`
`Gs
`
`HANWHA1026
`
`HANWHA 1026
`
`

`

`89048209449
`
`b89048209449a
`
`AION
`
`
`
`

`

`Journal
`
`of
`
`Applied Physics
`
`Eimer Hurcnisson, Editor
`
`Earte C. Greoe, Assistant Editor
`
`Maurice A. Brot
`Ray E. Bouz
`Watuace J. Eckert
`
`Board of Editors
`Rosert D. HEIDENREICH
`Louis A. Press
`Roman SMOLUCHOWSKI
`
`W. R. SmytTHe
`Artaur VY. ToBoLsky
`A. M. WEINBERG
`
`VOLUME 25
`
`JANUARY-DECEMBER, 1954
`
`Published by the
`AMERICAN INsTITUTE oF PHysics
`Incorporated
`
`

`

`INFORMATION FOR CONTRIBUTORS
`
`Send manuscripts to Elmer Hutchisson, Case Insti-
`tute of Technology, Cleveland 6, Ohio. Textual ma-
`terial (abstract, the body of the paper, tabular mate-
`rial, references, and figure captions) must be in Eng-
`lish,
`typewritten (except mathematical expressions)
`double spaced with wide margins.on a durable, white
`paper, preferably letter size. The original copy, not
`the carbon, should be submitted. For general style,
`spellings, abbreviations, and form, consult recent
`is-
`sues of
`this journal. Authors’
`institutions are re-
`quested to pay a publication charge of $8 per page
`which entitles them to 100 reprints without further
`charge.
`
`It
`An abstract must accompany each aarticle.
`should be adequate as an index and as a summary.
`
`expressions
`simplest mathematical
`the
`All but
`should be carefully printed in by hand. Be sure to
`distinguish between capital and small
`letters.
`Indi-
`cate clearly superscripts and subscripts by inverted
`carets and carets,
`respectively. Avoid complicated
`superscripts and subscripts. Avoid repetition of a
`complicated expression by representing it by a sym-
`bol.
`Identify in the margin Greek letters and unusual
`symbols. Use fractional exponents instead of root
`signs. Use the solidus (/) for fractions whereverits
`use will save vertical space.
`
`References should appear as footnotes, numbered
`consecutively, and arranged thus:
`1A, B. Smith, Phy. Rev. 41, 852 (1932).
`2H, Lamb, Aydrodynamscs (Cambridge University Press, Tedding-
`ton, England), sixth edition, pp. 573, 645.
`
`Tables may be typewritten on sheets separate from
`the running text. Each table must have a caption
`which will make the data in the table intelligible with-
`out reference to the text. Avoid complicated column,
`headings.
`
`Wherever possible figures should be planned for
`the width of a column (3 in.). A convenient reduc-
`tion ratio is about 2:1 or 3:1. Lettering should be
`large enough onthe original that when reducedit shall
`not be less than about Ys
`in. (~1.5 mm). Line draw-
`ings must be in Jndia ink on white paper or tracing
`cloth. Coordinate paper is not desirable but if used
`must be blue lined and the coordinate lines which are
`to show must be inked over. Photographs should
`have maximum black and white contrast.
`
`Identify each original drawing or photograph on the
`back with the author’s name and the figure number.
`Figures should be numbered consecutively. Captions
`should be typed on a separate sheet of paper.
`
`A limited number of alterations in proof are un-
`avoidable, but
`the cost of making extensive altera-
`tions after the article has been set
`in type will be
`charged to the author.
`
`Proof and all correspondence concerning papers in
`the process of publication should be addressed to the
`Publication Manager, American Institute of Physics,.
`57 East 55 Street, New York 22, New York.
`
`
`
`INFORMATION FOR SUBSCRIBERS
`
`SUBSCRIPTION PRICE
`
`The Journat or APPLiep PHysIcsis devoted to physics
`U. S. and
`in its role as the science basic to other natural sciences
`Canada=Elsewhere
`and to the arts and industries, Previous to 1937 this pub-
`To members of the American Insti-
`lication was known as Prystcs. It aims to be of service
`tute of Physics and Affiliated So-
`not only in physical laboratories, but also in the labora-
`CHOCES sees caress sare eenes ere ini ieee ee $10.00
`tories of
`industrial,
`chemical, geological, geophysical,
`‘To all others a. cscs ceicancnre ss
`12.00
`meteorological, radio, and similar concerns. The JourNAL
`or Appiien Puysics publishes editorials and reviews, as
`well as technical papers of applied physics.
`
`$11.00
`13.00
`
`BACK NUMBER PRICE
`
`Yearly back number rate when com-
`plete year is available: $14.00.
`Single copies: $1.25 each.
`
`Subscriptions and orders for back numbers should be
`addressed to the American Institute of Physics, 57 East
`55 Street, New York 22, New York.
`
`Advertising rates will be supplied on request.
`
`57 East 55 Street
`
`American Institute of Physics
`
`New York 22, New York
`
`

`

`OF
`
`AP
`
`PLI
`
`|4
`
`y
`
`) P
`
`ycE|®ie
`
` 2|
`
`JOURN
`r|Ne
`
`@
`
`
`
`

`

`ELMER HUTCHISSON,
`op
`Eanve C. Greece,
`Assistant Editor
`Board of Editors
`suStor
`Ray E. Bouz
`Watuiace J. Eckert
`Rosert D. HEmENREICH
`Lours A. Prees
`Roman SMOLUCHOWSKI
`W. R. SmyTHEe
`Artuur V. Tospotsky
`Atvin M. WEINBERG
`Publication
`of
`th
`ublication of
`the
`Ameri an
`.
`.
`£
`| Institute of Physics
`
`|
`
`|
`
` |
`
`|
`
`|
`i
`
`!
`i
`
`|
`|
`|
`|
`
`JOURNAL
`oO F
`APPLIED PHYSICS
`
`
`
`-
`Vol. 25, No. 5
`
`MAY, 1954
`—$—$$—______—.
`
`
`
`I
`I Th °
`rt
`ts £88Ue
`
`Fase
`Cireumferential Gap In a Circular Wave Guide Excited bya Dominant Cicular-Eleetrie
`E. Storer 545
`ave
`wzconard §.
`Sheingold
`and
`James
`Impedance of a Top-Loaded Antenna of Arbitrary Length over a Circular Grounded
`Sereen
`James R. Wait and Walter J. Surtees 553
`Plastic Deformation In Beams under Distributed Dynamic Loads
`J, A. Seiler and P. 8. Symonds 556
`Tube Noise under Large Transit-Time Conditions
`S. V. Yadavalli 564
`OFFICERS
`Some Phenomena Associated with Supersonic Liquid Jets
`Hoenn te
`Brian Dunne and Benedict Cassen 569
`REDERICK SEITZ,,
`Transverse Motion of an Electron in a Constant Wave Speed Section of a Linear Acceler-
`Chairman
`=
`ator
`Ira Jacobs and E. S. Akeley 572
`Georce B. Pecnam,
`Friction of Clean Metals and Oxides with Special Roferenes ty Saberlage
`;
`Treasurer
`4.
`8. Machlin and W. R. Yankee 576
`Onthe Band Width of Cavity Antennas
`Marshall H. Cohen 582
`GOVERNING BOARD
`Multiple Shock Reflection in Corners
`Joseph B. Keller 588
`siaia s Bacher
`Measurement of the Effect of Chforine Treatment on the Work Function of Titanium
`=
`UCHTA
`and Zirconium
`Herbert Malamud and Aaron D. Krumbein 591
`Karu K, Darrow
`Delayed Yield and Strain Rate and Temperature Dependenceof Yield Point in
`5. A. GoupsMiT
`4
`Iron
`Takeo Yokobori 593
`DeaneB. Jupp
`Negentropy and Information In Telecommunications, Writing, and Reading
`.
`Hua8. Know.es
`L. Brillouin 595
`C. C. LAuRrrsEN
`Harry Letaw, Jr., and John Bardeen 600
`Electrolytic Analog Transistor
`W. H. Mankwoop, Jr.
`R. W. Balluffi and L. L. Seigle 607
`Diffusion in Bimetal Vapor-Solid Couples
`Wiuuram F. Meacers
`L. R. Walker 615
`Stored Energy and Power Flow In Electron Beams
`Pamir M. Morse
`D. K. C. MacDonald 619
`Information Theory and Knowledge
`aoe O’Brien
`E. V. Laitone 623
`Limiting Pressure on Hydrofoils at Small Submergence Depths
`G ARRY #: ae
`Fast Time Analysis of Intermittent Point-to-Plane Corona in Air. III. The Negative
`ses Ro
`BOnAM
`Point Trichel Pulse Corona
`M.R. Amin 627
`RIC
`LLODGERS
`Volume and Surface Recombination Rates for Injected Carriers in Germanium
`Raven A. SAWYER
`J.P. McKelvey and R. L. Longini 634
`FREDERICK SEITZ
`Propagation of Microwaves between a Parallel Pair of Doubly Curved Conducting Sur-
`WILLIAM SHOCKLEY
`faces
`K.S. Kunz 642
`Ricuarp M. Surron
`Errors in Power Spectra Due to Finite Sample
`Lee M. Spetner 653
`J.H. Van VieEck
`Approximate Emissivity Calculations for Polyatomic Molecules. I. CO:
`8.8. Penner 660
`Mark W. ZeEMANSKY
`Cross Section of Pulled Crystals
`Robert G. Pohl 668
`oe
`Some New Magneto-Mechanical Torsion Experiments
`A.W. Cochardt 670 mae .L Printing
`Letters to the Editor:
`ELVIN
`LOOS
`Electron Bunches of Short Time Duration
`E. W. Ernst and H. VonFoerster 674
`ADMINISTRATIVE
`Effect Produced by Oblique Impact of a Cylinder on a Thin Target
`STAFF
`William A. Allen, Joe M. Mapes, and Wesley G. Wilson 675
`New Silicon p-n Junction Photocell for Converting Solar Radiation into Electrical
`Henry A. Barron,
`Power
`D. M. Chapin, C. 5S. Fuller, and G. L. Pearson 676°
`Director
`Dissected Amplifiers Using Negative Resistance
`W. Shockley and W. P. Mason 677
`WALLACE WATERFALL,
`On the Interaction of Two Plane Shocks Facing in the Same Direction
`Executive Secretary
`Otto Laporte and E. B. Turner 678
`WHepon VGnEunGEn
`High-Speed Magnetic Pulsing of Ferrites
`R. Conway LeCraw 678
`Aclnaniising Aa
`i
`Crossed Electron Beam Technique for Measuring Space-Charge Effects In Beams
`F 13 Tung
`Paul D. Coleman, Murray D. Sirkis, and Irving Kaufman 679
`Ruta f.
`DRYANS,
`Combined Slow-Wave Circuit and Beam Focusing Structure
`Philip Rice 680
`Publication Manager
`Announcements
`681
`Epitu I. NEFTEL
`|
`Index to Advertisers
`xxii
`_ CirculationManager
`KATHRYN SETZE
`
`Assistant Treasurer
`
`|
`
`Subscriptions, renewals, and orders for back numbers should be ad-.
`dressed to the American Institute of Physics, 57 East 55 Street, New
`York 22, New York.
`
`The JourNAL oF APPLIED Puysics, published monthly at Prince and
`Lemon Sts., Lancaster, Pa., is devoted to physics in its role as the science
`basic to other natural sciences and to the arts and industries. Previous
`Subscription Price
`U.S. and
`to 1937 this publication was known as Puysics. The JOURNAL OF APPLIED
`Canada
`Puysics publishes editorials and reviews, as well as technical papers of
`To members of the American Institute of
`applied physics. Articles appearing in it are indexed regularly, according
`Physica and Affiliated Societies............ $10.00
`to their field, by Physics Abstracts, Chemical Abstracts, Engineering
`TOsGUE oi. o bcia.aserece te, atecnceoi ave
`12.00
`Index, Inc., Industrial Arts Index, etc.
`Back NumberPrices
`Manuscripts should be submitted to Elmer Hutchisson, Editor, Case
`Yearly back number rate when complete year is available: $14.00.
`Institute of Technology, Cleveland 6, Ohio, Unless otherwise stated,
`Single copies: $1.25 each.
`submission of a manuscript is a representation that it has been neither
`Changesof address should be sent to the American Institute of Physics;
`copyrighted, published, nor submitted for publication elsewhere. The
`corrected proofs should be sent to the Publication Manager, American
`authors’ institutions are requested to pay a publication charge of $8 per
`Institute of Physics. Advertising
`rates supplied on request. Orders,
`advertising copy, and cuts should
`be sent to the American Institute of
`page which,
`if honored, entitles them to 100 reprints without further
`Physics.
`charge.
`Instructions will be sent with galley proofs,
`Entered as second class matter January 22, 1937, at the Post Office at Lancaster, Pennsylvania, under the Act of March 3, 1879.
`mailing at the special rate of postage provided for in paragraph (d-2), section 34.40, P. L. & R. of 1948, authorized May 2, 1932.
`
`Elsewhere
`$11.00
`13.00
`
`Accepted for
`
`

`

`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`

`

`LETTERS: TO
`
`THE EDITOR
`
`677
`
`mum sensitivity is arbitrarily taken as unity. Ideally the maximum
`should appear near 1.2 microns with half this sensitivity at 0.6
`and one-quarter at 0.3 micron. CurveB is the normal distribution
`of full sunlight! with the maximum expressed as unity. Curve C
`is the product of Curve A and Curve B again reduced to unity
`for maximum. This curve shows which part of the sun’s radiation
`is most useful for this particularcell.
`The photocells described here have been madeto deliver power
`from the sun into a resistance load at the rate of 60 watts per
`square meter of photocell surface. This is approximately 6 percent
`efficiency and compares with a measured value of 0.5 percent on a
`commercially available photocell. The greatest over-all efficiency
`previously reported for direct conversion of solar radiation into
`electrical power is that of Telkes* using thermoelectric junctions
`and amounts to 1 percent.
`Wewish to thank H.B. Briggs who madethe spectral measure-
`ments shown in Curve A.
`LW, E. Forsythe, Measurement of Radiant Energy (McGraw-Hill Book
`Company, Inc,, New York, 1937),
`* Maria Telkes, J. Appl. Phys. 18, 1116 (1947).
`
`
`Dissected Amplifiers Using Negative Resistance
`W. SHocKLey AND W. P. Mason
`Bell Telephone Laboratories, Inc., Murray Hill, New Jersey
`(Received December 28, 1953)
`
`r is the purpose of this communication to point out a line
`of attack upon the problem of making high-frequency semi-
`conductor amplifiers. This attack may be described as the dissec-
`tion of the amplification mechanism into its two constituents:
`negative resistance anddirectionality.
`Negative resistance is a common feature of all amplifying
`devices operating from a dc power source. Thus the net ac power
`from the device when operating as an amplifier is positive. Since
`this poweris the net flux of the ac part of Poynting’s vector,it
`follows that Poynting’s vector must have a positive divergence
`throughout certain “negative resistance regions.” For a vacuum-
`tube triode amplifier, the grid-plate space is such a region, the
`current being high when the voltage is low and vice versa. Simi-
`larly, the space-charge region of the collector junction of a junction
`transistor amplifier has negative resistance. It should be noted
`that these negative resistances are not characteristic of the regions
`per se but instead of their behavior in an amplifier circuit.
`This method of analysis enables one to evaluate quickly such
`proposals as the making of a junction transistor from isolated
`p-n junctions (impossible because no negative resistances can
`occur) or an amplifier from a nonlinear dielectric and dc power
`source.
`Directionality arises in vacuum tubes and transistors because
`of the separation of input and output circuits. This leads to
`asymmetry in the current voltage matrix and permits the achieve-
`mentof high gain by cascading stages withoutresultantinstability.
`Because the drift velocity of electrons (or holes) in semiconduc-
`tors is smaller than the average transit velocity in vacuum tubes,
`the physical dimensions must be smaller for the same transit
`times. Although the solid nature of transistors gives them an
`inherent structural advantage over vacuum tubes,
`the smaler
`scale required for them poses serious problems of construction
`for devices in the highest-frequency ranges.
`One wayaroundthis difficulty is to make two-terminal negative
`resistance devices in which only one dimension need be small.
`A variety of operating principles are possible for the operation
`‘of such devices."
`Two-terminal negative resistance devices have the disadvan-
`tage, however, that they do not have directionality; consequently,
`high gain is associated with narrow margins of stability.
`This limitation may be overcome by adding directionality
`through passive elements* such as Hall effect couplers or gyrators*
`
`Ya
`
`Yel;
`
`Vala
`
`yy
`
`Ys
`
`—
`
`Fic. 1. Hall effect plate and three resistances. (Power supplies and con-
`nections necessary to power the negative resistances are not shown.)
`
`in wave guides.‘ Such combinations can simulate conventional
`amplifying devices such as a vacuum-tube triode. In Fig. 1 the
`hexagon represents a germanium plate with a magnetic field
`perpendicular to its plane. It contributes in Eqs. (1) and (2) the
`terms involving the conductance yo and the dimensionless quantity
`a which is an odd function of the Hall angle with |a| <1:
`(1)
`1, =(yotyit+y2) 1-41 —a) votye]V2,
`(2)
`In=—[4 (+a)yo+y2]Vi+ (vot yey) V2.
`The conductances 1, y2, and ys may correspond to negative
`resistance elements. For example,if y:, v2, and ys have the values
`y=Jr= — (1+a)y0/2,
`(3)
`ya= (a—1)90/2,
`(4)
`
`then Eqs. (1) and (2) reduce to
`(5)
`1,=0 and f2=ayoV 1.
`This corresponds to an ideal pentode for a>0 and to such a
`pentode with voltage reversing transformer for a<0. Thus the
`circuit can simulate conventional amplifiers and in addition
`produce new types of devices.
`Values of a as high as 0.2 can be obtained in n-type germanium
`at room temperature atfields of 17 500 oersteds. Similar reasoning
`to that discussed here applies also to the four-terminal Hall
`effect plate or balanced gyrator.*
`A simple test® of this principle was made at low frequencies
`by putting two gas diodes with negative resistance characteristics
`on either side of the balanced gyrator of reference 3. With this
`arrangement a gain of 6 db was obtained in onedirection, and a
`loss of 46 db in the other. By more careful balancing of the negative _
`resistances with the input and output resistances of the gyrator,
`higher gains are possible without instability.
`1A guide to one method of designing such semiconductor devices is
`furnished by the work of F. B. Llewellyn on transit time vacuum-tube
`diodes: Bell System Tech. J. 13, 59-101 (1934); 15, 575-586 (1936).
`Methods of extending this approach to semiconductors have been indicated
`by W. Schockley, U. S. Patent 2,623,102 sees Transistor). Another
`physical effect considered by one of us (W-S.) in unpublished work is the
`negative differential mobility to be expected for holes in very high electric
`fields, Interest in this predicted effect was the provo
`cause of the
`research on high field mobilities, see E. J. Ryder and W.
`Shockley, oe
`Rev. 81, 139 (1951) and E. J. Ryder, Phys.
`Rev. 90, 766-769 (1953),
`The
`t
`of this effect has been independently discovered oy H.
`er, Z.
`Physik 134, 435 (1953), who calls it the “Staueffect" but
`does not consider
`its w
`ion as a power source,
`? Passive elements can have the desired directionality,i.e., lead to unsym-
`metrical current-volt:
`matrices, only in the presence of magnetic fields.
`For a proof see H, B.
`G. Casimir, Revs. Modern Phys. 17, 343-350 (1945).
`+ Mason, Hewitt, and Wick, J. Appl. Phys. 24, 166-175 (1953). Thefact
`that transmission through a crystal is nonreciprocal in the presence of a
`magnetic field was first pointed out by E. M. McMillan, J. Acoust. Soc.
`Am. 19, 922 (1947).
`4C. L. Hogan, Bell io Tech, J. 31, 1-32 (1952).
`+ This test was made by W. H. Hewitt.
`
`