EP2342762A1 - Method for producing an optoelectronic semiconductor chip and optoelectronic semiconductor chip - Google Patents

Abstract

The invention relates to a method for producing an optoelectronic semiconductor chip, comprising the following steps: A) forming a semiconductor layer sequence (200) having at least one doped functional layer (7), which comprises bonding complexes with at least one doping agent and at least one codoping agent, wherein one is an electron acceptor selected from the doping agent and the codoping agent and the other is an electron donator, B) activating the doping agent by breaking open the bonding complexes by introducing energy (14, 15), wherein the codoping agent remains at least partially in the semiconductor layer sequence (200) and at least in part forms no bonding complexes with the doping agent. Furthermore, an optoelectronic semiconductor chip is provided.

Description

description

Method for producing an optoelectronic semiconductor chip and optoelectronic semiconductor chip

This patent application claims the priority of German Patent Application 10 2008 056 371.4, the disclosure of which is hereby incorporated by reference.

A method for producing an optoelectronic semiconductor chip and an optoelectronic semiconductor chip are specified.

In the production of highly doped and at the same time crystalline high-quality semiconductor layers in semiconductor chips, codoping with a second material is required for many materials, in particular for broadband semiconductors, in addition to the actual doping. For example, in the case of a desired high p-type doping to generate an increased charge carrier concentration, ie an increased hole concentration, an electron acceptor material is introduced. In order to simultaneously counteract a degradation of the crystal quality, an electron donor material is additionally introduced as codoping, whereby, however, the electrical neutrality of the crystal is at least partially restored. The codoping may therefore be undesirable but required by the manufacturing process. However, in the case of such p-doped and n-codoped layers according to the cited example, the codoping results in only a small p-doping or even an intrinsic charge carrier concentration or even an n-doping. For a high hole concentration for the aspired in the example mentioned high p-conductivity, the compensating effect of the codopant must be repealed, which is referred to as so-called activation of the electrical conductivity, in the given example, the p-conductivity, or as activation of the dopant.

The electrical activation of such codoped semiconductor materials is usually achieved by activation in the form of a purely thermal annealing step. This requires that the Kodotiers _v toff more volatile than the dopant and that the co-dopant can be driven by the thermal annealing step to a certain extent or fully, so for example from 0.001% up to 100%, from the doped semiconductor layer. This method is necessary, for example, for the activation of the p-side of GaN-based light-emitting diodes (LEDs). Established methods exist for this activation, for example, in particular on the basis of so-called RTP ( "rapid termal processing" -). Processes under special atmospheres conventional activation processes take place at high temperature of 700 to 1000 ⁰ C in the wafer composite, in the form of RTP processes or at a lower temperature at 500 to 600 ⁰ C in the wafer assembly in the tube furnace at a comparatively much longer duration and another gas mixture instead.

However, the known methods do not function adequately if, for some reason, the diffusion of the codopant from the doped semiconductor material is prevented, as is the case, for example, with so-called buried p-doped layers. There are clear differences in the achievable degree of activation for codoped p- Layers that are exposed, that is, that lie near a surface of the crystal and for which the described conventional methods work, and for p-doped layers that are buried under one or more layers, in particular n-doped layers. The latter can be activated only slightly or not at all with the known activation methods. The measurable operating voltage of components, such as LEDs, is thereby significantly increased.

An object of at least one embodiment is to specify methods for producing an optoelectronic semiconductor chip which has at least one doped functional layer. An object of at least one further embodiment is to specify an optoelectronic semiconductor chip.

These objects are achieved by the method and subject matter of the independent claims. Advantageous embodiments and further developments of the method and the subject matter are characterized in the dependent claims and will become apparent from the following description and the drawings.

In accordance with at least one embodiment, a method for producing an optoelectronic semiconductor chip comprises in particular the steps:

A) forming a semiconductor layer sequence having at least one doped functional layer which has binding complexes with at least one dopant and at least one codopant, one selected from the dopant and the codopant being an electron acceptor and the other an electron donor, B) activating the dopant by breaking the binding complexes by introducing an energy, wherein the codopant at least partially in the

Semiconductor layer sequence remains and forms at least partially no binding complexes with the dopant.

In accordance with at least one further embodiment, an optoelectronic semiconductor chip comprises in particular a semiconductor layer sequence with at least one doped functional layer with a dopant and a codopant, the semiconductor layer sequence comprising a semiconductor material having a lattice structure, one selected from the dopant and the codopant an electron acceptor and the other an electron donor is, the codopant is bound to the semiconductor material and / or arranged on interstitial sites and the codopant at least partially forms no binding complexes with the dopant.

The embodiments, features and combinations thereof described below relate equally to the optoelectronic semiconductor chip and to the method for producing the optoelectronic semiconductor chip unless otherwise explicitly stated.

The fact that a layer or an element is arranged or applied "on" or "above" another layer or another element can mean here and below that the one layer or the one element directly in direct mechanical and / or electrical contact is arranged on the other layer or the other element. Furthermore, it can also mean that the one layer or the one element indirectly on or above the another layer or the other element is arranged. In this case, further layers and / or elements can then be arranged between the one and the other layer or between the one and the other element.

The fact that a layer or an element is arranged "between" two other layers or elements may mean here and in the following that the one layer or the element directly in direct mechanical and / or electrical contact or in indirect contact with one of the other two Layers or elements and in direct mechanical and / or electrical contact or in indirect contact with the other of the two other layers or elements.In this case, in indirect contact further layers and / or elements between the one and at least one of the two other layers or be arranged between the one and at least one of the two other element.

In the present case, the term "doped functional layer" always denotes a layer with a dopant and a codopant in the sense described above.

By a selective introduction of energy in method step B, the binding complexes in the doped functional layer, which are present for example in the form of atomic bonds between the dopant and the codopant or in the form of binding complexes between the dopant, the codopant and the semiconductor material of the doped functional layer, can to this effect be manipulated that the compensating effect of the codopant can be canceled on the doping properties of the actual dopant. By suitable Process conditions can be avoided by the method described here, a direct restoration of these broken bonds or binding complexes compared to conventional methods. In particular, it can be achieved that the codopant is bound at a different point, that is not at the dopant, in the crystal lattice of the semiconductor material of the doped functional layer or another layer of the semiconductor layer sequence or stored in the intermediate lattice, where it can no longer have a compensating effect on the dopant , As a result, the number of charged, that is not compensated, charge carriers introduced by the dopant increases without the codopant having to be expelled from the doped functional layer or from the semiconductor layer sequence. Due to the achievable higher conductivity and the operating voltage of such an activated optoelectronic semiconductor chip decreases.

The fact that the codopant forms at least in part no binding complexes with the dopant can mean here and below in particular that at least a part of the codopant is present in the doped functional layer which does not form binding complexes with a part of the dopant, so that this part of the dopant can contribute to increasing the density of free charge carriers in the doped functional layer. In this case, the term "free charge carriers" in a p-doped layer includes in particular holes, ie locations where electrons are missing and which contribute significantly to the electrical conductivity of p-type semiconductors, and electrons in an n-doped layer. In particular, the optoelectronic semiconductor chip can be produced as a light-emitting diode (LED) or as a laser diode and can have at least one active layer with an active region which is suitable for emitting electromagnetic radiation. Here and below, "light" or "electromagnetic radiation" may equally mean in particular electromagnetic radiation having at least one wavelength or a wavelength range from an infrared to ultraviolet wavelength range of greater than or equal to 200 nm and less than or equal to 20,000 nm. In this case, the light or the electromagnetic radiation may comprise a visible, ie a near-infrared to blue wavelength range with one or more wavelengths between about 350 nm and about 1000 nm.

The semiconductor chip can have, for example, a pn junction, a double heterostructure, a single quantum well structure (SQW structures) or a multiple quantum well structure (MQW structures) as active region in the active layer. In the context of the application, the term quantum well structure includes in particular any structure in which charge carriers can undergo quantization of their energy states by confinement. In particular, the term quantum well structure does not include information about the dimensionality of the quantization. It thus includes quantum wells, quantum wires and quantum dots and any combination of these structures. The semiconductor layer sequence may comprise, in addition to the active layer with the active region, further functional layers and functional regions selected from p- and n-doped charge carrier transport layers, ie electron and hole transport layers, p-, n- and undoped confinement, cladding and waveguide layers, barrier layers, planarization layers, buffer layers, protective layers and electrodes and combinations of said layers. The electrodes may each comprise one or more metal layers comprising Ag, Au, Sn, Ti, Pt, Pd, Cr, Al and / or Ni and / or one or more layers comprising a transparent conductive oxide such as zinc oxide, tin oxide, cadmium oxide, titanium oxide , Indium oxide or indium tin oxide (ITO). In addition, additional layers, for example buffer layers, barrier layers and / or protective layers can also be arranged perpendicular to the arrangement direction of the semiconductor layer sequence, for example around the semiconductor layer sequence, ie approximately on the side surfaces of the semiconductor layer sequence.

The semiconductor layer sequence or the semiconductor chip can be formed as epitaxial layer sequence, ie as an epitaxially grown semiconductor layer sequence. In this case, the semiconductor chip or the semiconductor layer sequence can be formed in particular as a nitride semiconductor system or be. The term nitride semiconductor system includes all nitride compound semiconductor materials. It may be a semiconductor structure of a binary, ternary and / or quaternary compound of elements of the III main group with a nitride. Examples of such materials are BN, AlGaN, GaN, InAlGaN or further III-V compounds. In this sense, the

Semiconductor layer sequence or the semiconductor chip based on InAlGaN be executed. InAlGaN-based semiconductor chips and semiconductor layer sequences are in particular those in which the epitaxially produced semiconductor layer sequence generally comprises a layer sequence having different individual layers containing at least a single layer, comprising a material from the III-V compound semiconductor material system In _x Al y Ga ₁ _ _x _ _y N with 0 ≤ x ≤ 1, 0 <y <1 and x + y ≤. 1

For example, semiconductor layer sequences comprising at least one InGaAlN-based active layer may preferentially emit electromagnetic radiation in an ultraviolet to green or green-yellow wavelength range.

Furthermore, the semiconductor layer sequence can be embodied, for example, on the basis of AlGaAs. Among AlGaAs-based semiconductor chips and semiconductor layer sequences fall in particular those in which the epitaxially produced semiconductor layer sequence usually has a layer sequence of different individual layers containing at least one single layer, the material of the III-V compound semiconductor material system Al _x Gai_ _x As with 0 ≤ x ≤ 1. In particular, an active layer comprising an AlGaAs-based material may be capable of emitting electromagnetic radiation with one or more spectral components in a red to infrared wavelength range. Furthermore, such a material may additionally or alternatively have the elements In and / or P.

Alternatively or additionally, the semiconductor layer sequence can also be based on InGaAlP, that is to say that the semiconductor layer sequence can have different individual layers, of which at least one individual layer is a material composed of the III-V compound semiconductor material system In _x Al _y Gai. _{X -y} P with O ≦ x ≦ l, o ≦ y ≦ l and x + y ≦ 1. Semiconductor layer sequences or semiconductor chips, at least For example, an active layer based on InGaAlP may preferably emit electromagnetic radiation having one or more spectral components in a green to red wavelength range.

Alternatively or additionally, the semiconductor layer sequence or the semiconductor chip can also have II-VI compound semiconductor material systems in addition to or instead of the III-V compound semiconductor material systems. An II-VI compound semiconductor material may include at least one element of the second main group or the second subgroup such as Be, Mg, Ca, Sr, Cd, Zn, Sn, and a sixth main group element such as O, S, Se , Te. In particular, an II-VI compound semiconductor material comprises a binary, ternary or quaternary compound comprising at least one element of the second main group or second subgroup and at least one element of the sixth main group. Such a binary, ternary or quaternary compound may additionally have, for example, one or more dopants and additional constituents. For example, the II / VI compound semiconductor materials include: ZnO, ZnMgO, CdS, ZnCdS, MgBeO.

All materials specified above need not necessarily have a mathematically exact composition according to the formulas given. Rather, they may include one or more other dopants as well as additional ingredients that do not substantially alter the physical properties of the material. For the sake of simplicity, however, the formulas given include only the essential constituents of the crystal lattice, even if these may be partially replaced by a small amount of other substances.

The semiconductor layer sequence may further comprise a substrate on which the above-mentioned III-V or II-VI compound semiconductor material systems are deposited. The substrate may comprise a semiconductor material, for example a compound semiconductor material system mentioned above. For example, the substrate may include or be GaP, GaN, SiC, Si and / or Ge or sapphire. The substrate may be formed as a growth substrate, which means that the semiconductor layer sequence has grown epitaxially on the substrate and that the functional layer of the semiconductor layer sequence farthest from the substrate is the layer that is the uppermost in the growth direction. Alternatively, the substrate can also be embodied as a carrier substrate onto which a semiconductor layer sequence which has previously grown on a growth substrate is transferred, for example, by bonding such that the layer of the semiconductor layer sequence lying on top of the growth substrate in the growth direction lies closest to the carrier substrate after the bonding. The growth substrate may be partially or completely removed after the transfer step, so that the layer of the semiconductor layer sequence first grown on the growth substrate can be exposed. In particular, an optoelectronic semiconductor chip having a carrier substrate can be or are designed as a thin-film semiconductor chip. In particular, thin-film semiconductor chips are characterized by at least one of the following characteristic features:

A reflective layer is applied or formed on a first main surface of a radiation-generating epitaxial layer sequence which faces toward a carrier and which reflects back at least part of the electromagnetic radiation generated in the epitaxial layer sequence;

- The epitaxial layer sequence has a thickness in the range of 20 microns or less, in particular in the range of 10 microns; and

the epitaxial layer sequence comprises at least one semiconductor layer having at least one surface which has a mixing structure which, in the ideal case, leads to an approximately ergodic distribution of the radiation in the epitaxial layer sequence, i. it has as ergodically stochastic scattering behavior as possible.

A basic principle of a thin-film light-emitting diode chip is described, for example, in I. Schnitzer et al. , Appl. Phys. Lett. 63 (16), 18 October 1993, 2174 - 2176, the disclosure of which is hereby incorporated by reference.

In method step A, for example, the doped functional layer may be provided as a semiconductor layer sequence in the form of a single layer or in the form of a doped functional layer stack. Furthermore, the semiconductor layer sequence can be provided or formed as a partially finished part of the semiconductor chip. This may mean, for example, that the semiconductor layer sequence comprises a substrate, such as a growth substrate, on which a plurality of functional layers including the doped functional layer with the dopant and the codopant being grown in the growth direction to uppermost layer. Alternatively, the substrate may be a carrier substrate, to which a semiconductor layer sequence having a doped functional layer is transferred and the growth substrate is subsequently removed, so that the doped functional layer is exposed.

Furthermore, the semiconductor layer sequence in method step A can be formed with a plurality of functional layers, wherein the doped functional layer with the dopant and the codopant is arranged between two further functional layers, so that the doped functional layer is formed as the layer lying on top of the semiconductor layer sequence is. If in each case one or a plurality of further functional layers are arranged in the growth direction above and below the doped functional layer, wherein at least the layers directly adjacent to the doped functional layer are different from the doped functional layer, in particular differently doped, then the doped functional layer here and below also be referred to as a so-called "buried" layer.

In particular, the semiconductor chip can already be completed in method step A, which means that the semiconductor layer sequence after method step A already has all the functional layers of the semiconductor layer sequence required for the operation of the semiconductor chip. The semiconductor layer sequence can be formed, for example, in a wafer composite. The In this way, the semiconductor layer sequence finished in this way can be formed and provided in accordance with the method step A in the wafer composite or furthermore already individually according to individual semiconductor chips.

The semiconductor chip or the

For example, a semiconductor layer sequence can have an active region between a p-doped layer and an n-doped layer following in the direction of growth such that the polarity in the growth direction is opposite that of a conventional semiconductor chip in which the p-doped region follows the n-doped region in the growth direction , is inverted. Depending on the design of the semiconductor chip with a growth substrate or a carrier substrate, the n-doped layer or even the p-doped layer may be formed as the buried doped functional layer.

A buried doped functional layer can furthermore be formed, for example, in an optoelectronic semiconductor chip having at least one tunnel junction with at least one n-doped ("n-type") tunnel junction layer and at least one p-doped ("p-type") tunnel junction layer doped functional layer with the dopant and the codopant can by the at least one n-doped tunnel junction layer or by the at least one p-doped

Tunnel junction layer to be formed. In this case, an active layer with an active region can be arranged downstream of the tunnel junction within the semiconductor layer sequence in a direction away from a substrate. In this case, an undoped one can exist between the at least one n-type tunnel junction layer and the at least one p-type tunnel junction layer Area, be arranged at least one undoped interlayer, so that the n-type tunnel junction layer and the p-type tunnel junction layer are not directly adjacent to each other, but separated by at least one undoped interlayer. The term "tunnel junction layer" is used to distinguish it from the other functional layers of the

Semiconductor layer sequence or the semiconductor chip used and means that the so-called n-type tunnel junction layer or p-type tunnel junction layer is disposed in the tunnel junction.

By separating the n-type tunnel junction layer and the p-type tunnel junction layer from each other by the undoped region, adverse compensation of the different charge carriers at the interface that would otherwise occur due to diffusion of carriers across the interface is prevented when the p -Type tunnel junction layer and the n-type tunnel junction layer would be directly adjacent to each other. Although the insertion of the undoped region between the n-type tunnel junction layer and the p-type tunnel junction layer creates a region with only a low carrier density within the tunnel junction, it can also be achieved in the form of one or more undoped intermediate layers In addition, the undoped region has less adverse effect on the electrical properties of the tunnel junction, in particular on the forward voltage, than a region at the interface between an n-type tunnel junction layer and an immediately adjacent p-type tunnel junction layer in which carriers are due to diffusion across the interface compensate each other. Furthermore, the semiconductor layer sequence or the semiconductor chip can be formed as a stacked LED having a plurality of active layers grown on top of each other, wherein each of the active layers is in each case arranged between at least one n-doped layer and one p-doped layer, also in combination, for example with tunnel junction layers , As a result, a semiconductor layer sequence designed as a stacked LED can have at least one buried doped functional layer.

The structures described above for the

Semiconductor layer sequence or the semiconductor chip in the form of regular or inverted polarities, tunnel junction layers and stacked active regions are known in the art and will not be further elaborated here. All of these structures have in common that they may have buried doped functional layers that may include a dopant and a codopant and that are adjacent to layers that may act as diffusion barriers for the codopant. With such buried doped functional layers, therefore, conventional activation of the dopant, for example by means of a conventional RTP method described above, may be difficult or impossible.

The dopant for a p-doped functional layer, that is to say at least one suitable electron acceptor, can generally not be introduced in a pure form into the semiconductor material of the doped functional layer, at least at high doping strengths to be achieved. Instead, the dopant is in a complex with at least one other substance, the codopant. This further substance often acts as an electron donor for the semiconductor material, which is the

Electron acceptor material, so the dopant compensated in its electrical effect. In particular, the doped functional layer may thus be a p-doped layer in which the dopant comprises or is an electron acceptor material, while the codopant has or is an electron donor material. The activation step according to method step B is suitable for permanently producing the electrical effect of at least part of the dopant within the semiconductor material, that is, it is suitable for increasing the p-type conductivity.

More preferably, in a broadband semiconductor system, such as in a nitride compound semiconductor system for a p-doped functional layer, the dopant comprises magnesium or is magnesium. The magnesium is usually incorporated in a complex with hydrogen as a codopant in the semiconductor material. The activation step according to method step B produces the electrical effect of at least part of the magnesium as p-dopant, which is compensated by the hydrogen. For example, in an II-VI compound semiconductor system such as ZnSe, the dopant has nitrogen or is nitrogen. The codopant may also be hydrogen preferably here.

Alternatively, the doped functional layer may also be an n-doped layer, that is, the dopant is electron donor material while the codopant is an electron acceptor material. This can be, for example particularly suitable for relatively low bandgap semiconductor materials, such as CdTe or GaAs interconnect semiconductor materials.

Since the expulsion of the codopant is not required in the method described here, it is particularly suitable for doped functional layers from which the codopant can not diffuse out for fundamental reasons, for example because the doped functional layer is a layer buried in the above sense. The method described here can therefore for the first time offer the possibility of activating these layers as well. For PILS ("polarity inverted LED structures") described above and stacked LEDs, for example based on GaN or other of the above-mentioned compound semiconductor materials, this type of activation is essential since in these cases a p-doped functional layer, for example with Mg can not be activated as dopant and hydrogen as codopant by an RTP-based annealing step, which aims exclusively at the expulsion of the codopant.

To achieve a targeted and permanent as possible disruption of the passivating the dopant bond of the codopant to the dopant or to crystal atoms, the supply and the introduction of energy is necessary. In this case, in process step B, the energy can be introduced by generating a current in the doped functional layer. This can also be referred to here as "electrical activation." In the case of electrical activation, at least the doped functional layer can be electrically connected to an external current source Semiconductor layer sequence already finished and trained optoelectronic semiconductor chip are electrically operated for a certain period, that is to be connected to an external power and voltage source. The optoelectronic semiconductor chip can still be located in the wafer composite, so that a plurality of optoelectronic semiconductor chips or semiconductor layer sequences can be activated simultaneously. Alternatively, the semiconductor chip may already be singulated in method step A and thus be detached from the wafer composite, so that the semiconductor chip can be activated individually and independently of another semiconductor chip of the wafer composite. With regard to the required current densities described below, an electrical activation of a isolated

Semiconductor layer sequence or a scattered optoelectronic semiconductor chip may be advantageous, since scaling of this method can be technologically limited to larger wafer wafers.

Alternatively, the current can be generated contactlessly by induction by means of an external suitable coil arrangement. In this case, in a plane parallel to the plane of extent of the doped functional layer, at least in the doped functional layer or additionally also in further layers

Semiconductor layer sequence, a circular current or a plurality of circular currents are generated perpendicular to the direction of growth of the semiconductor layer sequence and thus directed perpendicular to the operational current direction of the semiconductor chip. If the semiconductor layer sequence in method step A is formed, for example, on an electrically insulating sapphire substrate as a growth substrate, the electrical activation can take place by electrical connection after the semiconductor layer sequence has been bonded to an electrically conductive carrier substrate and after removal of the growth substrate. Alternatively or additionally, the electrical activation can be carried out by means of induction already before a bridging step that may no longer be required.

In the electrical activation, both by electrical connection as well as by induction, a steady drop in the required operating voltage can be observed to a minimum value, which remains permanently. The generated current density can be greater than or equal to 50 A / cm ² , with higher current densities can accelerate the activation.

Furthermore, particularly preferably in addition to the generation of the current, a heat energy can be supplied, so that the temperature of the semiconductor layer sequence or of the optoelectronic semiconductor chip, or at least the temperature of the doped functional layer, is increased. The temperature of the doped functional layer should be greater than or equal to about 8O ⁰ C, and more preferably greater than or equal to 100 ⁰ C. Furthermore, the generated current density at such temperatures may be greater than or equal to 10 A / cm ² . In comparison to known activation methods, the temperature in the activation methods described here may be less than or equal to 400 ^° C. and moreover less than or equal to 300 ^° C. With increasingly higher temperatures, the activation can accelerate almost exponentially, which at the same time requires less Current densities can allow. It was determined by measurements that at about 300 ⁰ C, for example, after about 1 minute, a saturation in the fall of the operating voltage can be set. The activation time must be chosen very precisely, that is to say in particular not too long, since otherwise an additional drop in the light emission due to aging of the semiconductor layer sequence can occur. However, there is a parameter space within which aging begins much later and is slower than saturation occurs at a reduced level of operating voltage. In particular, the activation time may be less than or equal to 10 minutes both in the electrical activation and in the alternative and additional activation processes described below, and more preferably less than or equal to 5 minutes.

The heat energy can be supplied by an external heat source, such as a heater. Alternatively or additionally, the heat energy can also be supplied by the generated current itself due to ohmic losses. Due to the influence of the increased temperature generated by the introduced heat energy in combination with the generated current flow, the codopant, so for example, the above-mentioned hydrogen, so rearranged that the actual dopant, so for example, the above-mentioned magnesium is activated.

Alternatively or in addition to generating a current and / or for introducing a heat energy, in method step B, the introduction of energy can be effected by irradiation of an electromagnetic radiation. This can be referred to here and below as "electromagnetic activation" be designated. An electromagnetic activation can mean that the semiconductor layer sequence formed in method step A can be irradiated with electromagnetic radiation which is resonant or non-resonant with absorption wavelengths or absorption bands of the doped functional layers and / or further layers of the semiconductor layer sequence.

By the irradiation of an electromagnetic radiation, it may be possible, for example, that additional charge carriers are generated which, in conjunction with the above-mentioned electrical activation, permit a larger induced current. This may also be advantageous in particular when, for example, intrinsically only very few or no free charge carriers are present in the doped functional layer. Furthermore, in the case of resonant irradiation, charge carriers can be excited in a targeted manner in the layers in which the activation is to take place, ie, for example, in the doped functional layer. Furthermore, the activation of the doped functional layer can be carried out solely by electromagnetic activation.

The frequency of the radiated electromagnetic radiation determines the type of electromagnetic activation. When using microwave radiation, ie electromagnetic radiation having a wavelength of greater than or equal to about 1 millimeter and less than or equal to about 1 meter or a frequency of about 300 MHz to about 300 GHz, the activation is typically not resonant in typical semiconductor materials. The transfer of energy to atomic bonds can take place, inter alia, by excitations of rotons and / or phonons. phonons can typically have excitation energies of a few 10 meV in the doped functional layer, and rotons can have typical excitation energies of less than 1 meV up to a few milli-electron volts. Rotons may include intrinsic rotations of atoms as well as complexes such as excitons. When so-called terahertz radiation is used, ie electromagnetic radiation having a wavelength of greater than or equal to about 100 micrometers and less than or equal to about 1 millimeter or a frequency of about 300 GHz to about 3 THz, conventional semiconductor materials are generally one resonant activation, in which lattice vibrations, ie phonons, can be generated directly.

The process conditions, such as frequency, power, atmosphere, time, additional susceptors that can absorb the electromagnetic radiation, define the degree and success of the electromagnetic activation. In particular, the electromagnetic activation can also be effected by means of a mixture of resonant and non-resonant activation. For example, for a buried doped p-GaN functional layer, the frequency of the radiated electromagnetic radiation may be between 5 and 10 GHz at a power of 100 to 4000 watts. The irradiation can preferably take place over a period of 10 seconds to one hour. In order to avoid so-called "not spots" and a so-called "arcing", ie local overheating and flashovers, the frequency can also be varied. Furthermore, additionally or alternatively, the semiconductor layer sequence relative to the device for irradiation of the electromagnetic Radiation or vice versa rotated and / or translated.

Activation by means of electromagnetic radiation offers the additional advantage of being applicable throughout the chip cycle. With conventional thermal activation, the activation step is always one of the first

Processing steps in the context of a chip run, since this typically requires a very high temperature, for example above 700 ⁰ C, as described above; These high temperatures generally damage many of the "follower components" or "follower layers" which are successively applied to the semiconductor layer sequence in a chip pass within process step A. A targeted coupling of the electromagnetic radiation to the material used, for example a doped functional Layer of p-GaN, on the other hand, makes it possible to integrate this process step at a later stage in the process run, since the wave properties of the electromagnetic radiation can be adjusted so that the activation energy couples very selectively precisely there, and almost only where it is "used", namely, for example, in the dopant-Kodotierstoff- or dopant-codopant-semiconductor crystal-binding complexes to be activated. This allows more freedom in design and in the so-called chip flow, for example with regard to the possible sequence of the individual processes. In addition, the activation efficiency can be increased, for example, by performing the activation after the mesa etching, that is, at a time when a larger open crystal area through the generated mesameters is present and so the codopant can be better removed.

The semiconductor layer sequence or the optoelectronic semiconductor chip formed in method step A can also have a plurality, ie at least two doped functional layers arranged directly adjacent to one another or several doped functional layers, between which further functional layers are arranged. The activation of the plurality of doped functional layers can take place simultaneously in process step B. Alternatively, each of the doped functional layers can be activated in a respective process step B adapted for activation with respect to the parameters mentioned above.

The detection of a change in the local binding states of the codopant can be carried out or carried out in various ways. A particularly sensitive method is spin resonance, as described, for example, in the publication Zvanut et al. , APL 95, 1884 (2004), the disclosure content of which is hereby incorporated by reference. An altered binding of the codopant effectively results in a changed g-factor of a charge carrier to be considered in the vicinity of this bond, the g-factor denoting the so-called gyromagnetic factor or the so-called landing factor. The changed g-factor manifests itself in a changed resonant frequency.

In addition, the binding states of the codopant can also be detected directly via their characteristic vibration frequency in the crystal lattice. Have so For example, Mg-H and NH bonds in GaN as a function of their position and bonding states in the crystal lattice vibrational modes with energies between 2000 and 4000 wavenumbers, which are detectable by Raman spectroscopy and infrared Courier) spectroscopy, as in Neugebauer and van de Walle, PRL 75 , 4452 (1995), Van de Walle, Phys. Rev. B 56, 10020 (1997), Kaschner et al. , APL 74, 328 (1999), Harima et al. , APL 75, 1383, (1999) and Cusco et al., APL 84, 897 (2004), the disclosures of which are hereby incorporated by reference.

Further advantages and advantageous embodiments and developments of the invention will become apparent from the embodiments described below in conjunction with FIGS.

Show it:

Figure IA to ID schematic representations of

Method steps of methods for producing optoelectronic semiconductor chips according to various embodiments,

Figures 2 to 4 are schematic representations of

Method steps of methods for producing optoelectronic semiconductor chips according to further embodiments and

5 shows a measurement of the operating voltage as a function of the activation time.

In the exemplary embodiments and figures, identical or identically acting components may each be provided with the same reference numerals. The illustrated elements and their proportions with each other are basically not to be regarded as true to scale, but individual elements, such as layers, components, components and areas, for better representability and / or better understanding exaggerated be shown thick or large.

In the figures IA to ID are different

Exemplary embodiments of method steps A of methods for producing optoelectronic semiconductor chips are shown. A semiconductor layer sequence 100, 200, 300 and 400, which has at least one substrate 1, one doped functional layer 7, one active region 8 and one further functional layer 9, is formed in each case according to FIGS. For electrical contacting are on a side facing away from the doped functional layer 7 of the substrate 1 and on a side facing away from the substrate 1 surface of the respective semiconductor layer sequence 100, 200, 300, 400th

Electrode layers 10, 11 are applied, which may have one or more metals and / or one or more transparent conductive oxides as described in the general part. The semiconductor layer sequences of the exemplary embodiments shown here are embodied purely by way of example as nitride compound semiconductor layer sequences. Alternatively or additionally, the semiconductor layer sequences may also have other compound semiconductor materials described in the general part.

Furthermore, in the following, semiconductor layer sequences 100, 200, 300, 400 which have already been finished, with regard to the semiconductor chip to be produced in each case, are formed in the following, ie semiconductor layer sequences which, with regard to FIG their respective layer structure already correspond to the finished semiconductor chip. Alternatively, only partially completed

Semiconductor layer sequences are formed in step A, which have at least the doped functional layer 7. Furthermore, the semiconductor layer sequences in method step A can still be formed and provided in a wafer composite before a singulation step to be subsequently performed.

The following description is purely illustrative of a doped functional layer 7, which is p-doped, and on further functional layers, which are then formed corresponding n- or p-type. Alternatively, the polarities of the doped functional layer 7 and the further functional layers or the polarities of their dopants and optionally their codopants may also be reversed, that is, inter alia, the doped functional layer 7 is formed n-doped.

The substrate 1 of the semiconductor layer sequence 100 according to the exemplary embodiment in FIG. 1A is a growth substrate on which the overlying layers 7, 8, 9 are epitaxially grown in the course of method step A. The growth direction is indicated by the arrow 99 in FIG. 1A as well as in the following figures IB to ID.

The growth substrate used in this embodiment is preferably an n-conducting substrate. Possible n-conducting substrates are, for example, n-GaN, n-SiC, n-Si (III). But it is also possible that an electrically non-conductive substrate such as sapphire is used, wherein Here then the electrode layer 10 is disposed on the side facing the layers 7, 8, 9 of the substrate 1.

The functional layer 9 is an n-conducting layer, which in the embodiment shown is formed as a silicon-doped gallium nitride layer. Over the functional layer 9, the active layer 8 is grown, which has a radiation generating single or multiple quantum well structure as the active region. The active layer 8 is preferably based on the III-V semiconductor material system Ga _y In _y N, where 0 <y <1, with alternately arranged optically active layers and barrier layers. Preferably, the active layer 8 is suitable for generating electromagnetic radiation in the ultraviolet, blue, blue-green, yellow or red spectral region, wherein the wavelength of the emitted electromagnetic radiation is adjustable by means of the composition and structure of the active layer 8. The indium concentration in the active layer is preferably between 10 and 60 percent.

On the active layer 8, the doped functional layer 7 epitaxially grown GaN or AlGaN has as a semiconductor material and as a dopant magnesium for p-doping and hydrogen as a co-dopant to degradation of the crystal quality of the semiconductor material, for example by the incorporation of intrinsic defects counteract the incorporation of the dopant during crystal growth. The dopant and the codopant form binding complexes, whereby the free charge carriers actually generated by the dopant are compensated and the electrical neutrality of the dopant is compensated Semiconductor crystal is at least partially restored

The structure of the semiconductor layer sequence 100 corresponds to the arrangement of the n-type functional layer 9 between the substrate 1 and the active layer 8 and the p-type doped functional layer 7 of a conventional light-emitting diode (LED) formed in the growth direction 99 on the active layer 8. and may have other functional layers such as buffer, barrier and / or diffusion barrier layers, which are not shown for clarity.

The semiconductor layer sequence 200 according to the further exemplary embodiment in FIG. 1B has a polarity reversed compared to the semiconductor layer sequence 100, the p-type doped functional layer 7 between the growth substrate 1 and the active layer 8 and the n-conductive further functional layer 9 in the growth direction 99 are formed on the active layer 8. The respective layer composition of the layers 7, 8 and 9 corresponds to the previous embodiment.

Furthermore, the semiconductor layer sequence 200 between the substrate 1, which is formed as in the previous embodiment as an n-type growth substrate, and the p-type doped functional layer 7, an n-type further functional layer 2 of silicon-doped GaN. For the effective electrical connection of the p-type doped functional layer 7 to the n-type further functional layer 2, a tunnel junction 3 with a highly doped n-type tunnel junction layer 4, a diffusion barrier layer 5, is interposed therebetween and a heavily doped p-type tunnel junction layer 6. The tunnel junction 3 is designed as described in the general part, wherein the p-type tunnel junction layer 6 as the p-type doped functional layer 7 as a dopant magnesium and as a codopant hydrogen. Thus, the highly doped p-type tunnel junction layer 6 as well as the doped functional layer 7 is designed as a doped functional layer to be activated in the sense of the present description.

In contrast to the semiconductor layer sequence 100 of the exemplary embodiment according to FIG. 1A, the doped functional layers 6 and 7 are in the

Semiconductor layer sequence 200 formed as a so-called buried doped functional layers, which are arranged between other functional semiconductor layers. An activation of the layers 6 and 7 by a known activation method by expelling the Kodotierstoffs is therefore not possible for the semiconductor layer sequence 200.

The semiconductor layer sequence 200 may include further functional layers (not shown) such as a buffer layer between the substrate 1 and the functional layer 2 and / or a diffusion barrier layer between the doped functional layer 7 and the active layer 8.

FIG. 1C shows a semiconductor layer sequence 300 embodied as a thin-film semiconductor chip, which likewise has a buried doped functional layer 7. The layers 7, 8 and 9 correspond to the layers 7, 8 and 9 in Figure IA, which after growing on a Growth substrate, such as sapphire, were transferred by Umbonden on a carrier substrate 1, which is why the growing direction 99 in the direction of the carrier substrate 1 shows. The growth substrate was removed after bonding. The semiconductor layer sequence 300 may have further functional layers, for example a reflective layer between the carrier substrate 1 and the p-type doped functional layer 7, and / or further features of thin-film semiconductor chips described in the general part. As a result of the bonding, the doped functional layer 7 likewise exists as a buried layer, which can not be activated after the bonding by means of known activation methods based on expulsion of the codopant. In the known activation methods, the activation would preferably have been carried out at the time before the bonding, to which the doped functional layer 7 was still exposed.

Alternatively to the embodiment as a thin-film semiconductor chip on a carrier substrate, the layer sequence of the layers 7, 8 and 9 with the doped functional layer 7 between the substrate 1 and the active layer 8 can also be formed by epitaxial growth on a p-type growth substrate. In this case, the p-type substrate may be, for example, p-GaN, p-SiC or p-Si (IIl), in which case the growth direction 99 would be directed away from the substrate 1.

FIG. 1D shows a semiconductor layer sequence 400 which has an inverted structure according to FIG. 1B, which is furthermore designed as a stacked construction with a further active layer 8 '. The doped functional layer 7 'corresponds to the doped functional layer 7. The further functional layers 3 'and 9' correspond to the layers 3 and 9, the tunnel junction 3 ', like the tunnel junction 3, having the tunnel junction layers 4, 6 described in connection with FIG. 1B and the diffusion barrier layer 5 (not shown).

In the activation methods described in the further method step B in advance in the general part and in the following according to the exemplary embodiments, the point in time in the production process at which the activation according to method step B is carried out is independent of the formation and the manufacturing process of the respective semiconductor layer sequence. By way of example, the exemplary embodiments for method step B are shown with reference to the semiconductor layer sequence 200 according to FIG.

In method step B according to the exemplary embodiment of FIG. 2, the doped functional layer 7 as well as the highly doped p-conducting tunnel junction layer 6 are activated by introducing energy in the form of electrical energy. For this purpose, the semiconductor layer sequence 200 is connected to an external power and voltage supply 12. In this case, a current density of about 50 A / cm ² in the semiconductor layer sequence 200 or in particular in the doped functional layer 7 and in the highly doped p-type tunnel junction layer 6 is generated in the illustrated embodiment. Furthermore, the semiconductor layer sequence 200 is brought by supplying heat energy 13 to a temperature above the usual ambient and operating temperature. In the exemplary embodiment shown, the semiconductor layer sequence 200 is heated to a temperature of at least 80 ^° C. by an external heater (not shown). At least part of the supplied Heat energy can also be caused by ohmic losses of the impressed current.

It has been shown experimentally that the required operating voltage, which is provided by the current and voltage supply 12 for operating the semiconductor layer sequence 200 under the aforementioned conditions, continuously decreases with time and reaches a saturation value which remains permanently. This means that the current-voltage characteristics of the

Semiconductor layer sequence can be improved by the method step B and can be permanently maintained after step B. FIG. 5 shows a measurement of the operating voltage U (in arbitrary units) to be applied for a specific operating current as a function of

Activation time t (in arbitrary units) of the electrical activation shown. Furthermore, it has been found that increasing the current density and / or the temperature can cause an acceleration of the voltage drop and the reaching of the saturation.

By virtue of the aforementioned activation operation of the semiconductor layer sequence 200, the dopant codopant and the dopant codopant semiconductor crystal bond complexes which have formed in method step A during the production of the semiconductor layer sequence 200 can be broken up in the layers 6 and 7. It can additionally be achieved in comparison to conventional activation methods that at least part of the codopant is bound at other locations, that is to say not in a manner forming the dopant binding complex, in the semiconductor crystal of the layers 6 and 7 or stored in the intermediate grid. That's it in the case of the electrical activation shown here, it is not necessary to expel the codopant at least partially from the semiconductor layer sequence, as is absolutely necessary in the case of the known purely thermal activation methods.

Due to the above-mentioned advantageous for the process step B current density, the electrical activation by Stromaufprägung by the external power and voltage supply 12 shown is particularly suitable for already isolated semiconductor layer sequences with at least one buried doped functional layer 7, where a conventional activation method technically hardly or not at all is feasible. In addition, the application of the process step B shown here to semiconductor layer sequences in the wafer composite is by no means excluded.

In particular for semiconductor layer sequences in the wafer composite but also for already isolated semiconductor layer sequences, method step B according to the exemplary embodiment of FIG. 3 is suitable. In this case, by means of a device 14 by induction, a current in the semiconductor layer sequence 200 or at least in the doped functional layer 7 and also to be activated highly doped p-type tunnel junction layer 6 generates and so energy to break the binding complexes with the dopant and the Kodotierstoff out. The induction device 14 is realized purely by way of example in the illustrated embodiment by coils, in which case any device that has a sufficient Induktionsström in the

Semiconductor layer sequence 200 may be suitable. Through the device 14, circulating currents are induced in the layers 6 and 7 by means of the free charge carriers, by means of which the activation effect described above in connection with FIG. 2 can be achieved. The circulating currents are generated perpendicular to the growing direction 99 and parallel to the plane of extent of the functional layers of the semiconductor layer sequence 200. In addition, the semiconductor layer sequence 200 can still be supplied with thermal energy in the form of an external heater (not shown) and / or by ohmic losses of the circulating currents.

Furthermore, in the exemplary embodiment shown in FIG. 3, the semiconductor layer sequence 200 is irradiated with electromagnetic radiation 15 which is resonant or non-resonant with the absorption wavelengths of the functional layers and in particular of the layers 6 and 7 to be activated. By the irradiation of the electromagnetic radiation 15 additional free charge carriers are generated, which allow a greater current intensity of the induced circular currents. This is particularly advantageous if intrinsically after method step A only very few or no free charge carriers are present in the layers 6 and 7 to be activated. In particular, in the case of resonant irradiation, further free charge carriers can be specifically excited in the layers 6 and 7 to be activated, whereby the efficiency of the activation can be increased.

As shown in FIG. 4 according to a further exemplary embodiment of method step B, the activation, ie the breaking up of the complexes with the dopant and the codopant, can also be effected only by supplying energy in the form of electromagnetic radiation 15. The frequency of the electromagnetic radiation 15 determines the type the activation. When using microwave radiation, the electromagnetic activation is not resonant, when using terahertz radiation is a resonant electromagnetic activation. The process conditions such as frequency, power, atmosphere, time and / or additional absorption centers for the electromagnetic radiation 15 define the degree and success of the activation. In the non-resonant case, the transfer of energy to atomic bonds occurs, among other things, by excitations of phonons and rotons. In the resonant case, lattice vibrations, ie phonons, are generated directly, which can break up the binding complexes with the dopant and the codopant.

The method steps B according to FIGS. 3 and 4 can advantageously be used throughout the production process of the semiconductor chips within the scope of a chip pass on the wafer level or also after singulation, since these take place without contact and, as also does the method step B according to the exemplary embodiment in FIG. require no exposed doped functional layer 7 to be activated. This allows more freedom in design and in the so-called chip flow with regard to the possible sequence of individual processes. Furthermore, the activation efficiency can be increased, for example, by performing the activation after etching of mesen, so that a larger open semiconductor crystal surface is present and thus at least part of the codopant can also be removed.

The embodiments of method steps B described here as well as the further embodiments according to the general part are also in combination or successively applicable for individual layers to be activated as well as for a plurality of layers to be activated within a semiconductor layer sequence.

If necessary, further known method steps for completing the semiconductor chip with now activated semiconductor layer sequence can follow the method step B.

The invention is not limited by the description based on the embodiments of these. Rather, the invention encompasses any novel feature as well as any combination of features, including in particular any combination of features in the claims, even if this feature or combination itself is not explicitly stated in the patent claims or exemplary embodiments.

Claims

claims 1. A method for producing an optoelectronic semiconductor chip, comprising the steps: A) forming a semiconductor layer sequence (200) with at least one doped functional layer (7) having binding complexes with at least one dopant and at least one codopant, one selected from the dopant and the codopant being an electron acceptor and the other an electron donor, B) activation of the dopant by breaking the Binding complexes by introducing an energy, wherein the codopant remains at least partially in the semiconductor layer sequence (200) and at least partially does not form binding complexes with the dopant. 2. The method of claim 1, wherein - The semiconductor layer sequence (200) is formed in step A such that the doped functional layer (7) between two further functional layers (2, 8) is arranged. 3. The method according to claim 1 or 2, wherein - In process step A, the semiconductor layer sequence (200) finished in a wafer composite and then separated. 4. The method according to any one of the preceding claims, wherein - The dopant magnesium and the codopant has hydrogen. 5. The method according to any one of the preceding claims, wherein - In process step B, the introduction of the energy Generation of a current in the doped functional layer (7) takes place. 6. The method of claim 5, wherein - The current is generated without contact by induction. 7. The method of claim 5, wherein - The current by electrical connection of the at least doped functional layer (7) to an external power source (12). 8. The method according to any one of claims 5 to 7, wherein - In addition to the generation of electricity, a heat energy (13) is supplied. 9. The method of claim 8, wherein - The heat energy (13) is at least partially supplied by the stream. 10. The method according to any one of the preceding claims, wherein - In process step B, the introduction of the energy Radiation of electromagnetic radiation (15) takes place. 11. The method of claim 10, wherein - The electromagnetic radiation (15) is at least partially resonant with an absorption wavelength of the doped functional layer (7). 12. The method according to claim 10 or li, wherein - The electromagnetic radiation (15) is at least partially non-resonant with an absorption wavelength of the doped functional layer (7). 13. An optoelectronic semiconductor chip, comprising a semiconductor layer sequence (200) with at least one doped functional layer (7) with at least one dopant and at least one codopant, wherein the semiconductor layer sequence (200) has a semiconductor material with a lattice structure, one selected from the dopant and the codopant is an electron acceptor and the other is an electron donor, - The codopant is bound to the semiconductor material and / or arranged on interstitial sites and - The codopant at least partially none Forms binding complexes with the dopant. 14. The semiconductor chip according to claim 13, wherein - The doped functional layer (7) between two further functional layers (3, 8) is arranged. 15. The semiconductor chip according to claim 13 or 14, wherein - The dopant magnesium and the codopant has hydrogen.