oRESEARCH ON

MICROWAVE WINDOW MULTIPACTORAND ITS INHIBITION

"4

REPORT NUMBER 1

SCONTRACT NUMBER DA36-039-SC-90818DEPARTMENT OF THE ARMY

TASK NUMBER OST 76-10-318-28-03

C,

f7_

FIRST QUARTERLY PROGRESS REPORT1 July 1962 through 30 September 1962

1U. S. ARMY ELECTRONICS RESEARCH AND DEVELOPMENT LABORATORY

FORT MONMOUTH, NEW JERSEY

II ASTIA

EITEL-McCULOUGH, INC. " ' nSSan ear~los, California 1), 1T17

S "APR 15 1963

TISIA00v

Qualified requestors may obtain copiesof this report from the Armed ServicesTechnical Information Agency (ASTIA),Arlington Hall Station, Arlington 12,Virginia. ASTIA release to OTS is notauthorized.

1 RESEARCH ONMICROWAVE WINDOW MULTIPACTOR

AND ITS INHIBITION.

I REPORT NUMBER 1

SCONTRACT • DA36-039-SC-90818

- .Tk _U,_, _ _, OC T 7 =•61 310 2- - 3--.

1 MW QUARTERLY PROGRESS REPORT. Po,S1 July `- ---k-30 Sep 2

I Object of Research: (1) Attai" a morecomplete theoretical understandingof microwave tube window multipactor,its initiation and sustainment, andthe tube performance limitations itimposes; (2) Investigate techniquesto inhibit multipactor without de-grading tube performance.

-,aped--by: 0,•,y- /41 er'l(Oskar Heil

'v '4nald PreistiThis research is a part of Project DEFENDER, sponsoredby the Advanced Research Projects Agency, Departmentof Defense, under ARPA Order Number 318-62, ProjectCode Number 7300, and is conducted under the technicalguidance of the U. S. Army Electronics Research andDevelopment Laboratory, Fort Monmouth, New Jersey.

I

i

TABLE OF CONTENTS

Section Page

1 Purpose 1

2 Abstract 2

3 Publications, Lectures, Reportsand Conferences 3

4 Task A 8

4.1 Factual Data. Progress to Date 9

4.2 Conclusions 14

1 4.3 Program for Next Interval 14

5 Task B 16

1 5.1 Factual Data. Progress to Date 17

1 5.2 Conclusions 33

5.3 Program for Next Interval 34

6 Identification of Key TechnicalPersonnel 34

7 Abstract Card 48

IIIIII

iiiII

1. PURPOSE

The broad purpose of the study is to provide a deeper under-standing of multipactor effects at waveguide windows used withhigh power microwave tubes which will lead to practical methodsfor preventing or eliminating multipactor, thereby raising thepower handling capacity of windows. Throughout the investi-gation emphasis will be placed on a scientific approach andunderstanding of the phenomena involved so that solutions maybe obtained in the most general terms and can therefore beexpected to be applicable over a wide range of conditions.

At present, the method of eliminating multipactors is theapplication of an evaporated film of titanium in order toreduce the secondary emission coefficient of the window andadjacent surfaces, using techniques which have already beendeveloped to a certain point as a result of the company-1 sponsored research at Eitel-McCullough on cylindrical windowsprior to the inception of the present contract. This workhas been summarized in two published articles . In the courseof this work klystrons were made with coated output windowsand adjacent metal parts. These windows without coatingsexhibited severe multipactor, and with coatings exhibitedno multipactor. We believe this is the first time this hasbeen done. At the time of the inception of the contract ithad not been done anywhere else, to our knowledge. It seemsreasonable to suppose that these techniques would be equallyeffective if applied to waveguide windows at which multipactordischarges are occurring. Another possible method would bethat originally proposed by 0. Heil in 1960 which is the directapplication of titanium suboxides, rather than titanium metal.

An alternate method of eliminating multipactor is provided byphase or space defocusing of the electron cloud near thewindow. If the electric fields in the window region areproperly shaped, or if the window configuration is suitablydesigned, the secondary electrons produced will not beentrapped in a resonant field. This method does not, inprinciple, require a reduced secondary emission coefficientat the window surface. However, these two approaches aremutually beneficial, and together should provide a veryeffective means of suppressing multipactor.

Because these approaches are quite different in application,the work has been divided into the following two tasks, withseparate investigators:I

l"on the Heating of Output Windows of Microwave Tubes by ElectronBombardment," by D. H. Preist and R. C. Talcott, IRE Trans.PGED Vol. ED-8 No. 4, July 1961, and "The Effects of TitaniumFilms on Secondary Electron Emission Phenomena in ResonantCavities and at Dielectric Surfaces," by R. C. Talcott, IRETrans. on Electron Devices, Vol. ED-9, No. 5, Sept. 1962, pp 405-410.

S~- 1-

TASK A: An Experimental Study of Electron BombardmentPhenomena at the Output RF Windows of HighPower Microwave Tubes.

Phase 1. Experimental and analytical study ofmultipactor effects at waveguide windowsunder high power conditions.

Phase 2. Development and application of evapo-ration coatings and techniques appliedto the window and surrounding metalparts in order to reduce secondaryemission coefficient to less than unity.

TASK B: Study of the Inner-Window Surface and Con-figurations Affecting Power Handling Capabilitiesof High Power Microwave Tubes.

Phase 1. Analysis of various means of obtainingspace and phase defocusing of electrons

Sby shaping the fields and window surfaces.

Phase 2. Study of materials and coatings in con-junction with shaped fields to developwindows capable of handling higherpowers without multipactor.

2. ABSTRACT

TASK A: An Experimental Study of Electron BombardmentPhenomena at the Output RF Windows of High

\Power Microwave Tubes

This s-qi__u*Gu.Im.. ---report on the study andelimination of multipactor discharges at waveguidewindows of high power microwave tubes • I_ H, "O

dkI progress on the two major phases, which are:

|1) Xn analytical and experimental study of multi-pactor effects at ywveguide windows using afrequency of 2850/2 under both CW andpulse conditions, and

(2) i#he development and application of evaporativeI 'coatings and techniques to waveguide windowsand the testing of these coated windows undermultipactor conditions to evaluate the effectsof the coatings in preventing multipactor.

So far.,progress on Phase 1 has consisted of obtaining, 'I0setting up and checking out the # equipment required ..

-2-

I

to provide necessary power and instrumentationand also the design and partial fabrication ofthe window box itself. On Phase 2, progress hasbeen confined to a study of various methods ofimproving the control of evaporative coatingsand to devising methods of applying such coatings

to waveguide windows.

jWýý do h */nn *;ýLow 16rag e de n a e h, er -- Wro es

'The gliding type tnultipactor with practically nophase restrictions is discussed. Methods ofreducing or preventing multipactors are described.These are concerned with the lowering of the rateof secondary emission and by phase and space de-focusingt. Special emphasis is given to the electronaccelerating forces resulting from field inhomo-geneities including those of standing waves. Twomethods for instantaneous multipactor detection aregiven. Expected effects from a grooved windowsurface are discussed. Titanium monoxide sputteringexperiments using Hg ions to get low secondaryemission coatings are discussed, and their resultsare described.

3. PUBLICATIONS, LECTURES.EPORTS. AND CONFERENCES

TASK A:

3.1 Publications

"The Effects of Titanium Films on Secondary EmissionPhenomena in Resonant Cavities and at DielectricSurfaces," by Ruth Carlson Talcott, Trans. IRE, PGED,Vol. ED-9, September 1962.

1 3.2 Lectures

"Multipactor Motions in Microwave Tubes," by D. H.Preist, paper read at the International Congress onMicrowave Tubes, the Hague, Holland, September 1962.

3.3 Conferences Held

1. At Stanford University, August 30. 1962. Thosepresent: Don Preist, Ruth C. Talcott. JohnSoderatrum, from Eimac and John JasbergandDr. Pedro Szenti of Stanford.

The Stanford experiments on windows at S-band incavity resonators were described and discussed.

S- 3-

IThese experiments were done with a higher windowRF electric field strength than in the past onaccount of the lower loss of the cavity resonatorthan that of the ring resonator. They have shownrepeatable damage to windows which appears as aninternal physical change to the window materialin the form of black dots or streaks. The powerlevel at which this damage is observed appears tobe a function of the gas pressure and becomeshigher as the pressure is reduced. This work hasbeen done with O-ring seals and a vacuum between10-4 and 2 x 10-5 Torr but more work is plannedwith brazed joints and a better vacuum. Thepossibility that the damage could be caused bymultipactor discharges was discussed but it wasconcluded that there was insufficient evidenceto prove the existence of multipactor.

2. At The Hague, Holland between Preist of Eimac andDr. G. Schaffer of the Hamburg Linear AcceleratorProject (DESY), September 7, 1962.

Dr. Schaffer has some very interesting colorphotographs of discharges on the vacuum side of"a window feeding a linear accelerator cavity at"a frequency around 500 megacycles. The arrange-ment is shown in Figure 1, from which it can beseen that che window is a round disc waveguidewindow. When the power level was raised highenough (well beyond the designed level for theaccelerator) bright spots began to appear andstreaks of light on the vacuum side. When it wasraised further, the window would crack. This wasrepeated several times. Examination of the windowshowed that the bright spots and streaks wereaEsociated with permanent damage to the interiorof the window surface very similar to that observedat Stanford. Once again it was noted that the gas

-* pressure on the output (vacuum) side of the windowhad an important bearing on the power level at whichfailure would occur. A calculation showed thatthe damage occurred at just above the criticalfield strength which we have previously shownto exist for a single surface multipactor (1), andit was agreed that in this case there were reason-able grounds for thinking that the damage resultedfrom a single surface multipactor discharge similarto the ones we have found at Eimac in the past.

4•~- 4-

!

I

rectangular w/gSection of

linearaccelerator

(evacuated)

Klystron

Amplifier

circul Viewingdisc Window

window(alumina)

Schematic Diagram of Dr. Schaffer's Set-Up

at Hamburg (DESY)

Fig. 1

[ " -5--

i3. At EMI. Haves, Middlesex. England between Preist

of Eimac and Dr. Kreuchen. Mr. Dixon and Mr. Barnesof EMI, September 12, 1962.

Their work with quartz-tipped glass conical wave-guide windows was discussed. This work had beendescribed in a paper by Mr. M. J. Smith of EMI atthe Microwave Tube Congress at The Hague, Holland,September 5, 1962. Briefly, these windows appearto be capable of passing several megawatts and atleast 50 kilowatts of average power at S-band inring resonators. They exhibit evidence of singlesurface multipactor of the kind described in refer-ence (1) at times, but this multipactor is notdestructive and disappears when the power level israised. EMI believe that this is due to the factthat at the higher power levels the electronvelocities are such that 6 for quartz would beless than unity, preventing multiplication ofsecondary electrons. Also, the shape of thewindow and its surrounding transition betweencircular and rectangular waveguide is such asto produce a force sweeping electrons out of thewindow and, therefore, also limiting multipactoraction.

The method of analysis of the single surfacemultipactor given in reference (1) has been usedby EMI, and they find it adequate to explainquantitatively the effects they observe. Theyhave calculated the magnitude of the electricfield gradient force at the window, and showthat this force will drive secondary electronsemitted from the window back to the window asrequired to sustain the multipactor 2 .

TASK B:

Conferences Attended by Oskar Heil

The American Physical Society Conference in Seattle,Washington, was attended. A paper presented by H. Seiwatzand C. M. Groom on the nondestructive breakdown on dielectricsI led to the idea for window protection as described in thetext.

2EMI "Progress Report No. 2 on Research Proj. RP 5-9 -

E.M.I. Research Labs. Ltd. (Gr. Brit.), November 1961,"by M. J. Smith and A. Bamford.

!. -6-

LBOTH TASKS:

Conferences Held

1. At Eitel-McCullough, San Carlos, California withLt. Col. W. B. Lindsay of ARPA, August 23, 1962.

The general philosophy, purposes and orientation ofthe program were discussed and plans reviewed.

2. At Eitel-McCullough, San Carlos. California withMr. Louis Heynick of the U. S. Army ElectronicsResearch and Development Laboratory, July 2 andJuly 19, 1962.

Plans and progress were reviewed.

7

I

1

iI

II[-7

4. TASK A

AN EXPERIMENTAL STUDY OF ELECTRON BOMBARDMENT PHENOMENA

AT THE OUTPUT RF WINDOWS OF HIGH POWER MICROWAVE TUBES.

Prepared

VDonald H. Preist

I

I 8-

I4.1 FACTUAL DATA. PROGRESS TO DATE

4.1.1 Design of Window Box and RF Plumbing Run andAcauisition or Fabrication of These Items

This arrangement is shown schematically in Figure2. The RF driver will be either an Eimac 4KM70SKCW klystron rated at 20 KW CW output or an EimacX632G pulse klystron rated at 10 megawatts peak and10 kilowatts average power output. The frequencyin both cases will be 2850 Mc. Following the driveris the power divider and waterload, which will ensurethat the klystron is loaded adequately during alladjustments of frequency, etc., which might otherwiseproduce a large and damaging electric field strengthat the klystron output window. Following this aretwo directional couplers in series to give simul-taneous measurements of forward and backward powerat high power levels.

The window box itself, which is shown in detail inFigure 3, will be, for the first experiment, aresonant cylindrical cavity containing the ceramicwindow and operating in the TE1 1 1 mode. Powerwill be fed through a circular coupling iris whichis adjusted to be overcoupled in the condition whereonly dielectric losses and copper losses are present,and to be somewhere near a match if and when multi-pactor discharges occur. The window box is de-mountable and water-cooled, and equipped with ahole for insertion of an electric field probe. Also,there are holes in the end face on the air side ofthe window for viewing and for insertion of contactpyrometers for measuring window temperature immedi-ately after turning off the power. The window boxis designed so that the window and the minimum numberof other parts attached to the window are expendablein case of failure. The only vacuum seals are thebraze between the window and the surrounding coppercylinder and a braze betweee- this cylinder and thesealing ring, and a final heliarc take-apart jointbetween the two sealing rings. The RF current willflow through compression seals made by squeezing thethinned edges of the copper cylinder against the endflanges by means of the bolts shown in the figure.The cavity is equipped with a connection to a getter-ion pump gauge.

A subsequent model of the window box will be fittedwith a viewing window through which the vacuum side

I9-9

N N

-10-

HOLE FOR ELECTRICRUBBER "0" RINGS FIELD PROBE

// VIEWING

<1/1111HO•LES

HELIARC SEAL CERAMICWINDOW

R.F. i

COMPRESSION

RECT.- , WAVEGUIDE

TO GETTER-ION

S.S. CLAMP RING WATER CONNECTION

I

""I "WINDOW BOX" ASSEMBLY

F-11- 3

iof the window may be observed during operation.This should be of particular interest during thepulse tests.

Parts for the window box are being fabricated andare expected to be assembled shortly.

There has been considerable delay in obtaining someof the components in the RF plumbing run in Figure2. However, all these parts should be availableearly in the next quarter.

4.1.2 Fabrication and Testing of Driver Klystron

A 4KM70SK klystron has been manufactured and testedin our High Power Microwave Division. Specialattention has been paid to operation of the tubearound 2850 megacycles, which is the frequency wepropose to use for testing windows. In particular,a thorough examination of the ghost-mode pattern ofthe klystron window has been made to ensure that wewill not damage it by operating at ghost-mode fre-quencies, if these exist. In the course of thesetests the whole equipment, including power supplies,driver, etc., has been set up and checked out toensure reliable and, hopefully, trouble-free opera-tion during the testing of windows. It is expectedthat these tests will be finished early in the nextquarter.

4.1.3 Work on Control of Evaporated Coatings on Windows

During this program we propose to use evaporatedtitanium coatings on both the window and surround-ing metal parts as a means of controlling orinhibiting multipactor discharges when these arefound to exist without such coatings. This approachhas been found to be very successful in previousEimac-sponsored work carried on before the inceptionof the present contract and has been rath r completelydescribed in a recently published article . Briefly,the method is to evaporate titanium metal from ahot filament onto the target in a bell jar in a goodvacuum. The purpose of this is to obtain, if possible,a pure titanium surface having the secondary emissioncharacteristic of titanium, particularly the lowvalue of 6 max. The previous work showed that this

" "The Effects of Titanium Films on Secondary Electron EmissionPhenomena in Resonant Cavities and at Dielectric Surfaces," byRuth Carlson Talcott, Trans. IRE PGED, Vol. ED-9 No. 5,

SSeptember 1962.

- 12 -

could be achieved, apparently in a reproduciblemanner, both on metal surfaces and ceramic sur-faces. The thickness of the films used on theceramic was on the order of 100langstrom unitsand at the frequency used (650 megacycles), thisthickness of film gave only a small and insignifi-cant increase in the apparent dielectric losses ofthe windows. Such coated surfaces could be bakedout in vacuum in normal fashion without deteriora-tion and have been successfully used in a number ofhigh power CW klystrons having cylindrical ceramicwindows inside the output cavities. This work andits results have also been described in reference(1). It is reasonable to suppose that similarcoatingA will have similar beneficial resultswhen applied to waveguide windows suffering frommultipactor effects.

A study has been made of various possible methodsof improving the control of coatings. In the previouswork we were concerned with cylindrical windows whichhave the advantage of two parallel ends, whichsimplifies the measurement of the coating resistances.This measurement of the coating resistance can bemade during and after the coating process and can beused as a means of controlling the thickness of thecoatings. With round disc windows we are, of course,unable to used this method. In addition, we were notsatisfied with the method as it required somewhatJ undue skill on the part of the operator.We have considered the use of three basic methods

of control, which may be described as:

1. Metering the amount of material evaporated.

2. Control by optical measurements on a coatingdeposited on a transparent material simul-taneously with the window.

3. Control by secondary emission measurementduring coating.

The first method, while simple in concept, isvery difficult to carry out in practice becausewe are dealing with exceedingly small amounts ofI" deposited material. It should be remembered thatthe required thickness of material on the windowdielectric surface is on the order of 100 A only.

1

S- 13 -

Such films are nearly transparent and are alsotoo thin to give interference or color patterns.

The second method, using optical absorption orreflection, has some advantages and severaldisadvantages; one is that a separate glasstarget has to be used in the evaporating set-upand the evaporating filament must be designed tocoat both this and the window. Optical measure-ments have to be made of the transmission orreflection of light from the glass target duringthe coating process. Once again, because thefilms are so thin, it would be difficult to makethe measurements, and the films are sensitive tosuch things as the rate of evaporation and thequality of the vacuum.

The third method appears to be the most promisingat present. This method is essentially to measurethe value of 6, the secondary emission coefficientof the surface during the evaporation process sothat the operator can tell when 6 has fallen belowunity on account of the coating. A pulse techniqueis essential to prevent significant static chargesbeing built up by the electron beam, which isnecessary to measure 6, and to prevent the crackingof hydrocarbons at the window with a resultantcarbonaceous deposit. we are planning to pursuethis approach experimentally during the nextperiod.

Having established a satisfactory method of control,we plan to apply titanium coatings to the windowsto be tested in the window box shown in Figure 3to determine their effect on multipactors. We donot propose to limit our studies to titanium alone,but will be in a position to try other elementswhich appear to be favorable as a result of analyti-cal study.

4.2 CONCLUSIONS

At this stage no conclusions have been reached about eitherof the phases of this task.

4.3 PROGRAM FOR NEXT INTERVAL

During this period we expect to test under CW conditionsa sufficient number of windows to determine the conditionsunder which multipactor occurs and to make significantquantitiative measurements on the power loss in the multi-

Ii -14-

I pactor discharge as a function of window electric fieldstrength.

On coating control, we plan to continue with experimentsbased on the secondary emission measurement techniquedescribed earlier, and if this is successful, we shallattempt to apply evaporative coatings to round discwindows and test these windows in the window box describedherein.

i

I

I1• -15 -

S5. TASK B

S TUDY OF THE INNER-WINDOW SURFACE AND CONFIGURATIONS

AFFECTING POWER HANDLING CAPABILITIES OF HIGH POWER

MICROWAVE TUBES.

Prepared by

I Oskar Hell

Ii - 16-

5.1 FACTUAL DATA, PROGRESS TO DATE.

5.1.1 Introduction

There exists the possibility that many highpower window failures by cracking or arcingwithin the material originate from a surfacemultipactor. In the case of cracking, it isthe heating by the multipactor which causesthermal stress leading to fracture. In thecase of arcing, the heating produces gas andthe resulting ion space charge compensationincreases multipactoring, leading locally toa runaway condition ending in internal windowarcing. However, there is also the possibilitythat arcing is independent from multipactoringand is caused within the window material itself.In this case arcing may be favored by the windowheating caused by multipactoring.

The aim of this study is to fully understandthe effects of field configurations and materialson multipactor to control and prevent it.

5.1.2 Multipactors

The multipactor appears in many varieties: thedouble surface and the single surface, the loworder and the higher order multipactor (accord-ing to the number of cycles between impacts),on conducting or insulating surfaces. Forinsulating materials the electric field neednot be perpendicular to the surface, but canhave any angle, even to the extent of beingparallel to the surface. A multipactor occurringwhen the electric field is not normal to the planeof the window is known as a "gliding" multipactor.This gliding type of multipactor can exist betweentwo sur-faces or on a single surface. All singlesurface multipactors, the gliding type as well asthe perpendicular, require a force driving theelectrons back to the surface. This force cancome from an electric field, or a magnetic field,or can be a force resulting from field inhomogene-ities. This restoring force can be rather weakfor the gliding multipactor, but must be of thesame magnitude as the ac field in the perpendicularmultipactor.

- 17 -

II

U

The gliding multipactor is the most common andeasiebt to start on windows for the followingreasons:

1. Since the electric field of an electro-magnetic wave stands perpendicular to thedirection of energy propagation, and sincethis is also the direction of the usualwaveguide window surface, the gliding multi-pactor is most common on window surface.

2. The gliding multipactor is the most stableand persistent multipactor because it haspractically no phase restriction for themultiplying impact of electrons. It isalways possible to have a free accelerationof emitted secondary electrons from the acfield no matter in which direction theelectric field points. This is not so ona surface perpendicular to the electricfield where half of the electrons are inter-cepted by the surface before gaining energyfor multiplication, because half of the timethe field points towards the surface.

3. The restoring force, the force driving theelectrons back to the surface, is not criticaland can have a wide range of values becausethe phase of arri.val on the surface is unim-portant. Only a force which is too strong andleaves the electron no time to pick up energyof the ac field would kill the multipactor.

4. Sliding incidence of the electrons on thesurface increases the rate of secondariesand moves the voltage of the second cross-over-point to higher values. On high powerpulses the electrons oscillating in front ofthe window reach velocities higher than thesecond crossover, and by increasing the voltageof the second crossover, due to grazing inci-dence, the multipactor is more likely to occur.

5.1.3 Reduction and Prevention of Multipactors

There are three ways to reduce or prevent multi-pactors:

1. Lower the secondary emission coefficient ofthe surfaces (coating)as described below.

I

22. Space defocusing of electrons, described below.

3. Phase defocusing of electrons as described in5.1.7.

1. Secondary emission coefficients of insulatingmaterials as a rule are higher than those ofconductors. An explanation for this can befound in the microfield of the surface. Thepotential of a good insulator is not uniform,but is in the form of a random distribution.The local fields favoring electron emission domore than the fields counteracting emission.This favored emission can be reduced by a smallamount of surface conductivity. There existno low secondary emission materials withoutan electric conductivity. On window surfaceswe cannot tolerate any appreciable conductivitybecause of ohmic losses. The window coatingshould have high specific resistance. Thethermal coefficient of resistivity should nothave a highly negative value, as found in semi-conductors, to prevent runaway condition for thewindow temperature. The coating must be thinand, therefore, thermally and chemically stable.Titanium suboxides seem to fulfill these condi-tions 4 . This material was proposed and stronglyrecommended at Eimac by the author as a windowcoating for secondary emission prevention inlate 1960 as a result of the data in reference(4). It is very likely that the 100 angstromthick titanium coatings used by Ruth Talcottand Don Preist might easily have changed intotitanium suboxides due to the chemical activityof this metal. This subject will be investi-gated further during the course of this study.Titanium monoxide is similar in property to ametal, having a positive temperature coefficirntof resistivity. The resistivity is 2.8 x I0-(0cm). (TiC and TiN have very similar prop-perties.) With higher degrees of oxidation thetemperature coefficient becomes gradually nega-tive and the resistivity goes up. The temper-ature coefficient for Ti 2 0 3 at room temperatureis about -8 x 10-3 per degree C with a resistivity

4 L. E. Hollander and P. L. Castro, "Anisotropic Conduction inNon-Stoichimoetric Rutile (Ti02 )" in Sept. 1960, Phys. Rev.,Vol. 119, Pg. 1882.

-19-

of about 3 (acm). The small negative temper-

ature coefficient still promises stabilityand the resistance value is not too low andwould permit coating layers of 10-5 cm thick.After sputtering with TiO and sending thecoating through a vacuum bakeout the approxi-mate composition of the materials is Ti 2O3.

2. Space defocusing of electrons in order toprevent multipactor may be achieved by electricdc fields pushing electrons away from thewindow surface. We have, however, no freechoice for the potential of the insulatingsurface. It has the tendency to become posi-tively charged with a secondary emission ratiogreater than one, and attracts electrons. Wehave to use other forces for moving electronsaway from windows such as an inhomogeneousconstant magnetic field or the "disquiestforces" making the electrons spin or oscillatemore violently near the window than at greaterdistances. In the latter case, any inhomogeneityof the high frequency electric field pusheselectrons into the regions of smaller fieldstrength. The dc energy gained between twopoints on the electron trajectory equals thedifference of the two peak oscillatory energiesof the electron at these two points. We candiscriminate three different inhomogeneitiesof the electric field. The resulting threetypes of inhomogeneity pumping of electronsare illustrated in Figure 4 with correspondingapplications to windows.

a. Inhomogeneity in the direction of diverginglines of force.

The cyclical acceleration away from the strongfield is always greater than the accelerationbackwards, resulting in a net outward electronacceleration. For example, a concentric linewindow drives electrons from the inner .othe outer conductor. A slight conical windowshape lifts electrons from the window surfaceon the vacuum side.

b. Inhomogeneity perpendicular to the curvedlines of force.

Centrifugal force makes electrons overshootthe lines of force radially, resulting in

1-~- 20 -

I

I

C)

rrWI

Fig. 4 Inhomogeneity Acceleration of Electrons

- 21-

electron acceleration. A waveguide taperedin height with a window near the narrow endgives the electric field lines the requiredcurvature to accelerate electrons away fromthe window.

c. Inhomogeneity perpendicular to the straightlines of force (standing wave).

An inhomogeneous electric field with straightlines of force is obtained dynamically in astanding wave, for instance in a half-wavesection of a waveguide or in a cylindricalresonance cavity. Also, in this inhomo-geneous field electrons move toward thequiet area. The motion comes from magneticdeflections in the magnetic, high-frequencyfield. Electrons have their peak velocitywhen the electric field is zero. At thismoment the magnetic field is at its peakvalue, which results in an outward de-flecting force. When reaching the resonatorwall the radial velocity of the electronequals the peak oscillating velocity, whichit had when starting its motion.

This electron acceleration exists wheneverthere is a standing wave energy. For instance,the propagation of a wave in a waveguide canbe interpreted as a superposition of twosymmetrical planar waves moving at an angleto the waveguide axis, which is identicalto a propagating wave in the direction ofthe axis. This standing wave will accelerateelectrons according to its electric fieldinhomogeneity, which in the TE 10 mode isa motion away from the center to the sidewalls of the waveguide.

The electron accelerations obtained from the threekinds of inhomogeneities can, of course, appearsimultaneously and superimpose their effects. Forinstance, a tapered waveguide (came b) can cntainsome longitudinal standing wave energy (case c) ofthe proper phase location and the two effects combine.The acceleration due to a standing wave can also beused to reduce multipactor on flat disc windows onconcentric lines. A slight standing wave, properlypositioned, in conjunction with the radial acceler-ation of the concentric line fields, will drainthe electrons away from the flat window. Alter-natively, a wrong phase location of some standing

-22-

wave energy near the window can drive electronsto the window surface and favor multipactoring.

The electron accelerating field inhomogeneitiesneed not be of the same frequency as the workingfrequency. Also, a different mode or plane ofI polarization can be used to clear the window ofelectrons. All these methods have one thing incommon, they increase the specific dielectric loadon the window. The price paid for the liberationof possible multipactors is higher dielectric lossescaused by the purposefully added field inhomo-geneities. The inhomogeneity forces reverse theirdirection when going from free to bound electrons.The electron is considered bound when its naturaloscillation or rotation frequency is higher thanthe frequency applied. This would take place forelectrons under the influence of a constant magneticfield of greater value than the gyromagnetic reso-nance field. Such fields are rather high forpractical application.

5.1.4 Phase Dependent dc Velocity for Fast MultipactorDetection and Window Protection

Until now we were concerned with dc velocitieswhich were obtained by acceleration in either dcfields or in ac field inhomogeneities. In addition,electrons have certain dc velocities in a directionparallel to the window, superimposed on theiroscillating motion, which depend on the phaseangle at which they are first exposed to the influ-ence of the high frequency field. Electrons releasedat the peak value of the field have no dc velocity,except in the direction of the axis of the guide,Figure 5a. Those released at zero field have thehighest dc velocities in a direction parallel tothe window. The velocities of these electrons,Figure 5b, equal the peak ac velocities of theundisturbed electron, Figure 5a. The dc velocityis proportional to -cosine-of the phase angle ofthe release time. Figure 5 illustrates how this dcvelocity of electrons can be utilized for detectionof a multipactor or any other kind of discharge onthe window surface within a time of only a few highfrequency cycles. A slot (c) in the waveguide walldoes not interfere with its high frequency propertysince it is not crossed by any high frequency current,but allows the fast electrons to strike the electrode"(d). This signal on (d) can be amplified and usedto either turn the power klystron off or a windowprotecting auxiliary frequency on, as discussed

I2Io~- 23 -

L

Fig. 5 D.C. Velocity of Electron and Multipactor Detection

1 24

I'.-

i

below. It is known and was described in a paperby H. Seiwatz and C. M. Groom at the 1962 AmericanPhysical Society Conference in Seattle, Washington,that nondestructive dc breakthrough tests ondielectrics can be made by early interruption ofthe beginning discharge. It should therefore bepossible with the above described method to protecthigh frequency windows from destruction.

5.1.5 Multipactor or Arc Detection by Rotation-of thePlane of Polarization

In addition to the fast multipactor detection des-cribed above, which utilizes the fast escapingelectrons, another detection method is possible.It is based on the rotation of the plane of oscil-lation of electrons in a weak magnetic field. Ifa magnetic field in an axial direction of a circularwaveguide is arranged near the window, the multipactorelectrons or any other free electrons will rotatethe direction of their oscillation and thereby coupleenergy into a wave with a perpendicular plane ofpolarization. This can be detected with a suitableprobe or coupling loop. The energy on the detectoris directly proportional to the number of freeelectrons near the window.

5.1.6 Hard X-rays Near Power Windows

The highest velocity of electron (b) in Figure 5is twice that of electroil 1a) and it has, therefore,four times the energy. If it would experience, atthe moment of its highest energy, a reflection with-I out energy loss at the waveguide wall, the phase ofthe high frequency field is such that an additionalacceleration would bring its highest velocity tofour times and its energy to sixteen times the valueof electron (a).

This process is not very probable, but possible,and can explain the observed hard X-rays nearwindows. Phase dependent dc velocities tend tospread electrons away from the window surface and,as we have shown above, additional dc velocitiescan be produced by electron acceleration in fieldinhomogeneities, which can help prevent multipactor

by space defocusing. We come now to the third methodfor multipactor prevention by phase defocusing.

Ii- -25 -

5.1.7. Phase Defocusing

If electrons arrive at the wrong phase at thesecondary emitting surface the multipactor can-not build up with exception of the gliding type,as explained above. Phase defocusing can beobtained by shortening or lengthening the transi-tions of electrons between impacts. Shorteningis to be preferred because lengthening can leadto a higher order multipactor and is therefore notso safe. The arrangement of an open metal venetianblind, touching the inner surface of the window,with blades running perpendicular to the electriclines of force will kill any multipactor if thedistance between blades is smaller than the distancerequired for the lowest order two-surface multipac-tor. The electrons do not reach their full velocityand the secondaries are released at a moment whenthe field brings them right back to the emittingsurface. The venetian blinds can be consideredI as nontransparent for electrons because under thepresence of the high frequency field practicallyno electron can pass through, whereas the electro-magnetic wave passes almost undisturbed. For areasonable ratio of blade thickness to bladedistance (about 5), the resulting dielectricconstant of the venetian blind is only 1/5. Theblades could be slightly embedded in the dielectricwindow material for window cooling. Hollow bladescould have internal liquid cooling. A major dis-advantage, besides its complexity, is the follow-ing: The blades would have to follow ratherprecisely the equipotential surfaces, or excessivelength currents, and possible arcing on the bladeedges might result.

A compromise solution for the venetian blind ideahas been prepared for experimental testing. Insteadof metal blinds, the surface of the dielectridmaterial has v-shaped parallel grooves runningapproximately perpendicular to the electric field.The field configurations between the grooves andinside the dielectric have been plotted using aresistance network for a dielectric constant of 8and 4, and for groove angles of 600, 900 and 1200.A dielectric constant of 8 is close to that ofalumina, and 4 close to that of silica, the twomaterials we intend to test. The resulting fieldsare illustrated by equipotential surfaces in Figure6 for alumina, and in Figure 7 for silica. The

-26

IDielectric Vacuum

a (900);~ b (1200).

-27-

Dielectric Vcu

moouIWOM0-

aig ___edramge

(6 0 ) b________A r o u nd__________A

l u m in a )

F28

I

strongest fields exist near the bottom of thegrooves where the dielectric loading of thewindow material will be heaviest. For smallelectron amplitudes the field inhomogeneitywill drive the electrons out of the grooves,thereby destroying the multipactor throughspace and phase defocusing. In the case ofelectron amplitudes, which are large when com-pared to the groove width, a kind of periodicfield focusing, produced by the periodic dis-tortion of the field by the teeth, will repelelectrons away from the surface. The fieldstrength variations along the surface was plottedin Figure 8. In addition, the field variationat the bottoms of the grooves is shown and givesa measure for the nonuniformitg of dielectricloaing. The curves are for 90 grooves and adielectric constant of 8. For an electronoscillating across the crests of the ridges,the high frequency field strength variation isabout Z 30%. Because of the rather high frequencyof this disturbance the energy in this jitteroscillation remains low and therefore the repellingaction on the electrons stays small. The presenceof the grooves prevents the gliding type of multi-pactor because there exists no tangential incidenceof electrons. Electrons can strike only near thecrests of the ridges. The valleys should be freeof any multipactors because of lack of space.Therefore, if a low secondary emission coatingis applied to the window surface it need only coverthe crests of the ridges in the form of narrowstrips. This is done either by evaporating fromtwo sides at a shallow angle, which leaves thevalleys with wires. The narrow stripes of coatinglack coherence in the electric field direction,which results in less ohmic losses for a givencoating thickness or in thicker coatings for agiven amount of ohmic loss. Another advantageof the grooving is the following: A possiblesmall arc across the window is prevented fromgrowing larger because the anchoring points ofthe arc cannot freely travel across the surface.They will be anchored to the bottom of the groovesdue to the field strength variation. From this itseems that there are quite a few favorable condi'-tions connected with the grooving of the windowsurface.

5.1.8 Manufacturing of Grooved Windows

Grooved alumina windows were made by Western Gold,

f - 29-

I

0

to0

-,4S4-)

0

"0 "0C4)

.. 0

cr E

0H4.)

r4 )

WW

3I

S~- 30 -

I

Belmont, California in close collaboration withEimac. A specific amount of alumina powder waspressed between grooved hardened and polishedsteel plates. The outer cylindrical edge wasground to size after firing but the groovedsurfaces did not require any finishing. Thewindows varied less than 0.33% in weight. Figure9 shows a steel tool and a finished window disc.Discs with 600, 900 and 1200 groove angles havebeen made. The finished groove width is 1.25 mm,the disc diameter 76 mm (3") and the thickness3.5 mm. Tests with and without coatings will bemade using linear accelerator test equipment atStanford University. Since these windows willhave a vacuum on both sides, the grooving willbe performed on both sides.

Silica grooved windows will also be made and tested.This grooving will be performed by grinding thesurface with diamond wheels followed by fire polish-ing.

5.1.9 Window Coating

Window coating was performed on test pieces bysputtering titanium-monoxide with mercury ions.A glass bell jar was pumped with a mecury diffusionpump and the mercury pressure was kept at a constantvalue by holding a trap between pump and bell jarat an exact temperature of 18 0 C. A plasma wasgenerated by a high frequency coil wound aroundthe bell jar, which was energized by a 27 Mcoscillator of several hundred watts. The ionbombarding velocity was 1500 volts. During thefirst experiments the connecting wires to thetitanium monoxide and to the target were hiddeninside graphite tubing to prevent the sputteringof impurities from the wires themselves. Graphitewas used because it is supposed to sputter verylittle. However, some chemi-•puttering apparentlytook place and the carbon compound seems to havereacted with the freshly depositied TO to form acarbon containing titanium compound, as a resistivelayer. In the next experiments graphite was replacedwith quartz glass as a shielding material and theresults became more consistent. Relatively thickcoatings Vere put on alumina ceramics and on silica.The temperature coefficient of the resistivity wastaken as an indication of the chemical compositionof the coating. This varies in a known manner

-31-

[-

D4

0

0

X.

44

32)

I

continuously with the degree of oxidization5 ' 6 .In a series of experiments the distance betweenthe titanium monoxide and the ceramic discs wasvaried and the sputtering rate was kept constant.The targets at greater distance showed systemati-cally a higher degree of oxidization. Becausethe coverage rate becomes smaller with distance,more oxygen or other impurities are built intothe coating and the observations indicate that thequality of vacuum is not sufficient for reproduc-ible coatings. However, the bell jar did not gothrough a bakeout, which can explain the results.

An additional titanium evaporator unit was builtto go between the mercury pump and the bell jar.With the heavy ionization produced by the highfrequency, impurities should be pumped away withhigh speed.

Coatings in an actual tube have to go through abakeout process. Several samples of coatings,varying greatly in resistance value and also intemperature coefficient, were sealed in a glassvacuum envelope and submitted to a one-hour bake-out. The resistance values changed but the temper-ature coefficient of resistivity was practicallythe same for all samples and correspond to that ofTi 2 03 . If this can be confirmed, the exact initialcomposition of the coating is of no great importance.

Coating by sputtering does not require that thewindow be heated to obtain good adherence of thecoating. In contract to vapor deposition the atomsand molecules arrive with a high velocity, whichassures a good bond.

5.2 CONCLUSIONS

There are a great variety of window and electromagneticfield geometries which show advantages for use in multi-pactor reduction or prevention. Our experimental workwill restrict itself to grooved windows with partial coatingon the crests, with total surface coating, and with nocoating.

5A. D. Pearson, "Studies on the Lower Oxides of Titanium,"J. Phys. Chem. Solids, 1958, Vol. 5 p. 316

f 6W. D. Fuller, "Titanium Integrated Electronic Componrts."

-33

S5.3 PROGRAM FOR NMXT INTERVAL

Tests on alumina and silica grooved windows with differentgroove angles with and without titanium suboxide coatingsare planned. Further studies of the coating process willbe performed.

6. IDENTIFICATION OF KEY TECHNICAL PERSONNEL

The hours worked by all those participating in the programare:

TASK A:- Manhours

D. Preist 135

R. Talcott 84J. Soderstrum 192

J. Zegers ---

B. Hill 35.5

K. Scholz 52

J. Leidigh 54.5

TASK B:

0. Heil 385G. Brauns 250.5B. Morozovsky 424

S. Zott 422

3I

I.- 34 -

I.

DR. OSKAR HEILASSOCIATE DIRECTOR OF RESEARCH (ADVANCED RESEARCH)

Dr. Heil received his Ph.D. in Physics at the University ofGottingen in 1932. He did research work at the University ofGottingen, in the Cavendish Laboratory, Cambridge, England, atStandard Telephones and Cables, England, and with C. H. Lorenz,A.G., Julius Pintech, K.G., and Telefunken of Germany. Priorto joining Eimac, he was employed by the Wright Air DevelopmentCenter and was engaged in research and development in the tubelaboratory at Ohio State University.

Dr. Heil is credited with the discovery of the basic velocitymodulation principle and with building of the-first tube of thistype. An article published in Zeitschrift fur Physik, Vol. 95,page 762, 1935, and later translated, Electronics, Vol. 16, page164, July, 1943, gives the first published theory of velocitymodulation of an electron beam and describes a tube utilizingthis principle. Patents based on this concept were issued toDr. Heil in the United States and Germany. All present daykiystrons utilize the principles described by Dr. Heil in thisearly work. Another significant patent of Dr. Heil's was issuedin 1950 for development of the Heil Electron Gun which, becauseof its high area convergence, is now being used in many tubedesigns. Other work of Dr. Heil includes harmonic oscillators,plasma studies, electron beam optics and diagnostics, a semi-conductor, amplifier, the floating-drift-tube klystron, and morerecently, a technique for quartz-to-metal sealing and a new highefficiency bunching technique for high power tubes.

3I

F

- 35 -

6

DONALD H. PREISTASSOCIATE DIRECTOR OF RESEARCH (SPECIAL STUDIES)

Mr. Preist received his Bachelor of Science degree andDiploma in Engineering from King's College, London University,in 1936, and subsequently entered the British Government Serviceas a member of the first radar team in England under Sir RobertWatson-Watt. Prom then until 1946, he was associated with variousaspects of radar development; in particular, the early experimentson detection of ships, development of high power ground radartransmitters, and development of the MKV IFF and beacon systems,including a period with the Combined Research Group at the NavalResearch Laboratory, Washington, D. C., from 1943 to 1945. During1946, he served in the British Ministry of Supply, London, on theapplication of radio and radar to civil aviation, and representedthe Ministry at PICAO (Provisional International Conference aero-nautical Organization) International Conferences as a scientificadvisor.

During the early part of the war he served as a flight lieu-tenant in the Royal Air Force in connection with the establishmentof radar in France, and later with Combined Operations Headquarters.

Mr. Preist joined the Research staff of Eitel-McCullough, Inc.,in 1946 and has been responsible for the design and developmentof many high power triodes, tetrodes and circuits for CW andpulse applications. Later, Mr. Preist was instrumental in thedevelopment of the Eimac line of external cavity high poweramplifier klystrons and other microwave tubes, and was responsiblefor the design of the first super power klystrons for BMEWS radar.

More recently he has concentrated on research on tubesleading to higher power levels, and under his direction substantialadvances have been made in the understanding of window phenomenaand improvements to windows, in extended interaction klystrons,and in the evolution of new concepts for the design of super powernegative grid tubes.

He has served as Klystron Project Coordinator, as ChiefResearch Engineer and presently is Associate Director of Research.

Mr. Preist has published numerous papers concerning electrontubes and their applications and holds many patents in the field.

He is a full member of the Institute of Electrical Engineersof Great Britain.

3. - 36 -

I

RTH CARLSON TALCOTCONSULTANT

Mrs. Talcott received her Bachelor of Arts (Chemistry)degree from the University of Colorado in 1948.

Mrs. Talcott joined Eitel-McCullough, Inc., in 1950 andhas been continuously working on experimental physical and chemicalprocesses and materials, and tube technology. Specific projectshave included grid and heatet coatings, fabrication and evaluationof oxide-coated cathodes, metallizing, plating and brazing of highalumina and beryllia emission, and thermal stresses in ceramiccylinders.

More recently, Mrs. Talcott has been working on the problemsassociated with dielectric windows in high power klystrons. Thiswork has already resulted in the use of sintered beryllium oxidewindows in several Eimac klystrons which are now in production,and in the isolation and analysis of certain electron bombardmentphenomena of great significance at windows of high power micro-wave tubes. She has also originated and developed successfulcoatings for both metals and dielectrics which have very lowsecondary emission yield resulting in greatly improved tube per-formance.

Mrs. Talcott has published several papers and is a memberof the Vacuum Society of American and the International Organi-zation for Vacuum Science and Technology.

In February, 1962, Mrs. Talcott relinquished her positionas Senior Research Scientist in order to take a research appoint-ment at the University of California. She is serving the companyas a Consultant.

!-7

II

I-II.

- 37 -

DISTRIBUTION LIST

" Constract DA 36-039 SC-90818 No. of CoDies

OASD (R&B)Attn: Technical LibraryRm. 3E1065, The PentagonWashington 25, D. C. 1

CommanderArmed Services Technical Information AgencyAttn: TIPCRArlington Hall StationArlington 12, Virginia 10

Advisory Group on Electron Devices346 BroadwayNew York 13, New York 2

DirectorU. S. Naval Research LaboratoryAttn: Code 2027Washington 25, D. C. 1

Commanding Officer & DirectorU. S. Navy Electronics LaboratorySan Diego 52, California

Chief, Bureau of ShipsDepartment of the NavyAttn: 681A-1Washington 25, D. C. 2

CommanderAeronautical Systems DivisionAttn: ASAPRLWright-Patterson AFB, Ohio 1

Commander, AF Command & ControlDevelopment DivisionAir Research & Development CommandAttn: CCRR (1 cy)

CCSD (1 cy)CRZC (1 cy)

USAF, L. G. Hanscom FieldBedford, Massachusetts 3

CommanderAir Force Cambridge Research LaboratoryAttn: CRXL-R, Research LibraryL. G. Hanscom FieldBedord, Massachusetts 1

-38-

No. of copies

CommanderRome Air Development CenterAttn: RAALDGriffiss Air Force Base, New York 1

AFSC Liaison OfficeNaval Air R & D Activities CommandJohnsville, Pennsylvania 1

Chief of Research and DevelopmentDepartment of the ArmyWashington 25, D. C. 1

Chief, U. S. Army Security AgencyArlington Hall StationArlington 12, Virginia 2

Deputy PresidentU. S. Army Security Agency BoardArlington Hall StationArlington 12, Virginia 1

Commanding OfficerU. S. Army Electronics Research UnitP. 0. Box 205Mountain View, California 1

Commanding OfficerDiamond Ordnance Fuze LaboratoryAttn: Library, Rm. 211, Bldg. 92Washington 25, D. C. 1

CommanderArmy Missile CommandAttn: Technical LibraryRedstone Arsenal, Alabama 1

Commanding OfficerU. S. Army Electronics CommandAttn: AMSEL-RDFort Monmouth, New Jersey 1

Commanding OfficerU. S. Army Electronics Materiel Support AgencyAttn: SELMS-ADJFort Monmouth, New Jersey 1

Corps of Engineers Liaison OfficeU. S. Army Electronics R & D LaboratoryFort Monmouth, New Jersey 1

-39-

I

No. of Copies

Marine Corps Liaison OfficerU. S. Army Electronics R & D LaboratoryAttn: SELRA/LNRFort Monmouth, New Jersey

Commanding OfficerU. S. Army Electronics R & D LaboratoryAttn: Director of ResearchFort Monmouth, New Jersey

Commanding OfficerU. S. Army Electronics R & D LaboratoryAttn: Technical Document CenterFort Monmouth, New Jersey

Commanding OfficerU. S. Army Electronics R & D LaboratoryAttn: Technical Information Division

(FOR RETRANSMITTAL TO ACCREDITED BRITISHAND CANADIAN GOVERNMENT REPRESENTATIVES)

Fort Monmouth, New Jersey 3

Commanding OfficerU. j. Army Electronics R & D LaboratoryAttn: SELRA/PR (Mr. Garoff) (1 cy)

SELRA/PR (Mr. Hanley) (1 cy)SELRA/PRG (Mr. Zinn) (1 cy)SELRA/PRM (Mr. Hersh) (1 cy)

Fort Monmouth, New Jersey 4

Commanding OfficerU. S. Army Electronics R & D LaboratoryAttn: Logistics Division (For: SELRA/PRT

Project Engineer)Fort Monmouth, New Jersey 2

Commanding OfficerU. S. Army Electronics R & D LaboratoryAttn: SELRA/PRT, Record File CopyFort Monmouth, New Jersey

Stanford UniversityMicrowave Laboratory, W. W. Hansen LabsStanford, CaliforniaAttn: Prof. S. Sonkin

Commanding OfficerU. S. Army Electronics Materiel Agency

225 So. 18th StreetPhiladelphia, Pennsylvania

Attn: SELMA/R2A

F -40 -

I

No. of Copies

Director of Defense Research and EngineeringOffice of Secretary of DefenseWashington 25, D. C.Attn: Mr. James M. Bridges 1

Advanced Research Projects AgencyOffice of Secretary of DefenseWashington 25, D. C.Attn: Program Manager, AO 318 3

Chief, Bureau of ShipsDepartment of the NavyWashington 25, D. C.Attn: Code 680 1Attn: Code 335 1Attn: Code 670B 1Attn: Code 361B 1

Chief of Naval OperationsDepartment of the NavyWashington 25, D. C.Attn: OP-07 1

Chief, Bureau of Naval WeaponsDepartment of the NavyWashington 25, D. C.Attn: RRRE 1Attn: RAAV-4423 1Attn: RREN-3 1Attn: RNWC 1

Applied Physics LaboratoryAttn: Mr. William DobbinsHoward County, MarylandVia. Bureau of Naval Weapons RepresentativeSilver Spring, Maryland 1

Chief of Naval ResearchDepartment of the NavyWashington 25, D. C.Attn: 461 1Attn: 427 1

DirectorU. S. Naval Research LaboratoryWashington 25, D. C.Attn: Dr. S. T. Snith, Code 5240 1Attn: Mr. R. C. Guthrie, Code 5300 1

- 41 -I

No. of Copies

CommanderNew York Naval ShipyardNaval Material LaboratoryBrooklyn 1, New YorkAttn: Code 920 1

Commander, Air Force Systems CommandR & T DivisionAttn: Mr. S. Tepper, RTHBolling AFB, Washington 25, D. C. 1

Commander, Air Systems DivisionAttn: ASWRNETWright-Patterson AFB, Ohio 1

Commander, Aerospace Systems DivisionAttn: ASRNETWright-Patterson AFB, Ohio 1

Commander, Electronic Systems DivisionAir Force Systems CommandAttn: ESRDEBedford, Massachusetts 1

Commander, Rome Air Development CenterGriffiss Air Force Base, New YorkAttn: RCLTT 1Attn: RCLS 1Attn: RCLC 1Attn: RCTTM 1Attn: RALTP 1

CommanderAeronautical Systems DivisionAttn: ASDL-9Wright-Patterson AFB, Ohio 1

Commander, Air Force Ballistic Missile DivisionAir Force Unit Post OfficeAttn: WDZRSjLos Angeles 45, California 1

AMC Liaison OfficerBell Telephone LaboratoriesAttn: Lt. Col. Lee G. JonesWhippany, New Jersey

4I

-4- '

S

No. of Copies

Commanding OfficerU. S. Army Electronics Researchand Development LaboratoryFort Monmouth, New JerseyAttn: SELRA/SR (Dir., Radar Div.) 1Attn:- SELRA/S (Dir., Cm Div.) 1Attn: SELRA/PE (Dir., EP&M Div.) 1Attn: SELRA/N (Dir., Comm Dept.) 1Attn: SELRA/SC 1Attn: SELRA/PEM 1

CommanderU. S. Army Materiel CommandWashington 25, D. C.Attn: AMCRD-RS-PE-E 1

Commanding OfficerFrankford ArsenalAttn: ORDBA-FELPhiladelphia 37, Pennsylvania 1

CommanderArmy Missile CommandAttn: ORDXM-RMPRedstone Arsenal, Alabama 1

General Electric CompanyPower Tube Department1 River RoadSchenectady, New YorkAttn: Mr. E. D. McArthur, Knolls Research Lab 1

Eitel-McCullough, Inc.301 Industrial WaySan Carlos, CaliforniaAttn: Dr. George Caryotakis 1Attn: Mr. Earl Shelton 1

Radio Corporation of AmericaLancaster, PennsylvaniaAttn: Mr. E. E. Spitzer

David Sarnoff Research CenterRCA LaboratoriesPrinceton, New JerseyAttn: Dr. L. S. Nergaard 1

-43-

I

I

No. of Copies

General Electric CompanyTraveling Wave Tube Products Section601 California AvenuePalo Alto, CaliforniaAttn: Mr. S. E. Webber

Kane Engineering Laboratories845 Commercial StreetPalo Alto, CaliforniaAttn: Mr. John Kane

Sylvania Electric Products, Inc.Chief Engineer of Microwave Device Division500 Evelyn AvenueMountain View, CaliforniaAttn: Dr. Rudy Hutter

Sylvania Electric Products, Inc.East 3rd StreetWilliamsport, PennsylvaniaAttn: Dr. John Whitmore

Sperry Electronic Tube DivisionSperry Rand CorporationGainesville, FloridaAttn: Dr. A. D. Sutherland

Manager Applied Research DepartmentElectron Tube DivisionWestinghouse Electric CorporationP. 0. Box 746Baltimore 3, MarylandAttn: Mr. G. R. Kilgore 1

Raytheon CompanyAttn: Mr. William C. BrownWaltham, Massachusetts

Sperry Gyroscope CompanyAttn: Dr. V. R. LearnedGreat Neck, Long Island, New York

Varian AssociatesAttn: Dr. T. MorenoAttn: Dr. A. StapransAttn: Dr. E. W. Herald611 Hansen WayPalo Alto, California

I4. - 44 -

I,No. of Copies

SFD LaboratoriesAttn: Dr. J. Feinstein800 Rahway AvenueUnion, New Jersey

Litton IndustriesAttn: Dr. J. F. Hull960 Industrial RoadSan Carlos, California

Hughes Aircraft CompanyAttn: Dr. L. M. Field, Microwave Tube DivisionCulver City, California

Watkins-Johnson Company3333 Hillview AvenuePalo Alto, CaliforniaAttn: Dr. Rolf Peter

Field Emission CorporationMcMinnville, OregonAttn: Dr. F. M. Charbonnier

Sperry Gyroscope CompanyDivision of Sperry Rand CorporationAttn: D. Churchill, J. McLinden (1 cy ea)Great Neck, Long Island, New York 2

General Electric CompanySuperpower Microwave Tube Laboratory1 River RoadSchenectady, New YorkAttn: R. Bondley, Knolls Research Laboratory

Eitel-McCullough, Inc.Attn: L. ReedSan Carlos, California

Stanford Research InstituteAttn: L. FeinsteinMenlo Park, California

General Telephone and Electronics Corp.Bayside LaboratoryAttn: B. W. LeavittBayside, New York 1

I!~- 45 -

I

I

No. of Covies

Mitrouics, Inc.132 Floral AvenueAttn: S. S. ColeMurray Hill, New Jersey 1

General Electric Company3001 East Lake RoadAttn: J. J. Cacciotti

¶ Erie, Pennsylvania

Ford Instrument Company31-10 Thomas AvenueAttn: W. Franklin,Long Island City, New York

Director, Lincoln LaboratoryAttn: Dr. R. ButmanP. 0. Box 73Lexington 73, Massachusetts

Stanford UniversityAttn: Prof. M. Chodorow, DirectorMicrowave LaboratoryStanford, California

Aerospace CorporationAttn: Dr. I. GettingLos Angeles 45, California

RAND Corporation1700 Main StreetSanta Monica, California

Project MStanford UniversityStanford, CaliforniaAttn: Mr. J. Jasberg

Polytechnic Institute of Brooklyn55 Johnston StreetAttn: J. W. GriemamannBrooklyn 1, New York

Physics LaboratoryUniversity of UtahSalt Lake City, UtahAttn: P. GibbsAttn: G. Baker

-46-

go. of Copies

Laboratory of Insulation ResearchMassachusetts Institute of TechnologyCambridge, MassachusettsAttn: D. A. Powers 1

Battelle Memorial Institute505 King AvenueColumbus 1, OhioAttn: Defender Library

Cornell Aeronautical Laboratory, Inc.4455 Genesee StreetBuffalo 21, New YorkAttn: Mr. R. C. Beitz

MITRE CorporationBedford, MassachusettsAttn: Dr. R. F. Naka, Associate Technical

Director, Bldg. 2A-251 1

Chief, Army Research OfficeWashington 25, D. C. 1

This contract is supervised by the Techniques Branch, ElectronTubes Division, ECD, USAELCTRDLAB, Fort Monmouth, Now Jersey.For further technical information contact Mr. Louis Heynick,Project Engineer, Telephone 201-59-61414.

I.

- 47 -

F-

II 0 S I A0 0 *

41 0 * 0 436

M004 H 45

4.4 0 004000

0 14

16 -4. 14 a 0 0 0 4400.4 'A 1. 0 . 0 k. 0

a .* C A4.4 6. 40 f6.4 0 F.. * * *4 040 0 0 F.6 a .

-14 06 00 0.-0 040 -.43

~ ~: ~ 1..0. -.0 og:: \.-s 0 ýa1'4' AW S.6 4 1*04 a AM 10.4 -A C ?p

Is4~ L 410 -9.0 06 1. ..A. 3. I 41 0 . .5.4

0 ~ ~ aC h~ 0.41,.4 .q 14

16 3 4. 03 . 14-.1 0

.1 14 4A L S A.

1.30 40 t I ( 0

.. 4 0 0 4 to.

* .It I . 01 ; I "IW

a - .3 1 4 14 A 1-4

1041

06.4.

ý .N' M goa4 30.4 1M 0 00 46

'A 4 04 0 -m3-.4 u1r1k 191 Aii II a .1 r4.. _._0