US 3699404 A
An electron emitter comprising a body of gallium phosphide having a thin surface layer of cesium. The gallium phosphide is doped with a deep acceptor such as iron. Interaction between the cesium layers and the semiconductor surface results in ionization of the deep acceptor impurities in a small region near the surface. The ionization of deep acceptors at the cesiated surface results in a graded concentration of ionized impurities through the gallium phosphide layer, which establishes an internal electric field for impelling electrons toward the cesiated emitting surface.
Description (OCR text may contain errors)
United States Patent Simon et al.
[151 3,699,404 [4 1 Oct. 17, 1972  NEGATIVE EFFECTIVE ELECTRON 3,105,166 9/1963 Choyke et a1 ..313/310 AFFINITY EMITTERS WITH DRIFT 3,150,282 9/ 1964 Geppert 13/346 FIELDS USING DEEP ACCEPTOR 3,422,322 1/1969 Haisty ..'317/235 DOPING 3,478,213 11/1969 Simon et a1 ..250/207 1 1  Inventors: Ralph E. Simon, Trenton; Brown F. g ,3 X22 22 at a 22 22 Williams, Princeton, both of NJ.
 Assignee: RCA Corporation OTHER PUBLICATIONS Feb. 24, Eigeggert, Proceedings VIC]. 54, NO. I, Jan.
 App]. No.: 118,491 .1. Scheer, Philips Res. Reports, 15, 584 (1960).
Relaed Us Application Data I Sgigeergg; a1., SOIld State Communications, 3, 189-  Continuation of Ser. No. 751,862, Aug. 12,
' 1968, abandoned. Primary Examiner-Martin H. Edlow Attorney--Glenn H. Bruestle  US. Cl. ....3l7/235 R, 317/235 N, 317/235 UA,
317/235  ABSTRACT  Int. Cl. ..H0ll 15/00 An electron emitter com prislng .a body of gallium  new of gkgg 5 2 6 phosphide having a thin surface layer of cesium. The
gallium phosphide is doped with a deep acceptor such 56 R fare c Cited as iron. Interaction between the cesium layers and the 1 e n es semiconductor surface results in ionization of the deep UNITED STATES PATENTS acceptor impurities in a small region near the surface,
The ionization of deep acceptors at the cesiated sur- 3,422,322 H1969 Haisty .317/235 face results in a graded concentration of ionized impu 3,458,782 7/1969 Buck ......317/235 rities through h gallium phosphide layer, which 3,121,809 2/1964 Afana "307/885 establishes an internal electric field. for impelling e1ec- 3,478,213 1 81111011 trons toward the cesiated emitting surface.
3,387,161 6/1968 Van Laar ..313/94 2,960,659 1 H1960 Burton ..330/65 6 Claims, 6 Drawing Figures f PATENTEDUBI 1'1 am sum 20F 3 IITOIII'Y NEGATIVE EFFECTIVE ELECTRON AFFINITY EMITTERS WITH DRIFT FIELDS USING DEEP ACCEPTOR DOPING This application is a continuation of Ser. No. 751,862, filed 8/12/68, now abandoned.
BACKGROUND OF THE INVENTION This invention relates to electron emitters or cathodes, and more particularly to an electron emitter which comprises a body of semiconductor material.
A particular type of cold cathode semiconductor electron emitter has recently been discovered in which electron emission is obtained by manufacturing a semiconductor structure such that the bottom of the conduction band in the semiconductor bulk lies at an energy level above the vacuum energy level at the emitting surface. Thus, conduction band electrons may drift to the emitting surface with a residual energy above that of the vacuum energy level, so that these electrons may be emitted from the surface without the necessity of supplying additional energy. Such a structure is known as a negative electron affinity emitter, and is described in the following U. S. Patent Applications:
i. Ser. No. 668,130; filed Sept. 15, 1967; Entitled: Semiconductor Electron Emitter now issued as US Pat. No. 3,478,213
ii. Ser. No. 665,511; filed Sept. 5, 1967; Entitled: Transmission Type Secondary Emission Device with Semiconductor Dynode.
SUMMARY OF THE INVENTION An electron emitter is provided which comprises a P type semiconductor body including a deep acceptor impurity. An electropositive work function reducing layer is disposed on a given surface of the semiconductor body. The ionization energy of the deep acceptor does not exceed the difference between the energy gap of the semiconductor body and the work function of the electropositive layer on the semiconductor surface.
DRAWINGS FIG. 1 shows a cross-sectional view of a secondary emission dynode according to a preferred embodiment of the invention;
FIG. 2 shows a cross-sectional view of an injection type electron emitter according to an alternative embodiment of the invention;
FIG. 3a shows a partial cross-sectional view of the active portion of the dynode shown in FIG. 1;
FIG. 3b shows an energy band diagram corresponding to the partial cross-sectional view of FIG. 30;
FIG. 4a shows a partial cross-sectional view of the active portion of the injection emitter shown in FIG. 2; and
FIG. 4b shows an energy band diagram corresponding to the partial cross-sectional view of FIG. 4a.
DETAILED DESCRIPTION A secondary emission dynode l utilizing the negative electron affinity principle, as shown in FIG. 1, comprises an alumina substrate 2 upon which is disposed a thin layer 3 comprising an (shallow) acceptor impurity material such as beryllium. A gallium phosphide P type semiconductor layer 4 is disposed on the beryllium layer 3. A thin cesium layer 5, which may have a thickness on the order of a few atomic diameters, is disposed on the gallium phosphide layer 4.
The gallium phosphide layer 4 is doped with a deep acceptor impurity material such as iron, chromium or copper. Interaction between the electropositive cesium layer 5 and the gallium phosphide P type layer 4 results in ionization of the deep acceptors near the interface between these layers.
As a result, the surface of the gallium phosphide layer 4 adjacent the cesium layer 5 behaves as highly doped (i.e., degenerate or nearly degenerate) P type semiconductor material. The energy band structure of the dynode 1 is such that electrons in the conduction band of the semiconductor material comprising the layer 4 can reach the exposed surface 6 of the cesium layer 5 with sufficient residual energy to surmount the potential barrier at this surface and be emitted therefrom.
The dynode 1 may be employed as a secondary emission type electron multiplier by bombarding the exposed surface 6 of the cesium layer 5 with primary electrons in a direction such as that indicated by the arrow 7. Secondary electrons which are emitted from the surface 6 may be drawn off by a suitable electric field in the direction indicated by the arrow 8. Typically, the primary electrons may have energies in the range of 300 to 10,000 electron volts.
By reducing the thickness of the alumina substrate 2 to a value on the order of 500 Angstroms, the dynode I may be employed as a transmission type secondary emission electron multiplier. In this structure, primary electrons are directed at the exposed surface of the alumina layer 2; the primary electrons penetrate the alumina layer 2 and the beryllium layer 3 to reach the gallium phosphide layer 4, where secondary electrons are produced and emitted from the exposed surface 6 of the cesium layer 5. Since approximately 40 percent of the energy of the incident primary electrons is lost in passing through the alumina and beryllium layers, relatively high primary electron energies, on the order of 3,000 to 5,000 electron volts or more, are desirable when the dynode l is operated as such a transmission type electron multiplier.
Typically, the beryllium layer 3 may have a thickness on the order of a few atomic diameters, and the gallium phosphide layer 4, which may be monocrystalline or polycrystalline, may have a thickness on the order of 0.2 to 1 micron. During fabrication of the dynode 1, heat treatment causes beryllium from the layer 3 to diffuse a short distance (on the order of 10 to I00 Angstroms) into the gallium phosphide layer 4 to form a thin highly doped P type surface region.
The use of a deep acceptor impurity in the gallium phosphide layer 4 results, as will hereafter be explained in detail, in an internal electric field (drift field) within the layer 4 which impels electrons toward the cesium layer 5, thus increasing the external quantum efficiency of the dynode I.
In the structure shown in FIG. 1, electrons are introduced into the conduction band of the gallium phosphide layer 4 by bombarding this layer with primary electrons.
Alternatively, electrons may be introduced into the conduction band of the gallium phosphide layer 4 by injection, utilizing a forward biased P-N junction. Such a structure is exemplified by the injection type negative affinity electron emitter 10, as shown in FIG. 2.
Theinjection type electron emitter 10 comprises a highly doped N type gallium phosphide substrate 11, which may be doped with a suitabledonor impurity such as tellurium. Ohmic electrical contact to the substrate 1 l is provided by means of a tin electrode 12. An insulating layer 13 comprising, e.g., pyrolytically deposited silicon dioxide is deposited upon the upper surface of the gallium phosphide substrate 11. The insulating layer 13 has an aperture 14 exposing a portion ,of the gallium phosphide surface.
A P type gallium phosphide layer 15 is disposed on the insulating layer 13 and in the aperture 15 so that the P type layer 15 forms a P-N junction 16. with the underlying portion of the N type gallium phosphide substrate 11. A nickel electrode layer 17 provides ohmic electrical contact to the P type gallium phosphide layer 15. The layer 15 includes a deep acceptor impurity such as iron, chromium or copper.
A thin cesium layer 18 is disposed on the exposed surface of the P type layer 15, the layers 18 and 15 cooperating to form a negative affinity electron emitting structure which operates in similar fashion to that provided by the layers and 4 of the dynode l. Electrons are introduced into the conduction band of the P type gallium phosphide layer 15 by injection from the N type substrate 1 1 across the forward biased P-N junction 16. Injection biasing is provided for the P-N junction 16 by means of a battery 19 and a series resistor 20 connected between the electrodes 12 and 17.
The injected electrons diffuse or drift through the P type layer 15 and are emitted from the exposed surface of the cesium layer 18.
The operation of the dynode 1 will be better understood by reference to FIGS. 3a and 3b.
As shown inpartial cross-sectional view in FIG. 3a, the active portion of the dynode 1 comprises the P type gallium phosphide layer 4 and the cesium surface layer 5. A thin P region 21 is provided by the diffusion of beryllium into the gallium phosphide layer 4 from the metallic beryllium layer 3. The total thickness 11 of the gallium phosphide layer 4 may typically be on the order of 0.2 to 1 micron, while the thickness a of the P re gion 21 may be on the order of to 100 Angstroms, i.e. substantially less than the total thickness of the semiconductor layer 4. The thickness c of the cesium layer 6 may be on the order of l to 10 Angstroms.
The energy band diagram for the active portion of the dynode 1 is shown in FIG. 3b, which is vertically aligned with FIG. 3a. The energy gap between the valence and conduction bands has a value E,, while the ionization energy of the deep acceptor (iron, chromium or copper) impurity in the gallium phosphide layer 4 has a value E An "intrinsic line is drawn on the diagram midway between the valence and conduction bands. As is well known, portions of the semiconductor layer in which the Fermi level lies below the intrinsic line exhibit P type conductivity, while portions of the layer in which the Fermi level is above the intrinsic line exhibit N type conductivity.
By the ionization energy I5 of the deep acceptor impurity is meant the energy which must be imparted to each impurity site to generate a hole thereat.
By the term deep acceptor impurity is meant an acceptor impurity which is not normally ionized at the operating temperature of the semiconductor body. Since a thermal energy of kT electron volts is available at any particular absolute temperature T (k being the Boltzmann constant), the ionization energy of a deep acceptor impurity must be greater than this value. We prefer to employ deep acceptor impurities having ionization energies on the order of 4 kT or more. Since the value of kT at room temperature (300 C.) is 0.026 electron volt, deep acceptors used in conjunction with the structure described herein should preferably have room temperature ionization energies on the order of at least 0.1 electron volt.
The electropositive cesium layer 5 pins the bottom of the .conduction band to the Fermi level at the emitting surface of the P type layer 4, as illustrated in FIG. 3b. The work function "I at the emitting surface 6 may be defined as the difference between the vacuum energy level and the Fermi level at this surface.
The pinning of the bottom of the conduction band at the Fermi level at the emitting surface necessitates a sharp bending of the valence and conduction bands in the immediate vicinity of the emitting surface. The electropositive cesium layer 5 introduces electrons into the adjacent portion of the gallium phosphide layer 4, so that a thin N type inversion region exists at the emitting surface. This inversion layer has a thickness 8, and terminates at the point where the Fermi level crosses the intrinsic line.
In the region where the Fermi level and the deep acceptor level cross, the deep acceptors are ionized. These deep acceptors are ionized with an energy spread of a few times kT from the Fermi level. The ionization of these deep acceptors contributes to the sharp band bending at the emitting surface.
Typically, for gallium phosphide containing iron as the deep acceptor, with an iron concentration on the order of 10" to 10 atoms/emf, the deep acceptors are ionized for a distance of approximately Angstroms from the interface with the cesium layer 5.
The proportion of the deep acceptor impurities which are ionized decreases with distance away from the emitting surface, resulting in a varying distance between the deep acceptor level and the Fermi level. Since the valence and conduction bands are necessarily parallel to the deep acceptor level, these bands are therefore sloped, resulting in an internal potential gradient, i.e. an electric field oriented in a direction which impels electrons toward the emitting surface.
Electrons in the conduction band of the semiconductor layer 4 (having been introduced into the conduction band by secondary or avalanche generation, photon generation of hole-electron pairs, injection, tunneling, or any other suitable means) are prevented from diffusing toward the surface opposite the emitting surface (where the electrons may recombine and cease to act as charge carriers) by the presence of the P layer 21, which as previously mentioned is doped with a high concentration of a shallow acceptor such as beryllium. Other shallow acceptors, such as zinc or cadmium, may also be employed, The ionization energy of the beryllium acceptor impurity is less than 0.020 electron volt, so that the beryllium acceptors are virtually completely ionized at room temperature (kT being 0.026 electron volt at room temperature). The highly doped P layer 21 therefore pins the Fermi level near the top of the valence band, as shown in FIG. 3b. The resultant potential barrier 22 prevents electrons from reaching the P layer 21.
Instead of employing a highly doped P type layer to provide the requisite potential barrier, a suitable high work function metal (substituting for the layer 3) may be disposed adjacent the surface of the gallium phosphide layer 4 which is opposite the emitting surface, so as to provide a Schottky barrier which likewise prevents electrons from diffusing away from the emitting surface. A suitable metal for provision of such a Schottky barrier is platinum.
In order for the dynode l to exhibit a negative electron affinity, it is necessary that the bottom of the conduction band within the semiconductor layer 4, at the point where sharp band bending commences, be
disposed at an energylevel above the vacuum energy level. In order to meet this condition, the ionization energy E, of the deep acceptor impurity should not exceed the difference between the energy gap value E, and the work function I. Since it has already been stated that the ionization energy E, should preferably be at least equal to a value on the order of 4 kT, the deep acceptor impurity, the semiconductor material, and the electropositive surface layer should be chosen such that Preferably, the parameters of the dynode 1 should be chosen so that the distance d from the emitting surface at which the bottom of the conduction band crosses the vacuum energy level does not exceed a value cor responding to a few times the mean free path for excited conduction band electrons within the semiconductor material.
When the structure described is employed, the distance A from the emitting surface within which sharp band bending takes place is less than a value on the order of a few times the mean free path for excited electrons, so that electrons in the conduction band may diffuse to the electropositive layer 5 with a residual energy above the vacuum energy level, and be emitted from the surface 6; no additional energy need be supplied to enable these electrons to surmount the potential barrier at the emitting surface. The resultant effective electron affinity, corresponding to the difference between the vacuum energy level and the height of the conduction band bottom above the Fermi level, is therefore negative, as indicated in FIG. 3b.
Iron, the preferred deep acceptor impurity, has a room temperature ionization energy on the order of 0.8 electron volt. The work function I of cesium on the P type gallium phosphide layer 4 is approximately 1.3 electron volts, while the energy gap E, for gallium phosphide is 2.3 electron volts. The ionization energy for the iron acceptors is therefore substantially greater than 4 kT (0.1 electron volt) and less than the difference between E, and I (1.0 electron volt).
With the aforementioned parameters, a beryllium impurity concentration on the order of l0"/cm. in the p layer 21, and a thickness of 0.5 micron for the layer 4, the energy bands are sloped so that a potential difference of approximately 0.2 volt exists across the portion of the layer 4 denoted by the dimension e in FIG. 3a. The resultant electric field within the bulk of the layer 4 has a relatively high value on the order of 6 X 10 volt/cm., since the bulk of the P type layer 4 (i.e.- the portion where the energy bands are. essentially straight lines) is essentially insulating, having only on the order of 10 charge carriers per cm."'.
Although we have shown cesium as the material comprising the electropositive work function reducing layer 5, a cesium-oxygen composite layer may also be employed. Such a cesium-oxygen layer is known in the art as a work function reducing material and is described, e.g., in the following reference:
A. A. Turnbull and G. B. Evans, Photoemission from GaAs-Cs-O, BRITISH JOURNAL OF APPLIED PHYSICS, Series 2, Volume I, Pages -160, Feb. 1968.
A partial cross-sectional view of the active portion of the injection emitter shown in FIG. 2 appears in FIG. 4a, with a corresponding energy band diagram shown in FIG. 4b, FIGS. 4a and 4b being vertically aligned.
As in the structure of the dynode l, the deep acceptor (iron) doped P type layer 15 interacts with the thin electropositive work function reducing cesium layer 18 to provide a negative effective electron affinity at the exposed'surface of the cesium layer, and to establish a potential gradient (drift field) within the bulk of the layer 15 which impels electrons toward the emitting surface.
Electrons are injected from the heavily doped N type region 11 into the P type layer 15 across the forward biased P-N junction 16. Since a shallow donor such as 'tellurium is employed as the doping material for the N region 11, the donor impurities are essentially completely ionized at room temperature.
The deep acceptors in the portion of the P layer 15 adjacent the P-N junction 16, however, are only very slightly (l0 ionized acceptors per cm) ionized at room temperature. As a result, current flow across the forward biased P-N junction 16 is carried mostly by electrons (as opposed to holes), so that the P-N junction 16 serves as an efficient injector of conduction band electrons into the layer 15.
Electrons so injected into the layer 15 are impelled toward the emitting surface by the internal electric field indicated by the sloping energy bands, so that high efficiency of electron emission from the exposed surface of the cesium layer 18 is realized.
The various symbols in FIG. 4b correspond to similar symbols in FIG. 3b.
The dynode l, as shown in FIG. 1, may be fabricated by first evaporating the beryllium layer 3 onto the alumina substrate 2. The optical transmission characteristics of the beryllium layer 3 may be monitored while the evaporation progresses, so that the evaporation may be terminated when an optical transmissibility corresponding to the desired layer thickness is observed.
The iron doped gallium phosphide semiconductor layer 4, which in this example is: polycrystalline (a monocrystalline layer may be employed), is deposited onto the beryllium layer 3 by, e.g., vapor phase reaction of gallium chloride (Gal,)|, ferrous chloride (FeCI and phosphine (H5), at a temperature on the order of 800 C. The deposition is continued for several minutes to deposit the polycrystalline gallium phosphide layer 4 to the desired thickness. Under some conditions, the gallium phosphide layer 4 may be essentially monocrystalline, even though grown onto the crystallographically incompatible beryllium layer 3.
, During the growth of the gallium phosphide layer 4, beryllium diffuses from the layer 3 a short distance (10 to 100 Angstroms) into the gallium phosphide layer 4 to form the P* region 21 (see FIG. 3a).
The iron concentration in the gallium phosphide layer 4 may typically be on the order of 10" to 10 atoms/cm, as previously mentioned. However, high iron or other deep acceptor impurity concentrations are desirable, provided that they do not introduce excessive crystallographic strains into the gallium phosphide layer 4.
After cleaning the exposed surface of the gallium phosphide layer 4, a thin layer 5 of cesium or cesiumoxygen is evaporated onto the gallium phosphide surface. The thickness of the layer 5 is monitored during the evaporation process by observing the photoemission current from the layer 5 resulting from the illumination of the layer with light. The evaporation of the layer 5 is terminated when the photoemission current reaches a peak value. The resulting layer thickness, as previously mentioned, is typically on the order of l to Angstroms.
Where it is desired to employ the dynode 1 as a transmission type secondary emission structure, the alumina substrate may have a thickness on the order of 500 Angstroms, so that primary electrons incident on the exposed substrate surface may penetrate to the gallium phosphide layer 5. Techniques for manufacturing selfsupporting alumina'layers of this type are well known in the art, and generally involve anodizing an aluminum film to provide an alumina layer of the desired thickness, and subsequently dissolving the aluminum film in a liquid, so that the alumina layer remains floating on the liquid surface.
The injection type electron emitter 10 shown in FIG. 2 may be manufactured by providing a monocrystalline N type gallium phosphide substrate 11, masking one surface of the substrate with an insulating layer 13 comprising a material such as, e.g., silicon dioxide and depositing a deep acceptor doped gallium phosphide layer 15 from the vapor phase onto the masked surface, so that the layer 15 grows epitaxially on the exposed gallium phosphide substrate surface and forms the P-N junction 16 at the interface between the layer 15 and the substrate 11. The electrodes 12 and 17 may be evaporated ,onto the substrate 10 and layer 15, respectively, a suitable mask being employed for evaporation of the electrode layer 17.
The electropositive cesium or cesium-oxygen layer 18 may then be evaporated onto the exposed surface of the P type gallium phosphide layer 15, in the manner previously described. The battery 19 and resistor 20 are connected to the electrodes 12 and 17 by wires soldered or otherwise bonded thereto.
l. A negative effective electron affinity electron emitter having an internal drift field, comprising:
a P type semiconductor body;
a deep acceptor impurity at least in a region of said body att'lijacent a given electron emitter surface of said bo said impurity being present in a substantially uniform concentration of atleast 10 atoms per cubic centimeter,
a thin layer of work function reducing material coated on said electron emitter surface, said layer having a thickness on the order of from one to ten Angstroms, and
means for exciting electrons into the conduction band of said P type semiconductor body,
said semiconductor body, said acceptor impurity, and said work function reducing material being chosen so that the ionization energy of said acceptor impurity in said semiconductor body is greater than one-tenth electron volt, but does not exceed the difference between the energy gap of the semiconductor body and the work function of the work function reducing material.
2. The emitter defined in claim 1 wherein said work function reducing material comprises a material selected from the group consisting of cesium and oxygen.
3. The emitter defined in claim 1 wherein said semiconductor is gallium phosphide.
4. The emitter defined in claim 3 wherein said acceptor impurity comprises iron and said concentration is on the order of 10 to 10 atoms per cubic centimeter.
5. The emitter defined in claim 1 comprising a potential barrier region adjacent a surface of said body opposite said emitter surface, for preventing diffusion of electrons away from said emitter surface toward said opposite surface.
6. The emitter defined in claim 1 wherein said barrier region comprises a P region having a thickness substantially less than the thickness of said semiconductor body between said emitter surface and said opposite surface.
t i i I