011dC04.1LE Cr

1 downloads 0 Views 4MB Size Report
1.3 ProQ:ram Objectives. The aim of this research program ...... C2F6 in Cl14 are given in Figs. 49a and 49b, respectively. ..... 147 (1982). 15. L. G. Christophorou ...
AD-A 198 641 ORNL/TM-10844

011dC04.1LE

Cr

OAK RIDGE NATIONAL LABORATORY Basic Studies of Gases _for

Fast Switches

Final Report

S. R. Hunter L. G. Christophorou

OTIC S

,,.t,,

AUG 9 1988

ofnent has been approvedi wn

f>r public reiccsa and sale; its ORDctribution is unlimited.

OPERATED DY! MARTIN MARIETTA ENERGY SYSTEMS, INC. FOR THE UNITED STAES DEPARTMENT OF ENERGY

Printed in the United States of America. Available from National Technical Information Service U.S. Department of Commerce 5285 Port Royal Road, Springfield, Virginia 22161 NTIS price codes-Printed Copy: A06- Microfiche A01

This report was prepared as an account of work sponsored by an agency of the United States Government Neither the U nited 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

ORNL/TM- 10844

Interagency Agreement DOE No. 1246-1246-Al Navy No. NOO014-82-F-O123

Office of Naval Research Physics Division Arlington, Virginia 22217

BASIC STUDIES OF CASES FOR FAST SWITCHES

Final Report by S. R. Hunter and L. G. Christophorou Health and Safety Research Division Oak Ridge National Laboratory P. 0. Box 2008 Oak Ridge, Tennessee 37831-6123 and Department of Physics The University of Tennessee Knoxville, Tennessee 37996

Date Published - August 1988

OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37831 Operated by MARTIN MARIETTA ENERGY SYSTEMS, INC. for the U.S. Department of Energy Under Contract No. DE-AGO5-81OR21400

Contents Page

Section I. 1.1 1,2 1.3 1.4 1.5 II. 2.1 2.2 2.3 2.4 2.5 III. 3.1 3.2 3.3 3.4 3.5 3.6 3.7

3

Background ............................................. Operating Parameters for a Diffuse Discharge Opening Switch ......................................... Program Objectives .................................... Summary of Accomplishments ............................. Publications and Presentations .........................

3

14 17

Experimental Approach ..................................

23

High Pressure Swarm Experiment ......................... High Temperature Swarm Experiment ...................... Pulsed Townsend Experiment ............................. Gas Ionizing W Value Experiment ........................ High Voltage Breakdown Experiment ......................

23 26 26 29 31

Experimental Techniques and Selected Results ...........

33

Room Temperature Electron Attachment Rate Constants and Attachment Cross Sections .......................... Elevated Temperature Electron Attachment Rate Constant Measurements .................................. Room Temperature Electron Drift Velocity Measurements, Room Temperature Electron Attachment and Ionization Coefficien ts ........................................... Elevated Temperature Gas Mixture Studies ............... W-Value Measurements ................................... High Voltage Breakdown Field Strength Studies ..........

5

33 38 42 50 64 75 84

Relevance to Switching Experiments ..................... 91

IV 4.1 4.2 4.3 4.4 4.5 V.

Introduction ...........................................

C 2 F6 /Ar Gas Mixtures ................................... C 2 F6 /CH4 Gas Mixtures .................................. Enhanced Discharge Current Characteristics ............. Operation at Elevated Gas Temperatures ................. Diffuse Discharge Closing Switch Experiments ...........

Conclusions ....................... References ...................... .

.

.. ..

...

.....

......

95 95 99 101 103

.

....

108

...

....

110

3

I.

Introduction This final report covers the contract period October 1, 1982 to March 31,

1988, and includs an outline of the operation of diffuse discharge switches in pulsed power energy storage circuits, the methods by which the operation of these switches may be optimized, and the measurements we have performed on several gas mixtures which exhibit these optimized characteristics.

The goal

of the program has been to identify these optimization procedures, and using basic data on electroratom and electron~molecule collision processes, to find gas mixtures in which the transport properties of the electrons in the gas discharges are "tailored"

to match these optimization requirements.

Using

these criteria, we have found several gas mixtures which have very desirable characteristics

for

use

in

these

switches.

The

potential

of

these

gas

mixtures has been demonstrated in recent switching experiments which have shown considerable operational improvements when these gases were used. )' The

following outline for

this report has

accomplishments of this program are highlighted.. I, a brief background description of switch

in an inductive

transport

and

of

in which

the

For the remainder of Section

the operation of a diffuse discharge

energy storage circuit

other parameters

been adopted

"

the

gas

is outlined.

mixtures

to

be

The electron optimized

are

identified, and the specific objectives of this program are discussed along with a summary of

the accomplishments and publications which have resulted

from this contract.

In Section II, the experimental techniques used to obtain

the electron transport and breakdown measurements are outlined;

the basic

experimental measurements are presented in Section III, while in Section IV, we discuss some of

the implications of these data in the operation of both

diffuse discharge closing and opening switches.

The results of

switching

experiments using our new gas mixtures will also be presented along with a discussion of the strengths and limitations of these switches in pulsed power energy storage circuits.

Finally in Section V some general conclusions are

presented. 1.1

Background

There has been considerable interest in recent years in the possibility of using inductive energy storage devices as a means of storing and rapidly transferring electrical energy in repetitive pulsed power applications.

The

primary advantage to be gained from the use of these energy storage devices is

4

that they have potential energy storage densities 100 to 1000 times

that of

comparable capacitive storage systems and that this energy can be transferred 1-3 to the load on the very short time scale of nanoseconds. One of the major problems

to be faced with this technology before

it can be

introduced

in a

number of applications is that these inductive energy storage systems require a switching device that can repetitively switch (repetition rates > 104 pps and more

than

106

shots)

high currents

confinement) with opening times of withstanding breakdown. 3

high

voltages

(e.g.,

100 kA for

inertial

fusion

the order of a few ns and being capable of

(>100

kV)

during

the

opening

stage

without

One of the most promising contenders for fast repetitive switching is an externally sustained diffuse gas discharge operating at gas pressures of one to several atmospheres.

Two different types of external electron sources have

been proposed for the control of the discharge current. volume

gas

controlled"

ionization

5 '6

)

or

by

by

a

pulsed

resonance

high

ionization

They are by means of

energy

electron

of

gaseous medium

the

beam

("e-beam using

a

pulsed high power laser ("optically controlled" 7'8) The first experiments using e-beam controlled discharges were performed in the early 1970s in an effort to improve the efficiency of high pressure CO 2 discharge lasers, where the efficient operation of these devices requires that they

be

voltage. fields

operated 9

at

Early

electric

researchers

fields in

below

this

field

the

self-sustained

noticed

that at

discharge

high

electric

the

discharge becomes unstable, and oscillations in the discharge 10 current occur. These oscillations have been shown to be due to an electron attachment

instability because

the

electron attachment

coefficient

in

02

increases in magnitude with increasing field strength over a range of E/N 11 values. A small, local reduction in the electron concentration causes a small

increase

diffuse

in

the

discharge,

local

which

electric

results

field

in

in

the

space-charge

dominated

electron attachment and a 11 further reduction in the electron concentration, etc. As a consequence, the

discharge

becomes unstable

electrodes e-beam schemes

in

the

sustained for

applications.

and waves

discharge. lasers,

rapidly Early

this

increased

of

plasma current

Although

this

mechanism

extinguishing

the

experiments using

between

is an undesirable

forms

the

discharge

e-beam

travel

basis in

of

process

most

the in

present

opening

switching

sustained diffuse

discharges

5 12 were mainly conducted in Russia for use in fast closing switch applications.

Hunter 5 and Koval'chuk and Mesyats 13 were the first discharges

sustained diffuse

could also

applications where switching times 10- 1 5 V cm 2 ) while the switch is opening. These

conditions

operating

allow

us

desirable

several

define

to

characteristics of the gaseous medium in the conducting (low E/N) and opening (high E/N)

stages

of

In

switching action.

the

conducting stage,

the

the

requirements are: 1.

Maximum electron drift velocity w - the larger w is, conductivity of the discharge and the greate

-

the higher the

the current density in

the diffuse discharge and the lower the switch resistivity. 2.

Minimum e-beam "ionization energy" W - the smaller W is,

the greater

the current gain in the discharge with a consequent increase in the efficiency

of

the coupling

of

the e-beam

to

the

discharge and

a

greater control of the resultant discharge current. 3.

Minimum electron loss terms k of

the

,

discharge drastically

increase

when

the

highly

kR1, k

and k

decreases and

mobile

relatively immobile negative ions.

- the conductivity

space-charge

electrons

are

problems

converted

into

Similarly, the conductivity will

decrease if the electron and negative ion-positive ion recombination in the discharge is large due to the loss of Another problem

that results

from large

the charge carriers.

recombination coefficients

is that the current gain in the switch will decrease, and the energy released in the recombination process will

result in increased gas

kinetic energy causing heating problems in the gas under repetitive operation. 4.

Minimum ionization rate constant k. required source,

to

be

completely

otherwise

considerably

the

e-beam is switched off.

the conductivity of the gas is

controlled

opening

increased due

-

by

time

the

of

to additional

external

the gas

ionization

switch

ionization

will

be

when

the

6

In the opening stage the requirements of the gas mixture as as follows: 1.

Minimum mobility mixture.

electron and

drift

hence

velocity

lower

the

w

-

i.e.,

electron

reduce

conductivity

the in

electron the

gas

11

2.

Maximum electron attachment rate constant ka

i.e.,

-

lower the gas

conductivity by converting highly mobile electrons into relatively immobile

negative

ions

and

by

removing

free

electrons

from

the

discharge, reducing the current density due to additional ionization processes as the E/N increases. 3.

High breakdown strength (E/N)e.m (for uniform fields defined as the -15 2 E/N at

> 10- 1 5 V cm

which k i = ka)

(E/N)eim,

the

faster

electron conductivity

the

the higher

-

permissible

rate

of

the

value

decrease

in the discharge, and hence

the

of

in

the

shorter

the

opening time of the switch. 4.

Self-healing gas mixtures

-

for

closed cycle

operation without

time dependent degradation in the performance of

a

the switch, it is

required that the gas mixture composition not change with time.

The

molecular gases in the switch can be fragmented either by collisions with

high

energy

electrons

impact-induced dissociation occurring

during

the

from

the

e-beam

and dissociative

diffuse

discharge,

or

by

electron

attachment

processes

particularly

during

the

opening phase where the E/N quickly rises to very large values (E/N > 10- 1 5 V cm 2). attach

This problem can be reduced by using gases

electrons nondissociatively

at

energies well

in

that

excess

of

thermal energy and also have low electron impact dissociation cross sections

and

large

neutral-neutral

and

negative

ion-positive

ion

recombination coefficients at high E/N values. 5.

In

photoexcited

and

photoionized

laser-controlled discharges) attaching

gas

in

which

it

electron

gas

discharges

is desirable attachment

(required

to have can

an

electron

be

increased 7 8 , radiation. laser the by molecules the of photoexcitation

The desirable characteristics for the E/N dependence of w and ka

for

by

for

the

gas mixture in the diffuse discharge are shown in Fig. 2. 1 4 ,15 When switch S is closed or conducting (i.e., when the electric field across

the the

switch electrodes is small), electron conduction in the discharge is optimized by using gas mixtures which have electron

loss processes

(e.g.,

large electron drift velocities and small

electron attachment and

recombination)

at

the

12

ORNL-DWG 85-9330

CONDUCTING STAGE

L)

I

OPENING STAGE

I

>

,

>

0

/

0

0-%120

3 E/N (10-17 V-cm 2 )

Fig.

2.

Schematic k (E/N)

illustration and

the

of

electron

required of a gas mixture switch.

Approximate

the

electron

drift

of

conducting and opening stages of (from Ref. 14).

velocity

for use

values

attachment w(E/N)

rate

characteristics

in a diffuse discharge E/N

for

the

constant

discharge

opening in

the

the switch are shown in the figure

13

electric field strengths characteristic of this stage of the switching action (indicated by the shaded region in Fig. 2). stage (i.e.,

Conversely, during the opening indicated by the

when the electric field strengths are large,

shaded region in Fig. 2). the electron drift velocity should now be low, and the electron attachment rate constant large in order to reduce the electron conduction in the discharge as quickly as possible. 1.3

ProQ:ram Objectives

The aim of this research program has been to find gas mixtures which would optimize

the gas discharge requirements outlined above

conduction and

opening

stages

of

the

switching action.

in both

This

has

the been

accomplished by performing the following measurements on potential gases and gas mixtures. 1.

Measurements constant k

a

of

the

room

temperature

electron

attachment

rate

were performed on selected electronegative gases in both

N 2 and Ar buffer gases as a function of E/N (or mean electron energy

()) and gas pressure in an effort to find the energy dependence and electron attachment mechanisms of these candidate gases.

If any of

these gases possessed the required characteristics (i.e.,

low ka at

thermal and near thermal electron energies,

and high ka at (E) > 2

eV) then the following measurements were performed. 2.

Room

temperature

electron

drift

velocity

w

measurements

were

performed on the candidate gases in gas mixtures with both Ar and CH, over a wide range of concentrations (typically varying from 0.1% to 100% of the attaching gas)

to observe any enhancement

may occur at low E/N values,

and to give additional

in w that

insights into

the electron scattering processes which may be occurring in these electronegative gases. 3.

Electron

attachment

coefficient

T1/N

and

electron

ionization

coefficient a/N measurements were also performed on a few selected gas

mixtures

with

Ar

and

concentrations in an effort

CH4

over

a

similarly

wide

range

of

to understand how these gases would

perform in switching gas mixtures.

14

4.

High voltage

breakdown

field

(EIN)tim measurements were

strength

performed on selected gas mixtures in order

to determine if these

gases were able to withstand the transient high voltages that occur when

the

switch

determine

is

opened.

These

measurements

the minimum concentration of

enabled

the attaching gas

us

to

that was

required to sustain these voltage transients without breakdown. 5.

For repetitively operated switches, the gas temperature within the switch is expected to rise several hundred degrees centigrade, and under

these

circumstances,

the

electron

transport

and

rate

coefficients are expected to be functions of the gas temperature T. Knowledge of

these parameters as a function of T is desirable for

modeling the operation of the diffuse discharge switch in practical applications. w (E/N).

Consequently, the above measurements (namely k a(()),

a/N (E/N),

r1 N (E/N)

and

(E/N)tim)

were

performed

as a

function of T on selected gas mixtures to simulate the environment found in repetitively operated switches. 6.

Finally,

the

W value

positive ion pair particle) effort

of

to

current minimize

average energy

to produce an electron-

from the energy decay of a high energy ionizing

several

find

gas

density

in

the

(the

gases

and gas

mixtures the

which

switch

"resistivity"

of

mixtures would

during the

was

measured

maximize

switch

the

electron

conduction

discharge

and

in an

(i.e.,

consequently

maximize the current switching efficiency). 1.4

Summary of Accomplishments

The following is a list of the most significant accomplishments achieved during the course of this program. 1.

Measurement

of

the

constant

as

a

ka

(Z 0.04 eV)

to

room

function

(C) Z 5 eV

temperature of

(E)

electron

from

thermal

attachment electron

in N 2 and Ar buffer gases

rate energy

for several

molecules (many of which have been found to possess very desirable electron attachment opening switches). 1-CF

6

characteristics

for

use

in

diffuse

discharge

These molecules include CF4, C 2 F6, CF,. c-C3 F 6 ,

, c-C 4F8 . n-C 4 Fo, n-C 5 F1 2 , i-C 5 F1 2, n-C 6 F,4 , GCIF 3 , (CF3 )2 0,

(CF,) 2 S. NF3 , N2 0 and S02.

15

2.

Measurement of the room temperature electron drift velocity over a wide E/N range (0.002 < E/N < 400 x 10- 17 V cm 2 ) in the following CF 4. C 2 F 6 , C 3 F8,

pure gases;

n-C 4 Fo, (CF3 )2 0. SiF 4 . BF3 . Ar. CH 4

and N 2 . 3.

Measurement of the room temperature w in the following gas mixtures which were proposed for use in diffuse discharge opening and closing Ar/CH 4 , CF 4 /Ar.

switches;

CF 4 /CH 4 ,

C 2 F6 /Ar,

C 2 F6 /Cli 4,

CaF,/Ar,

C3 F9 /CH 4. (CF3 )2 0/Ar. (CF3 ) 2S/CH4, SiF 4 /He. 4.

Developed a new procedure for the measurment of electron attachment and ionization

coefficients

in a pulsed Townsend experiment where

both electron ionization and attachment processes are present. 5.

Measurement of the electron attachment coefficient rI/N as a function of E/N and gas pressure CaF.,

n-C 4 F1 0,

was measured all

these

SiF

02.

to test

molecules

in

the

following pure gases;

and BF3 .

4

With the exception of 02

the accuracy of only

attach

CF 4 , C 2 F6,

the experimental

electrons

(which

technique),

efficiently

at

high

electric field strengths. 6.

Measurement of the electron ionization coefficient a/N as a function of E/N in the following gases; CF 4 , C 2 F6 , C3 F9 , n-C 4 F,

SiF

4

, BF,

and CH 4 . 7.

Room temperature il/N and a/N measurements were also

performed

in

C 2 F 6 /Ar and C 2 F 6 /CH 4 gas mixtures, as these mixtures were chosen to represent

the expected

coefficients

in

E/N and

typical

concentration dependence

diffuse

discharge

of

opening

these switch

applications. 8.

Room temperature a/N and 7/N measurements were also performed in the following mixtures; mixtures

were

SiF 4/He. SiF 4 /He/TEA and SiF 4 /He/TMAE.

chosen

for

study

as

they

possessed

almost

These ideal

characteristics for self-sustained diffuse discharge closing switch applications. 9.

The

high

voltage

performed at room

temperature

field in

the

strength

measurements

following mixtures;

The average energy to produce an electron-positive measured for CH 4 , Ar,

were

C 2 F6 /Ar,

C3 F9/Ar, CaFa/CH 4 and c-C 4 F,/Ae.

C 2 F 6 /CH4,

10.

breakdown

N

2

the

, C 2H 2

following pure gases; and 2-C 4 H8 .

ion pair W was

CF 4 , C 2 F 6 , CaFa.

n-C 4 F1 0 ,

16

11.

The W values were measured in the following binary gas mixtures: CF./Ar, C 2 F 6 /Ar, C 3 Fs/Ar. CF4 /I

4

, CF 4 /C 2 H 2 , C 2 F 6 /C 2 H 2 , C 3 F,/C 2 H 2 ,

C 2 F6 /2-C4 H,, CF,/C2H 2. Ar/C 2 H 2 and Ar/2-C4 Hs.

12.

The

W

were

values

also

the

C 2 F6 /Ar/C 2 H 2 ,

CF 4 /Ar/C 2 H 2 ,

mixtures;

in

measured

gas

ternary

following

and

C 2 F 6 /Ar/2-C 4HO

These mixtures have not only been found to possess

C 3 Fq/Ar/C 2 H2 .

excellent w, a/N and rjN characteristics, but also for specific

C 2 H2

or 2-C4 He concentrations, can also considerably enhance the electron production in the discharge by the external electron beam, thereby maximizing the current switching efficiency. 13.

Electron attachment rate constant k

a

measurements were performed

over the gas temperature range 300 < T < 700 K in OCIF 3 . C2 F6 . CFs, n-C 4 F1 o, c-C4 Fs, c-C 4 F 6 , and c-C6 F 6 .

These measurements helped to

elucidate the electron attachment mechanisms in these molecules. 14.

Electron

drift

performed

in

C 2 F 6 /CH 4

at

measurements opening

velocity,

a/N

and

pure C2 F6 and CH4 . the gas

switches will

not be

also

were

the gas mixtures C2 F6 /Ar

and

500 K and 700 K.

These

the operation of repetitively

pulsed

temperatures

have shown that

and

7/INmeasurements

300 K,

seriously degraded by operation at

these gas temperatures. 15.

High voltage breakdown

field strength measurements

(E/N)Rim were

performed in the following gases over the temperature range 300 K < T < Z 600 K at atmospheric gas pressures; CF 3 CI, C 2 F 6 , C 3 F8 , I-C3 F6,.

c-C4 Fs and n-C4 F1 o.

These measurements have shown that for C2 F,

(which is considered a prime candidate for use in diffuse discharge opening switches) operation of the switch at these temperatures does not degrade the dielectric properties of this gas, and in fact, the dielectric strength increases slightly over this temperature range.

17

1.5 A.

Publications and Presentations

Journal Publications

The following journal papers were published under full or partial support of this contract. 1.

S. M. Spyrou, S. R. Hunter. and L. G. Christophorou, "Studies of Negative Ion Formation in Fluoroethers and Fluorosulphides Using Low-Energy ((10 eV) Electron Beam and Electron Swarm Techniques." J. Chem. Phys. 81,

4481-4493 (19S4). 2.

K. Nakanishi. L. G. Christophorou, J. G. Carter, and S. R. Hunter, "Perning Ionization Ternary Gas Mixtures for Diffuse Discharge Switching Applications," J. Appl. Phys. 58. 633-641 (1985).

3.

S. R. Hunter, J. G. Carter, and L. G. Christophorou. "Electron Transport Studies of Gas Mixtures for Use in e-Beam Controlled Diffuse Discharge Switches," J. Appl. Phys. 58. 3001-3015 (1985).

4.

S. M. Spyrou and L. G. Christophorou, "Effect of Temperature on the Dissociative Electron Attachment to OClF 3 and C 2 F 6." J. Chem. Phys. 82, 2620-2629 (1985).

5.

S. R. Hunter. J. G. Carter, and L. G. Christophorou, "Electron Transport Measurements in Methane Using Improved Pulsed Townsend Experimental Techniques," J. App[. Phys. 60, 24-35 (1986).

6.

G. F. Reinking. L. G. Christophorou, and S. R. Hunter. "Studies of Total Ionization in Gases/Mixtures of Interest to Pulsed Power Applications," J. Appl. Phys. 60, 499-508 (1986).

7.

P. G. ]htskos and L. G. Christophorou, "Variation of the Electron Attachment to n-C 4 F 1 o with Temperature," J. Chem. Phys. 86, 1982-1990 (1987).

S.

S. R. Hunter. "Comment on Shortening of Electron Conduction Pulses by Electron Attachers 02, N 2 0, and CF4," J. Appl. Phys. 60, 4335-4337 (1986).

9.

S. R. Hunter. J. G. Carter, and L. G. Christophorou. "Electron Attachment and Ionization Processes in CF4, C 2 F6 , CaFe, and n-C4 F 1 o," J. Chem. Phys. 86, 693-703 (1987).

10.

S. R. Hunter, J. G. Carter, and L. G. Christophorou, "Electron Motion in CF 4 , C 2 F 6 . C 3 F9, and n-C 4 F1 o," Phys. Rev. A, in press.

11.

L. G. Christophorou, R. A. Mathis, S. R. Hunter, and J. G. Carter, "Effect of Temperature on the Uniform Field Breakdown Strength of Electronegative Gases," 1. AppI. Phys. 63, 52-59 (1988).

12.

S. R. Hunter, J. G. Carter, Attachment. Ionization and Drift preparation.

and L. G. Christophorou, in SiF 4 and BF3 ," J. AppI.

"Electron Phys., in

18

"Electron

Attachment

to

NF3 ."

J.

Appl.

Phys.,

in

13.

S. R. Hunter, preparation.

14.

S. R. Hunter, J. C. Carter, and L. G. Christophorou. "Electron Transport in C 2 F 6 /Ar and C 2 F/M4 Gas Mixtures at Elevated Temperatures," 1. Appl. Phys., in preparation.

B.

Conference Proceedinqs

The following papers were published in conference proceedings under full partial support of this contract.

or

1.

L. G. Christophorou, S. R. Hunter, J. C. Carter, and S. M. Spyrou. "Gases for Possible Use in Diffuse-Discharge Switches," In Proceedings of the Workshop on "Diffuse Discharge Opening Switches" (M. Kristiansen and K. M. Schoenbach, Eds.), Texas Tech University, Lubbock. Texas, 1982, pp. 236-251.

2.

J. C. Carter, S. R. Hunter, L. C. Christophorou, and V. K. Lakdawala, "Electron Drift Velocity and Ionization and Attachment Coefficients in Gases/Mixtures for Diffuse-Discharge Opening Switches," In Proceedings of the 3rd International Swarm Seminar (W. Lindinger, H. Villinger, and W. Federer. Eds.), Innsbruck, Austria. 1983, pp. 30-36.

3.

L. G. Christophorou. S. R. Hunter, J. V. K. Lakdawala. "Basic Studies of Gases Applications," In Proceedings of the Externally Controlled Diffuse Discharges, Texas, 1983, pp. 104-133.

4.

L. G. Christophorou, S. R. Hunter, J. G. Carter, S. M. Spyrou, and V. K. Lakdawala. "Basic Studies of Gases for Diffuse-Discharge Switching Applications." In Proceedings of the 4th IEEE Pulsed Power Conference (T. H. Martin and M. F. Rose, Eds.), The Texas Tech Press, Lubbock, Texas, 1983, pp. 702-708.

5.

S. R. Hunter. J. G. Carter, L. G. Christophorou, and V. K. Lakdawala, "Transport Properties and Dielectric Strengths of Gas Mixtures for Use in Diffuse Discharge Opening Switches," In Gaseous Dielectrics IV (L. G. Christophorou and M. 0. Pace, Eds.), Pergamon Press, New York, 1984, pp. 224-237.

6.

K. Nakanishi, L. G. Christophorou, J. G. Carter, and S. R. Hunter, "Penning Ionization Ternary Gas Mixtures for Diffuse Discharge Switching Applications," In Proceedings of the 5th IEEE P'ilsed Power Conference (M. F. Rose and P. J. Turchi, Eds.), IEEE, New York, 1986, pp. 40-43.

7.

S. R. Hunter, J. G. Carter, L. C. Christophorou, and S. M. Spyrou, "Temperature Dependent Electron Transport Studies for Diffuse Discharge Switching Applications," In Proceedings of the 5th IEEE Pulsed Power Conference (M. F. Rose and p. J. Turchi, Eds.), IEEE, New York, 1986, pp. 402-409.

C. Carter, S. M. Spyrou, and for Diffuse-Discharge Switching U.S.-F.R.G. Joint Seminar on Texas Tech University, Lubbock,

19

8.

L. G. Christophorou, S. R. Hunter. J. G. Carter, and S. M. Spyrou, "Effects of Temperature on Dissociative and Nondissociative Electron Attachment," In Swarm Studies and Inelastic itectron-Molecule Collisions (L. C. Pitchford, V. McKoy. A. Chutjian, and S. Trajmar. Eds.), Springer-Verlag. New York. 1986, pp. 303-308.

9.

S. R. Hunter, J. G. Carter, and L. G. Christophorou. "Electron Drift Velocity and Attachment and Ionization Coefficients in CH4, CF 4 , C 2 F6 , CF,. and n-C 4 F," In Swarm Studies and Inelastic Electron-Molecule Collisions (L. C. Pitchford, V. McKoy, A. ChuLjian. and S. Trajmar, Eds.), Springer-Verlag, New York, 1986, pp. 93-94.

10.

S. R. Hunter. "Gas Engineering for Pulsed Power and Switching," In Gaseous Dielectrics V (L. G. Christophorou and D. W. Bouldin. Eds.), Pergamon Press. New York. 1987, pp. 363-372.

11

S. R. Hunter. L. G. Christophorou. and J. G. Carter, "Gas Engineering Studies for High Pressure Self-Sustained Diffuse Discharge Closing Switches," In Gaseous Dielectrics V (L. G. Christophorou and D. W. Bouldin, Eds.), Pergamon Press, New York, 1987, pp. 404-411.

12.

J. G. Carter, S. R. Hunter, and L. G. Christophorou, "Temperature Dependent Electron Transport and Rate Coefficient Studies for e-BeamSustained Diffuse Gas Discharge Switching," In Gaseous Dielectrics V (L. G. Christophorou and D. W. Bouldin, Eds.), Pergamon Press, New York, 1987. pp. 47-54.

13.

L. G. Christophorou, R. A. Mathis. S. R. Hunter, and J. G. Carter, "Effect of Temperature on the Uniform Field Breakdown Strength ot Electronegative Gases," In Gaseous Dielectrics V (L. G. Christophorou and D. W. Bouldin, Eds.). Pergamon Press. New York. 1987, pp. 88-95.

14.

S. R. Hunter, J. G. Carter, and L. G. Christophorou, "Electron Transport Studies in Gaseous Media for Diffuse Discharge Closing Switches," In Proceedings of the International Conference on Phenomena in Ionized Gases (ICPIG XVIII), Swansea, United Kingdom, July 13-17, 1987.

15.

S. R. Hunter, L. G. Christophorou, J. G. Carter, and P. G. Datskos, "New Concepts in High Current Self-Sustained Diffuse Discharge Closing Switches," In Proceedings of the 6th IEEE Pulsed Power Conference, Arlington, Virginia, June 29-July 1, 1987. pp. 1-8.

C.

Book Chapter

The following book chapter was published under contract.

the partial

support

of

this

L. G. Christophorou and S. R. Hunter, "Electrons in Dense Gases." In Swarms of Ions and Electrons in Gases (W. Lindinger, T. D. Mark. and F. Howorka, Eds.). Springer-Verlag/Wien, Austria, 1984, pp. 241-264.

20 D.

Patents and Patent Applications

The following patents and patent disclosures were developed under partial support of this contract. 1. L. G. Christophorou, S. R. Hunter. and J. G. Carter. Gas Mixtures for Diffuse-Discharge Switches, Patent No. 4,490,650 (December 25, 1984). 2.

L. G. Christophorou and S. R. Hunter, Penning Ionization Ternary Gas Mixtures for Diffuse-Discharge Switch (granted).

3.

S. R. Hunter and L. G. Christophorou, Binary and Ternary Gas Mixtures for Diffuse Glow Discharme Closing Switches.

4. L. G. Christophorou and S. R. Hunter, Binary and Ternary Gas Mixtures with Tempe:ature Enhanced Characteristics for Use in Diffuse Glow Discharge Closing Switches. E.

Presentations

The following invited and contributed talks and lectures were presented on the work developed during this contract. 1. L. G. Christophorou, S. R. Hunter, J. G. Carter, and S. M. Spyrou (invited), "Gases for Possible Use in Diffuse-Discharge Switches," Workshop on "Diffuse Discharge Opening Switches," Tamarron, Colorado. January 13-15, 1982. 2.

L. G. Christophorou (invited), "Gases for High Voltage Insulation and Pulsed-Power Technologies." Sandia National Laboratories, Albuquerque, New Mexico, January 21, 1983.

3.

S. R. Hunter (invited), "Electron Transport Studies for Diffuse Discharge Switching Applications," Chemical Physics Seminar. Oak Ridge National Laboratory, Oak Ridge, Tennessee, May 9, 1983.

4.

L. G. Christophorou. S. R. Hunter, J. G. Carter, S. M. Spyrou, and V. K. Lakdawala (invited), "Basic Studies of Gases for Diffuse-Discharge Switching Applications," 4th IEEE Pulsed Power Conference, Albuquerque, New Mexico, June 6-8, 1983.

5.

J. G. Carter, S. R. Hunter, L. G. Christophorou, and V. K. Lakdawala (contributed), "Electron Drift Velocity and Ionization and Attachment Coefficients in Gases/Mixtures for Diffuse-Discharge Opening Switches," 3rd International Swarm Seminar, Innsbruck, Austria, August 3-5, 1983.

6.

L. G. Christophorou (invited), "Development of New Gases for Diffuse Discharge Opening Switches," Workshop on "Externally-Controlled Diffuse Discharges," Bad Honnef, West Germany, August 15-19, 1983.

21

7.

L.

G.

Christophorou,

S.

R.

Hunter,

J.

G.

Carter,

S.

M.

Spyrou.

and

V. K. Lakdawala (invited), "Basic Studies of Gases for Diffuse-Discharge Seminar on Externally Joint U.S.-F.R.G. Switching Applications," Controlled Diffuse Discharges. Bad Honnef, West Germany, August 15-19. 1983. 8.

L. G. Christophorou (invited). "Electron-Molecule Interactions and the Development of Gases for High Voltage Insulation and Pulsed Power Department of Electrical Engineering. Old Dominion Technologies," University, Norfolk. Virginia, March 23. 1984.

9.

S. R. Hunter, J. G. Carter. L. G. Christophorou, and V. K. Lakdawala (contributed), "Transport Properties and Dielectric Strengths of Gas Fourth Mixtures for Use in Diffuse-Discharge Opening Switches," International Symposium on Gaseous Dielectrics, Knoxville. Tennessee, April 29-May 3. 1984.

10.

S. R. Hunter (invited), "Electron Transport Studies of Gas Mixtures for Pulsed Power Diffuse Discharge Switching Applications," Health and Safety Research Division Information Meeting, Oak Ridge National Laboratory. Oak Ridge, Tennessee. October 3-5, 1984.

11.

L. G. Christophorou (invited). "Basic Physics of Gases for Pulsed Power." Ecole Superieure d'Electricite de France, Gif-sur-Yvette, France, July 5, 1984.

12.

L. G. Christophorou (invited), "Basic Physics of Gases for Pulsed Power Technologies," Naval Research Laboratory, Washington. D.C., October 24. 1984.

13.

S. R. Hunter. J. G. Carter, L. G. Christophorou, and S. M. Spyrou (invited), "Temperature Dependent Electron Transport Studies for Diffuse Discharge Switching Applications," 5th IEEE Pulsed Power Conference, Arlington, Virginia, June 10-12, 1985.

14.

R. Nakanishi, L. G. Christophorou, J. G. Carter, and S. R. Hunter (contributed), "Penning Ionization Ternary Gas Mixtures for Diffuse Discharge Switching Applications," 5th IEEE Pulsed Power Conference, Arlington, Virginia, June 10-12, 1985.

15.

L. G. Christophorou, S. R. Hunter. J. G. Carter, and S. M. Spyrou (invited), "Effects of Temperature on Dissociative and Nondissociative Electron Attachment," Joint Symposium on Swarm Studies and Inelastic Electron-Molecule Collisions, Tahoe City, California, July 19-23, 1985.

16.

S. R. Hunter, J. C. Carter, and L. G. Christophorou (contributed), "Electron Drift Velocity and Attachment and Ionization Coefficients in CH4 , CF4 . CF 6,. CaFa, and n-C4 F1 0 ," Joint Symposium on Swarm Studies and Inelastic Electron-Molecule Collisions, Tahoe City, California, July 19-23, 1986.

22

17.

J. G. Carter, S. R. Hunter, and L. G. Christophorou (contributed), "Electron Drift Velocity and Attachment and Ionization Coefficients in C2 F6 /Ar and C 2 F6 /CH4 Gas Mixtures at Elevated Gas Temperatures," 38th Annual Gaseous Electronics Conference, Monterey, California, October 15-18. 1985.

18.

L. G. Christophorou (invited), "From Basic Research to Application: Gas Engineering for High Voltage Insulation and Pulsed Power Technologies," Department of Electrical Engineering, Mississippi State University, Mississippi State, Mississippi, November 21, 1985.

19.

S. R. Hunter (invited). "Electron Transport Studies in Vibrationally and Electronically Excited Gases/Mixtures for Use in e-Beam Controlled Diffuse Gas Discharge Switches," Department of Electrical Engineering, Old Dominion University, Norfolk. Virginia, March 21, 1986.

20.

S. R. Hunter, L. G. Christophorou, and J. G. Carter (contributed), "Electron Drift Velocity and Attachment and Ionization Coefficients for Gases/Mixtures for Use in Diffuse Discharge Switching Applications," 1986 IEEE International Conference on Plasma Science, Saskatoon, Saskatchewan, Canada, May 19-21, 1986; abstract published in IEEE Conference Record Abstracts, IEEE Catalog No. 86CH2317-6, IEEE. New York, 1986, pp. 7-8.

21.

S. R. Hunter (contributed), "Electron Attachment to NF3 Revisited," Thirty-ninth Annual Gaseous Electronics Conference, Madison, Wisconsin, October 7-10. 1986.

22.

S. R. Hunter and L. G. Christophorou (contributed), "Gas Engineering Studies for High Pressure Self-Sustained Diffuse Discharge Closing Switches," Fifth International Symposium on Gaseous Dielectrics, Knoxville, Tennessee, May 3-7, 1987.

23.

L. G. Christophorou, D. L. Mcorkle, and S. R. Hunter (contributed), "Gas Mixtures for Spark Gap Closing Switches with Emphasis on Efficiency of Operations," Fifth International Symposium on Gaseous Dielectrics, Knoxville, Tennessee, May 3-7, 1987.

24.

S. R. Hunter (invited), "Gas Engineering for Pulsed Power Switching," Fifth International Symposium on Gaseous Dielectrics, Knoxville, Tennessee, May 3-7, 1987.

25.

J. G. Carter, S. R. Hunter, and L. G. Christophorou (contributed), "Electron Transport and Rate Coefficient Studies for e-Beam-Sustained Diffuse Gas Discharge Switching." Fifth International Symposium on Gaseous Dielectrics, Knoxville. Tennessee, May 3-7, 1987.

26.

S. R. Hunter, L. G. Christophorou, J. G. Carter, and P. G. Datskos (invited), "New Concepts in High Current Self-Sustained Diffuse Discharge Closing Switches," 6th IEEE Pulsed Power Conference, Arlington. Virginia, June 29-July 1, 1987.

23

27.

S. R. Hunter, J. G. Carter, and L. G. Christophorou (contributed), "Electron Transport Studies of Gaseous Media for Diffuse Discharge Closing Switches," International Conference on Phenomena in Ionized Gases (ICPIG XVIII), Swansea, Wales, United Kingdom, July 13-17, 1987.

28.

L. G. Christophorou, S. R. Hunter, L. A. Pinnaduwage, J. G. Carter, and P. G. Latskos (contributed), "Electron Attachment Properties of Excited for Pulsed Power Dielectric-Gas Molecules and their Possible Use IX International Conference on Gas Discharges and their Switching," Applications, Venice, Italy, September 19-23, 1988.

II.

Experimental Techniques Several

different

experimental

data discussed in this report. commencement of

Most of

to obtain

were employed

the

these experiments existed before the

this program, but some were modified during the course of the

research program.

present

techniques

experiments are described briefly

These

in

this

section. High Pressure Swarm Experiment

2.1

rate constants in 16 1 and Hurst. ' 7 Bortner by the present swarm studies was originally devised The technique used

the electron attachment

to obtain

In Fig. 3 is shown the schematic diagram of the present apparatus.

Electrons

are produced in a plane at the source electrode, perpendicular to the applied electric

field,

particles.

E,

by

the

passage

through

the

gas

of

high

energy

alpha

The alpha particle trajectories are well collimated such that the

electron swarms produced by the energy decay of the alpha particles lie in a Each alpha particle (produced by the decay

well-defined plane at the source. of Cf

2

6

2

,

energy Z 6.1 MeV) produces a swarm of Z 2.3 x 105 electrons in N 2

within Z 5 x 10- 9 s at electrons then drift

133 kPa to 2 x 10- '° s at 3.2 MPa.

to the anode under

These swarms of

the influence of a uniform electric

field established between the cathode and anode. The change in the number n(t) of electrons in the swarm during the time interval dt is dn(t) dt

where

T1

wn(t)

_ -

(11)

-

is the average number of attaching collisions per centimeter of drift,

and w is the drift velocity of the swarm.

The measurements were performed at

high total gas pressures (133 kPa < PT < 3.9 MPa) such that the influence of

24

PROGRAMMABLE SUPPLY(O-3OkV) S

VOLTAGE DIVIDER AND FILTER

HIGH PRESSURE CERAMIC FEEDTHROUGHS

PRESSURE GAGES, GAS HANDLING AND

PURIFICATION SYSTEMS

CATHODE D

VIDEO TO UHV PUMPING

SOURCE

GOLD PLATED ELECTRODES

STAINLESS STEEL HIGH PRESSURE

VESSEL

x Av1

i

PRINTE-R

!,,CROCOMPUTER MULTICHANNEL ANALYZER

PREAMPLIFIER LINEAR AMPLIFIER

Fig. 3.

Schematic diagram of conditioning

the electron attachment apparatus and signal

electronics

used

in

the

room

temperature

attachment rate constant measurements (Ref. IS).

electron

25

diffusion upon the drift of

the swarm was negligibie.

Equation (11)

has the

following solution

(12)

n(t) = n o exp[-(/N a)Na wt]

where 77/Na

is defined as the attachment coefficient with the units of cm2 , Na

is the partial attaching gas number density, and n0 is the initial number of electrons produced by each a-particle at the beginning of the drift space. To follows:

obtain

the attachment coefficients,

the

experiment

A pure buffer gas sample is admitted

into

is performed

as

the vessel at a given

pressure, and a distribution of voltage pulse heights are obtained at a given E/N, which are accumulated in a multichannel analyzer (MCA) operating in the pulse height analysis mode. E/N

The most probable pulse height is recorded.

is changed and a series of most probable pulse heights

function of E/N.

The

is obtained as a

If the pressure dependence of the attachment process is to

be studied, then the above set of measurements are performed over a range of buffer gas pressures as well. to the drift vessel in the present

A small quantity of attaching gas is then added

(usually one part in 105 to 108 of the total gas pressure

study),

the chamber

refilled with

the buffer gas, and a new

series of most probable puise heights obtained as a function of E/N and PT. The ratio of the pulse heights with and without the attaching gas is obtained. The electron attachment coefficient 9/N is then obtained from this ratio. a The drift region electrodes of the experimental apparatus shown in Fig. 3 are housed in a stainless steel high pressure vessel, capable of operating at total

gas pressures up

to 10 MPa.

The chamber

system with a base pressure of a few parts -

of

10 3 Pa per hour.

enameled

insulators

All are

used

reactive attaching gases. Cf

2 52

electrodes

reduce

in 10 6 Pa and an outgassing rate the

chamber are gold plated, and

possible

surface

reactions

with

The alpha paticle source used in these studies was

which was electroplated

the source electrode.

to

in

is pumpcd by a UHV pumping -

onto a thin platinum ribbon and housed within

This source produced

Z104

detectable swarms per second.

The induced anode voltages, produced by the motion of the swarm, are detected and

amplified

microprocessor

by

a

based,

preamplifier 2048

channel

and

linear

amplifier

multichannel

and

analyzer

passed

(MCA),

into

where

a the

26

The pulse height

distribution of the resultant peak pulse heights are stored.

histogram is analyzed by the microprocessor to find the most probable pulse is then used

which

height,

programmable voltage supply and hence obtained,

the

and

procedure

whole

the

resets

then

microprocessor

The

procedure.

iterative

numerical

using a simple

thus ka

i/Na , and

to obtain

the E/N value, and a new histogram is of

details

Further

repeated.

the

experimental technique are given in Ref. 18. High Temperature Swarm Experiment

2.2

The high temperature swarm apparatus is shown schematically in Fig. 4. of

The operation except

that

number

of

were

modifications

the measurements at high gas

facilitate other

a

swarm

distances of

the

apparatus,

similar

is very

this experiment

swarm

that described

to

made

to

chamber

made

was

-100

at a

much

design

cm

long,

to

to our

contrast

the chamber

T7 avoid leaks at the flanges when heating or cooling, and was kept

basic

In

temperatures.

the two end flanges from the middle of

flanges was water cooled

the

above,

and

the

are --50 cm.

the region around the

lower

temperature

than

collision region (the region between the anode and cathode: see Fig. 4).

the This

arrangement also allowed the high voltage and signal feedthroughs to be kept low T.

at a

Special

care was also

taken

in

the construction of

supporting stands holding the anode and the cathode (the Cf

2s

2

the

long

alpha particle

source used to produce the electron swarms by the energy decay of the alpha particles was mounted on a plate supported on the cathode), which consist of a stainless steel rod with insulating rods at its two ends. A

furnace

and

Applied Test System, Inc.) The

temperature

(chromel-alumel,

in

control

temperature

the

system

(0-10000 C,

resolution

1C,

was used to heat the central region of the chamber.

collision region was

type KX) located as close

measured

by

six

thermocouples

to the drift region as possible.

With the aid of two independently controlled heating elements embedded in the insulating walls of

the

furnace, 0

was controlled to within 1-2 C.

the gas

temperature between the electrodes

Further details of the experimental technique

are given in Ref. 19. 2.3

Pulsed Townsend Experiment

The electron drift apparatus used to obtain the room temperature and high temperature electron drift velocity and attachment and ionization coefficient data is shown schematically in Fig. 5.

The high pressure chamber was designed

27

C)

c-

-------

-

----

L.

biD

;S15 i

C)

n

28

0

00

I-

1 atm) electron attachment

High pressure alkanes

and several fluorinated ethers

molecules possess electron attachment

26

the perfluoro-

of

studies

have shown that several of these

rate constants which have

desirable

energy dependences for diffuse discharge switching applications (i.e.,

they

attach electrons efficiently at high energies and have much reduced electron These measurements are

attachment rate constants at near-thermal energies).

summarized in Fig. 8a and have been obtained using a high pressure swarm technique. 18 '26 -2 7

The

molecules

C2 F6

and

interest

for opening switch applications due

thermal

electron

attachment

and

that

the

CF3 OCF 3 to peak

are

their in

comparatively high electron energies for both molecules.

considerable

of

very

low rates of

ka((E))

occurs

at

These two molecules,

along with C4 and CF3 SCF 3 attach electrons dissociatively.

On

the other

hand, molecule C3 Fg (and n-C4 Fo) is particularly noteworthy in that electron attachment to this molecule at atmospheric pressures and ambient temperatures 18

is predominantly

by

parent

negative

ion

stabilization,

and

thus

this

molecule could possibly be used in closed-cycle switches. The electron attachment rate constant for NF3 is plotted in Fig. 9 as a of (C) function measurements.28,29 comparatively 10-

°

-

10- 9

high cm3 S-I]

switch applications.

electron previous several along with This molecule was previously thought to thermal

electron

attachment

rate

constant

attachment possess a [(ka)th

and consequently to be unsuitable for use in opening The present measurements show that the thermal electron

attachment in NF3 is up to two orders of magnitude lower than these estimates [(ka)th

3 x 10- 1 1

cm3 S-1], making this molecule a suitable candidate for

study in switching applications.

34

000 co

0

00 0 0~

0

o

0.

U, WJ

0

(A

>

U.

-- "I

/-

a:

11

x

4--1

'

wr

L.J

z

(.)

so

0 C,,

0~~V

0

o~

wi

00

0

(U~

V

Ln

u

>0

0

09

0)

*

>4

C

LD

0

0

V

mo i

v

Sc

U

~~~~~

SLI.

0

(s

LuD)

-

4

f

CDV

V

o00

0 r4

0

zS.4

U*Q

4

00

W)

"

.

-

-

0C.

SA4

V

I

*-

'M 1NVSNOO ]1V8 IN3Y4HOViIV IVIOI

0

0

0~'S-

35

QSNFiNF3 r

~1o'°

0

N

N

NF3 in Ar

10-

f

1

0

2

0-

and - N " a LAKDAWALA MORUZZI (1980) 2 nLAKDAWALA and MQRUZZ 1 (1950)

a

E-

CHANTRY (1979)

0 SPRESENT

Z

He

-

*

MEASUREMENTS

A

PRESENT MEASUREMENTS

rU

-

N2

-

Ar

£-2

10

Fig. 9.

id

100 10(E) (eV) ENERGY ELECTRON MEAN

The electron attachment

rate constant k

as a function of

mean

a

electron

energy

(C)

for

NF3

in

comparison

with

the

previous

measurements of Chantry (Ref. 28) and Lakdawala and Moruzzi 29).

(Ref.

36

The

total

electron attachment

section a (E) for

cross

of

each

these

as a function of (E)

molecules is related to the attachment rate constant k a (or

E/N)

and

distribution

energy

the electron

of

f(E,E/N)

function

the gas

mixture by

ka(E/N) = rI(E/N)aw(E/N)

where

r1/N a

is

electron mass.

the

normalized

0f

E Y2 cra(E) f(E,E/N) dE

electron attachment

coefficient

,

(13)

and m is

the

The electron energy distribution function is normalized by

.f

(14)

f(E.E/N) dC

0 If both k a(E/N) and f(E,E/N) are known over a wide range of E/N values, then

a(E) can be determined over a wide range of electron energies using a swarm 30 ,iven in The attachment rate constpnrt meg'rc~ment unfolding technique. Figs.

8a

and

9

have

been

obtained

for

these

electronegative

performing the measurements in gas mixtures composed of 106

to

108) of

traces

gases

by

(one part in

the attaching gas in the high pressure buffer gas of argon.

The electron energy distribution functions f(EE/N) and w(E/N) values used in 18 as we have made the justifiable assumption Eq. (13) are those for pure Ar, are unaffected by 18,26 (Any the addition of minute quantities of the attaching gas to the Ar. that

the electron

transport

parameters

and hence

f(EE/N)

observed dependence of k a on the attaching gas partial pressure can be removed 18,26 by extrapolating the measurements to zero attaching gas pressure.) The electron attachment cross sections a (E) obtained from these analyses are plotted in

Figs.

8b and

10 and

values given in Figs. Sa and 9, electronegative electron

gases

energies

which

well

in

indicate

of

indeed have

that we

attach electrons excess

along with the ka((E))

these measurments,

over

thermal

a

wide

energies

found several

energy (E

)

range 0.3

at

eV).

Consequently, the electron attachment processes in a potential switching gas

37

22q

o PNF

3

oS1.8

0

MEASUREMENTS

A,

b 1.6

zo

........ PRESENT

A

1.4

HARLAND and

A

U

A

V)

CHANTRY (1979)

__

,,

iFRANKLI

N (1974)

12

and

.HARLAND

FRANKLIN (19-4) RENORMALIZED

1.0 A

z

~

0.8

A

0.6 0.4

Z&-

0A

0

0

Ae

i0.

E-4

L) f

0~

0

s.

0.0 0.0

1.0

2.0

3.0

4.0

ELECTRON ENERGY

Fig.

10.

The electron attachment

cross

section

5.0

6.0

7.0

8.0

E (eV)

a (E)

for

NF,

in

comparison

with the previous measurements of Chantry (Ref. 28) and Harland and Franklin (Ref. 31).

38

mixture using these gases can be positioned at appropriate (E) or E/N values by choosing a gas which efficiently attaches electrons in that E/N range. Elevated Temperature Electron Attachment Rate Constant Measurements

3.2

energy

storage

circuit is fired, the gas temperature within the switch is expected

to rise

several

hundred

Whenever

several

the

diffuse

discharge

centigrade, and

degrees

in an

switch

inductive

temperatures

operating

An understanding of

likely for repetitively operated switches.

degrees are

of

the factors that affect the operation of the switch under these conditions can be

obtained

by

the

performing

electron

transport

and

rate

coefficient

measurements at gas temperatures above room temperature. has been measured for C 2 F 6

The electron attachment rate constant k ((E)) a over

the

temperature

< T

300

range

< 750 K

in

order

to

investigate

the

influence of gas heating on the electron attaching properties of this molecule (Fig.

11).

the gas

As

temperature

increases,

k a((E))

increases, and

this

increase is progressively larger at lower energies such that the threshold and the k a((E))

the peak in

shift to lower energies at higher gas temperatures.

increases by Z30% over this temperature range near its peak at (E)

The ka((E))

3 eV (Fig. 11). Measurements of ka((E)) have also been performed in CF, and n-C 4 F1 o as a function of

gas

temperature

to 750 K

up

in Ar

electron energy range 0.76 < (E) < 4.8 eV). Figs.

12 and

slightly up

13 and show T - 400 K,

to

to

be

of

(E),

rapidly decreases with

then

gas

the

(over

mean

These measurements are given in

that at a given value

T - 450 K, and finally significantly temperature.

buffer

increases with

k a

decreases only

increasing

T up

increasing T above

to

this

The lower temperature measurements (T < 450 K) have been found

strongly

dependent

on

total

gas

pressure.

indicating

that

parent

negative ion formation processes are significant electron attachment processes at

these

dependent

temperatures. electron

At

higher

attachment

gas

temperatures

processes

are

(T > 450 K)

negligible

pressure

indicating

that

electron attachment to CFa and n-C 4 F1 o at these temperatures is predominantly dissociative. the

gas

These measurements

kinetic

energy

indicate

(and hence

in

the

that

relatively small

vibrational

changes in

populations

of

the

attaching gas) can have a large influence on the electron attaching properties of CFg and n-C 4 Fo which could, in turn, significantly affect the performance

39

ORNL-DWG 84-15088R ,

/

,

/750

K

Z5

400 K

300 K

04 30

E-3 2 F6

zC

IN ARGON

120 0.5

1

1.5

2

Z5

3

3.5

4.5

4

5

5.5

MEAN ELECTRON ENERGY, (E) (eV)

Fig. 11.

The total electron attachment rate constant k

mean

electron

as a function of the

a



energy

(E) for

C2 F6

between 300 K and 750 K (Ref. 19).

at

several

gas

temperatures

40

0tL-DWG

S5--94,24R

t0

300 K .

-..-.

.

0'.0" 425 K

.... . -.

l .a

S

S

5.'

-

z

2

U

12

0-

$w

,-..

--

450 K /

**

,'

wC3F8

,

i

-

,.".T

IN Ar i

i

,

0

E

750 K 0

8

.

"

2 75 K

3.

3

4..

3.0

3.5

4.0

E-

0D

'A".

10~

di ssion su

0.5

at th4a ofeeto

Fig.

12.

1.0

1.5

2.0

2.5

4.5

5.0

eprtrsgvnintefgr. MEAN ELECTRON ENERGY, (E) 300se Tecrea K ef.•1.orfl (eV) atcmn atti temertue

Electron attachment

rate constant

kI 1 for NT -

(and N

i

- 0)

for

CF, as functions of mean electron energy ()

in a buffer gas of Ar

at the gas temperatures given in the figure.

The curve at 300 K in

Fig. 12b is the dissociative attachment component to the total rate of

electron attachment

discussion).

at

this

temperature

(see

Ref.

19 for

full

41

ORNL-DWG 86-12903 30

n-C 4 F 10 IN Ar

(a) -

25

300 K 350

V)

S 0

15

1

450

&H 5

z Z 0

0 14

(b)

H L2

z

"10 w 10

H

~

6

606

400

0.5

1.0

1.5

2.0

as5

3.0

as5

40

4.5

S

MEAN ELECTRON ENERGY, (z~) (eV)

Fig.

13.

Electron attachment rate constant kIfor n-C.F10 obtained under the same conditions as those for GaF8 in Fig. 12.

42

of repetitively operated switches operating at elevated gas temperatures using C3F. or n-C 4 Fjo. Room Temperature Electron Drift Velocity Measurements

3.3

Pure Cases

A.

in several pure gases

The electron drift velocity w has been measured using our pulsed Townsend experimental apparatus.

The w measurements in gases

which are potentially useful in diffuse discharge opening and closing switches are

in

given

Figs.

14 and

15,

(with

these gases

All

respectively.

the

exception of n-C 4 Fo) possess significantly enhanced w values (i.e. regions of conductivity

negative differential

at

[NDC])

low E/N

comparatively

values,

which as indicated in Fig. 2 is a very desirable characteristic of gases for use in diffuse discharge opening switches. Gas Mixtures

B.

The electron drift velocity w has been measured in several gas mixtures composed of gases

buffer

the attaching gases shown in

an

attempt

to

in Fig. 8a using either Ar or CH 4 as

combine

the

electron

mobility

enhancement

afforded by

the NUC effect at low E/N observed in gas mixtures using Ar and

CH 4 ,

the

with

electronegative

very gases

desirable given

in

electron Sa

Fig.

attaching at

E/N

high

of

properties (or(C)).

the These

measurements are plotted in Figs. 16 to 19 over the concentration range of 0.1 to 100% of the attaching gas in the buffer gas. accuracy of

the

The measurement technique and

resultant data have been discussed

in Ref.

32.

These gas

mixtures all possess pronounced regions of NDC over a range of E/N values. The

NDC

effects

observed

in

several

of

these gas

mixtures

are among

the

largest that have been observed in any gas mixture and are the result of large vibrational inelastic energy loss processes in these electronegative gases at comparatively low electron energies [0.1 < C < 1.0 eV]. Electron drift velocity measurements in several SiF 4/He gas mixtures are plotted in Fig. 20.

These gas mixtures have been found to possess the very

desirable characteristics outlined discharge closing switches.

later in

this section for use in diffuse

They have the added advantage of possessing high

w values at E/N values near the operating E/N value of the discharge when the switch is closed.

43

ORNL-DWG 86-13258

11

U

i

I

Z

10

I

1

OI

1

10

.E-.

a

C3F 8 n-C4Fo

,-4 .&.%J10-"~

t

10 0

E/N (V cm

Fig. 14.

2

Electron drift velocity w versus E/N for n-C4+F 1 o.

-I

CH4, CF4,

C2F6,

CF, and

44

ORNL-DWG 87-10824

K

>--

/

XN

//,' m, SiF4

E/6

id'

4

4

o ,

/CF CH4

?t

10-

i0

1

-16

C

0-

E/N (Vom2)

Fig. 15.

Electron drift velocity w versus EI

for CH4, CF4, SiF4 and BF,.

45

AA U

0.

0

Sk. 0.

0

o Zo

a

.

00

10

00

p.

C

(ISUJ

0L

iO03

0

-4

!

i1C]NK31

00

V46

0-0

CA) 1

C4 (N

(N

Wo

(vL-s

10

0

g900 M

kil:O13A

N

ildO NOdO313W

LAJJ

I.--

00

(f L.JN

(D 1 U

I 'oW

0c

L

I to~ 10 10 C*4 C

C>

(N

L c,, C4

L. C-4C14CL)

0

P

0

0 -

010(

LWC- go90)M Xi 10013A .LJI80 NO LLOJ13

0

47

cv,

C)

0.

U')'

Cza 00

4.

It.*

oO

L

*00 cN

c0

(N

00

W))

C

co

I-S

U-1

900

M

pn

to

iO-1Ai-G

V

r4

O

3

0c

48

V0

af)

-

Lz'K

Z:k U)IOU

0.A

"I pe) PO pn

L

LL

t t

oo 10o3

~

.

~

(I-S

*-

L

)

.'0Ui

9W

.....

~

0

cm.

uG

'010

.-)N

~

N0

(/)

1-

w

000

1*

L

4

idl

....

O

37

biOO3

49

ORNL-DWG 87-6072 |*

I I

|

I

I

I II

II|

I

1

l

0

7

~10 SiF4 /He rin

S

GAS MIXTURES

A"I

0i > 4PERCENTAGE

OF SiF4 £

7~

10 %

0507%

N.100%

d

E/N (V cm 2 )

Fig. 20.

Electron mixtures.

drift

velocity

w

versus

E/N

for

several

SiF 4 /He

gas

50

Room [emperature Electron Attachment and Ionization Coefficients

3.4

The electron attachment and ionization coefficients can be obtained from the pulsed Townsend (PT) experiment using the following analysis. A schematic 20 . diagram showing the operating principle of the PT technique is given in Fig. 21.

In

the

PT method, a

photoelectrically

from

a

small

group

source

on

or

the

swarm

of

cathode

electrons

of

a

drift

is

produced

chamber

by

irradiating the cathode with a short duration ultraviolet (UV) laser radiation light

pulse.

influence

of

The electrons diffuse as the

electronegative

applied

gas

or

at

uniform high

they drift

electric

fields,

to

field.

the

the

In

swarm

anode under

the

may

the

presence of

experience

an

electron

attachment and/or ionization collision processes with the gas. The

present

experiments

were

performed

operating in the voltage integrating mode.

with

the

The motion of

detection

circuit

the electrons and

ions in the drift gap induces a charge on the anode, and hence across C, and thereby establishing an Lime

constant

of

the

increasing potential V(t) preamplifier

input

(T

=

across R (Fig. 21).

RC Z

1 s)

experiments was much greater than those of the electron (Te positive and negative T e (< T++

T,-'

negative

ions

ion

transit

times

(T+ Z T_

10- 4

in 10

Z

the

The

present

-10

-

_l0 - 2 s).

s) or

5

Since

the voltage drop across R due to the drift of the positive and is

negligible

during

the

electron

swarm

transit

time

T

e

Consequently, a break will cccur in the voltage transient as shown in Fig. 22 for w

CF 4

,

allowing

Te

,

and

hence,

the

measured

electron

drift

velocity

= md/Te e , to be obtained from the discontinuity in the waveform. The absolute values of the attachment and ionization coefficients may be

obtained from the ratio of the voltage transients shown in Fig. 22. time

constant

of

the detection circuit

T = RC >> T- Z T+

(i.e.,

When the the

charge

depletion from C is negligible during the positive and negative ion transit times),

then the potential drop across R will rise

to a constant value when

all of the ions have drifted to the anode of the drift gap, i.e.,

VT(t > T_ and T+)

ne -C T

,when

RC >> T_ or T +

(15)

51

ORNL-DWG 81-17850 NEGATIVE HIGH

VOLTAGE SUPPLY

I

CATHODE

PULSED LIGHT SOURCE

'

ANODE

IMPEDANCE PREAMPLIFIER [/ HIGH

ANODE i""O

.......

.... .....

Fig. 21.

I TRANSIENT

>

[>I

RECORDER

I

Schematic diagram showing the principle of the pulsed Townsend (PT)

technique for measuring

the electron drift velocity and electron

attachment and ionization coefficients.

52

ORNL-DWG 85-15787 1.2

I

(a)

1.0 0.8

2 E/N=25.0x10'17 V cm

0.6

0.4 -CF

4 Nd=5.45xl1

0.2 C/)

E-

0.0

_

_

_

_

_

8

_

cm 2 _

_

_

_

_

_

_

_

_

0.8 0.6

17

V cm 2

17

V c2

,E/N=30.0x10-

< 0.4 O

0.2

1.0

(C)

0.8

/N=50.0xl10

0.6 0.4 0.2

0.0

2.0

4.0

50.0

100.0

150.0

200.0 250.0 300.0

TI ME ()us) Fig. 22.

Digitized voltage waveforms obtained in CF, for three values of E/N. corresponding

to

(a)

electron attachment.

low, The

(b) firs-

moderate, segments

and

(c)

high

rates

of

the

waveforms

of were

digitized at 10 ns/channel. and Lhe second segments (occurring after the vertical dajhed lines) were digitized at 50 ns/channel.

53

We define the voltage ratio as eV(Te) + V_(t) + V+(t) V(T)

R

In the presence of voltage ratio R

significant

when t > T

(16)

orT

electron attachment and

ionization,

the

is given by

- r1d

R

- ad exp[ad]

(17)

(1 - exp[ad])

v

ild exp[r7d] - ad exp[ad] -

(lS)

(exp[rtd] - exp[ad])

When electron ionization is negligibly small, Eq. (17)

R

pd

v

1-

(19)

exp[-d](

or when electron attachment is negligible, Eq. (17)

R

R-

v

A.

further reduces to

reduces to

ad ad

(20)

1 - exp[-ad]

Pure Cases

We have performed an extensive series of measurements of ri/N and a/N 02,

kPa.

CF4, C 2 F6 , CF

8

and n-C 4 Fjo over the gas pressure range 0.03 < P < 100

The present measurements in 3

the previous

in

literature data.i '

34

02

are shown in Fig. 23 in comparison with Low energy electron attachment to

02

is

due to three body parent anion formation, which is very strongly dependent on gas

pressure.

changes

in

present

and

the

The measured

ri/N values are

consequently very

experimental

parameters and

the good

previous

measurements

in

this

gas

agreement

indicates

that

sensitive between the

to the

present

technique is capable of obtaining reliable measurements with good accuracy.

54

G 85-18989

-17ORNL-DW

02

z V

I-A

Y(0

DENS

7

n3

A

AIL-

GA0 NUBERG(9)

A

*.3 PRESENT 267 CHANINEta (196 )

LOU

o

2.7t 1.98

z•z-

CCIF 3

~ 1.96

U

L94 1.92

I

I

250

300

350

400

450

500

GAS TEMPERATURE T (K)

Fig. 48.

(E/N)e m versus T for C 2 F6 and OC1F,.

550

600

91

each value of T, the measurements in Figs. 45 and 46 (and similar measurements on the rest of the gases investigated) were fitted to the expression

(E/N)m = (E/N)'im + A/Nd

(24)

(where A is a constant) and the limiting value, (E/N)eim, of E/N was obtained. The values of (E/N)eim at various T are plotted in Figs. 47 and 48 for all six gases studied. These measurements indicate that the effect of increasing gas temperature on the hold-off voltage in opening switch applications will be minimal for gas temperatures T < 600 K for most of these gases. C2 F6 , the gas dielectric properties

increase slightly

respectively) with increasing T, while for slightly

1-C 3 F6

there are

larger decreases

n-C4 Fo (Z 8%) (and c-CF

and C3 F8 , (E/N)em decreases

6

in (E/N)Rm for

be

used

to

c-C4 F8

[?]) over this temperature range.

reductions in (E/N)pim are not beneficial can

(Z 3.6% and Z 1.0%,

(Z 2.6% and 3.0%. respectively) over this temperature range.

other hand,

they

Specifically, for OCIF 3 and

enhance

the

On the

(Z25%) and

Although these

for opening switch applicaitons,

conductivity

in

self-sustained

closing

switches, and thereby increasing the current switching efficiency of this type of switch.4 IV.

Relevance to Switching Experiments From

developed

the measurements given in Section III, several

gas

mixtures

characteristics shown in Fig. 2.

with

the

it is clear

optimized

w(E/N)

that we have and

ka(E/N)

These gas mixtures contain a low percentage

perfluorocarbon (0.1% to 30% in the total gas mixture) and either Ar14 .15 cr 20

CH,4

as a buffer gas.

The E/N dependence of w and the density normalized

electron attachment coefficient q/Na of gas mixtures containing C 2 F6 in Ar and C 2 F6 in Cl14 are given in Figs. 49a and 49b, respectively. (e.g.

the 2% C2 F 6 mixture

in

Fig.

49a and

the 10% C 2 F6

These measurements

mixture

in

Fig.

49b)

indicate that these mixtures have the near ideal characteristics required to optimize Jsw(E/N)

[and hence

Jsw(t)]

shown in

Fig.

2.

Small

scale

e-beam

sustained switching studies have been performed using these gas mixtures and 43 4 6 have shown that switch opening times > 50 ns are achievable.

92

ELECTRON ATTACHMENT COEFFICIENT 2 77/Na (10-17 cm ) 00 0

W

:

0

~Au, E

(3

cc,~ 00 00c

-

0

uj

ELECTRON ATTACHMENT COEFFICIENT 2 7 'Na (10-17 cm ) a,

(

(U

0

0D

Li

-D

0 Go

(i-s wo 9 01.)

m

,kIIDO13A

0

v

Iildal

NOK3313

0

N

4

93

i

0

0

@0

0-0

Ln

LL

bc

'u-

u

i

C4

C

U

m

(-WO V)IM*e A.LISN30 ±N3uuIfo

94

In

z o

w

a~ ~

u %%MOO-

C;.

Cm.L

U))

0 13.

w 00 en

(WO WLqO)/d A.LIALLSISMU

c

ILl

>b

95

4.1

CF/Ar Gas Mixtures

The current density JSW and discharge resistivity p measurements obtained as a

function of

colleagues several

E/N

in the e-beam switching experiments of

(Ref. 44, 45) are given

concentrations

of

C2 F6

in Figs. 50

in

Ar.

A

and 51,

strong

Schaeffer and

respectively.

negative

for

differential

conductivity is seen in the measurements given in Fig. 50, with the peaks in

JSW occurring at E/N z 2-3 x 10- 11 V cm2 (1 Td = 10-11 resistivity measurements given in Fig. 51 2.5 < E/N ( 5 x over

an

order

-

10 £7

of

indicate

2

V cm . the resistivity of

magnitude.

These

characteristics at decrease

low E/N

in conductivity

2

Similarly, the

).

that over

the E/N

indicate

that

increase

which occur during switch opening.

a

mixture

the best electrical conductivity

(i.e. during switch conduction), and

(i.e.

range

the discharge increases by

measurements

composed of approximately 2% C 2 F 6 in Ar has

V cm

the

in resistivity) at high E/N

largest values

The reduction in conductivity was found to

be primarily due to the onset of electron attachment at E/N = 2-3 x 10cm

2

(Fig. 49a) rather than due to the reduction in w at high E/N.

17

V

At low E/N,

the discharge is recombination dominated and the magnitude of the peak ii-jSW was found to be controlled by the maximum in the electron mobility 4(= w/E) at these E/N values (Fig. 5, Ref. 44).

Studies of the switching characteristics

of opening switches using a UV light source to provide a source of electrons in the discharge by photoionization of 45 results. 4.2

the

--As mixture have obtained similar

CF&/CH, Gas Mixtures

The transient electron beam current Ib' the switch discharge current ISW and the switch voltage VSW obtained in the e-beam switching measurements of Commisso and colleagues (Ref. 43) using a mixture composed of 1% C 2 F 6 in C- 4 are given in Fig. 52. Fig.

49b

give

rapid

These measurements indicate that the mixtures shown in switch

opening

upon

termination

of

the

discharge

sustaining source, with switch opening times an order of magnitude faster than those

for pure CH 4 .

This can be seen in Fig. 53 where

the switch voltage

obtained in pure N 2 , CH 4 and a 1% C 2 F 6 in CH4 gas mixture are compared (Ref. 43).

Even at the lower total gas pressure, the C 2 F6 /CH 4 gas mixture exhibits

a larger voltage transient with a much faster decay than pure CH 4. opening

time

in

porc

N,

is

also

orders

of

magnitude

slower

The switch

than

for

the

96

ORNL 87-7883

15

'

'

'

300

'

5otm 99% CH4 -1I% C2 F6 L= 1.5,H A = 500 cm2 R=2cm 10

-200 v

is

w

// I

I-

zwJ

/ /

>.

_

/>

W

5-

-00

// / D/

/

-

/

V(0) VC

b 00

0

Fig. 52.

0.5 TIME (O=sec)

The diffuse

discharge

current

and

Ib

,

switch

the voltage

current

1.0

ISW,

the

sustaining

transient developed across

e-beam

the switch

VSW during switch opening using a gas mixture of CH, containing 1% of C 2 F6 (Ref. 43).

97

300

i

i

'

,

1% C2 F6 -CH 4 (5 atM)

200

-

00% CH4

Cl

(10 otto)

.

I I

(/1

>I too-

1100%

-

-

Fig. 53.

-

N2

.

-

-

--

-(6otm}

40C 300 200 00 0 I00 CLOSURE (ns) TIME RELATIVE TO DIVERT SWITCH

Comparison of

switch voltage VSW for the switch opening times and

pure N2 . pure CH4 and a

(Ref. 43).

500

gP

mixture of CH4

containing

1% of C2 F6

98

0

b

LU

C

Le

0

Z

Li)

ul/ (z

j-0 -:

)'N

-..-

)

d

3'(

M

S

-(JO

99

C 2 F 6 /Ca 4 gas mixture.

The calculated discharge parameters (density normalized and E/P) in the C 2 F 6 /CH 4 gas mixture are shown in

electron mobility pN, TI/Na

These calculations show

Fig. 54 in comparison with the discharge current ISW.

(i.e. rapid increase in the

that the rapid decrease in the discharge current

discharge resistivity) in this mixture is due initially to the reduction in siN.

the

times,

later

and at

increase

rapid

at

in rj/t a

switch

the higher

voltages (i.e. high E/P values) causes a further rapid reduction in ISW* 4.3

Enhanced Discharge Current Characteristics

Although

the discussion given above has gas

diffuse

sustained

externally

using

the

indicated

an

as

discharges

feasibility of opening

switch

concept, several problems have been identified which may potentially limit the usefulness of these switches. e-beam

electrons

efficiency) and switch.

in the

the

Of primary concern are the energy losses by the gas

switching

lack of

(i.e.

timing

precise

reduced

control

of

current the

switching

opening of

the

Possible gas engineering solutions to these problems are outlined in

the following discussion. The

electron

source

function

S

in

an

sustained

e-beam

diffuse

gas

discharge is

S= where

(dE/dx) is

the

)JBW

(25)

average energy deposited

electron in the e-beam, JB

in

the gas

primary

is the flux of the e-beam, and W is the average

energy required to produce an electron-positive ion pair by electrons in the e-beam.

by each

the high energy

The W value of a gas is usually approximately twice

the value of the ionization onset energy, with the remaining electron energy being

used

processes

in in

vibrational the

gas.

(for molecular

The

excitation

gases) energy

and is

a

electronic excitation loss

process

in

the

discharge, but a sizeable fraction of the electronic excitation energy can be recovered (i.e., mixture with a

J

increased) d,,ring sw~itclh 1nduction by

low ionization onset gas additive.

If

the

sceding the gas ionization onset

energy is less than the energy of any metastable or resonant electronic states of the predominant gas then extra electrons can be generated in the discharge

100

1.00

c'J %C

0

E

2

0