DE1231353B - Electroluminescent semiconductor diode and process for their manufacture - Google Patents

Description

FEDERAL REPUBLIC OF GERMANY

GERMAN

PATENT OFFICE

EDITORIAL

Int. Cl .:

Number:
File number:
Registration date:
Display day:

H05b

German KL: 2If-89/03

J 26962 VIII c / 21f
November 24, 1964
December 29, 1966

The invention relates to a method for producing an electroluminescent semiconductor diode, preferably with a negative characteristic, and a semiconductor diode manufactured according to this method.

Because of the increasing working speeds of the data processing machines, the The importance of light as a carrier of logically linked information is growing. Through this The need for very fast switching elements and carriers or converters were in the field of Electro-optical and opto-electrical converters made many attempts. So was z. B. by the German patent specification 1 048 346 serving as a light source semiconductor arrangement for current electroluminescence, in which the recombination energy of electrons and holes at the transition points between zones of different conductivity types is used, known, which is characterized in that a for the radiation to be generated permeable semiconductor crystal with at least three zones of different conductivity types rindet, which are applied to an alternating voltage that each loaded in the forward direction Transitions between zones of different conductivity types radiate, while those do not radiate in the forward direction lying transitions also serve as capacitive current limitation. The one that serves as a light source Semiconductor arrangement according to this invention consists of a semiconductor body, preferably a semiconductor crystal, in particular a single crystal, in which, analogously to a transistor arrangement, at least two transitions between zones of different conductivity, preferably opposite one another Line type, are connected in series, in which also the recombination lights that are connected to the Transitional zones occurs, is exploited.

The German patent specification 1052 563 also discloses an arrangement and a manufacturing method for injection luminescent lamps in which crystals that are contacted on both sides and flowed through directly by direct or alternating current, but preferably by direct current, are used as the carrier of the lighting. The arrangement according to this patent specification is characterized in that the charge carriers recombining to generate light are injected into the high-ohmic crystal by two thin, opposing surface layers of different conductivity types and the recombination of the electrons and holes injected by the surface layers in the high-ohmic volume inside the crystal, which causes the light generation is effected. Prak electroluminescent semiconductor diode and
Process for their manufacture

Applicant:

International Business Machines Corporation,

Armonk, N. Y. (V. St. A.)

Representative:

Dipl.-Ing. H. Böhmer, patent attorney,

BÖblingen, Sindelfinger Str. 49

Named as inventor:
Ralph Sidney Levitt, Ossining, NY;
Kurt Weiser, Millwood, NY;
William Paul Dumke,
Chappaqua, NY (V. St. A.)

Claimed priority:
V. St. v. America November 26, 1963
(326 171, 326 114)

This arrangement is carried out table according to the patent specification by a mixture of 50% purest Zinc sulfide and 50% purest zinc selenide, put together in a suitable crucible in a protective gas atmosphere melted down at around 1650 ° C and brought down into colder parts of the furnace by slowing down solidified over the course of many hours. By sawing, etching and diffusing of foreign atoms, the arrangement according to the invention is created from the single crystal obtained in this way.

Another semiconductor element has become known through the German patent specification 1 054 179, which thereby is characterized in that the semiconductor body consists of a first zone in which charge carriers are located recombine with the emission of a recombination radiation, and from a second zone which has a forbidden energy band, the width of which is smaller than the quantum energy of the recombination radiation and into which the recombination radiation from the first zone enters, so that the electrical Conductivity in this second zone changes depending on the intensity of the recombination radiation will.

U.S. Patent 3,064,132 discloses another electroluminescent semiconductor device known. In this patent specification is

609 749/158

wrote that indium, arsenide, indium-antimonide, gallium-arsenide are particularly suitable for production and GaUium antimonides and other n-p-n-p or p-n-p-n combinations are suitable. the Melting takes place at a temperature of 1100 to 1200 ° C in an atmosphere of an acceptor material, the Z. B. can be indium, gallium, aluminum or boron. Another one afterwards Diffusion takes place at temperatures between 600 and 1200 ° C. In between is the semiconductor device ground and etched. It is also in a publication in "Journal of the Electrochemical Society", Vol. 110, No. 11 of November 1963, on p. 1153 to 1159, a diffusion process is described. After this publication it is known to be used as an acceptor for gallium arsenide crystals, zinc and as a donor tellurium. It will pointed out that these two elements are particularly useful in the manufacture of electroluminescent Semiconductor diodes in gallium arsenide form are suitable.

The invention is now based on the object of creating a component that, in addition to the conversion From optical to electrical and electrical to optical energy, there is also the logical connection and storage of information.

The electroluminescent semiconductor component, which is used to solve this problem, is characterized by that in a semiconductor body, a first region of lower resistance is made smaller by means of acceptors Energy level and an immediately adjacent second area of high resistance by means of acceptors higher energy level is formed, which is at least a third, preferably from the unchanged Starting material of existing low resistance area containing donors, and that the two areas of low resistance are provided with conventional ohmic contacts.

A particularly suitable method for producing the electroluminescent according to the invention Semiconductor component is characterized in that first the second layer is made by diffusion on all sides of manganese, cobalt or chromium in an n-type galum arsenide body and then the first Layer by subsequent all-round diffusion, of zinc, cadmium or magnesium in such a way that only a smaller depth of penetration than that achieved in the first diffusion is produced, and that then the areas produced by the diffusion with the exception of one side on the rest Sides of the gallium arsenide body are removed.

A component consisting of three areas and having two ohmic connections can according to FIG the invention as a light amplifier, as a light converter and as a quantum amplifier use.

The invention will then be explained in more detail with reference to the figures. It shows

F i g. 1 shows a circuit with an electroluminescent diode with negative resistance to explain the idea of the invention,

F i g. 2 shows the characteristic curve of an electroluminescent diode with a negative characteristic curve range according to FIG Invention,

3a and 3b show the spectral distribution of the in FIG. 1 circuit shown in its two stable States

Fig. 4 shows the energy bands of the in F i g. 1 shown diode,

F i g. 5 and 6 sectional views of the diode according to the invention in different steps of its manufacturing process,

F i g. 7 shows another embodiment of the arrangement according to the invention and its use in FIG a circuit as a photon-coupled switch, FIG. 8, 9 and 9 a characteristic curves of the in F i g. 7 shown Arrangement,

F i g. 10 the circuit of one of the invention

ίο light or quantum amplifier containing semiconductor component,

11 shows sectional views of an embodiment of the arrangement according to the invention in different Steps of the manufacturing process,

12 and 13 are sectional views of the inventive Arrangement in different steps of the manufacturing process.

The present invention is based on an electroluminescent semiconductor diode with at least three regions of different conductivity types and a light-sensitive junction. It is claimed as inventive that a first region (P) of low resistance by means of acceptors of a low energy level and an immediately adjacent second region (P ⁰ ) of high resistance by means of acceptors of a higher energy level is arranged in a semiconductor body, to which at least a third, preferably from the unchanged Starting material of existing, donor-containing region (N) of low resistance connects and that the two regions (P and N) of low resistance are provided with conventional ohmic contacts.

At room temperature, the current-voltage characteristic of this diode is the same as the current-voltage characteristic common gallium arsenide diodes with acceptors having shallow energy levels, for example Zinc or cadmium in the area of the p-type layer. At temperatures in the range of the liquid nitrogen these diodes show a very high series resistance at voltages up to about 1 V, in which the well-known diodes begin to conduct. Once the breakdown voltage has been reached, then the diode in a region of negative resistance, in which the current increases with falling voltage, where the characteristic breakdown voltage is a function of the diffusion profile.

In both areas, i.e. below and above the breakdown voltage, electroluminescence occurs on. It should be noted that the spectral intensity distribution in both states, d. H. very different below and above the breakdown voltage is. In addition, the overall intensity of the light changes with the transition from one state very considerable in the others. The light intensity can be in the range of the higher current by 10- to 100 times greater than in the low current range. There is another property of this diode in that when irradiated with a certain wavelength the current-voltage characteristic curve during the time of illumination changed in such a way that the breakdown voltage is changed. These Property can be used to make the diode by irradiation from its low conductivity in its to transfer highly conductive state. The switching speed of the diode from its low conductivity in their highly conductive state and vice versa is on the order of 10 nanoseconds. on Due to its properties, the diode is suitable as a

Light amplifier, so that it can be used as an image converter in a suitable modification. Since the The diode can be switched optically and has two different optical states, a large number of optical logic circles can be implemented with it.

The in F i g. 1 shown diode 10 consists of an η-conductive gallium arsenide body, the one p has a high resistance conductive layer 14 and a low resistance p-type layer 12, generated by diffusion with manganese and zinc. Ohmic contacts 18 and 20 are on the n-type Area 16 and arranged on the p-conducting area 12. The in F i g. 1 includes circuit shown the variable bias source 22, a load resistor 24, and a switch 26.

As is readily apparent, the actual interface lies between the p ° -conducting layer 14 and the η-type layer 16. The low-resistance p-type, with zinc doped layer 12 plays, as will be explained in more detail below, together with the Manganese-doped p ° layer plays a decisive role in relation to the characteristics of the arrangement. Preliminary it should be pointed out briefly that the diode by diffusion of manganese into an n-conducting Galliurn arsenide crystal and subsequent zinc diffusion that penetrates to a much shallower depth is realized in the same crystal. As a result of this diffusion, the next to the n-type Area lying p ° layer 14 is generated, the conductivity of which is determined essentially by manganese is, and then the p-conductive layer 12 is produced, the conductivity of which is substantially through Zinc is determined.

The mode of operation of the arrangement can be seen from the FIGS. 2 can be seen, which represent a current-voltage diagram. The solid line represents the diode characteristic without external influence by light energy, while the dashed line shows the characteristic curve irradiated by an external light source. If the in F i g. 1 shown arrangement of the switch 26 closed and the diode at the temperature of the liquid nitrogen, as kept at about 77 ⁰ K and the voltage of the source 22 slowly changed from 0 to upwards, it will first be observed that the current is quite slow with the Tension rises. This area of the curve is known as the area of low conductivity or high resistance. If the voltage rises above the point t, one arrives at the negative resistance range of the diode characteristic, in which the voltage drops with increasing current. Assuming a fixed resistance 24, the state of the diode will be established, which is formed by the intersection point c of the load line L1 with the solid curve. In other words, the points c and I, which represent the intersections of the load characteristic Ll with the characteristic of the diode, characterize the two stable states of the diode for a given resistance 24. It is readily apparent that the diode from point I to point . c can be switched by applying an additional voltage which is sufficient to run through the characteristic curve through point t. With a given bias voltage and by measuring the diode current, it can be determined whether it is characteristic of point c or point I , so that one can determine whether at any point in time an additional voltage has been applied which is suitable to turn the diode off to convert their through the point I into the state indicated by the point c.

The dashed curve in FIG. 2 another important property of the present diode is identified. The dashed curve represents the characteristic of the diode illuminated by an external light source. It can be seen from this curve that the diode resistance is reduced in certain areas in the case of external lighting and that the critical or switching point of the diode is shifted to point t '. From the comparison of the dashed and solid curves in FIG. 2 it can be seen that the diode at point I is stable in the non-illuminated state at a voltage Vl. If the diode is irradiated by an external light source, the characteristic curve changes into the dashed curve, in which the voltage Vl is greater than the voltage assigned to point f.

In this way the diode is switched over and becomes stable at point c in the region of high currents, in which the two curves essentially coincide. If the lighting is switched off, the diode remains at point c on the characteristic curve assigned to the non-illuminated state.
From the preceding explanations it can be seen that the diode according to the invention either by increasing the voltage above the critical point t or by illuminating the same so that the critical point t becomes smaller than the bias voltage of the diode from the state of low conductivity to the state of high Conductivity can be switched. Switching from the state of low conductivity to the state of high conductivity can therefore take place either by light or by increasing the voltage. On the other hand, it is easy to see that switching back to the state of low conductivity can only be carried out by temporarily interrupting the circuit by opening switch 26 or by reducing the voltage V _{v of} the partial current, so that the diode is transferred to its state indicated by point I. will.

In the F i g. 3 a and 3 b is the spectral distribution

the radiation of a diode according to the invention before and after switching from the lower state Conductivity shown in the state of high conductivity.

In Fig. 3 a shows the spectral distribution in the state of small currents. It can be seen that the intensity of the light of 0.89 μm wavelength, which is characteristic of the manganese-doped p ° layer is, is much larger than that of the radiation of 0.84 μm wavelength, which is characteristic of the zinc-doped p-layer.

After switching to the high conductivity state becomes, as shown in FIG. 3 b, the intensity of the radiation in the range of 0.84 μπι be greater than that of the radiation in the range of 0.89 μm wavelength.

By way of explanation it should also be pointed out that the scale of the light intensity in FIG. 3 a thirty times greater than that of FIG. 3 is b, d. that is, the intensity of the light in Figure 3b is thirty times greater. When recording the curves given above

7 8

was the voltage of the voltage source and the p ° and η denoted. The Fermi level is through

Working resistance chosen so that the diodes are shown in a dashed line passing through them

State of low conductivity 5 mA and in to- three areas passes through it.- With regard to charged

stand high conductivity 6OmA current. It and uncharged impurities becomes the p ° - is explained by the fact that the total light yield 5 layer by a constant density N _D of positive

after switching is about thirty times larger than charged donor ions, an equal density of

before switching. It should also be noted that negatively charged manganese ions and one not

the ratio of the change in intensity of the beam - uniform density of neutral manganese atoms

ment in the range of 0.84 μΐη wavelength significantly. In the p-layer, the N _{D are} positive

is greater than that of the radiation in the range of 10 charged donor ions compensated by the

0.89 μΐη wavelength. 'equal number of negative zinc ions and all

Therefore, a radiation detector with a filter, manganese centers are neutral. The zinc atoms in

that only radiation with 0.84 \ tm wavelength is through-excess to N ₀ , cause the conductive ones

lets to determine if the diode has holes in the area. At the interface between the p- and p ° -

low or high conductivity works, especially 15 layers there is a comparable number of

advantageous. By changing the layer thickness and negative manganese and zinc ions. The specified

the zinc and manganese impurities in the diode acceptor levels of 0.4 electron volts for Zn

the specified ratio of the spectral supply and for manganese of 0.1 electron volts come from

division also changed. But it can be all-publications.

What can be said in common is that the zinc treatment 20 In F i g. 5 are the different steps of the invented

impurities generated spectral lines a more accurate production method according to the invention schematically

Means for determining the switching state are shown as. According to Fig. 51, a block of n-line

the gallium arsenide caused by the manganese contamination in the desired form

Lines. shown. In the second step, a

Although the processes inside such diodes have 25 p ° conductivity causing contamination in the

are theoretically not yet completely clarified, material of the block is diffused in. As special

The following is a brief overview of the procedure advantageous for the second process step

given these diodes. The layer proved to be high in resistance to manganese. This is followed by a second

As of p ° 14, the introduction of manganese diffusion, in which a different p-conductivity

(Mn) as the substance that determines the conductivity type in the semiconductor block.

causes impurity. It is diffused by the two layers. For this purpose, in particular

th low "resistance 12 and 16 included. Which zinc proved to be particularly advantageous.

Layer 12 contains zinc as the p-type penetration depth of the second diffusion is not critical, only

State-causing pollution. The den-lei- must be taken to ensure that it is lower

from 35 than that of the first diffusion, since a p ° layer is obtained.

Gallium arsenide existing starting crystal can ten must remain. By changing the width of the

by any of the substances Te, Se or Si erp ° layer, the characteristic properties

be procreated. The layer of high resistance p ° is determined by the diode. It is still

exists because at 77 ⁰ K practically all man- care must be taken that the p-conductivity of the

Whole centers which are in excess of those which determine the 40 second layer of impurities

originally in the η-conducting gallium arsenide acceptors with shallower energy levels than the

Compensate donors contained in crystal, have new acceptors with lower energy levels,

are neutral. At low currents from the battery 22 generated by manganese. After the second

the F i g. 1, before the negative resistance area diffusion, the arrangement has the shape of the

is reached, the electrons cross the p ° - 45 Fig. 5ΙΠ. The next step will be the back

Layer and combine with the holes in and sanded down the side surfaces of the assembly

Neighborhood of the pp ° interface. Investigations or otherwise removed so that the in Fig.5IV

about the origin of the light in such diodes shown pp ° n-arrangement results. In addition,

during the change of the voltages and the ohmic contacts in and of itself known

Currents have shown that the thickness of the p ° layer 50 is wise at the p- and the η-layer of the device

is about 30 μm. appropriate.

If the critical threshold voltage has been reached, which in FIG. 6 become the steps of a something

for example at 4 V for a given load- modified manufacturing process schematically

resistance is the case, the negative resistance continues. In this procedure, an or

until the voltage drops to about 2.2 V. 55 several areas of the semiconductor body with masks

The curve in the current-voltage diagram is covered in, so that the subsequent distance

positive again at this point, and the spectral desired diffused layers cease to exist. In

division and the intensity of the light as a radio- Fig. 61 is a gallium arsenide body of the

tion of the current before and after the area negative desired shape shown with a mask M.

Resistance shows that the light emission accompanies 60. A first layer is used to generate the p ° layer

is made up of one filling of the zinc and manganese diffusion with manganese and then a second

Centers with holes. Diffusion with zinc to create a P-layer.

In Fig. 4 is a conventional energy band diagram. The results of the individual process steps are shown for an electroluminescent diode with negative denomination in FIGS. 611 and 6III. On-resistance range shown at 77 ^° K, which after 65, the side surfaces of the arrangement were produced using the method according to the invention. removed in a suitable manner so that, according to the way in this diagram, the conduction and valence assumptions of the mask M result in the bands shown in FIG. 6IV corresponding to the three layers of the diode p, arrangement. It is of course also possible

9 10

in addition to the rear side of the arrangement, there is also the lateral tration in the relevant p ° and p-layers

surfaces with masks or other diffusion fluctuates.

the preventive protective layers so that the same or similar results are obtained by grinding, etching or polishing aims when one instead of manganese chrome and on the arrangement proceeded after the end of the diffusion of zinc, cadmium or magnesium completely omitted. turns.

The in connection with the F i g. 5 and 6. A modification of the prescribed diffusion steps described above are shown, for example, for the production of a modified half in a tightly sealed and evacuated quartz conductor element according to the invention is shown in FIG. 11 tube or furnace carried out, which is shown with suitable io. An n-conducting heating winding is provided with I in FIG. 11. In order to carry out semiconductor bodies with a suitable donor concentration process, the crystal body is referred to as cleaned tration. It has proven to be particularly advantageous to etch and then in the quartz tube or in the gallium arsenide for this purpose, which in the quartz furnace is denoted by I and η together with both sides of the crystal in FIG. 11. Subsequently arranged, but somewhat separated from one another, a material generating p-conductivity, containing zinc and manganese, is accommodated. For example, manganese diffused in, whereby a p-conductive layer is created which, as shown in FIG holder surrounding the gallium arsenide crystal and the. Subsequently, the arrangement, made of manganese, is heated in a suitable manner so that, as can be seen in FIG. 11III, one side is ground where the first diffusion takes place. Then the lapped, etched or otherwise treated in order to heat the earthed manganese layer on one side of the wafer, so that the second diffusion layer, as in Fig. 4, is removed. Subsequently, FIGS. 5III and 6III are shown by diffusion. another p-type layer is produced, so that

In the practical implementation of the method 25 the in F i g. 11IV results in the arrangement shown. the

it is expedient to proceed in such a way that the diffusion takes place expediently in this case

zinc impurities that cause η conductivity. The depth of the second diffusion becomes like this

of the gallium arsenide nucleus selected a concentration of that the desired properties of the

2 to 4 x 10 ¹⁷ cm " ³ and the crystal very final product can be obtained. The second p-type

finely polished and sealed in an evacuated quartz tube. The layer contains an impurity, the acceptors

together with a single manganese metal snippet with a flatter energy level than the acceptance

is housed. The first diffusion step will gates the area doped with manganese, for which purpose

is suitable for a period of about 4 hours with, inter alia, zinc, for example. That

a temperature of about 900 ⁰ C carried out and the resulting platelets is then to the

results in a layer of about 100 μm penetration depth. 35 side surfaces by grinding, etching, lapping or the like.

The pn interfaces are then treated in such a way that the diffused layers

electrolytic etching in a KOH solution can be removed and an arrangement according to

paves. The platelet thus obtained will be shown in Fig. 11V. The element thus obtained exists

Enclosed again in a quartz tube, but this time from a p-conductive layer P, from a second

together with a snippet of zinc metal and 40 layer of semi-insulating material, which as

a second diffusion takes place during a period of high resistance layer P ⁰ , an inner one

from 15 to 45 minutes at a temperature of η-conductive layer JV and finally a p-conductive

about 900 ⁰ C instead. the layer P, which in its composition of the

The time and temperature of the second diffusion with the first-mentioned p-type layer is the same. The zinc were chosen so that the penetration depth of the 45 GaAs platelets would be much less than that of the manganese after the zinc described above, i.e. activation in parallelepiped-shaped bodies of about 70 μm. The second diffusion changes 250 · 120 · 120 μηα ³ split. Ohmic contacts of the relative position of the first interface between the then with the help of tin on the η-conductive gallium arsenide and the p ° layer, the η-conductive side of the element and with the help of manganese doping is not created, so that 50 indium on the p-conductive side of the element results in a p ° layer of about 30 μm thickness. That posed.

The platelet is then split into parallelepipeds of a variation of the above-described Veretwa 250-120 μm ² area. Ohmic driving for the production of a PP ° NP semiconductor contacts are on the η-conductive sides with the help of element is explained with reference to FIG. In the case of tin and on the p-conducting sides with the aid of this process, a mask M is used, made of indium. In summary, it should also be established by a grinding, lapping or etching step according to that the p-layer 12 is about 70 μm thick, which Hill can be avoided. After the termination layer p ° high resistance about 30 μm and the supply of the second diffusion, the side surfaces, remaining η-conductive layer of the gallium arsenide corresponding to the crystal described with reference to FIG. 11 is about 75 μm thick. The concentration of 60 nen process, removed, so that a semiconductor zinc results in the p-type layer 12 of this arrangement element, as shown in Fig. 6V.
is about 10 ²⁰ atoms cm ^-3 and the concentration. Another modified method for producing Magan in the p ° layer 14 is about 10 ¹⁸ cm ^-3 Instead of the diffusion times, diodes are obtained which in the first diffusion have essentially the same properties on the back of the platelet, but which can be freed from the manganese-imprecise spectral intensity distribution by etching and lapping slightly with a smaller layer, changing the entire platelet the manganese and zinc atomic concentrations cut in two parts, as shown in Fig. 13III

represents is. After this process step, those in connection with FIGS. 11 and 11 follow 12ΙΠ to V process steps shown.

With the aid of the semiconductor component shown in FIG. 7, large currents can be switched very quickly. The semiconductor element 2 consists of a GaUium arsenide crystal, on one side of which there is a luminescent injection diode and on the other side of which there is a light-sensitive PP ° N diode with negative resistance, so that the photons emitted by the luminescent diode to control the resistance of the PP ° N diode are used. The coupling is achieved in that both elements have a common η-conductive layer JV. In the production of the gallium arsenide component described, a layer 6 is obtained which contains zinc or another material for producing acceptors with a flat energy level, for example cadmium, and a p-conductive layer P with low impedance within the gallium arsenide plate forms. The layer 8 forms an insulating layer or a high impedance layer and is denoted ^{by P 0.} It contains manganese or other substances used to generate acceptors with low energy levels, for example chromium as the most important impurity. The area 10 consists of the η-conductive gallium arsenide and is designated with JV ". 12 is a further area with low impedance, which is designated with P and contains the same impurities as the layer 6, namely zinc.

The battery 14, which is connected to the component 2 via the resistor 16 and the closed switch 18 provides a forward bias for layers 6, 8 and 10 of FIG PP ° N diode. The battery 20, which is connected to the component 2 via the closed switch 22 is, supplies the luminescent injection diode consisting of layers 10 and 12 with an in Forward voltage sufficient to cause the diode to luminesce allow.

A reflective layer 24 is disposed on the outer surface of the P-layer 12, the one Reflection of the luminescent light in the direction of the P ° layer 8 causes. The arrangement would, however work even without said reflective layer 24.

The operation of the in F i g. The arrangement shown in FIG. 7 is illustrated below with reference to the Z ₁₄ -F ₁₄ diagrams in FIG. 8 and 9 explained in more detail. Areas 6, 8 and 10 form an arrangement which, when _{forward-biased by the voltage of battery F 14} , exhibits negative resistance. When the switch 18 is closed and at low values of + F, the impedance of the PP ⁰ N diode is very high due to the properties of the area 8, so that the point I on the curve J _2n = O in FIG. 8 shows the stable, high impedance state of the part of the device that can be addressed as an electroluminescent diode with negative resistance. If the voltage F _{14 is} increased to point t , which corresponds to the threshold voltage of the electroluminescent diode with negative resistance, the latter is switched over and the i ₁₄ -F ₁₄ curve I ₂₀ = O shows the properties in the negative resistance range. A steady state indicated by point B is referred to as the low impedance state of the negative resistance electroluminescent diode.

This also has the property that the threshold voltage is lowered when energy in the form of light is supplied to ^{the region of high resistance P 0.} The curve Z ₂₀ > 0 of FIG. 8 is a representation in the z _'14 -F ₁₄ diagram when light is generated by the current Z ₂₀ , the

ίο falls through the PN interface 10 to 12 on the area P ^{0 of} the diode 6-8-10. It can be seen from this that the switching of the electroluminescent diode with negative resistance takes place at a lower voltage t ' when light is applied. The point B essentially indicates the same stable state with a low impedance of the diode, regardless of whether the diode is operated under the action of light on the region P ⁰ or without this action.

The diode 6-8-10 can be transferred to its first stable state (point I) by removing the lighting and briefly opening the switch 18. It is of course also possible to switch to the first state by applying negative pulses that briefly lower the bias voltage V _u so far that the voltage is less than V _x , which causes a return of the diode 6-8 in their Low impedance condition (point I) . L is the load characteristic of the electroluminescent diode with negative impedance.

In summary, it should be stated once again that for the case in which the diode 6-8-10 is in its low impedance state identified by B and is to be converted into the high impedance state identified by point I, the diode in Forward voltage must be reduced to below the value V _x. The area PP ⁰ JV in FIG. 7 is important because it acts as a switch with negative resistance. The electroluminescent properties are only a by-product of the switching process, but this property is important as an optical means of nondestructively determining the state of the photon-coupled switch. The luminescent light can be sensed without affecting the switching state of the diode.
An essential feature of this arrangement, in which two diodes with a common base NB area 10 are created, is the very close coupling between a luminescent injection diode (areas 10 and 12) and an electroluminescent diode with negative resistance (areas 6, 8 and 10 ), so that rapid switching of the electroluminescent diode with negative resistance is possible. Switching times of at least 10 ~ ⁷ seconds can be achieved in this way. The arrangement according to FIG. 7 acts as a very fast three-electrode switching component or relay in which the current z ' _an serves as a control current. The switch-on time of the light generated by the diode 10 to 12 is a maximum of 10 ~ ⁹ seconds, and its intensity is over a wide range the current i,

proportional.

In Fig. 9 a, the current-voltage characteristic of the arrangement shown in FIG. 7 is shown for different values of Z ₂₀ , in which z ' _2O in steps

is increased by 5 mA. In this circuit, the resistance of R16 is 200 ohms, and batteries 14 and 20 have been replaced by time-varying generators. As mentioned earlier, the arrangement is operated at a temperature of 77 ^° K. The threshold voltages vary between 3.8 and 1.5 volts, and the current i varies from 1.0 to 2.5 mA. As shown in FIG. 9, the gain or gain of the arrangement PPW is shown in the form

or the ratio between the current through diode 6-8-10 in its low impedance state and the current flowing through the diode in its high impedance state. This gain can also through the relationship

are shown, where R _{16 is} the value of the resistor 16 of Figure 7 and R _P ^ _{N is} the impedance of the diode 6-8-10. The benefit of the whole arrangement

(PP ⁰ NP) can be written as

AIa

AL,

hi (in point B) - hi ( in point I)

Uo - 1 * 20

As can be seen from Fig. 9, at Z ₂₀ = O (no light) and 4Ό of the size of rnAzf J _{14 can be} in the order of a few 100 mA (due to the size of R ₁₆ ), which corresponds to a high gain .

In the circuit according to FIG. 10, the circuit through the areas 10 and 12 according to FIG. 7 formed light source replaced by an external light source.

In this way the arrangement can be used as a light amplifier. 10, areas 6, 8 and 10 PPW form an electroluminescent diode with negative resistance. 30 is an alternating current source £ (r) which supplies a forward bias voltage for each half cycle, but which is not sufficient to switch the diode via the variable resistor 36 without illuminating its P ° range. The light source 34 is focused on the η-conducting region 10 by means of the lens 36, while 38 is a dichroic filter which _{transmits the wavelength A 1 of} the light source 34, but reflects the wavelength A ₂ generated in the diode. Although the dichroic filter 38 as well as the reflective layer in FIG. 7, improves the efficiency of the arrangement, the arrangement can also work without this filter. The filter 40 is designed in such a way that the light of the wavelength A _{1 is} suppressed, but light of the wavelength A _{2 is} allowed to pass through. A lens 42 focuses the transmitted light onto the detector 44.

If the externally incident light I _{ex has} less energy per photon than the emitted light I _em , ie the energy from externally incident photons is less than the energy of the emitted photon, the arrangement acts as a quantum amplifier. When using manganese as an impurity to generate acceptors of flat energy levels in gallium arsenide and an ionization energy of 0.1 electron volts, a quantum gain of 10 is achieved because the energy of the emitted photon is 1.0 electron volts. By changing the depth of the impurity, one can achieve smaller quantum gains with deep impurities and larger quantum gains with shallower impurities. If the incident light is in the infrared range and the light emitted by the PP ° N diode is in the visible spectrum, a matrix-shaped arrangement of this type can be used

ίο Diodes can be used in the long-wave To transform an image lying in the infrared range into an image lying in the short-wave infrared range. By suitable doping of the gallium arsenide with compositions of groups III and IV with wider Band gap, for example with gallium phosphite, can be used in place of near infrared lying images in the visible spectral range lying images are generated.

If the arrangement, in its highly conductive state, emits more photons per second than are required to trigger the switchover from the light source 34, a light amplifier results. In order to be able to transmit a modulation of the source 34 within wide limits, it is necessary that the bias frequency E (t) of the current source 30 is greater than the modulation frequency of the light source 34. If the light source 34 is switched on when the potential of E (t) is positive, the threshold value of the diode 6-8-10 is reduced. The diode is thereby switched to its low impedance state, and relatively high values of the current Z _{30 result} , which result in the generation of light of wavelength A ₂ , which is detected by the detector 44.

If E (i) falls to a value below the said threshold value or assumes a negative value, the diode 6-8-10 falls back into its high impedance state, and the small values of Z ₃₀ flowing in the process cause minor or even negative values

no luminescence of the diode. As a result, no light will reach detector 44. As long as the frequency of the current source 30, which supplies the forward bias voltage for the diode PPW, is greater than the frequency at which the light source 34 is switched on and off, the circuit according to FIG. 10 as an amplifier that maintains the source 34 modulation characteristics. If the source 34 has a wavelength X ₁ which is equal to the wavelength X ₂ , then the filter 40 can be removed.

The arrangement acts as a converter when the light from the source 34 has a wavelength A ₁ greater than A ₂ , A _{2 being} the wavelength of the light generated in the diode during the switching process. For example, X _{1 could be} in the long-wave infrared range

lie and the detector 44 would detect infrared light whose wavelength is shorter than that of the Source 34 is.

Claims (13)

Patent claims: 1. Electroluminescent semiconductor diode with at least three regions of different conductivity types and a light-sensitive transition, characterized in that a first region (P) in a semiconductor body low resistance by means of acceptors with a small energy level and an immediately adjacent one second area (P ⁰ ) high resistance is arranged by means of acceptors of a higher energy level, which is followed by at least a third, preferably consisting of the unchanged starting material, donors-containing area (N) of low resistance and that the two areas (P and N) with low resistance usual Ohnischen contacts are provided. 2. Electroluminescent semiconductor according to claim 1, characterized in that its Base body consists of η-conductive gallium arsenide. 3. Electroluminescent semiconductor according to claims 1 and 2, characterized in that the acceptors with a low energy level for the first layer (P) made of zinc, cadmium or magnesium, the acceptors with a higher energy level for the second layer (P ⁰ ) made of manganese, There are cobalt or chromium and that the donors of the third layer (N) consist of tellurium ao and / or silicon. 4. Electroluminescent semiconductor according to claims 1 to 3, characterized in that the doping of the first, second and third layers zinc (for P), manganese (for P ⁰ ) and tellurium (for N) is used. 5. Electroluminescent semiconductor according to claim 4, characterized in that the doping of the first layer (P) is about 10 ²⁰ cm ^-3 zinc and that of the second layer (P ⁰ ) is about 10 ¹⁸ cm ^-3 manganese. 6. Electroluminescent semiconductor according to claims 1 to 5, characterized by its Use as one lying in a circuit with controllably variable voltage bistable module, which by voltage change and / or exposure to light from a stable state can be transferred to the other. 7. Electroluminescent semiconductor according to claim 10, characterized in that the im rising and falling characteristic curve range in different frequency ranges maxima (0.84 and 0.89 μΐη) having emitted luminescence radiation as a criterion for the Switching state of the arrangement is used. 8. Electroluminescent semiconductor according to claims 1 to 5, characterized in that a fourth layer (P 12) is additionally provided which follows the third layer (NlO) , which at least partially with a spherical regions having reflective layer (24) after is completed on the outside. 9. Electroluminescent semiconductor according to claim 8, characterized in that its N-region (10) is connected to a center electrode which is connected to the two respective outer electrodes (6 and 12) and each containing a power source (14 and 20) containing circuits is common in such a way that the controlled circuit (20) from the increasing characteristic curve range [(Z) in FIG. 8 and 9] merges into the falling characteristic range (B). 10. Electroluminescent semiconductor according to claim 1 to 9, characterized by its use as a light amplifier, light converter or quantum amplifier. 11. A method for producing an electroluminescent semiconductor according to claims 1 to 10, characterized in that first the second layer (P ⁰ ) by all-round diffusion of manganese, cobalt or chromium in an η-conductive gallium arsenide body and then the first Layer (P) is produced by subsequent diffusion on all sides, of zinc, cadmium or magnesium in such a way that only a smaller penetration depth than in the first diffusion is achieved, and that the areas (P ⁰ and P) generated by the diffusion are then also produced With the exception of one side, the remaining sides of the gallium arsenide body can be removed (IV in FIGS. 5 and 6 and V in FIGS. 11 and 12). 12. The method according to claim 11, characterized in that a fourth area (P 12) is generated from the opposite outer surface of the third area (N) by diffusion by means of acceptors of a low energy level and that the first (6), the third (10) and the fourth area (12) is provided with ohmic contacts in a manner known per se. 13. The method according to claim 12, characterized in that the first (P 6) and fourth area (P 12) is doped with the same substance. Considered publications:
German Auslegeschrift No. 1048 346,
1052563,1054179; U.S. Patent Nos. 2,817,783, 3,064,132; Journal of the Electrochemical Soc, Vol. 110, No. 11, November 1963, pp. 1153 to 1159, especially p. 1154. For this purpose 2 sheets of drawings 609 749/158 12.6 «© Bundesdruckerei Berlin