DE19963550A1 - Bipolar light source with temporally and spatially uniform current supply, based on e.g. aluminum gallium arsenide-gallium arsenide structure, includes luminescent and non-luminescent columns - Google Patents

Abstract

Optical plate (23) forms a mechanical bridge (21) for the uniformly-transparent semiconductor, with luminescent and non-luminescent columns (24, 34) at required mutual spacing. Luminescent column includes a layer structure (27) and luminescence active layer (26). All columns terminate below, in the same plane of a planar mounting surface (33). Columns taper conically away from the plate with its general, transparent, uniformly and highly conductive distribution layer (28), supplying the active layer (26) with charge carriers from an electrode (36). The distribution layer is connected by a metallic track (37) along the sloping surface of a non-luminescent column (34) to its contact (36) on the mounting surface (33). Each active zone on its mounting surface is provided with an electrode of opposite polarity (32) to supply charge carriers of a second kind. Adjacent mounting surfaces (33) are provided, having a similar metal coating (37). This permits simultaneous, identical introduction of a supply or control circuit.

Description

The invention relates to a bipolar illumination source from a one-sided, self-bundling Semiconductor body, its for the luminescence waves transparent substrate with a double hetero structure layers of ternary and binary III-V compounds, preferably AlGaAs-GaAs, connected and for one pseudo-vertical operation is established.

The lighting source is said to be used in service, protection, security and alarm devices such as

• a) remote control applications and light barriers, such as
• b) the stationary or mobile, pulse-like or continuous lighting one too monitoring site, room or Vehicle for the purpose of access, Presence or movement control or
• c) short-range data transmission and
• d) the presentation of information on a display

be used.  

Others come to the need for protection and security Barrier and property protection systems, such as walls, grilles u. a. Access blocks only with massive effort opposite.

With the help of surveillance systems and new ones Illumination sources can better meet this need be met. A special accent here Illumination sources that humans do not can be perceived. Until then preferably the use of motion detectors is maintained, which act as passive detectors on heat sources like birds, and responded to pets just as they responded to humans Intruders (e.g. US-A 3,476,946). Through camouflage of the heat sources with the help of an appropriate However, masking these detectors had to be outwitted. Light barriers with one or more pairs Luminescent diode and photo transistor are already heavier to deceive (US-A 3,875,403; US-A 4,239,961; US-A 4,514,625), but also speak with harmless ones Interruptions.

Sound and ultrasound signals were also added monitoring rooms transferred and on the changes checked the reflected signals, z. B. according US-A 4,499,564. Flora moved by the wind or too harmless animals could fake intruders.

Radar and microwave based surveillance by Air spaces and waters have been around for decades common, but much too for most users complex.  

With the help of a stationary or rotating infrared Laser diodes were moving objects reflected rays on suitable receivers determine somewhat cheaper (US-A 4,903,009, US-A 4,952,911 or DE-A 41 30 619). Regardless of that Effort for image evaluation with IR photo receivers, The relatively expensive IR lasers hinder the expansion use in other areas of application.

Illumination sources, their performance at lasers reaches, but where manufacturing and Operating costs lag significantly behind these, are also known. The invention aims for such lighting sources to improve the Yield both in reliable ecological Manufacturing, as well as a cheaper conversion electrical into optical energy regardless of the Operating time.

As with rectifying diodes, initially also the two for AlGaAs-based luminescent diodes opposite main surfaces of the Semiconductors as carrier levels one each Used contact, see US-A 3,677,836, DE-A 21 58 681, DE-A 197 41 609. In this Diode construction, the electric current flows from a contact area vertically through the semiconductor opposite contact surface. Already this Vertical structure brings about by covering the Luminescent region under the contact metal and the low lateral spread of the charge carrier current in the semiconductor body losses at the chipextern measurable luminescence. These losses are both at  AlGaAs structures from Kish-F-A and Fletcher-R-M, volume 48 of the series "Semiconductors and Semimetals", Vol. 48 (1997) pp. 172ff, as well as on quaternary AlGaInP diodes Lin-J-F et al., Electronics Letters 30 (1994) No. 21, P. 1794, on the intensity profiles of the Surface emission has been demonstrated. The current flows steeply and without much from the contact metal Wide spray in the underlying semiconductor. The As a result, current flow is in its lateral spread very limited. The decisive factor for this is relatively low dopability of the compound semiconductors and the resulting charge carrier concentration as well as the limited mobility of the charge carriers in some compound semiconductors. Hence the Atrium around the contact is also limited extended.

As a rule, the different Luminescent diode types based on AlGaAs GaAs substrates used. They are inexpensive and special Orientation of the disc normal brings Growth improvements for the layers. Of the emitted almost isotropically in all directions However, luminescence swallows up this GaAs substrate Because of its optical properties (high absorption) the downward emission and leads to Optical yield losses.

To remedy this, the Bragg effect can be used become. That of William Henry Bragg and William Lawrence Bragg found reflection condition for X-rays at network levels of single crystals also to be transmitted to radiation of wavelength λ. At  a distance a of two network levels occurs Bragg reflection only at the gloss angles δ (Angle between beam and interface ≠ from the optical Angle of incidence) for the directions for which the Path differences of the reflected partial waves have an integer multiple b of λ (2a sinδ = b λ). Bragg reflectors only work relatively in one limited wavelength range and are therefore on trim every other λ range.

The average refractive index n can be compared to Mixing ratio of AlAs and GaAs in the ternary Compound semiconductor GaAlAs are changed. With a multilayer GaAs-Ga0.7Al0.3As stack in which low and high refractive indices alternate and each Layer pair a quarter of the thickness of the advised Wavelength λ of the incident radiation can a Bragg reflector can be realized. Small values of Δn force many pairs of layers. The refinement of the Bragg reflectors can be used for substrates Molecular beam epitaxy (MBE) with fewer defects single crystal and with significantly less scatter than by evaporation. Such reflectors were according to US-A 5,406,095, UK-B 23 20 136 or DE-A 197 41 609 attached directly to the GaAs substrate.

The system complexity of the MBE and the limits of simultaneously editable number of disks or Disc diameters narrow the area throughput per However, the machine considerably.

While the vertical LED structures just mentioned  on the front radiation past the front contact recourse can increase the external Luminous efficacy with vertically operated AlGaAs Luminescent diodes also include the substrate.

According to Ishiguro-H et al. in Appl. Phys. Lett., Vol. 43 (1983), No. 11, pp. 1034-1036, is this luminescence usable if the substrate for the wavelength of the Emission is transparent. The emission zone must therefore to the front as well as to the back with Semiconductor layers covered with window properties become. GaAlAs heterostructures were therefore discovered generated transparent GaAlAs documents that previously GaAs wafers had been deposited. The GaAs Washer was therefore only used temporarily and was removed after completion of the forest.

This technological process is according to US-A 4,864,369 (Col. 3 lines 39-49) much more difficult and costly, because different transparent layers on one later generated unused substrate and this temporary Substrate must be removed again. In addition, must the transparent layer that later than Substitute substrate should serve, only tedious in one costly growth process.

According to US-A 4,719,497, the production of a "High efficiency light-emitting diode "for the light guide Telecommunications immediately with an n-AlGaAs substrate started on which then with conventional Epitaxial technology the other AlGaAs layers of the DH Structure. The shadow effect of the p- Contact with a front coupling (p-side up) and  the resulting relatively high losses are now by a Zn diffusion under the Al contact reduced. The external light output is by the factor 2-5 (p. 1, lines 59-63; p. 2, lines 59-61) with uncapped LED's increased.

According to EP-A 0 350 242, the "p-side up double heterojunction AlGaAs LED "the same Shift sequence scheme used. However, it turns out that one without a good substrate for the growth of the Shift package does not get along and that for the "n- side up DH structure "required starting substrate from p- GaAs type not so easy with a low one Obtaining dislocation density is like that corresponding n-type GaAs (p. 3, lines 61-65). Accordingly, the "p-side up DH structure" used, leave the GaAs substrate on the LED and only through a 10-30 µm thick AlxGa1-xAs layer for better reliability of the structure and before especially for a larger window for the light emission on the front contact side (p. 4, lines 59-61). The Back contact remains unchanged.

Nevertheless, this is a technical-technological process difficult and expensive. That different transparent layers on a later unused Substrate generated and the substrate removed must be, both in terms of material and complex with regard to the systems. Cutbacks on the Quality of the substituted substrates, however, are inadmissible. In addition, that must be transparent Layer that will later serve as a carrier tedious in a costly growth process  be formed on high quality epitaxial systems.

As further construction options for Luminescent diodes is the shape and arrangement of the Contacts on the same surface side that unobstructed light emission from a free Surface and the transition from vertical to Lateral operation known.

To suppress losses due to radiationless Recombination on the surface from one side contacted GaAsP-GaP or GaP-Planar- Diffusion diodes have been proposed in DE-A 27 12 412, the currents of the minority charge carriers Field plate effect in deeper areas oust. The majority bearers should conductive luminescent epitaxial layer over a highly doped n-type connection area in a higher n-type layer are introduced. This Layer structure and the realization of similar ones Diffusion zones do not succeed with AlGaAs-DH structures in the same way.

UK-A 2,019,643 proposes a luminescence diode with a layer sequence of Ga1-xAlxAs on a GaAs substrate, in which an increased external quantum efficiency and a reduced internal absorption can be achieved. For this purpose, an active n-luminescent layer made of Ga0.90Al0.10As: Te (5.1017 cm ³ ) with a degree of compensation ND / NA of 10 and the thickness of an absorption length z. B. 3 microns of two cladding layers of p-Ga0.70Al0.30As: Zn (5.1017 cm ³ ) and n-Ga0.70Al0.30As: (5.1016 cm ³ ) coated. A p + -conducting ring zone, doped with Zn (1019 cm ³ ) encompasses the active luminescent layer on the outside and creates a connection to the lower lying p + -conducting cladding layer. The p + -conducting ring is provided with a p- contact made of Au-Be or Au-Zn and the n-conductive cladding layer with an n-contact coplanar on the same surface. The peculiarities of the active luminescent layer (thickness relative to the absorption length, degree of compensation and doping) ensure photon recycling of the photons returned to the luminescent zone by multiple reflection. The photons are converted back into radiation-recombinable charge carrier pairs and can then leave the semiconductor as new photons with a more favorable angle to the exit surface.

In US-A 5,472,886 an LED structure for fiber optic communication technology has been specified, with which the emission in the near field of a precisely aligned light emitter made of InGaAsP on InP substrate can be coupled exactly into a fiber. The structure aims to overcome difficulties in mass production, especially adjustment, and to avoid wire bonding techniques. In order to achieve this goal, as shown in FIG. 2, an LED chip 1 is formed from an n-type and a p-type layer 3 . A pit 4 separates this p-type layer starting from the surface 5 of the underside of the p-type layer 3 . The pit 4 extends to the n-conducting layer 2 and has a first side wall 4 a and a second side wall 4 b on both sides of the bottom 4 c. It serves as a spacing element between the thick p-type bonding pad 13 to the p-contact 7 and the thick n-type metallic bonding pad 14 to the n-contact 6 . The bond pads are on thin metal surfaces 12 a and 12 b. The n-contact 6 consists of Au / Sb / Au and extends from the surface 5 over the first side wall 4 a of the pit 4 to the bottom 4 c of the pit 4 and contacts the n-conductive layer on the bottom 4 c of the pit 4 2nd A mask 8 made of dielectric material such as Si ₃ N ₄ , SiO ₂ or photoresist with a thickness of 200 nm limits the n-contact 6 . In the direction of the projection of the p-contact 7 behind the pn junction there is a lens 9 to improve the coupling of the emission into a fiber. The lens is inserted into a cavity similar to the Burrus structure. Contact is now made with solder bumps, a dielectric film preventing the solder from spreading beyond the metallic bonding pads 13 and 14 . The size relationships between the p-contact 7 and the pn-transition surface 11 indicate that only a small part of the DH structure in the p-conducting zone is required for the direct operation of the fiber behind the lens 9 . A similar procedure is followed for the Burrus structure.

The geometric shape of the semiconductor body in the scattered tiles and the individualization of the luminescence active zone before disassembly of the disks depends directly on the decision for one of the two Main variants, the edge or surface emitters, from. While the edge radiation predominantly at Laser diodes are used Luminescent diodes preferably on surface emitters  back. Laser diodes owe their efficiency to Attachment of gap areas on the exit surface of the Emission. In the case of luminescent diodes, the chip body through a separation process with the help of the scratch and Crushing technology, through a sawing process with the help of Diamond discs or wires or a separation etching process following a selective masking of one or shaped on both main surfaces.

The cuboid or columnar shape of the chips is thanks to them simple manufacturing very common. It is in the Rule of a two-sided (DE-C 21 58 681), in Special case but also a one-level contact with overturned chips (US-A 5,670,797) accessible. The Edge length of the chips could be reduced and the number of chips per slice can be increased like that Contacting the chips remained possible. The Wire bonding technology left edge lengths up to 250 µm to. With connections made of solder or conductive adhesive according to US-A 5,265,792, the edge length even to 125 µm be reduced.

It is part of a long cultivated assembly practice such or similar LED chips in To insert housing reflectors, see DE-A 20 62 209, US-A 4,013,915, DD-C 138 852, DD-C 232 377 or US-A 5,298,768, or, as described in US-A 5,606,181, to be placed next to it.

The multitude of the other geometry variants of the Luminescence diode chips could the economy of Material use does not account in the same way wear.  

The developed for fiber optic communication technology Burrus structure is through a DH-AlGaAs semiconductor chip with two-sided contacting and one hollowed-out GaAs substrate part directly under the optically marked active zone. The surface with the luminescence active layer carries a 50 µm large Current contact and is on this and the neighboring Surface in contact with a silver stamp. In the Hollow becomes a hybrid coupling in a fiber or made a ground lens (US-A 4,010,483; EP-A 0 047 591).

The equipment of the semiconductor chip with a Excavation for the insertion of ground Lentils were only a maturation of the burrus structure (US-A 5,472,886).

According to DE-A 25 42 072, the large-area direct contact of the BURRUS structure with the Ag stamp was abandoned and the material was removed down to the level of the active surface area with the formation of a mesa with bevelled edge surfaces. The radiation propagating parallel to the transition is emitted essentially perpendicular to the transition plane after reflection. For a maximum output radiation, a 45 ° inclined edge must be selected, from which one can deviate if necessary. If this structure is based on the layer thicknesses published by Burrus in 1971, it remains unclear how the 2.1-3.5 µm thin n-Ga _0.7 Al _0.3 As cladding layer, according to DE-A 25 42 072, Fig. 2, bipolar illumination source from a side contacted, even converging semiconductor body, other parts are not etched, overcomes the high saturation currents and can be uniformly distributed to the GaAs-luminescent layer. In addition to the weaknesses in the electronic properties of the layer, which are also indicated in US Pat. No. 4,010,483, the light dissipation in this layer and the resulting losses were also disregarded.

Spherically polished chip body with an active zone in the Focus of a semiconducting lens body after US-A 4,017,881 or DE-C 28 25 387 ensure a good one Decoupling over the entire surface and favor the lighting of a very large solid angle. she but require a high level in their elaboration individual effort per chip. Such or different Central symmetrical elaborations were themselves not with area-emitting laser structures required.

Mirror-like arrangements with the active zone in the Focus of the concave mirror by limiting etching the pn junction on one side and through manual removal of the semiconductor body from the Back and the formation of a parabolic reflector have also been manufactured suggested. With the insertion of Bragg reflectors the complexity increased on both sides of the active layer the product structure. In addition, the Surface emitter version deviated and for Edge heater returned.

According to US Pat. No. 5,498,883, a diode structure is manufactured, the one covered by two Bragg reflectors  Has structure and has a cathedral in focus. It creates a resonance structure from which the Edge radiation is emitted laterally. after the Radiation has left the cathedral, it strikes the Walls of a reflector made of the same semiconductor is built like the emitter region, but also with was coated with a reflective layer.

This results in a concave-like appearance Structure in which, despite the effort, only the Part of the emission lobe of the edge emitter is detected that is not laterally beyond the edge of the reflector device. Furthermore, it is a problem with this construction that the Radiation media limits e.g. B. pass semiconductor air must again with the semiconductor or with a comes into contact with the surface covering it.

A comparison of this laser structure with one conventional edge emitter consequently shows that the Edge heaters in terms of efficiency and the dependence of the optical power on the flux current regarding the threshold current and the steepness according to Jae- Hoon Kim et al. as published in Applied Physics Letters 57 (1990), H. 20, pp. 2048-2050, is clearly superior.

Cone-shaped chip body with a transverse to the cone Emission zone and deflecting lateral surfaces add Franklin-A-R et al. and publication in Journal applied physics 35 (1964), pp. 1153-1155 in GaAs zur Exploitation of primary reflection at. To DE-A 198 07 758 can also use the Secondary reflections in the side surface  light-emitting elements an inverted Pyramid stump similarly designed and with one uniform obtuse angle (β + 90 °) with β between 20 and 50 ° can be provided. This is one accompanied by large area losses Chip structure. To reduce this loss of material the embankment lengths would have to be shortened considerably without capturing the primary and secondary beams Reduce.

To reduce costs, instead of In-chip lenses often also include non-chip lenses and others Illustration elements used. However, this creates already losses when leaving the chip and at the Illustration in outer optical components. Sublime, over the general chip area and as a mesa designated luminescence zones became very early proposed and according to US-A 4,7427,378 DE-C 35 32 821 or according to US-A 5,349,211 Facilitation of the adjusted fastening of imaging beads for improving coupling used in fibers.

A chip-internal control of the luminescence is also possible the mesa structure according to US-A 4,742,378, where perpendicular to a chip is placed on a column in which there is a pn junction contacted in the form of a jacket continues. The luminescence is in the direction of optical axis of the column and from one radiated tubular pn junction. The metallic The jacket electrode shades the end face of the column not so that the luminescence easily in optical fibers can be taken over. One-sided contacting  and a simple manufacturing technology is this However, mesa structure of the laser or luminescence diodes inaccessible.

With the transition to one-sided contacting arises towards the bilateral the requirement of adequate, low-resistance supply of the active luminescent layer with charge carriers of both Species and specifically by an addition to installing distribution layer. This layer has from the same surface that has already been contacted off for the electrical supply with the second Load carrier type. Regardless of that the problem that initially emitted isotropically Luminescence from the active zone is directional bundle up. Thereby, the danger is simultaneous deriving of the photons from the active zone into one not for bundling and decoupling appropriate direction to eliminate. The distribution layer, which is necessary for the energy supply may consequently, on the other hand, not light-dissipating in Appear appearance and extraction of the transformed reduce optical energy.

It is no longer permissible that the desired one Area emitters in its emission area alone the chip edge length is defined. There must be one Restricted to a part of the chip area become. The isotropic emission distribution must nevertheless on-chip and completely on a single surface be redirected.

The invention is based on the object  to create bipolar structure that the electronic supply of the active zone in time and locally even and up to high current levels ensures that the luminescent surface limits the emission fully in one direction and a uniform, self-adjusting joining technology for everyone Allows contact electrodes.

This task is characterized by the characteristics of the Claim 1 solved. Appropriate configurations are the subject of the subclaims.

The invention is based on the creation of a Chip structure in which the relevant Product functions such as feeding the electrical Energy, its conversion into optical energy and the Decoupling this energy largely by itself and run with a high effect in the chip and hardly any Support from additional, external Picture elements need.

The basis of the claims The principle of formation consists in the subordination of Routing of the distribution layer and the conductor tracks for the supply of charge carriers from the cathodic polarized electrode under the optical requirements. The main idea here is to prevent derivation and the directional redirection of the in the Layer plane of laterally spreading part of the isotropic Luminescence distribution.

The desired effect is achieved through a spatial Limitation of the lateral extent of the optically  electronic layer package on the shape of pillars and a very special conical extension of the Diameter of the pillar towards one of the Optical plate connecting the pillars.

The height of the pillars must be the middle diagonal of the Consider the pillar. Require larger diagonals larger pillar heights. The optically usable pillar height can be graded in the layer profile of the Distribution layer can be increased.

Another variant uses a previously created one Protrusion on the optical disc, over which then the entire shift package is pulled.

An important feature of the invention is also that Division of the luminescent pillar into two or several pillars with a smaller diameter. With this Feature are both advantages in the Manufacturing costs, as well as with the optoelectronic Properties of the finished product.

The arrangement of an imaging lens Semiconductor material with increased focal length, or of Optical fibers with an adapted fiber diameter the optical disc in the optical section of the chip increase the bundling effect in the long-range order area.

The LED chips become simple Contact with your mounting surface on one Printed circuit board made of silicon. This circuit board can be very conveniently designed as a mounting tray.  

For the automatic control of the It has luminescence intensity of the illumination source proved to be very beneficial if one of the as Rectifier-operated DH layer structures as Photodiode is operated in the reverse direction and the measured photocurrent to the evaluation standard for the Maintenance or correction in the control loop of the Injection current is made.

Finally, a special frame-shaped Pillar shape, all the remaining pillars of a chip encloses coherently, to build one automatic hermeticization of optoelectronic active areas of the chip. Then can even an additional capping of the chip-chip carrier Systems are eliminated. Is there an adjustment to the Dimensions of the luminescent pillars to the Dimensions of silver cushions in depressions of the Respected chip carrier plate, the Very high thermal resistance of the entire lighting source lower advantageously.

Further advantages and details of the invention go from the following, explained with reference to the drawings Exemplary embodiments. In the drawings demonstrate:

Fig. 1 shows the division of the layer packet of a lighting source according to the invention with the two relevant functional sections in the inspection (a) and side view (b),

Fig. 2 shows an embodiment of Burrus structure according Dautartes (prior art)

Fig. 3 is an improved integration of the manifold layer in the optical section,

Fig. 4 shows the separation of the optical section of the illumination source into a plurality of parallel connected p-type regions to a common cathode,

Fig. 5 is a according to the invention designed lighting source with (a) incomplete and (b) with complete decoupling of the luminescence from the optical section in the desired direction,

Fig. 6 shows a variant with a surface-optimized illumination source,

Fig. 7 is a sectional view of an illumination source with on-chip deflection and bundling

Fig. 8, the attaching of an expanded transparent window by bonding disk provided with slots to reduce dilation and contraction stresses at the interface between the parts,

Fig. 9 shows a mounting structure of the illumination source,

Fig. 10 is a comparison of the flow stream Strahlungsleistung- characteristics of various types (?) Of a chip having a pillar emitting a) without b) with self-bunching effect,

Fig. 11 is an equivalent circuit diagram of a self-controlled illumination source and

Fig. 12 is a fabric structure of a light source with automatic hermetic sealing of the opto-electronically active zone.

1st embodiment

The improved decoupling tuning can best be taken into account with the design concept of a self-bundling isotropic radiator. In Fig. 1 (a), the inspection is displayed on the on-chip contacts of a bipolar Such illumination source 20. An optical section 22 is connected to the other chip elements via a mechanical bridge 21 . The optical section 22 consists of a luminescence-active pillar 24 with a conically widening cross section 25 . The pillar 24 consists of a luminescence-active layer 26 , which is encased on both sides by a heterojunction and a double heterojunction (DH) layer structure 27 is formed. A distributor layer 28 for the first type of electrical charge carriers is located in the direct vicinity of the DH layer structure 27 . The distribution layer 28 covers the entire surface of the mechanical bridge 21 . Of particular importance for the optical section 22 of the illumination source 20 is that the part of the mechanical bridge 21 above the pillar 24 is designed as an optical plate 23 . As a result, the luminescence from the luminescence-active layer 26 can leave the chip as radiation 29 . It is important for the optical section 22 that, in addition to the cross-sectional variation of the pillar 24, the closely adjacent distributor layer 28 is provided with a gradation 30 of the surface profile at the transition point between the pillar 24 and the optical plate 23 . This gradation 30 is formed by removing material from the distribution layer 28 in the entire area around the luminescence-active pillar 24 . This gradation of the distribution layer 28 limits the height 31 of the pillar 24 between a contact electrode 32 and the start of the optical plate 23 and increases it only in the optical section 22 . The consequence of this measure is an enlargement of the distance of the upper edge of the luminescence-active layer 26 from the plane of the distribution layer 28 obtained throughout. The pillar 24 ends with the later anodically poled contact electrode 32 on a flat mounting surface 33 .

A contact electrode 36 , which is later cathodically polarized, also ends in the same plane on the non-luminescent pillar 34 . The contact section 35 of the illumination source 20 includes both electrodes 32 and 36 , in particular also a metal coating 37 , which is a conductor path from the contact electrode 36 over part of the lateral surface of the pillar 34 to under the roof of the mechanical bridge 21 and into the gradation 30 of the Distribution layer 28 extends.

In the production of the bipolar illumination source 20 , a (100) -oriented, n-conducting, mechanical Al _x Ga _1-x As bridge 21 with an x value of 0.27 ± 0.12 and a doping of <3 10 ¹⁷ cm ^-3 an n-conducting distribution layer 28 with a significantly higher doping of 1 ± 0.3 10 ¹⁸ cm ^-3 , a layer composition of Al _x Ga _1-x As with an x value of 0.2 ± 0, 08 and a layer thickness of 30 microns. This is followed by the Al _x Ga _1-x As layer structure 27 with a luminescence-active GaAs layer 26 , which is enveloped by an n- and a p-conducting ternary layer with an x value above 0.11. The DH layer structure 27 has a total thickness of 30 microns.

The definition of the lateral size of the luminescence-active layer 26 and its combination with a simultaneous integration in deflection elements in a truncated pyramid-like pillar 24 under the optical plate 22 is carried out by targeted etching. In order to obtain a conical pillar 24 , use can be made of the increase in the etching speed with decreasing Al content in Al _x Ga _1-x As. Accordingly, an x-value gradient is built up in the DH layer structure 27 in each layer. A small x value must be maintained near the mounting surface 33 , where the greatest lateral etching removal is desired. In the direction of the optical plate 23 , the x value increases linearly with the distance from the mounting surface 33 . However, the scope for using this effect for crystallographic etching is essentially limited to the x value range between 0.01 and 0.12. For this reason, additional support for the oblique etching is additionally introduced when using larger x values for the layers. The optical plate 23 is coated on the surface side of the DH layer sequence with photo-positive lacquer and formed into a first etching mask. The first mask enables etching removal of only a narrow strip around the contact electrodes 32 and 36 . After removing the first mask, a second etching mask is applied and with the aid of it both the etched and parts of the previously unetched areas are subjected to a renewed etching attack. This step-by-step etching process is repeated once or twice with the masking area around the electrodes 32 , 36 becoming ever smaller. The step pyramid thus formed is brought into the desired pillar shape with a cross-section 25 widened towards the optical plate 23 by smoothing-out, leveling etching. The gradation 30 of the surface profile of the distributor layer 28 in the vicinity of the pillar 24 is preferably carried out after smoothing the pillar jacket. The remaining thickness of the distributor layer 28 is then at least 13 μm. The pillar 24 above the mounting surface 33 has a height 31 of 35 μm at the end of the etching process. Only now is the metal coating 37 applied by vapor deposition of an Au-Ge alloy. The complete, finished disk can be separated into individual platelets for the illumination sources 20 by suitable sawing with wires or diamond disks or by scratching and breaking from the free surface side of the optical plate 23 . If necessary, the two surface areas must be protected with cover foils.

2nd embodiment

A second possibility for designing the pillar structure can be seen in FIG. 3. A columnar arrangement of trapezoidal projections 41 is attached to the optical plate 23 made of Al _x Ga _1-x As on a surface side. They are formed by wet chemical etching after photolithographic masking and jump 30 µm from the surface. Their area dimension is about 220 × 170 µm ² and the distance from each other is 75 µm. A 25 μm thick distributor layer 42 and then all other Al _x Ga _{1-x As} layers of the DH structure are then drawn over this arrangement. The layer region above the protrusion is covered with an etching mask and then the layer components of the DH structure around the pillars are removed again by etching up to the distributor layer 42 . A vapor-deposited contact layer forms a cathode contact 43 , the metal covering for an interconnect route 44 and creates the connection of the contact on the cathodically polarized pillar to a collar 45 on the distributor layer 42 near the luminescence-active pillar. The optical section 22 of this structure is now significantly improved in terms of the power supply and the effective deflection of the luminescence by integrating the undiminished distributor layer 42 into the pillar structure. The height 31 between the lower edge of the distributor layer 42 and the mounting surface of the pillar is 50 μm here. An anode contact 46 is located at the foot of the pillar.

3rd embodiment

A new form of development can be created if the definition of the size of the luminescence-active zone is combined with the integration of deflection units for the luminescence from this zone. The multiple stages of the etching process for shaping the cone-like pillar cross section, in particular for large pillar heights L, is undesirably complex. In detail, the pillar height L follows the diameter D or the largest diagonal of the pillar according to equation (1) at a fixed critical reflection angle θ _{crit (n} ).

L = F (θ _{crit (n),} D) = D (tg θ _crit / (1 + 2tgθ _crit) (1)

The pillar height L can only be reduced if the diameter D of the luminescence-active layer is reduced. In the case of the luminescent diodes 50 in FIG. 4 (a), the emitted optical power depends on the size of the emission area. Smears on the surface are therefore hardly permissible. According to FIG. 4 (b), however, the emission area can be broken down into three partial diodes 51 with partial areas of the same size. Fig. 4 (c) shows an associated chip layout 52 with three anodically polarized contacts 53, 54, 55 next to a competent cathodically polarized contact 56. The anode contacts 53 , 54 , 55 lie side by side and are characterized by the truncated pyramid-shaped outer surfaces on the pillar walls. By dividing the area into three, the required pillar height L is reduced to less than 60% of its original height. If the area is divided into four equal partial areas, only half of the otherwise required height is required. FIG. 4 (d) shows a layout 57 of the arrangement of the chip elements in the wafer assembly and the mirrored disposition of the elements of the adjacent rows. This disposition is advantageous for the final separation of the chips, since the parting line between the cathodically polarized contacts of two neighboring chips is not subject to any requirements for cone-like expansion and can therefore be carried out in a space-saving manner. To facilitate separation, the chip lines are arranged parallel to a crystallographically oriented chamfer 58 .

A deviation from the in-line arrangement of the three anodically polarized pillars is shown in the chip layout 59 according to FIG. 4 (e). This also enables the return to square chip contours. In addition to the square ones, round or sickle-shaped pillar contours can also be produced.

In FIG. 5, the operation of the invention designed according to the illumination source is played back in consideration of the decoupling of the luminescence from the optical section in a desired direction. The perspective view of an incomplete chip structure 61 in FIG. 5 (a) shows a cathodically polarized pillar 62 and two anodically polarized pillars 63 , 64 . A luminescence active layer 65 is very close to the lower edge of the optical disk. As a result, a beam 69 emanating from the point of intersection 67 runs fully against the cone-like wall of the pillar 64 and is deflected from there to the optical plate. The same still applies to the limit case of a beam 68 . The deflected luminescence is represented in a half-cone 70 . The undirected luminescence is e.g. T. represented by a semi-cone 71 , in which rays 72 a-d or rays 73 are transmitted through the light guide effect in the distributor layer 66 . The radiation in the semi-cone 71 is lost for use as an illumination source and reduces the external optical efficiency.

In contrast, a chip structure 81 as shown in FIG. 5 (b) is more advantageous, in which the cone-like pillar structure extends deeper into the optical disk. This ensures that a luminescence-active layer 85 in anodically poled pillars 83 and 84 occupies a central level and parts of a distributor layer 86 have become a wall element of the anodically poled pillars 83 , 84 . A radiation cone 90 starting from the point 87 with the marginal rays 88 and 89 is now fully gripped by the cone wall of the pillars 83 , 84 and directed in the desired direction.

4th embodiment

In the fourth exemplary embodiment, the optimization of the semiconductor outer surfaces of the illumination source is described, see FIG. 6. In the optical section 22 , the optical plate 23 on the trapezoidal projections 41 is equipped with at least one anodically polarized pillar 98 . A uniformly thick, n-conducting distribution layer 92 is supplied with charge carriers by a cathode contact 93 of a pillar 91 via an interconnect route 94 . A luminescence-active layer 95 emits at 810 nm in operation. A cladding layer 96 lying above the luminescence-active layer 95 in the pillar 98 is separated from an anode contact 99 by a p-type Bragg reflector 97 . In order to avoid alloying of the Bragg reflector 97 by the Au-Ge contact metal of the anode contact 99 , the Bragg reflector 97 is coated with a highly doped GaAs layer before the metal alloy is vapor-deposited. Such reflectors are formed from a stack of layer pairs of GaAs-AlxGa1-xAs by the target JP and Ilegems-M, as published in 1975 in the magazine Applied Optics Vol. 14, Issue 11, pp. 2627-2630, whereby the x value jumps back and forth between two fixed values 0 and 0.38. The number of layer pairs and layer thickness of the individual layers is adapted to the wavelength of the reflection band and the desired degree of reflection. In the photon energy range of 1.53 ± 0.019 eV, the refractive index difference Δn is 0.25 between the two layer partners. The thickness of the binary layer d1 is 55 nm and that of the ternary layer d2 is 60 nm. With a layer stack of 25 layer pairs, a degree of reflection of over 85% is achieved with the selected luminescence band.

These Bragg reflectors 97 can be produced by molecular beam epitaxy (MBE), by atomic layer epitaxy (ALE), by epitaxy with phase-changing deposition and interruption (phase-locked epitaxy), but also by chemical vapor deposition with the aid of organometallic compounds (MOCVD).

To facilitate the decoupling of the luminescence deflected by the shape of the pillars 91 , 98 and the Bragg reflectors 97 , the exit side of the optical plate 23 is covered with an anti-reflection layer 100 . The anti-reflection layer 100 consists of aluminum oxide. Optionally, silicon oxide can also be used.

5th embodiment

In Fig. 7, an illumination source is to be displayed with improved range. In order to create a far field chip 101 according to FIG. 7, an optical disk 102 is increased in terms of its disk thickness. The assembly level of cathodic and anodic contacts 103 , 104 , 105 is retained. A luminescence-active layer 106 also maintains its position in the anodically poled pillars. As a result of the envisaged and implemented increase in thickness of the optical plate 102 , the distance 111 between the luminescence-active layer 106 and the exit surface of the radiation from the chip body increases. To increase the concentration of the radiation, the radius of curvature of the lens shape on the exit side is also increased as the distance 111 is increased. In this way, the angular distribution of the luminescence can be gradually focused from a wide lobe 108 to a narrower lobe 109 to a pointed lobe 110 .

If one relies solely on epitaxial coating processes when increasing the thickness of the optical plate 102 , the outlay for producing the starting materials for the far field chips 101 increases too quickly. It is therefore proposed according to FIGS. 8 (a) and (b) to provide an optical plate 121 designed according to the invention with a pillar relief 122 on a freely expandable expansion level 123 with an optical supplement. If semiconductor supplementary disks 124 with significantly different coefficients of thermal expansion are used, then a joining plane 125 on the supplementary disk 124 outside the optical section is to be provided with blind-hole-shaped recesses 126 . In the joining process of the known pane bonding, the tensions and curvatures of the panes are reduced and the technological manufacturing process is not hindered. On the supplementary disc 124 there are already curved projections which, as concentrators 127, contribute to the bundling.

According to FIG. 8 (b), the supplementary disk 124 can also be equipped with bundled optical fibers 128 .

In Fig. 9 a housing structure is shown for the final assembly of the illumination source. Here, for. B. the chip shown in Fig. 6 with the optical plate 23 upwards and the cathode and anode contacts 93 and 99 down in a prefabricated silicon housing 133 . In a mounting trough 134 of the silicon housing 133 , a solder connection 131 establishes a feed line to the cathode contact 93 and a solder connection 132 establishes a connection to the anode contacts 99 . The solder connections 131 and 132 wet conductor tracks 136 and 137 on the other side. Via individual through-contacts 138 , the conductor tracks 136 and 137 are linked to external connections 139 and 140 on the underside of the silicon housing 133 . The optical plate 23 is covered on the flat surface with the antireflection layer 100 with a sealing plate 141 which is transparent to luminescent light. The sealing plate 141 is hermetically connected to the inner wall 135 of the mounting trough 134 . An imaging bulge 142 is located in the sealing plate 141 above the anodically poled pillars.

The characteristics of an illumination source according to the invention with a single luminescence-active pillar and a luminescence-active layer made of Al _x Ga _1-x As with an x-value of 0.29 ± 0.02 and an emission wavelength of 880 nm were examined as a function of externally applied supply conditions . The forward DC current I _{F increases} strongly after the diffusion voltage of about 1.2 V is exceeded as shown in FIG. 10a. Despite the opposite silicon or other compound semiconductors relatively wide-band feed and the lateral arrangement of the contacts, a surprisingly steep I _{F F} -U arises -characteristic. The low resistance of less than 2 ohms that can be read from the increase is due to the functionally adapted, highly conductive distribution layer of the charge carriers of the first type.

The relationship between the radiation power P _opt (I _F ) and the flux current recorded at room temperature was determined for two different stages of the luminescence-active pillar and was shown in FIG. 10b. With approximately the same diameter of the anodically poled pillar of D ≈ 250 µm, the pillar height L was varied. One time, L was only extended to the beginning of the distribution layer. In the second case, the pillar height L was increased by 250 µm by providing the distribution layer with the gradation described above. With a fixed forward current of I _F = 250 mA, the outcoupled radiation power of the chip with the stepped distribution layer measured over the chip contacted on one side reaches an average value of 50 mW within curve 151 . If the luminescence-active pillar ends at the beginning of the distribution layer, then the measured radiation power within curve 150 comes to only 40 mW with the same flow current.

The chips used for this measurement can still be advantageous for a later in exemplary embodiment 7 Use the explained final assembly and thus enable increased thermal loads.

6th embodiment

In this exemplary embodiment, the possibility of automatically checking the operation of the lighting source is described. For this purpose, the chip structure 160 and the electrical circuitry are expanded with defined operating potentials. Fig. 11 shows that is defined by three diodes 163, 165 and 167, the first diode is short-circuited by a bridging resistor 162 and 163 to ground potential 161st The ground potential 161 thereby reaches a distribution layer 164 and all other diodes 165 and 167 . The DH structure of the diode 165 receives a pulsed or continuously equal supply voltage in the forward direction via the individual contact level of a contact 166 . As described, the injection current flowing thereby causes the luminescence in the luminescence-active layer of the diode 165 . The deflection on the lateral surface and the contact surface of the pillar leads the luminescence essentially in the direction of the exit surface. However, part of the luminescence strikes the exit surface at an unfavorable angle and remains in the semiconductor body as a result of the total reflection. This luminescence also reaches the diode 167 located in the vicinity. If this diode is not operated in the flow direction like the diode 165 , but receives a negative bias voltage via a contact 168 , the luminescence-active layer remains passive with regard to the emission. The photons originating from the adjacent diode 165 are finally absorbed in the DH structure of the diode 167 after penetrating the DH structure one or more times and converted into electron-hole pairs. The reverse polarity emanating from contact 168 creates an electric field in the DH structure that separates the charge carrier pairs. The current flowing thereby flows as a reverse current through the contact 168 and depends in a reproducible and linear manner on the intensity of the on-chip luminescence. Even with a low coupling factor between the luminescence intensity of the luminescence-active diode 165 and the photocurrent of the diode 167 , the luminescence intensity can be guided or tracked thanks to the linear relationship between luminescence and photocurrent. As a result, the lighting source designed and operated in this way can be easily inserted into a conventional electrical control loop of the lighting intensity.

7th embodiment

In the last exemplary embodiment, a self-hermetic luminescence chip structure 200 on a heat-dissipating chip carrier is explained. The chip structure 200 according to FIG. 12 (a) is formed by a closed, annular pillar 201 with external contact 204 and a square pillar 202 with a central contact 203 . Both contacts 203 and 204 end in a common plane and are covered with a solderable surface. The square pillar 202 in particular is equipped with a conical surface and serves to deflect the luminescence in the direction of the non-contact surface. This chip 200 is mounted on a thermally conductive mounting pad 210 made of n-type silicon, see Fig. 12 (b). The section along the section line AA 'in FIG. 12 (c) shows further details. To facilitate assembly, the assembly base 210 is smooth and flat on its rear side 211 and has recesses 213 on its contact side 212 in accordance with the size of the central contact 203 . The profiled contact side 212 is covered with a highly conductive, p-conductive layer 214 .

While the p-type layer 214 in the recesses 213 is freely contactable, an insulating dielectric layer 215 , e.g. B. from silicon oxide or silicon nitride, the flat regions of the profiled surface side. The free semiconductor surface of the p-type layer 214 in the depression 213 is covered with solid metal pads 216 made of silver. These metal pads 216 can be deposited by electroplating, e.g. B. according to DE-C 16 14 982 faster and more adhesive than by other methods. A thin solder deposit 217 on the metal cushion 216 creates the conditions for a mechanical, electrical and heat-dissipating connection between the central contact 203 and the metal cushion 216 . At the outer edge of the mounting base 210 , the dielectric layer 215 is locally removed again and the access to the p-type layer 214 is exposed. The p-type layer 214 can be connected to the operating voltage of the illumination source via the bond island 218 . The second connection is made via a strip conductor 219 deposited on the insulating layer. It too is covered with a thin solder deposit 217 and establishes the connection to the ring-shaped external contact 204 of the luminescence chip. The strip conductors 219 for the external contacts of the self-hermetic luminescence chips are drawn line by line to the edge of the mounting base 210 and there form a connection field 220 for the second connection terminal. The total intensity which can be achieved can be adapted to the requirements in the application situation by means of the number of luminescence chips used on the mounting base 210 . The soldered connection between the external contacts 204 of the individual chips and the adequately designed annular zones of the strip conductors 219 on the mounting base 210 separates the micro-atmosphere around the metal cushion 216 and the outer surface of the luminescence-active pillar 202 from the outside atmosphere in the long term. The entire illumination source according to FIG. 12 (b) can therefore be used without any additional capping.

Illumination sources in an aggressive environment are to be used, are previously chemically closed impregnate.

The chips of the lighting source can be electrical both individually X-Y addressed or as closed clyster field can be activated.  

List of the reference symbols used Fig. 2

1

LED chip

2nd

n-type layer

3rd

p-type layer

4th

pit

4th

a first side wall of the pit

4th

b second side wall of the pit

4th

c Bottom of the pit

5

Surface (of the p-layer)

6

n contact

7

p contact

8th

Mask (made of dielectric material)

9

lens

10th

Surface of the n-layer

11

pn junction area

12th

a,

12th

b thin metal surfaces

13

,

14

Bond island

Fig. 1

20th

Lighting source

21

mechanical bridge

22

optical section

23

optical disc

24th

(luminescent active) pillar

25th

conical cross section

26

luminescent active layer

27

Layer structure

28

Distribution layer

29

Radiation

30th

Gradation (of the profile)

31

Height of the pillar

32

(anodically polarized) contact electrode

33

Mounting surface

34

(not luminescent active) pillar

35

Contact section

36

(Cathodically polarized) electrode

37

Metal coating

Fig. 3

41

trapezoidal projection

42

Distribution layer

43

Cathode contact

44

Leitbahntrasse

45

collar

46

Anode contact

Fig. 4

50

Luminescent diode

51

Partial diodes

52

Chip layout

53

Anodic contact

54

Anodic contact

55

Anodic contact

56

Cathodic contact

57

layout

58

chamfer

59

Chip layout

Fig. 5 (a)

61

incomplete chip structure

62

(cathodically poled) pillar

63

,

64

(anodically poled) pillars

65

luminescent active layer

66

Distribution layer

67

Onset

68

,

69

beam

70

Semi-cone

71

Semi-cone

72

ad beam

73

(forwarded) beam

Fig. 5 (b)

81

Chip structure

82

(cathodically poled) pillar

83

,

84

(anodically poled) pillars

85

luminescent active layer

86

Distribution layer

87

Onset

88

,

89

Edge rays

90

Radiation cone

Fig. 6

91

(cathodically poled) pillar

92

(n-type) distribution layer

93

Cathode contact

94

Leitbahntrasse

95

Luminescence active layer

96

p-type cladding layer

97

(p-type) Bragg reflector

98

(anodically poled) pillar

99

Anode contact

100

Anti-reflective layer

Fig. 7

101

Far field chip

102

optical disc

103

(cathodic) contact

104

anodic contact

105

anodic contact

106

luminescent active layer

107

concave lead

108

,

109

,

110

Club

111

distance

Fig. 8 (a) and (b)

121

optical disc

122

Pillar relief

123

Expandable expansion level

124

Supplementary disc

125

Joining plane

126

Recess

127

Concentrators

128

Optical fiber

Fig. 9

131

,

132

Solder joint

133

Silicon housing

134

Assembly tub

135

Inner wall (of the assembly tub)

136

,

137

Conductor tracks (to chip contacts)

138

Plated-through hole

139

External connection (the anode)

140

External connection (the cathode)

141

Sealing plate

142

Imaging curvature

Fig. 10

150

P _opt

(I _F

) at room temperature without

151

P _opt

(I _F

) at room temperature with improvement
the decoupling

Fig. 11

160

Chip structure

161

Ground potential

162

Bridging resistance

163

Short-circuited DH structure

164

Distribution layer

165

Luminescence-active DH structure

166

Positive against ground biased contact

167

Photosensitive DH structure

168

Contact

Fig. 12 (a)

200

self-hermetic chip structure

201

(circular) pillar

202

(square) pillar

203

Central contact

204

External contact

Fig. 12 (b) and (c)

210

Assembly document

211

back

212

Contact page

213

Indentations

214

p-type layer

215

Dielectric layer

216

Metal upholstery

217

Thin solder deposit

218

Bond island

219

Stripline

220

Connector panel

Claims (17)

1. Bipolar illumination source from a self-bundling semiconductor body contacted on one side, which connects a substrate transparent to the luminescence waves with a structure of layers of ternary and binary III-V compounds, preferably AlGaAs-GaAs, and is set up for pseudo-lateral operation,
with a mechanical bridge ( 21 ) made of an optical plate ( 23 ) made of a semiconductor body which is uniformly transparent for luminescence, with which at least one luminescent and one associated non-luminescent pillar ( 24 , 34 , 62-64 , 82-84 , 91 , 98 , 201 ) are kept next to each other at a desired distance,
the luminescent pillars ( 34 , 62 , 82 , 91 , 202 ) having a layer structure ( 27 ) and a luminescent active layer ( 26 , 65 , 85 , 95 , 106 ),
all pillars ( 24 , 34 , 62-64 , 82-84 , 91 , 98 , 201 , 202 ) on the underside end in the same plane in a flat mounting surface ( 33 ),
the pillars ( 24 , 34 , 62-64 , 82-84 , 91 , 98 , 201 , 202 ) widen conically towards the optical plate ( 23 ) with respect to their cross-section ( 25 ) and into a continuous, transparent, uniformly conductive, highly conductive distributor layer ( 28 , 42 , 66 , 86 , 92 , 164 ) for supplying the luminescence-active layer ( 26 , 65 , 85 , 95 , 106 ) with charge carriers of a first type from a first electrode ( 36 , 43 , 93 , 204 ) pass over and this distributor layer ( 28 , 42 , 66 , 86 , 92 , 164 ) via a metallic interconnect ( 37 , 94 ) along the inclination of the lateral surface of a non-luminescent pillar ( 34 , 62 , 82 , 91 , 201 ) with its contact ( 36 , 43 , 204 ) on the mounting surface ( 33 ),
each luminescence-active zone is provided on its associated mounting surface ( 33 ) with an oppositely polarized contact electrode ( 32 , 46 , 53-55 , 99 , 203 ) for supplying charge carriers of a second type,
and the adjacent mounting surfaces ( 33 ) are provided with a similar metal coating ( 37 ) for simultaneous and identical insertion into a supply or control circuit ( 210 ). 2. Lighting source according to claim 1, characterized in that the distribution layer ( 28 , 42 , 86 , 92 ) has a gradation ( 30 , 41 ) of its profile. 3. Illumination source according to claim 1 or 2, characterized in that the distribution layer ( 28 , 42 , 66 , 86 , 92 , 164 ) has an n-conductivity and with a cathodically polarized pillar ( 34 , 62 , 82 , 91 , 201 ) is in electrical contact, while every other pillar ( 32 , 46 , 53-55 , 99 , 202 ) is anodically poled. 4. Illumination source according to claim 1 or 2, characterized in that the distribution layer has p-conductivity and an anodically polished pillar in electrical contact is made while everyone else Pillar is poled cathodically. 5. Illumination source according to claims 1 to 4, characterized in that the distribution layer ( 28 , 42 , 66 , 86 , 92 , 164 ) made of a ternary III-V compound semiconductor, preferably made of Al _x Ga _1-x As, with a x value of 0.19-0.33. 6. Illumination source according to claims 1 to 5, characterized in that the distribution layer ( 28 , 42 , 66 , 86 , 92 , 164 ) has a layer thickness of 15-30 microns and a doping of 1 ± 0.5.10 ¹⁸ cm ^-3 . 7. Lighting source according to claims 1 to 6, characterized in that in the distribution layer ( 28 ) in a narrow area around each luminescent pillar ( 24 ) a gradation ( 30 ) of the profile is formed by local removal of 33-50% of the layer thickness is and the cone-like extension of the pillar ( 24 ) continues continuously into this gradation. 8. Illumination source according to claims 1 to 6, characterized in that the gradation of the profile of the distribution layer ( 42 , 92 ) is caused by a trapezoidal projection ( 41 ) of about 20 microns thick on the optical plate ( 23 ), via which the distribution layer ( 42 , 92 ) spreads out as a uniformly thick layer and has a layer thickness of 5-25 μm, preferably 15 μm. 9. Illumination source according to claims 1 to 8, characterized in that the entire luminescence-active area is formed on a chip from a larger number of luminescence-active pillars ( 53 , 54 , 55 ) with a smaller cross-section and is arranged next to a non-luminescence-capable pillar per chip. 10. Illumination source according to claims 1 to 9, characterized in that the metallic interconnect on the inclined lateral surface of the non-luminescent pillar ( 34 , 62 , 82 , 91 ) is covered with a solder mask. 11. Illumination source according to claims 1 to 10, characterized in that the deflection of the luminescence takes place inside the chip by the pillar dimensioning with a height based on the diagonals and a shape with a conical cut and the bundling by an arrangement of an imaging element ( 107 , 124 , 141 , 142 ) occurs with an increased focal length on the optical disk. 12. Illumination source according to claim 12, characterized in that the bundling is carried out by an inseparably attached lens plate ( 124 ) which remains attached to each individual chip even during the separation and assembly process. 13. Illumination source according to claim 12, characterized in that the bundling is carried out by an inseparably attached optical fiber plate ( 128 ) with at least one fiber per luminescent element. 14. Illumination source according to claims 1 to 13, characterized in that in particular one of the luminescence-active pillars is operated temporarily or continuously like a photodiode ( 168 ) in the reverse direction and the removable photocurrent at the contact level of this pillar to measure the height of the other pillars ( 165 ) is made impulsive or permanent injection electric current. 15. Illumination source according to claims 1 to 14, characterized in that a coherent, ring-shaped or frame-shaped pillar structure which surrounds the luminescence-active pillar ( 203 ), a sealing zone for the hermetic connection between the individual chip and an assembly base ( 210 ) forms and a permanent mechanical and electrical connection between an external contact ( 201 ) of the chip ( 200 ) and adequate ring zones of a strip conductor ( 219 ) of the mounting base ( 210 ) is brought about by a solder or a conductive adhesive. 16. Illumination source according to claims 1 to 15, characterized in that an assembly base ( 210 ) at the chip receiving points contains depressions ( 213 ) which are filled with solid silver cushions ( 216 ) and a thermal and electrical connection to the luminescence-active pillars ( 203 ) produce. 17. Illumination source according to claims 15 or 16, characterized in that the mounting plate ( 210 ) consists of n-type silicon and is covered on the contact side by parallel arranged the surface profile of the silicon following p-type layers ( 214 ), which on a Bondinsel ( 218 ) can be connected to the operating voltage.