION AT THE 8521 A RESONANCE LEVEL IN CESIUM VAPOR

to the number of atoms in the excited state multiplied by the excitation energy.

Therefore,

P u-T 2 N = -r atoms excited in a volume of 1 cm xd gap (9)

ex _3

For the results shown in Figures 8 and 9 a value of 10 second for the capture

decay time and 1 mm Hg cesium vapor pressure were used.

It was observed that the maintenance voltage of low voltage, hot cathode

arcs in the alkali vapors decreased as filament temperature increased. This

appears to be in excellent correlation with the increase in excited state popula­

tion shown in Figure 9.

When the possibility of increasing the electrical conductivity of a gas by

means of photoionization is considered, it is generally concluded that the rate

at which positive ions are formed through the process of photon absorption by an

atomic gas in an inter-electrode space is bound to be insignificant when compared

to the rates of other ionization processes, such as electron-atom impact. Only

if a layer is used of wavelength resonant to an atomic transition crossing the

ionization "bottleneck" can significant ionization levels be produced. We can

distinguish three processes by which it is possible to create cesium ions:

(a) direct single-step collisional ionization of cesium ions

(b) two-step processes leading to the formation of cesium molecular ions

(c) multi-step collisional ionization forming atomic ions.

We can represent collisional ionization of the ground atate of the cesium

atom by

CS(6S., ,2) + e + 3.893 eV Cs" + e + e (10)

while the two-step col l is ional ionization leading to molecular ions follows:

1.386 eV CS(6P I / 5 ) Cs(6S,/^) + e + - '^'^ + e (11)

'^ 1.455 eV Cs(6P3/2)

Cs(6P) + Cs(6P) + Q -V Cs2 + e

where Q is between 0.05 eV i f the two col l id ing excited states are both Cs(6P.] ,2)

and -0.09 eV i f the two states are Cs(6P2/2)- The concentration of CS2 has been

demonstrated by Wilkins [6] to be important only for plasma desnities N < 10 cm

17

0 0

90

80

70

60

50

4 0

30

20

10

n

—

—

—

—

—

—

—

—

^

1 1 1 ' 1 1

CONSTANT CESIUM PRESSURE / OF 1 TORR. . /

/

/

/

/

/

/

/

/

^ ^

1 1 1

1 1 1

]

— j

— J

—

\

-\

— j

— ]

—

1 1400 1600 1800 . 2 0 0 0 2 2 0 0

EMITTER TEMPERATURE,(°K)

FIG. 9 . EXCITED POPULATION AT 8521 A RESONANCE LEVEL IN CESIUM VAPOR DUE TO RADIATION FROM TUNGSTEN

18

Moreover, the formation of atomic ions via one-step electron collisions appears

to be statistically unfavorable if the electron gas near the emitter is Maxwellian;

very few emitted electrons (i.e. the high energy tail) have enough energy to

cause ionization.

Each step in the formation of atomic ions by multiple collisions can be

presented by

Cs(state k') + M - Cs(state k") + m (12)

where M may be an electron, a ground state or excited atom or an ion.

Multi-step processes are believed to be controlled by electron neutral colli­

sions [7] but they are of course much more complex and difficult to unscramble.

We observe in Figure 10, energy level diagram for cesium, that the first few excited

levels of cesium are separated by several times the mean electron energy in typical

plasma conditions of our concern whereas the higher excited states are much more

closely spaced and would be expected to have populated densities in Saha equilibrium

with the free electrons. This implies that there exists an energy barrier, often

conveniently termed the "bottleneck," that must be crossed by an electron which

originating from the lower energy states is heading towards the continuum. There­

fore, ionization and recombination depend upon the rate at which transitions occur

between the higher levels, in equilibrium with the free electrons, and the lower

levels, which tend to equilibrate with the ground state. Both radiative and

collisional transitions must be considered. Several authors have made calculations

of this type (Eq. (8)-(14)).

We can surmise from collisional radiative theory that above some particular

level S the number densities of the excited states are in equilibrium with the free

electron density at the electron temperature T according to the Saha equation

Nk(Ng.Tg) = 2.06 x 10-"'\N^Tg-2/2 exp(Ej-E,^)/kTg, k > s (13)

where N. is the density of the kth state, N is the plasma electron density, g.

is the state degeneracy, and Ej, Ej, are the ionization potential and the energy

of the kth level respectively.

The density distribution of the lower states, which are in equilibrium with

the ground state determined by electron neutral collisions, would be given by the

Boltzmann equation

19

ENERGY LEVEL DIAGRAM OF CESIUM

Cs Z=55 GROUND S T A T E : IS^ 25^ 2p® Ss^Sp'^Sd'" ^s^ Ap^ 4d'° Ss^ 5p^ 6s • i > 0 , l / 2

ev 3.e93>-

2.48 —

2.23 —

1.98 —

1.74 —

1.49 —

1.2 4 —

*t/2.0 ^3/2 1/2 ' D V 2 ' D

- I

3^^ 7 / 2 9 / 2 cm

- ( 3 I 4 0 G . 7 I

I3p I3p iZd I2a 4f Q»

i2s : r ± ^ i 2 p • -p^ i2pnTr^ iod--nry ' iod-^^- . f t t ~ - .

MAJOR LINES OF INTEREST

.8f —30000

— 28000

— 26000

— 2 4 0 0 0

— 22000

— 2 0 0 0 0

— 18000

— 16000

— 14000

— 12000

—10000

— 8 0 0 0

— 6 0 0 0

— 4 0 0 0

— 2 0 0 0

— 0

FIG. 10. CESIUM 1

20

N^(NQ,Tg) = NQ(g^/gQ)exp(-E,^/kTg) (14)

where NQ is the ground state number density. Suppose intense line radiation corresponding to some particular excitation

k' ^ k" from a level k' < s to some level k" > s is utilized to pump the plasma, where k' is in equilibrium with the ground state while state k" is in equilibrium with the continuum. We expect the free electron density as well as the recombina­tion rate to exceed by a large amount the value sustained by collisional processes.

The free electron density can be related to the intensity of the incoming radiation according to a simple model by neglecting radiative decay and assuming that the net excitation rate must equal the net recombination rate, which is pro­portional to the frequency of electron collisions with atoms in the lowest state in equilibrium with the electron continuum.

Optical dN Excitation = -rr^ ^^N^N^v^a^ Rate ^ ^ ' ^ '

-Ne5jg-''exp(Ej-E3)/KTg 05)

here a is an average collision cross section for all transitions from the level s to lower states.

The transition k' •> k" effective in optically pumping the cesium ion density must satisfy three requirements: The oscillator strength must be large enough to provide sufficient absorption in the plasma. The excitation line must be broad enough to capture a significant portion of ground state or one of the lower excited states and the final state k" must be within a few kT of the continuum; this assures that the electron will be lifted across the "bottleneck" and will be close enough to the continuum to eventually become free.

The relationship between atomic, molecular, and ionic term levels can best be illustrated by means of a Franck-Condon diagram as in Figure 11. For an excited CSp molecule formed from a collision of two atoms in 2P states there exists a large density of states, consisting of a superposition of electronic and vibrational states, These states are characterized by a large variety of angular momentum and quantum mechanical symmetry properties, so that the requirements of energy coincidence and selection rules can be assumed to be satisfied for a significant number of these states.

In Figure 11 the total energy of a system composed of two cesium atoms in various states of excitation is plotted against the interatomic separation, r..

21

\ Cs (62s , / 2 )+C; ( 'Sp ) \

\ \ B

\

Cs 2

\ \

\ \

\ \

Cs(6^P|/2.3/2> + Cs(62p,/2,3/2)

Cs(6^S,/2) + Cs(6^S,/2)

r, INTERATOMIC SEPARATION

FIG. 11. POTENTIAL ENERGY CURVES OF THE CESIUM MOLECULE AND MOLECULAR ION

22

The zero of energy is taken as the energy of two ground state atoms at rest with

respect to one another and separated by an infinite distance. The solid and dashed

lines represent the potential energy of the systems as the two atoms approach one

another. The minima in the curves correspond to the fact that there is a net

attraction for two cesium atoms at distances in the vicinity of the equilibrium

bond distance of the diatomic molecule Cs^. The fact that repulsive states may

also result when two atoms approach one another is indicated by the dashed lines.

The energy level diagram is simplified in the sense that only two of the **

many possible resulting states of the excited molecule, CS2 , are included. If

it is assumed that one or the other of these two states is of the proper symmetry

to be perturbed by the ground state of the molecular ion, then in order for ionization

to occur it is only necessary to require that the potential energy curves of the

two interacting systems come into close proximity or cross one another as the pairs

of curves A and B, and B and C have been shown to do.

It should be emphasized that the details of the potential energy curves are

not known and Figure 11 is intended to indicate the general features of the rela­

tionships that must exist between terms of the molecule and of the molecular ion

in order that ionization (and by inference also two-step ionization) may take place.

(b) Experimental Method

The experimental apparatus for ionization measurements is very similar to that

for the absorption measurements shown in Figure 6 except there is no need for the

monohromator. Identical tantalum emitter and collector electrodes are used.

In operation the spacing between the emitter and collector will be about .020"

so that saturation thermionic emission can be drawn from the emitter at very low

values of voltage. The graph of Figure 12 indicates the magnitude of applied

voltage which is required to draw saturation emission for various mesh spacings.

Since we seek saturation emission the applied voltage necessary to eliminate space

charge in the cell should be between 0.2 and 1.0 volts. Once saturation current

is being drawn, the potential distribution between the electrodes should be

approximately linear and the external applied voltage should then be the voltage

appearing between the vacuum levels of the emitter and collector. This approach

should allow a close determination of E/p in volts per centimeter-mm Hg in the

region between the electrodes. The measurement of the growth of current in the

cell as a function of applied voltage then will lead to the determination of the

first Townsend ionization coefficient in ion pairs per centimeter-mm Hg as a func­

tion of E/p. The measurements will be made over a range of cesium pressures both

23

Arb

itra

ry

Un

its

en

ro

ro

.^

o

CO

1—t

-V

o

> o

o

m

m

o

m

jc

t-

3s

rr

i :x

3 O

CD

-

Hm

X

) o

r—

a

HH

mS

1—

1

CO

-H

•o

m

z cz

J= m

00

z:

J*

> IS "

O

m

r--I

3>

m

^ 73

m

73

rv3 b

r - —

—

^->

n o

o

31

re

< !

I ,

^>

"U

^"

1 !

1 1

t !

1

with and without injected resonance radiation in order to determine the role

which resonance radiation plays in the ionization of cesium vapor.

The electrodes serve as identical work function emitter and collector.

This latter property is due to the fact that the tantalum meshes are operated

at the same temperature and cesium arrival rate. The chamber body is constructed

from Kovar and the saphire windows are brazed to Kovar flanges as described in

Section 4.

c. Experimental Results

Typical values of the absorption coefficients that have been obtained for o

the broadened 8521 A cesium absorption line are included in Table 1 and Figure 13.

The absorption properties of a cesium vapor at a temperature, T, can be determined

by the equation:

l^ = I^ e"^(T)L (16)

where L is the length of the absorbing vapor. For example, at 318°C the value -1 °

of k is given as 1.25 cm for the wavelength 8579 A. Therefore, the attenuation A

of a beam of radiation at this wavelength passing through a cell length of 8mm

amounts to a factor of about 1/e, since

1 = I e-1-256 (.8) , 1 3-1 0 0

The absorption coefficients were obtained from photographic film which was

calibrated with the aid of a logarithmic sector disc, a 5000 RPM motor, a uniform

temperature tungsten ribbon lamp and a microdensitometer. These values of the

absorption coefficients are subject to verification by further measurements and

the studies are being extended to higher temperatures and shorter path length

absorption cells.

4. CESIUM PLASMA SUSTENANCE BY APPLICATION OF MICROWAVE P S W E ^

a. Experimental Measurements

In our effort to develop the second generation thermionic energy converter,

it becomes necessary to separate the emitter and collector electrode functions

from the ion generation task. The emitter at 1600°K can no longer supply sufficient

ionization generation levels. A very appealing concept in the generation and

25

in

CM

C»

^ o

CO

vo

cr>

^

00

vo

in

cvr

<

cc <

: cc

CT> C

O

cn r—

en

00

o o C

M

CT> in

CO

cn in

CO

CO

in

CO

I—

CO

C

O

O

o O

O

O

I—

o

oo

oo

oo

oo

^

CM

CO

in

CO

in

•!*•

—

lO

00

•—

ID

cri

in

CO

CT>

CO

r-

in

00

•—

r—

VD

in

CO

^

in

CM

in

CO

CO

CT>

00

<3-00

o

CM

r

^

00

CM

«3-in

^

CO

CO

ȣ>

CO

^

CO

in

CO

r—

^

CO

CTt I—

CM

lO

O

C

Ti CO

oc

Mi

nr

^i

—

cn

cT

ii—

oc

ov

o'

sa

-i

—

lo

oo

cn

«3

-c

oc

oc

oc

or

-o

cr

>

co

co

co

co

co

co

oo

r^

o

in

in

o

o C

D

•—

r>. r«.

in

CM CM CO

CM

r^ r— CM

r—

CO

O

"S- in cn

o

CO o

•a: <: in

CO

cn r

1^

1^

r«.

CO

ID

cn

CM

in

CO

r—

in

•si-CO

^

I—

«n CO CO

in o

in in

CO CO CM I— o

o

o

o

o

o

o

o

o

o

o

CO in

CO

^

CO O

CM

in

en CO 00 CM I— I—

I^

CM V£) CO o

CO in

o r>.

CO 1— o

CM vn

r^

ȣ> v

o vn lo in in

00 00

00 C

O

CO

C

O

00

cn

co

io

oo

oc

Mi

^m

i—

r

s>

cs

jco

<d

-C

Mc

nr

^L

n«

d-

cM

oc

oi

nc

n»

£)

r^

co

in

«a

-^

"«

-^

«a

-^

co

co

cM

CM

i—

o

co

co

co

oo

oo

co

co

oo

co

oo

co

co

co

OQ

•a:

in

wD

CM

cn

co

cn

o

oc

oo

co

V

DC

OI

-^

r—

C

Ml

O<

a:

e3

:0

"!

d-

^l

O

.—

CM

CM

^r

v.

cn

l—

(—

o

iO

'd

-c

o

in CO

m

^

cn O

lO

00

CM

CM

I—

O

O

o o

o o

o •—

I—

o

oo

oo

oo

ID

in

Cn

i—

C

M«

;:

i-

VD

00

Or

—

CO

in

CM

i—

cn

00

cn

a\ ID

00

o

r-~

ID

CO

CM

cn

in

CO

^

t^ cn I— CO

i~>. in CO CM

o

in in in in in

CO CO CO 00 CO

CO CO O

CM ^

CO

in ID in CO I—

•*

«^ ^- «d-

^

<^ CO

00 00 00 00 00 00

"sa- m

in vD

in ID

r>.

00

CM r— o cri

CO

CO 00 r^

:x

r

ID

II

u

1—

r— 1 E o f<

J)£ r<

CO

en

CM

o

ID

CO

in

<£> CO

VD

ID

O — •

o

o

CO

»D

CO

o

•*

CO

r"-.

o

•—

o o

<D

00

CM

00

CO

r— •

o

CO

CM

00

in

CO

o

«*

«a-CO •

o

in

«3-ID

in

00

in

in

t^ .

o

r

VD

«d-in

CO

•=r

1—

cn

00

CM

in

CO

<

1—

o ,

r—

in

00

o

cn

ID

CD

CM

CO

CTi

«;

CO

r-*~ 00

CO •

o

<3-

in

r»

^

CO

o

ID

CM .

O

ID

r

in

s

ID

O

CM •

O

CO

cn

CO

^

CO

^

00

^mm

•

O

en

,_

CM

«*

00

r^

CO

H— •

o

CM

1—

<a-o

•;rt-00

r-*

CO — •

o

CO

<

CO

CO

00

r

r

o

O

in

CM

r>.

f-m.

00

o

CO

«*

cn

cn

r>»

26

3.5

3.0

2.5

2 0

'E u

^ I 5

1.0

.5

o A 0 +

CESIUM TEMP f t M

•• •« M M

194'C 238 "C 277»C 318-C

7800 8000 8200 8400 8600 8800

X . ( A )

FIG. 13 . ABSORPTION COEFFICIENT AS A FUNCTION OF WAVELENGTH

27

sustenance of the thermionic plasma is the external application of microwave

power whose sole function is to support the plasma. It can thus be specifically

tailored to perform its assigned task in the most efficient manner. We expect

that highest efficiencies result when microwave power is fed into a resonant

system part of which is the thermionic plasma.

In order to examine this concept experimentally, we have built two microwave

coupling arrangements, a cavity and a waveguide insert shown in Figures 14 and 15

where the emitter and collector of a simple thermionic converter are part of two

of the flat walls of the cavity. The cavity is of the cylindrical reentrant type

with resonance microwave frequency about 1.5 GH . The emitter and collector

electrodes of the thermionic converter are housed in a pyrex glass envelope which

enters into the hollow inner cylindrical surface of the cavity. This inner surface

is completely enclosed except for a small gap as shown in the diagram. The size

of the gap can be adjusted continuously by sliding the inner metallic tube in

and out of the outer shell. This gap is the pathway through which the microwave

energy can be directed into the thermionic converter interelectrode spacing. The

length of the cylindrical cavity can also be varied continuously resulting in

changes of the resonant frequency of the microwave cavity. It is thus possible

to mechanically tune the cavity continuously over a wide range of frequencies and

so achieve a corresponding range of plasma densities. The plasma densities that

we can attain vary roughly at the square of the frequency of the applied microwave

power.

The coupling of power into the microwave cavity is accomplished by a magnetic

loop inserted from the top. In actual operation, we ignite the discharge by apply­

ing a DC voltage between the emitter and the collector. Subsequently, we provide

microwave power into the cavity. The resonant frequency of the cavity is strongly

affected by the presence of plasma. It is actually the system of plasma and

cavity which is resonant. Absorption of microv/ave power is thus maximum at the

particular resonant frequency of the plasma-cavity system and we can adjust the

microwave frequency to achieve maximum absorption. Coupling adjustments are needed

as well in order to match the loaded cavity impedance to the transmission line

impedance.

The microwave hardware required to transmit the power from the generator to

the cavity and to make the power measurements of input and reflected power are

shown in Figure 16. A sweep generator in the 1-2 GH feeds a 20 watt TWT amplifier

which provides power into the microwave cavity. Directional couplers and a slotted

line are utilized to measure incident and reflected power. Several parameters

28

r e

a l

r Oven

C a v i t y LciiRth C o n t r o l —>^^

n

' ^ Cavi ty

^ C c s l

yir ... — h , — : - ! Bakablf

um Reservoir (Vertical)

Valve

•< Microwave

Oven

ITX. Y "3 Flexible

Diffusion Pump

Emi I Feeil-Tluouij;h

•> T u b i n g

Va 1 vc V a l v e

Ion Pump CauLrcl-MilLcr

I o n Pump

F I G . 1 4 . VACUUM SYSTEM SET UP

n B

V c

E

L

V

B

TO PUMPING LINE

/ \

^ ^ ^ = ^

W

TO PUMPING LINE

FIG. 15.

30

1-2 GHz AND 2 -4 GHz

OSCILLATOR

^f

VARIABLE

ATTENUATOR

V

DIRECTIONAL COUPLER

1'

TWT AMPLIFIER

POWER METER

CIRCULATOR

V

MATCHED LOAD

POWER METER

iL

ATTEN­UATOR

1

J

i

POWER METER

a

ATTEN­UATOR

i i

DIRECTIONAL COUPLER

SWR INDICATOR

a

SLOTTED LINE

M

SPECTRUM ANALYSER

- * TUNER MICROWAVE CAVITY AND PLASMA LOAD

FIG. 1 6 .

are measured which include resonant frequency, frequency shifts due to the

presence of the plasma and the loaded and unloaded Q of the cavity.

We have measured the I-V characteristics of the thermionic converter with

and without the application of microwave power under various conditions.

We expected that the application of microwave power would raise the plasma

density and temperature and thus significantly alter the characteristic I-V curves.

Numerous curves were obtained but it was found that the important features of the

results could be demonstrated by a detailed analysis of the I-V traces obtained

at microwave power levels of 0, 1 and 2 watts.

If we study the I-V curves shown in Figure 17, we observe a most dramatic

effect displayed by the ignition traces (i.e. increasing voltage) of the thermionic

converter. When no microwave power is applied we get, as expected, the usual I-V

curves characterized by a breakdown voltage as signified by the knee of the curve.

However, upon application of microwave power, the knee of the I-V curve disappears

completely and no breakdown voltage is indicated anymore. Indeed, we observe

current conduction at zero voltage and even at negative applied voltage indicating

that there is conversion of microwave power in DC electrical output power.

The return traces of the I-V curves (decreasing applied voltage) shown in

Figure 17 also show important differences revealed in the amount of current that

can be carried by the plasma under the same applied voltage; considerably more

current can be carried when microwave power is applied and this current enhance­

ment is greatest at the low current levels. The high current ranges are affected

very little and,in fact, the value of the saturation current is not changed at all.

Moreover, it was also found that the knee of the ignition I-V curves indicat­

ing the presence of a breakdown voltage was not affected at all at applied microwave

power levels below 0.5 watts, then it continuously shifted to lower voltages as

the power increased from about 0.5 to 1 watt, and completely disappeared when the

level of 1 watt was reached. In order to appreciate the effects of varying

the applied microwave power level we show in Figure 18, a family of ignition and

return curves at cesium pressure P = 0.01 Torr. At higher microwave power levels

only small changes occur mostly in the lower current region. In these lower

ranges the current carried at the same voltage rises as a function of the applied

microwave power; in particular at zero voltage the transmitted current increases

indicating that more microwave power is converted into DC power.

b. Ttieoretical Considerations

A qualitative understanding of the above results can be gained by considering

the potential energy diagrams for the converter in the presence or in the absence

32

(Amperes)

watt 001 torr

1 2 3 4 5 6 7

(Amperes)

1 watt 0.03 torr

V(V3lts) VCVblts)

(Amperes)

1 watt 001 torr

CAmperes)

0 1 2 3 4 5 6 7 V(VDlts)

watt 0.03 torr

VCVolts)

FIG. 17.

33

V(Volts)

(Amperes)

-3 V (Volts)

FIG. 18.

34

of applied microwave power. In Figure 18 we show an ignition curve in the absence of microwave power; three main regions can be distinguished characterized by different potential energy diagrams, the unignited mode, the obstructed region, and the saturation region. In the unignited mode no current flows, while in the obstructed region, the double sheath erected in front of the emitter dominates the behavior of the converter. The voltage must thus be raised to the breakdown value in order to overcome this motive peak. The souble sheath disappears in the saturagion region. When microwave power of a sufficiently high level is applied, no such breakdown voltage threshold is observed which immediately indicates that no motive peak exists; the structure of the potential energy diagram under these conditions is very simple, as seen in Figure 19. An interesting method of evaluat­ing the performance of the thermionic converter is provided by Lam [1] who summarizes the plasma arc drop in terms on a single parameter, the normalized plasma resist­and R. According to this theory, the converter arc drop V , is related to the plasma resistance R. In the case of no motive peak

3 - ^v-riR 07) 1 + exp -^

normalized plasma resistance R is defined by the expression

^Xohmic = Jf ( 2)

where ^XQV,^.;- is the normalized (again in units of kTr/e) ohmic (momentum scattering) contribution to the total plasma arc drop. The plasma electron temperature T is also included in dimensionless form T given by

T = Tg/T^ (19)

In the obstructed region, we have

J^ = 0^ exp (-Ax) (20)

where Ax = Ax/(kTr/e). If the emitter motive peak is suppressed, i.e. if Ax = 0, then Jc = Jp. Without external heating, R cannot be determined experimentally because Jr and T both change simultaneously below the transition point of the I-V curve, where the emitter motive exists. This can be explained by Figure 20 which

35

FIG. 19.

UNIGNITED MODE

V

FIG. 20.

shows the potential diagram between the electrodes at four points on the I-V

curve. At point 1, the plasma is unignited and the current density is very low.

At point 2, the plasma is in the negative resistance region where the plasma is

not completely ignited and the emitter motive peak still exists. At either point

1 or 2, the emitter net current density Jr cannot be determined because Jr is a

function of T and Ax which are both varying during the transition from point 1 to

point 3. It is only when point 3 is reached which is the transition point or the

"knww" of the I-V curve, that the plasma is completely ignited, the emitter motive

peak disappears, and Jr = Jp becomes a constant. At point 4, the plasma is in

the saturation region. In Figure 19 we see the I-V curve and the potential

diagrams of the converter diode with microwave heating. The "knee" of the I-V

curve disappears and the plasma is sustained at high current density when both

AC and microwave power are applied while at low current density the plasma is

sustained only by microwave power. Because of the disappearance of the emitter

motive peak, there is no sudden jump of the diode current and JR = Jc = constant

can be assumed. As a result, Eq. 17 can be used to relate j = (J/Jp) and V . at

a certain value of R and x. The normalized I-V characteristics of the diode

with external heating are plotted as j versus V .. The best fit of these experimental

I-V curves with respect ta the parameters R and into the curves provided by Lam's

theory can provide us with the best values of R and T. Some of the data thus

analyzed are shown in Figure 21.

We have seen a significant rise in the thermionic converter current level

upon application of microwave power. In the experiments described in this work

the improvement is limited to the lower current ranges due to the shielding effects

of the plasma at the higher current values. At these high current ranges the plasma

density is high and microwave propagation is cut off if the frequency is not high

enough. It is expected through that higher frequency microwave power would suffer

no such limitations and would provide current enhancement at high current levels

which would be of great interest in practical applications.

A very attractive possibility also is the pulsed mode of thermionic converter

where the input is microwave power. The advantages over other modes of pulsed

power are simplicity, flexibility, minimum disturbances to thermionic surfaces,

and the possibility in some modes of operation to supplying power at very short

separations between emitter and collector, a distinct practical possibility.

It is thus believed that microwave power shows great promise as a source of

energy to sustain the cesium plasma in a thermionic converter. At the lower

38

J/J.

CO

-4.0

Tp=935K PC5=8x10 Torr

Lam's Theory

r=io

Vd (Volts)-1.0 2.0

-60-55-50 -45 -40-35 -30 -25 -20 -15 -10 -5 0 15 20 25 30 35 40 45 50 55 60 VjCkT^/e)

FIG. 21.

operating temperature of 1600°K the emitter in the advanced converter can no

longer supply sufficient ionization levels. An external source of ion generation

is needed which does not interfere with the emitter and collector electrodes.

Externally supplied microwave power may prove to be the best agent to perform the

task. It is attractive in many ways. There is considerable flexibility in that

we may adjust both the power and the frequency of the applied microwave power to

achieve the desired plasma condition. In supplying microwave energy we do not

interfere with the interelectrode spacing by the insertion of extraneous electrodes.

We may operate in the continuous mode in contrast to pulsed systems which would

not be available for power generation during the pulse on condition. Furthermore,

the geometrical size of the emitter-collector distance envisioned is the correct

order of magnitude to allow support of the plasma by microwave fields in a

resonant mode. The energy expenditure of microwave power at a resonant plasma

system is expected to be smaller in comparison to alternative energy sources.

Moreover, the power requirements in the applied microwave field seem rather modest--

a few watts—easily available in present day technology even for large thermionic

converters. Subsequent experiments should be performed with microwave power of

high frequency, i.e. at 10 GH (X-band) and 20 GH (K-band). The technology of

microwave power generation and transmission is well advanced with many off-the-

shelf items available to utilize in our systems.

More work is needed in this area to provide a comprehensive understanding

of the plasma in the thermionic converter. The simplicity of a microwave supported

plasma in an optimum diode could allow for easily interpreted data from which

conclusions may be drawn about the proper plasma density level for highest overall

efficiency. We thus believe that microwave power sustenance of a thermionic plasma

in a resonant configuration is indeed a rather attractive choice.

REFERENCES

[1] S. H. Lam, Report prepared for the U. S. Energy Research and Development Administration - Contract AT(11-1)-2533, March 1976.

[2] N. S. Rasor et al, Advanced Thermionic Conversion Progress Report, COO-2263-4 (NSR 2-4), ERDA Contract No. E (11-1) 2263, August 1975.

[3] C. N. Manikopoulos et al, The Advanced Thermionic Energy Converter with Microwave Power as an Auxiliary Ionization Source. To be sumitted to the J. Energy Conversion.

[4] H. S.- Chiu et al, Advanced Thermionic Covnerter Developments with Microwave External Pumping, Proc. 12th lEEC, Vol. 2, AICE, pp. 1575-1581.

[5] 6. Herzberg, Spectra of Diatomic Molecules, Van Nostrand, Appendix I, 1958.

40

[6] R.'Wilkins, Report LA-3143-MS, Los Alamos Scientific Laboratory, Los Alamos, New Mexico, 1964.

[7] E. W. McDaniel, Collision Phenomena in Ionized Gases, John Wiley & Sons, Inc., New York, p. 594 ff, 1964.

[8] V. B. Moizhes, F. G. Baksht and M. G. Mikiya, Sov. Phys.-Tech. Phys, 10, 1252, 1966.

[9] S. Byron, R. C. Stabler and P. I. Bortzt, Phys. Lett. 8, 376, 1962.

[10] J. V. Dugan, Jr., F. A. Lyman and I. V. Albers, Electricity from MHD, Vol. II, International Atomic Energy Agency, Vienna, p. 85, 1966.

[11] J. V. Dugan, Jr., Journal of Applied Physics, 37, 5011, 1966.

[12] J. M. Zgorzelski, Fluid Mechanics Laboratory Publication, No. 68-1, MIT, Cambridge, 1968.

[13] D. R. Bates, A. E. Kingston and R. D. McWhirter, Proc. Royal Soc. (London), A267, 1962.

[14] T. A. Cool and E. E. Zukoski, Recombination, Ionization and Nonequilibrium Electrical Conductivity in Seeded Plasma, Phys. Fluids, 9:780-796, April 1966.