US20020017652A1 - Semiconductor chip for optoelectronics - Google Patents

Abstract

An optoelectronic semiconductor chip has an active layer containing a photon-emitting zone. The active layer is attached to a carrier member at a bonding side of the active layer. The active layer has at least one recess therein with a cross-sectional area that decreases with increasing depth into said active layer proceeding from said bonding side.

Description

BACKGROUND OF THE INVENTION
• 1. Field of the Invention[0001]
• The present invention is directed to a semiconductor chip for optoelectronics of the type having an active layer with a photon-emitting zone and that is attached to a carrier member at a bonding side.[0002]
• 2. Description of the Prior Art[0003]
• Semiconductor chips of the above type produced in thin-film technology are disclosed in[0004]European Application 0 905 797. For manufacturing the known semiconductor chip, an active layer is usually applied on a substrate with an epitaxial process. A carrier member is subsequently secured to the upper side of the active layer, and the substrate is removed. A metallic reflection layer is located between the carrier member and the active layer, so that no light is absorbed by the carrier member. The known semiconductor chips are particularly suitable for applications wherein a high optical power is required.
• One disadvantage of the known semiconductor chips manufactured in thin film technology is that the metallic reflection layer arranged between the carrier member and the active layer generally does not exhibit a satisfactory reflectivity at short wavelengths. Particularly given a wavelength of less than 600 nm, gold becomes increasingly inefficient as the metallic reflection layer, since the reflectivity significantly decreases. At wavelengths below 600 nm, for example, the elements Al and Ag can be employed, their reflectivity remaining comparatively constant at wavelengths below 600 nm.[0005]
• Moreover, largesurfaces such as the metallic reflection layer aredifficult to bond. As a result of the bonding and the alloying of the metallic contact layer, moreover, the quality of the metallic reflection layer is generally degraded.[0006]
• German OS 198 07 758 discloses a truncated pyramid-shaped semiconductor chip that has an active, light-emitting zone between an upper window layer and a lower window layer. The upper window layer and the lower window layer together form a truncated pyramidal base member. The slanting alignment of the sidewalls of the window layers cause the light emanating from the active zone to be totally reflected at the side faces, so the reflected light is incident on the base area of the truncated pyramid-shaped base member, serving as luminous surface, substantially at a right angle. As a result, a part of the light emitted by the active zone emerges onto the surface within the exit cone of the semiconductor element. The term exit cone in this context means the cone of the light rays whose incident angle is smaller than the critical angle for the total reflection and that are therefore not totally reflected. In order to arrive at a significant enhancement of the light yield, this concept assumes a minimum thickness for the upper and lower window layer. In the known truncated pyramid-shaped semiconductor component, the thickness of the upper and lower window layer amounts to at least 50.8 μm (2 milli-inches). Such a thickness is still within a range allowing the layers to be produced without difficulty. If, however, the power of the known semiconductor chip is to be increased, it is necessary to scale all dimensions. Layer thicknesses thereby quickly derive that can be manufactured only given high outlay in an epitaxial layer. The known semiconductor chip is therefore not scalable without further difficulty.[0007]
SUMMARY OF THE INVENTION
• Proceeding from this prior art, an object of the invention is to provide a semiconductor chip that is manufacturable in thin-film technology that has improved light output.[0008]
• This object is inventively achieved in a semiconductor chip of the type initially described wherein a recess is introduced into the active layer proceeding from the bonding side, the cross-sectional area of this recess decreasing with increasing depth.[0009]
• As a result of recess, the bonding side of the semiconductor chip can be made significantly smaller, so that the bonding of the active layer on the carrier member can be unproblemmatically implemented. As a result of the recess, moreover, lateral faces are created at which a part of the photons emitted by the active layer is reflected such that the photons within the exit cone strike the exist face of the active layer lying opposite the bonding surface. In the semiconductor chip of the invention, so to speak, the reflection at the continuous reflection layer is replaced by the reflection at the lateral faces of the recesses.[0010]
• In one embodiment of the invention, the recesses are so deep that the active zone of the active layer is interrupted by the recess introduced into the active layer proceeding from the bonding side.[0011]
• It has been shown that semiconductor chips whose active zone is interrupted by a recess introduced into the active layer proceeding from the bonding side exhibit an especially high light yield. In this case, it is not only the photons emitted toward the bonding side but also the photons emitted toward the exit face of the active layer, that are caused to proceed at a large angle relative to the exit face by the reflection at the lateral faces of the recess.[0012]
• In a preferred embodiment of the invention, elevations on a connecting layer of the active layer are formed by the recesses.[0013]
• Such elevations act as collimators that align the trajectories of the photons emitted by the active zone nearly at a right angle relative to the exit face of the semiconductor chip. As a result, a majority of the emitted photons within the exit cone strikes the exit face and can exit the semiconductor chip.[0014]
• In another preferred embodiment, the connecting layer is fashioned such that at least one trajectory of the photons emitted by the active zone proceeds from the respective elevation to one of the neighboring elevations.[0015]
• Due to the optical coupling of the elevations, photons that have not been reflected at one of the lateral faces of the original elevation proceed into one of the neighboring elevations and are reflected such that at the lateral faces of the respective elevation so that they strike the exit face within the exit cone.[0016]
• In a further embodiment of the invention, the elevations are provided with concave lateral faces.[0017]
• As a result of these measures, photons that are initially reflected at the exit face are increasingly intensified with every further reflection at the lateral face of the elevations, so that they ultimately strike the exit face within the exit cone.[0018]
• In another preferred embodiment, the elevations are covered with a reflective layer.[0019]
• As a result of this measure, all light rays incident on the lateral face of the elevations are deflected in the direction toward the exit side of the semiconductor chip.[0020]
DESCRIPTION OF THE DRAWINGS
• FIG. 1 is a cross-section through an exemplary embodiment of a semiconductor chip of the invention.[0021]
• FIG. 2 is a cross-section through another exemplary embodiment of a semiconductor chip of the invention, wherein the active zone is arranged within truncated pyramid-shaped elevation.[0022]
• FIG. 3 is a cross-section through a semiconductor chip of the invention that is provided with elevations that have concave lateral faces.[0023]
• FIG. 4 is a diagram that shows the intensification of the light yield in a semiconductor chip of the invention, compared to a conventional semiconductor chips.[0024]
• FIG. 5 is a cross-sectional profile of an elevation in the inventive semiconductor chip that is composed of a lower, flat frustum and an upper, steep frustum.[0025]
• FIGS. 6[0026]athrough6drespectively show cross-sectional profiles of elevations in the inventive semiconductor chip, and a diagram that shows the dependency of the output efficiency on the radius of the boundary surface between the lower truncated pyramid and the upper truncated pyramid of the elevation of FIG. 5.
• FIG. 7 is a diagram that shows the dependency of the output efficiency on the reflectivity of a contact layer arranged on the tip of the elevation of FIG. 5.[0027]
• FIG. 8 is a diagram that shows the dependency of the output efficiency on the reflectivity of the lateral faces of the elevation of FIG. 5.[0028]
• FIG. 9 is a diagram from which the relationship between output efficiency and size of a luminous spot in the active zone proceeds.[0029]
• FIGS. 10[0030]athrough10drespectively show cross-sectional profiles of an elevation in the inventive semiconductor chip, wherein the height of the active zone is varied, and a diagram that shows the output efficiency dependent on the thickness of a lower limiting layer.
• FIG. 11 is a diagram that shows the dependency of the output efficiency on the sidewall angle of the lateral faces of an elevation with the cross-sectional profile shown in FIG. 10[0031]b.
• FIG. 12 is another diagram that shows the dependency of the output efficiency on the sidewall angle of an elevation having the cross-section profile from FIG. 10[0032]b.
• FIG. 13 is a diagram that shows the dependency of the output efficiency on the width of the active layer given constant height.[0033]
• FIG. 14 is a diagram that shows the dependency of the output efficiency on the thickness of a connecting layer uniting elevations with various profiles in the inventive semiconductor chip.[0034]
DESCRIPTION OF THE PREFERRED EMBODIMENTS
• The semiconductor chip shown in FIG. 1 for a light-emitting diode has a[0035]carrier member1 on which anactive layer2 is attached. For clarity, the thickness of theactive layer2 is shown exaggerated in relationship to the thickness of thecarrier member1. Theactive layer2 has a photon-emitting,active zone3 that is fashioned withelevations4 at a mid-height. Theelevations4 can be fashioned as a truncated pyramid or truncated cone.
• The[0036]elevations4 are arranged on a connectinglayer5 that has acentral contact location7 of the front side on a flatfront side6, thecontact location7 being formed by a metallization layer. Theelevations4 of the backside formed byrecesses8 are covered with a reflective layer that is composed of a dielectric insulatinglayer9 and ametallization layer10 applied thereon. The insulatinglayer9 is interrupted by through-contacts12 along abase area11 of theelevations4, the through-contacts12 being formed by metallic sections.
• For manufacturing the semiconductor chip shown in FIG. 1, the[0037]active layer2 is first epitaxially grown on a base substrate. Theactive layer2 can, for example, be manufacture on the basis of InGaAIP. The connectinglayer5 is thereby produced first on the base substrate and, subsequently, is doped with a concentration above 10¹⁶cm⁻³in order to assure a good conductivity of the connectinglayer5. Good conductivity of the connectinglayer5 is a pre-condition onecentral contact location7 on thefront side6 being sufficient for supplying theactive zone3 with current. Moreover, the composition of the connectinglayer5 is selected such that it is transparent for the photons generated in theactive zone3. This can usually be accomplished via setting the band gap on the basis of the composition of the material of the connectinglayer5.
• Subsequently, a further layer is applied on the connecting[0038]layer5, theelevations4 being introduced into this layer by suitable wet-chemical or dry-chemical etching processes. Such etching processes are known and are not the subject matter of the present application. Theelevations4 are preferably formed in the regions provided for the semiconductor chips. These are regions having typical outside dimensions of 400×400 μm². Theelevations4 have outside dimensions that lie in the range of the layer thickness of theactive layer2. The outside dimensions of theelevations4 therefore are in the range of 10 μm.
• The deposition of the insulating[0039]layer9 onto theelevations4 and the formation of the through-contacts12 ensue in a further method step. Subsequently, themetallization layer1 is applied.
• Subsequently, the[0040]active layer3 is divided (separated) according to the intended number of semiconductor chips. This ensues, for example, by wet etching.
• The separated active layers at the[0041]carrier member1 are then secured, for example by eutectic bonding, and the base substrate is removed by wet-etching. Finally, thecontact locations7 are formed at the exposed front side of theactive layer2, and the semiconductor chips are separated by dividing thecarrier member1.
• The semiconductor chip shown in FIG. 1 exhibits the advantage that the photons generated by the[0042]active zone3 do not strike components of the semiconductor chip that would absorb them. The photons are kept away from thecarrier member1 by themetallization layer10.
• Another advantage of the semiconductor chip of FIG. 1, is that a majority of the photons emitted from the[0043]active zone3 are totally reflected at lateral faces13 of theelevations4. The photons totally reflected at the lateral faces13 strike thefront side6 at a large angle. In particular, a part of the photons that would be totally reflected at the front side without reflection at the lateral faces13 strikes thefront side6 within the exit cone, and can therefore exit the semiconductor chip. In the semiconductor chip according to FIG. 1, therefore, the reflection at the continuous base area known from the prior art is at least partially replaced by the total reflection at the lateral faces13. The semiconductor chip of FIG. 1 compared to conventional semiconductor chips withoutrecesses8 therefore exhibits a light yield enhanced by nearly a factor of two.
• The described effect is explained in detail below with reference to the further exemplary embodiments shown in FIGS. 2 and 3.[0044]
• A number of light rays are considered in the following discussion, but the term light rays is not a limitation to a specific wavelength, but refers to the processes of geometrical optics, regardless of the wavelength.[0045]
• In the exemplary embodiments shown in FIG. 2, the[0046]elevations4 are fashioned as a truncated pyramid and are secured to thecarrier member1 via acontact layer14 only at thebase area11 of theelevations4. Theactive zone3 is supplied with current by thecontact layer14.
• Due to the large difference between the refraction indices of semiconductor material compared to casting resin of, typically, 3.5 to 1.5, only light rays that strike the boundary surface within an exit cone having an aperture angle of approximately 16° can be coupled out from the semiconductor at the boundary surface between the semiconductor and the casting resin. Given a uniform distribution of the angles of incidence of the light rays, this corresponds to approximately 2% of the light rays that are incident on a surface unit.[0047]
• Due to the[0048]elevations4, the light rays emanating from theactive layer2 are steered in the direction onto thefront side6. Theelevations4 therefore act as collimators in whose respective focal surfaces theactive zone3 is located. Theelevations4 causes the light rays that are incident on the lateral faces13 to be intensified in the direction toward thefront side6 so that they strike within the exit cone, so that they can exit the semiconductor chip. The light yield thereby can be optimized on the basis of a suitable selection of the dimensions of thebase area11, the angle of inclination of thelateral face13, and of the height of theelevations4, as well as the position of theactive zone3.
• FIG. 2 shows a[0049]light ray15 that is initially totally reflected at thelateral face13 and is steered therefrom to thefront side6. At thefront side6, thelight beam15 strikes the boundary surface within the exit cone and therefore can exit the semiconductor chip. Without the total reflection at thelateral face13, thelight wave15 would be totally reflected at thefront side6 and would have been deflected back to one of the reflection layers known from the prior art where it would have been reflected again. In this respect, the reflection at the conventional, continuous reflection layer is replaced by the reflection at the lateral faces13 given the exemplary embodiment shown in FIG. 2.
• This is also true for a[0050]light ray16 that is initially reflected at thebase area11 and then at thelateral face13. Thelight ray16 as well is deflected to thefront side6 after the second reflection, where it strikes within the exit cone. Without the reflection at thelateral face13, thelight ray16 would likewise be totally reflected at thefront side6 and would have been steered back to a back-side reflection layer.
• It is also advantageous that the[0051]elevations4 are optically coupled via the connectinglayer5. Optical coupling in this context means that at least one of the light rays emanating from theactive layer2 can proceed across acenter line17 from the regions of one of theelevations4 into the regions of one of the neighboringelevations4. Due to the optical coupling with the assistance of the connectinglayer5, alight ray18 that does not strike one of the lateral faces13 of therespective elevations4 can strike one of the lateral faces13 of one of the neighboringelevations4 and be deflected to thefront side6 where it is incident within the exit cone. Due to the optical coupling via the connectinglayer5, the light yield therefore is enhanced further.
• FIG. 3 shows a cross-section through a modified exemplary embodiment of the semiconductor chip wherein the[0052]elevations4 are fashioned as a concave cone with concave lateral faces13. The fashioning of the lateral faces13 causes alight ray18 to be reflected back and forth between thefront side6 and thelateral face13 and thus it is increasingly intensified as it approaches thecenter line27, until it strikes thefront side6 within the exit cone. The same is true oflight rays9 that initially proceed via the connectinglayer5 from oneelevation4 to the neighboringelevation4 and thus are directed at a large angle to thefront side6.
• FIG. 4 is a diagram wherein a measured[0053]curve20 shows the dependency of the light yield in relative units on the operating current given pulsed mode for a conventional light-emitting diode manufactured in thin-film technology. A further measuredcurve21 illustrates the dependency of the light yield in relative units dependent on the operating currentfora light-emitting diode according to the exemplary embodiment shown in FIG. 3. It can be seen from FIG. 4 that the light yield given the exemplary embodiments shown in FIG. 3 exhibits approximately twice the light yield of conventional semiconductor chips withoutrecesses8.
• In order to determine the most beneficial shape for the[0054]elevations4, a number of ray tracing simulations were implemented. The results of these calculations are presented in detail below with reference to FIGS. 5 through 14.
• First, the parameters varied in the calculations shall be explained with reference to FIG. 5. FIG. 5 shows a cross-sectional profile of one of the[0055]elevations4. In the illustrated case, theelevation4 is composed of a lowertruncated cone22 and of an uppertruncated cone23. The lowertruncated cone22 has abase area24 adjoining the connectinglayer5. Theactive zone3 is formed in the uppertruncated cone23. Moreover, acontact location25 arranged on thebase area11 of theelevation3 is provided in FIG. 5.
• The lateral faces[0056]13 of theelevation4 are composed of asidewall26 of the lowertruncated cone22 and sidewalls27 of the uppertruncated cone23. The geometrical dimensions of the lowertruncated cone22 along a sharedboundary surface28 are selected such that thesidewall26 merges directly into thesidewall27.
• Various dimensions of the[0057]elevation4 are referred to below as follows. The radius of thebase area24 of the lowertruncated cone22 is referenced r_n, the radius of theboundary surface28 is referenced r_t, and the radius of thebase area11 is referenced r_p. Further, theelevation4 can be divided into a lower limitinglayer29 between thebase area24 and theactive zone3 and an upper limitinglayer30 between theactive zone3 and thebase area11. The lower limitinglayer29 has a height h_u, and the upper limitinglayer30 has a height h_o. The overall height of theelevation4 is referenced H. This was consistently equated to 6 μm in all calculations. A value of 2 μm was selected for the thickness h_wof the connectinglayer5 in all calculations and the thickness h_wwas not varied.
• FIGS. 6[0058]athrough6dshow the result of a calculation wherein the radius r_pof thebase area11 was set equal to 5 μm, and the radius r_nof thebase area24 was set equal to 20 μm. The radius r_tof theboundary surface28 was varied between 6 and 18 μm according to the cross-sectional profiles shown in FIG. 6athrough FIG. 6c.
• In the calculations, a refractive index of 3.2 was set for the[0059]active zone3. The refractive index of the lower limitinglayer29, of the upper limitinglayer30 as well as of the connectinglayer5 was 3.3. Insofar as it was not varied, the reflectivity of thecontact location25 was set as 0.3. The reflectivity of thebase area11 not covered by thecontact location25, as well as the reflectivity of thesidewalls26 and27, was set to 0.8. As used herein, reflectivity means the reflection coefficient with respect to energy.
• Further, the self-absorption of the[0060]active zone3 was taken into consideration with an absorption coefficient of 10,000/cm. All calculations were implemented with photon recycling. An internal quantum efficiency of 0.8 was assumed for this. The quantum efficiency in the generation of photons by charge carrier recombination was not taken into consideration. The output efficiency η indicated in the diagrams is therefore equal to the ratio of the number of photons coupled out from the semiconductor chip to the number of photons actually generated. The values for the indicated output efficiency η therefore would also have to be multiplied by the factor 0.8 in order to arrive at the external efficiency.
• It was also assumed that the reflection at the[0061]contact location25 and thesidewalls26 and27 is angularly independent. The fact that the dielectric insulatinglayer9 is first applied onto theelevations4 and is supplemented by thereflective metallization layer10 is therefore underestimated in effect in the calculations, since the total reflection occurring in this case is not taken into consideration in the calculations.
• FIG. 6[0062]cshows a diagram wherein the output efficiency η is entered relative to the radius r_tin acurve31. For comparison, the output efficiency of a normal thin-film semiconductor chip is also entered, whereby the scatter is conveyed only via the photon recycling. This thin-film semiconductor having the edge length of 300 μm exhibits the same epitaxial structure as theelevation4 in the lowertruncated cone22 and uppertruncated cone23. It was assumed that the semiconductor chip is provided with a mirror at the p-side, the reflectivity of said mirror amounting to 0.72. This value is the average value—weighted with the degree of occupancy—of the reflectivity of a reflection layer and of a contact layer, whereby the value 0.8 is set for the reflectivity of the reflection layer and the value 0.85 is set for the occupancy of the reflection layer, and the value 0.3 for the reflectivity of the contact layer and 0.15 for the occupancy were employed.
• It can be seen from FIG. 6 that, given a very large pitch angle Φ[0063]_oof the uppertruncated cone23 according to the cross-sectional profile shown in FIG. 6a,the output efficiency η is hardly any better than the output efficiency η of a traditional thin-film light-emitting diode whose output efficiency is shown in FIG. 6dby thestraight line32. This is also easily understood since theelevation4 having the flat cross-sectional area profile shown in FIG. 6ahardly brings the light rays emanating from theactive zone3 into a steep angle relative to the light-emittingface6. It is precisely this result, however, that theelevation4 accomplishes with the cross-section profile shown in FIG. 6c,for which reason the output efficiency h_hin this case amounts to nearly double the output efficiency η of a conventional thin-film light-emitting diode.
• Further, the dependency of the output efficiency η on the reflectivity of the[0064]contact location25 was investigated. To this end, the output efficiency η was calculated dependent on the reflectivity of thecontact location25, whereby the cross-sectional profile of theelevation4 was same as the cross-sectional profile shown in FIG. 6b.It was also assumed that thecontact location25 covers theentire base area11. It can be seen from FIG. 7 that the output efficiency η is not significantly dependent on the reflectivity of thecontact location25. The semiconductor chips describedherewith elevations4 at the fastening side therefore seems significantly less sensitive to the poor reflectivity of thecontact locations25 than are the traditional thin-film light-emitting diodes, since only a tiny fraction of the multiplex reflections leading to the output apparently occur between thebase area11 and the light-emittingarea12 but ensue three-dimensional in theelevation4.
• The relative independence from the reflectivity of the[0065]contact relation25 is particularly advantageous since, in practice, a low ohmic resistance between thecontact location25 and the upper limitinglayer30 is generally linked to a poor reflectivity. A good ohmic contact, namely, requires the diffusion of atoms from the layer forming thecontact location25 into the material lying there below.
• In contrast to the dependency on the reflectivity of the[0066]contact location25, the dependency of the output efficiency η on the reflectivity R_sof the reflective surfaces on thebase area11 and on thesidewalls26 and27 is very pronounced. This is shown by the results of a calculation that was implemented with a model for the semiconductor chip whoseelevations4 have the radii r_p=5 μm, r_d=16 μm and r_n=20 μm.
• The[0067]elevations4 therefore have approximately the cross-sectional profile shown in FIG. 6b.
• The result of this calculation is a[0068]curve33 entered in FIG. 8 that rises monotonously with increasing reflectivity R_s. Apoint34 entered into the diagram from FIG. 8 represents the result of a calculation for a semiconductor chip on which no reflective layer was applied but that was embedded in resin as surrounding medium. However, total reflection occurs here, so that a greater output efficiency occurs compared to a semiconductor chip with a reflective layer. This would also be the case for the exemplary embodiment shown in FIG. 1 wherein the electrical insulating layer at which total reflection can likewise occur is arranged between themetallization layer16.
• FIG. 9 contains the result of a calculation that was carried out at a semiconductor chip with[0069]elevations4 for whose radii the following was valid: r_p=5 μm r_d=16 μm and r_n=20 μm. Theelevations4 therefore essentially have the cross-sectional profile shown in FIG. 6a.Theactive zone3 was thereby located at medium height between thebase area24 and thebase area11. In this calculation, the region wherein photons arise in theactive zone3 is constricted to a luminous spot whose diameter d_sis entered on the abscissa. It can be seen on the basis of the diagram in FIG. 9 that the output efficiency is especially high given a small luminous spot. This means that photons in the center of theactive zone3 are coupled out especially well. In this respect, a slight Weierstrass effect is present.
• Further, the influence of the position of the[0070]active layer3 was investigated. Various cross-section profiles are shown in FIGS. 1Oathrough10cwherein the thickness h_uof the lower limitinglayer29 and the thickness h_uof the upper limitinglayer30 were varied such that the overall height H of the elevation remained constant. The result of this calculation is shown in FIG. 10dwherein the output efficiency η is entered dependent on the thickness h_uof the lower limitinglayer29. This shows that the output efficiency η is only slightly dependent on the position of theactive zone3. Anactive zone3 that lies in the lower half of theelevation4 is to be preferred since the current density through theactive zone3 is then low and, therefore, the current load on theactive zone3 can be kept small, this avoiding aging and linearity problems.
• Further, the influence of the set angle Φ[0071]_oof thesidewall27 and of the set angle Φ_oof thesidewall26 was investigated. A cross-sectional profile was assumed wherein the lowertruncated cone22 and the upper truncated23 and exhibit the same value for the set angle Φ_uand Φ_o. The radius Φ_uof theactive zone3 was therefore kept constant at 10 μm and the set angle Φ=Φ_o=Φ_uwas varied. Two instances were considered. First, the output efficiency η was calculated for the periodic case wherein infinitely many elevations are arranged in a quadratic grid, whereby the distance of the foot point amounted to 10 ηm. The result is reflected in thecurve35 in the diagram in FIG. 11. Moreover, an aperiodic case was investigated. To that end, the output efficiency η of a semiconductor chip was calculated with asingle elevation4, whereby all photons entering into the connectinglayer5 are absorbed by the connectinglayer5. The aperiodic case is shown in FIG. 11 by thecurve36. It can be seen from FIG. 11 that the connectinglayer5 supplies a noticeable contribution to the output efficiency η.
• There is also an optimum range for the sidewall angle Φ. This can be seen with reference to FIG. 12. In the underlying calculation, the radius r[0072]_pwas set equal to 10 μm. The radius r_aof theactive zone3 and the radius r_nof thebase area24 were varied such that the set angle Φ of thesidewalls27 and26 covers the value range between 1.5° and 85°. As can be seen from FIG. 12, an optimum angular range exists for the set angle Φ. The sidewall angle Φ should lie between 5° and 60°, preferably between 10° and 400°. Especially good values for the output efficiency η arise when the set angle Φ lies between 15° and 30°.
• An investigation was subsequently carried out to determine the effect of varying the width of the[0073]elevations4 on the output efficiency η. In this case, the height H of theelevations4 was kept constant, and the radii r_p, r_aand r_nwere uniformly increased. Acurve37 in FIG. 13 illustrates the case where the reflectivity R_Kof thecontact location25 is equal to 0.3. Anothercurve38 is directed to the case where the reflectivity R_Kof thecontact location25 amounts to 0.8. Thecurve37 as well as thecurve38 show the dependency of the output efficiency η on the diameter 2 r_aof theactive zone3. Given good reflectivity of thecontact location25, the output efficiency η drops only slightly with increasing diameter of theactive zone3. Thecurve37, that illustrates the realistic case of a poor reflectivity R_Kof thecontact location25, however, shows that the output efficiency η decreases significantly with increasing diameter of theactive zone3. The output efficiency therefore becomes better as the lateral expanse of theelevations4 is made smaller.
• The thickness of the connecting[0074]layer5 is also of significance for the output efficiency η. FIG. 14 shows the output efficiency η for various cases dependent on the thickness h_wof the connectinglayer5. Acurve39 reflects the aforementioned periodic drop. Afurther curve40 is directed to the aperiodic case, and athird curve41 is directed to a case wherein square semiconductor chips having an edge length of 300 μm are connected to one another by a connecting layer. It can be seen from FIG. 14 that the connectinglayer5 is increasingly advantageous given increasing layer thickness. However, it also proceeds from FIG. 14 that a number of individual semiconductor chips, each having anelevation4, represent the best case, since the output efficiency is highest given the thickness h_w=0. Individual chips, however, have the disadvantage that their power cannot be arbitrarily increased, since the dimensions of the semiconductor chips must also be scaled with the power. For practical reasons, however, the thickness of epitaxial layers is limited. This means that individual semiconductor chips cannot be designed for arbitrarily high powers. The semiconductor chips presented in FIGS. 1 through 4, however, can be arbitrarily scaled for this purpose, since only the number ofelevations4 must be increased corresponding to the increasing area of the connectinglayer5 in order to increase the light power of the semiconductor chips.
• A further investigation was directed to determine whether the[0075]active zone3 could perhaps be arranged in the connectinglayer5. To this end, the output efficiency for a conventional thin-film light-emitting diode was calculated and this was set equal to 1 for comparison purpose. A semiconductor chip havingactive zone3 in the connectinglayer5 has an output efficiency of 1.5 compared to the conventional thin-film light-emitting diode. Finally, a relative output efficiency of 1.67 was derived for the semiconductor chips shown in FIGS. 1 through 4. This shows that an increase in the output efficiency η also can be achieved when theactive zone3 is arranged in the connectinglayer5.
• Although modifications and changes may be suggested by those skilled in the art, it is the intention of the inventors to embody within the patent warranted hereon all changes and modifications as reasonably and properly come within the scope of their contribution to the art.[0076]

Claims (14)

We claim as our invention:

1. An optoelectronic semiconductor chip comprising:

an active layer containing a photon-emitting zone, said active layer having a bonding side;

a carrier member to which said active layer is attached at said bonding side; and

said active layer having a recess having a cross-sectional area which decreases with increasing depth of said recess into said active layer proceeding from said bonding side.

2. An optoelectronic semiconductor chip as claimed inclaim 1 wherein said recess is disposed in said active layer so that said recess interrupts said photonemitting zone.

3. An optoelectronic semiconductor chip as claimed inclaim 1 wherein said active layer has a plurality of recesses each corresponding to said recess, and wherein said active layer has a connecting layer disposed at said bonding side, said connecting layer having a plurality of elevations formed by said plurality of recesses.

4. An optoelectronic semiconductor chip as claimed inclaim 3 wherein said photon-emitting zone and said elevations are disposed relative to each other so that at least one trajectory of photons emitted by said photon-emitting zone proceeds from one of said elevations to a neighboring one of said elevations.

5. An optoelectronic semiconductor chip as claimed inclaim 3 wherein said elevations are tapered toward said carrier member.

6. An optoelectronic semiconductor chip as claimed inclaim 5 wherein said elevations each have concave lateral faces.

7. An optoelectronic semiconductor chip as claimed inclaim 3 wherein said elevations each have a truncated pyramidal shape.

8. An optoelectronic semiconductor chip as claimed inclaim 3 wherein said photon-emitting zone is disposed in a portion of said elevations neighboring said connecting layer.

9. An optoelectronic semiconductor chip as claimed inclaim 3 wherein said connecting layer is transparent for photons emitted by said photon-emitting zone.

10. An optoelectronic semiconductor chip as claimed inclaim 3 wherein said connecting layer is highly doped.

11. An optoelectronic semiconductor chip as claimed inclaim 3 further comprising a reflective layer covering said elevations.

12. An optoelectronic semiconductor chip as claimed inclaim 11 wherein said reflective layer comprises a metallization layer adjacent an insulating layer.

13. An optoelectronic semiconductor chip as claimed inclaim 1 wherein said active layer has a thickness which is less than 50 μm.

14. An optoelectronic semiconductor chip as claimed inclaim 1 wherein said active layer has a thickness of less than 30 μm.

Images