JP2017050302A - Indirect transition type semiconductor light-emitting element - Google Patents

Abstract

PROBLEM TO BE SOLVED: To remarkably improve light emission efficiency by optimizing the distribution of dopant.SOLUTION: Heat is generated based on light absorption in a junction of a (p) layer 14 and an (n) layer 13 by applying a forward bias voltage and radiating a light of which the wavelength is shorter than an absorption end wavelength in a semiconductor layer and which has a linearly polarized component. Based on the generation of heat, the distribution of dopant in any one or more layers is iteratively changed. In a place where a proximate field light is generated based on the changed dopant distribution, a light is emitted by guiding and discharging electrons in a conduction band generated based on the bias voltage over a plurality of stages based on a non-heat-insulation process, and the changed dopant distribution is fixed by suppressing the generation of heat with the light emission. Both directions of dopants constituting the changed dopant distribution and being most neighboring to each other are vertical to an irradiation direction of the light to be radiated and a direction of a linearly polarized light.SELECTED DRAWING: Figure 7

Description

  The present invention relates to an indirect transition type semiconductor light emitting device, and more particularly to an indirect transition type semiconductor light emitting device suitable for further improving luminous efficiency.

  In recent years, light-emitting elements such as electroluminescence elements and LEDs have attracted attention as light-weight and thin surface-emitting elements. This light emitting element is an element utilizing the principle that a fluorescent substance emits light by recombination energy between holes injected from an anode and electrons injected from a cathode when an electric field is applied. The light emitting element is mainly used for a light emitting element of a display of various devices such as a mobile device. In particular, when this light emitting device is actually used, it can be said that it is more desirable to emit near infrared light with little optical loss.

  Among these, silicon is an inorganic semiconductor having a band gap in the near infrared. However, since it is an indirect transition semiconductor, it is not desirable as a near infrared light emitting element. In particular, in recent years, light-emitting elements using porous silicon in which silicon is microcrystallized have been proposed, but their light emission efficiency is also less than 1%. For this reason, a technique capable of improving the light emission efficiency in the near infrared in a light emitting element using silicon as a light emitting material has been desired.

  For this reason, the technique disclosed in Patent Document 1 has been proposed. According to the technique disclosed in Patent Document 1, a forward bias voltage is applied to a light emitting device including a semiconductor layer including a p layer and an n layer, and light having a wavelength shorter than the absorption edge wavelength in the semiconductor layer is irradiated. Keep doing. Thus, heat generation based on light absorption is generated at the junction between the p layer and the n layer, and the dopant distribution of any one or more layers is repeatedly changed based on the heat generation. As a result, near-field light is generated based on the changed dopant distribution, and light can be emitted by stimulated emission of electrons in the conduction band in a plurality of stages based on a non-adiabatic process. With this method, it is possible to improve the light emission efficiency in the near infrared in a light emitting element using silicon as a light emitting material.

  However, the disclosed technique of Patent Document 1 only describes a mechanism for using silicon as a near-infrared light emitting element, and specifically describes the spatial distribution of dopants for actually improving the light emission efficiency. It has not been.

JP, 2014-044836, A

  Therefore, the present invention has been devised in view of the above-described problems, and the object of the present invention is to optimize the dopant distribution when an indirect transition type semiconductor such as silicon is used as a light emitting element. It is an object of the present invention to provide an indirect transition type semiconductor light emitting device capable of dramatically improving the light emission efficiency than before.

  In a first aspect of the present invention, there is provided an indirect transition type semiconductor light emitting device including a semiconductor layer including a p layer and an n layer, a forward bias voltage is applied, a wavelength shorter than the absorption edge wavelength in the semiconductor layer, and linearly polarized light Irradiation with light having a component causes heat generation based on light absorption at the junction between the p layer and the n layer, and changes the dopant distribution of any one or more of the layers based on the heat generation. When the near-field light is generated based on the changed dopant distribution, the electrons in the conduction band generated based on the bias voltage are stimulated and emitted in multiple steps based on a non-adiabatic process. By making the light emitted, and suppressing the heat generation accompanying the light emission, the dopant distribution after the change is fixed, and the dopant distribution after the change is changed. Direction of dopant that is the closest to each other to formed is characterized in that there is a both a direction perpendicular to the irradiation direction and the direction of linear polarization of light that is the irradiation.

  According to a second invention, in the first invention, ΔE is an energy difference between the band gap of the junction and the energy of the light emission, ΔEf is an energy of an optical phonon localized in the distance between the adjacent dopants, and ΔE / ΔEf. When N is set to N, the most adjacent dopant spacing in the above-mentioned direction is N times the lattice constant a, which is the most distributed.

  According to a third invention, in the first or second invention, the direction of the nearest neighbor dopant is Δk, which is a wave number difference between the valence band and the conduction band in the junction, and is localized in the distance between the nearest neighbor dopants. When the wave number of the phonon is Δkf, Δkf = Δk / N.

  According to a fourth invention, in any one of the first to third inventions, the optical phonon is distributed so as to have a higher probability of existence between the adjacent dopants.

  In a fifth invention according to any one of the first to fourth inventions, the n layer is Si, the p layer is doped with B as a dopant to Si, and the dopant after the change is The interval between the adjacent Bs constituting the distribution of B is N (= 3) times as large as the lattice constant a, and is distributed most frequently.

  According to the present invention having the above-described configuration, it is possible to dramatically improve the light emission efficiency than before by optimizing the dopant distribution.

It is a figure which shows the structure of the indirect transition type semiconductor light-emitting device to which this invention is applied. It is an energy band figure of the semiconductor layer which consists of n layer-p layer. It is a figure which shows the detail of each energy band of a valence band and a conduction band in a joining layer. It is a figure which shows microscopic dopant distribution of the junction part of n layer and p layer before and behind Joule heat generation. It is a figure which shows the real space orientation of the reciprocal space of the silicon unit cell which comprises a semiconductor layer. It is a figure for demonstrating a non-adiabatic process with the model which replaced the coupling | bonding of atoms with the spring. It is a figure for demonstrating the example which carries out the virtual transition to the phonon vibration level located in the middle of a band gap, and discharge | releases an electron from there in the conduction band. It is a figure for demonstrating the manufacturing method of the indirect transition type light emitting element which can radiate | emit the light of a desired linearly polarized light direction. It is a figure which shows distribution of the boron (B) as a dopant actually measured by the three-dimensional atom probe method. (A) is a figure which shows the number distribution with respect to the distance L of the adjacent dopant (B) before irradiating light with respect to a semiconductor layer, (b) is after irradiating light with respect to a semiconductor layer. It is a figure which shows the number distribution with respect to the distance L of the nearest neighbor dopant (B) in. It is a figure which shows the result of having computed deviation (ratio) from the Weibull distribution actually shown with a straight line from the data of Drawing 10 (b). It is a figure which shows the result of having investigated the direction dependence of the dopant (B) which adjoins when the irradiation direction of light is made into the V direction in a figure.

  Hereinafter, an indirect transition type semiconductor light emitting device to which the present invention is applied will be described in detail with reference to the drawings.

  FIG. 1 shows a configuration of an indirect transition type semiconductor light emitting device 1 to which the present invention is applied. This indirect transition type semiconductor light emitting device 1 includes an N-type semiconductor layer (n layer) 13, a P-type semiconductor layer (p layer) 14 constituting a pn junction with the n layer 13, an n layer 13 and a p layer 14, And a bonding layer 35 formed therebetween. The layers from n layer 13 to p layer 14 are referred to as semiconductor layer 30. A power supply 2 is connected to the p layer 14, and in use, a bias voltage is loaded in the forward direction so that the p layer 14 side is a positive voltage and the n layer 13 side is a negative voltage. However, this bias voltage may be weak.

The n layer 13 is composed of a so-called silicon substrate or the like, but is not limited to this, and may be another indirect transition type semiconductor. Typical indirect transition semiconductor inorganic compounds include GaP, AlGaAs (depending on the mixed crystal ratio), AlP, AlAs, Ge, SiC, PbS, PbTe, TiO _2, GaS, AlSb, C (diamond). ), BN, etc., and this method can be applied to all of them. However, in the case of applying any semiconductor material, it is assumed that a dopant as an impurity is implanted on the p layer 14 side.

The p layer 14 is configured by implanting silicon, for example, with boron or the like as a p-type dopant with high density and high energy. The p layer 14 may be made of the same material as the indirect transition semiconductor inorganic compound used for the n layer 13 in addition to silicon. The p layer 14 may have an implantation energy of 700 KeV in ion implantation and a doping density of 2 × 10 ¹⁹ atoms / cm ³ in the vicinity of 1500 nm from the surface.

  When manufacturing such an indirect transition type semiconductor light emitting device 1, a forward bias voltage is applied to the p layer 14 and the n layer 13, and the semiconductor layer 30 is irradiated with light. As a result, the indirect transition semiconductor light emitting device 1 according to the present invention can be manufactured based on the following mechanism.

  FIG. 2 shows an energy band diagram of the semiconductor layer 30 including the n layer 13 to the p layer 14. When the semiconductor layer 30 is irradiated with light, if the irradiated light is shorter than the absorption edge wavelength corresponding to the energy difference between the valence band and the conduction band, the energy is greater than the energy difference. The semiconductor layer 30 absorbs light. In accordance with this light absorption, a number of electrons are generated in the conduction band and a number of holes are generated in the valence band.

  FIG. 3 shows the details of the energy bands of the valence band and the conduction band in the bonding layer 35. The electrons excited and generated in the conduction band move on the conduction band and have the most energy. It will concentrate on the lower level. Similarly, the holes generated in the valence band move on the valence band and concentrate to the highest energy level. Since the semiconductor layer 30 is an indirect transition semiconductor, the wave number of the conduction band with the lowest energy level and the wave number of the valence band with the highest energy level are different from each other.

  When such light absorption occurs, heat is generated as much as energy is absorbed. The part where the heat generation is particularly large is the surface of the bonding layer 35, the n layer 13, the p layer 14, or the like. Further, heat generation is promoted by increasing the intensity of light to be irradiated.

  As a result of the generation of heat due to light absorption, the fluidity in the bonding layer 35, the n layer 13 and the p layer 14 increases, and the distribution of the dopant changes. By continuing to apply the forward bias voltage described above, such a change in the distribution of dopants will continue to occur.

  FIG. 4A is an example of a microscopic distribution of the dopant 31 in the junction 35 between the n layer 13 and the p layer 14 before the generation of Joule heat. 4 focuses on the formation position 51 of the dopant 31 in the real space orientation of the reciprocal space of the silicon unit cell constituting the semiconductor layer 30 as shown in FIG. That is, it is a diagram in which only the formation position 51 is arranged side by side for each adjacent unit cell. As shown in FIG. 4A, as a result of the dopant (boron) 31 being implanted into the n-type Si substrate constituting the n layer 13 at the junction interface between the n layer 13 and the p layer 14, the dopant 31 is A state of being randomly distributed in any unit cell in the n layer 13 is shown.

  FIG. 4B is an example of a microscopic distribution of the dopant 31 at the junction 35 between the n layer 13 and the p layer 14 after heat generation due to light absorption occurs. Incidentally, in the example of FIG. 4B, the light irradiation direction is the depth direction of the paper surface in the drawing, and the light including the polarization component composed of the linear polarization direction indicated by the arrow in the drawing is irradiated. As a result of heat generation due to light absorption, the fluidity of the junction 35 between the n layer 13 and the p layer 14 increases, and as a result, the distribution of the dopants 31 such as the n layer 13 and the p layer 14 changes randomly. Become. As a result of repeated changes in the distribution of the dopant 31, for example, the interval between the dopants 31 that are closest to each other also continues to change. Among these, the directions of the adjacent dopants are gradually arranged so as to be most distributed in the direction perpendicular to the light irradiation direction and perpendicular to the direction of linearly polarized light. In the example of FIG. 4B, it is clear that the direction of the nearest neighbor dopant (the arrow direction of b in the figure) is perpendicular to the light irradiation direction and perpendicular to the direction of linearly polarized light. It is. The case where the distance between the adjacent dopants 31 is the distance L is formed by chance with a certain probability. This interval L can be represented by a lattice constant a × N (where N is a positive integer) in the unit cell. In the case of FIG. 4B, since the two unit cells are separated between the adjacent dopants, the distance L is the lattice constant a × N and N = 3.

  If the distance between the adjacent dopants 31 is the distance L, it can be said that the distance is more suitable for generating near-field light based on the incident light. The near-field light is not limited to the case where the near-field light is generated at the interface between the n layer 13 and the p layer 14, and the near-field light of any one or more layers of the dopant 31 constituting the indirect transition type semiconductor light emitting device 1. Any device that changes the distribution may be used.

  If the above-mentioned light irradiation is further continued when the distance between the adjacent dopants 31 is L, the near-field light is likely to be localized between the distances L between the adjacent dopants 31. This is because since the mass of boron as the dopant 31 is heavier than the mass of silicon, the region between the dopants 31 becomes a reflection surface for vibrations of phonons, the phonons are localized, and near-field light is likely to be generated. Since the near-field light here includes the meaning of a virtual electromagnetic field, the formation of a virtual electromagnetic field is understood as meaning the generation of near-field light. The generation of near-field light may occur as so-called induced light, or spontaneous emission of injected charge during forward current injection, even in the absence of induced light. And stimulated emission based on it. Generation of this near-field light causes a non-adiabatic process described below.

  This non-adiabatic process can be considered by a model in which the bonds between atoms are replaced by springs as shown in FIG. In general, the wavelength of propagating light is much larger than the size of the molecule, so it can be regarded as a spatially uniform electric field at the molecular level. As a result, as shown in FIG. 6A, adjacent electrons are vibrated by the spring with the same amplitude and the same phase. Since the nucleus is heavy, it cannot follow the vibration of this electron, and the molecular vibration is extremely difficult to occur in the propagating light. In this way, in propagation light, it can be ignored that molecular vibrations are involved in the excitation process of electrons, so this process is called adiabatic process (T. Kawazoe, K. Kobayashi, S. Takubo, and M. Ohtsu, J Chem. Phys., Vol. 122, No. 2, January 2005, pp. 024715 1-5).

  On the other hand, the spatial electric field gradient of near-field light drops very steeply. For this reason, near-field light gives different vibrations to adjacent electrons, and as shown in FIG. 6B, the heavy nuclei are also vibrated by the vibrations of the different electrons. The occurrence of molecular vibration in the near-field light corresponds to energy taking the form of molecular vibration. Therefore, in the near-field light, an excitation process (non-adiabatic process) via the vibration level is possible. In this way, the excitation process via the vibration level of the nucleus is called a non-adiabatic process because the nucleus moves in response to the adiabatic process, which is a normal optical response.

  Further, by continuing to apply the forward bias voltage described above, electrons move from the n layer 13 side to the p layer 14 side in the conduction band. As a result, on the basis of a non-adiabatic process using near-field light for electrons moving in the conduction band, as shown in FIG. Based on the above, it is possible to emit light by virtually transitioning electrons in the conduction band to the phonon vibration level 41 located in the middle of the band gap and emitting the electrons therefrom. Moreover, it can be made to emit light by carrying out the induced emission of the electron in a conduction band in several steps based on this near field light.

In the example of FIG. 7, the phonon vibration level 41 is formed over three stages. The electrons in the conductor fall in order from the top to the bottom of the phonon vibration level 41 over the three stages, and the ground level (highest energy in the valence band) from the bottom phonon vibration level 41. It emits light when it is released to the energy level. When the emission energy is E ₁ and the band gap of the junction (from the energy level with the highest valence band energy to the energy level with the lowest conduction band energy) is E ₂ , E ₁ <E ₂ When expressed by ΔE = E ₂ −E ₁ , the energy corresponding to ΔE is supplied by phonons from the viewpoint of the energy conservation law. In other words, what corresponds to this ΔE is the phonon vibration level 41 over three stages. Therefore, the energy per stage of this phonon vibration level 41 is equivalent to ΔE / 3, and it is just ΔE when these are composed of just three stages, which is realized by filling the energy conservation law with phonons. It will be possible.

  The horizontal axis of FIG. 7 is the wave number. The wave number at the highest energy level of this valence band energy is Γ, and when the wave number at the lowest energy level in the conductor is X, The wave number difference Δk between the valence band and the conduction band is represented by X−Γ. From the viewpoint of the wave number conservation law, the wave number difference Δk between the valence band and the conduction band also needs to be provided by phonons. In this example, phonon vibration levels 41 are formed in three stages. Therefore, the wave number per stage of this phonon vibration level 41 is equivalent to Δk / 3, and when these are constituted by just three stages, it becomes exactly Δk, which is realized by filling the wave number conservation law with phonons. It will be possible.

  That is, the phonon vibration level 41 has an energy per stage of ΔE / 3 and a wave number per stage of Δk / 3, so that the phonon vibration level is composed of three stages. Thus, both the energy aspect and the wave number aspect satisfy the conservation laws. As a result, even when the semiconductor layer 30 is composed of an indirect transition type semiconductor, it is possible to realize light emission by electron emission.

Note that the formation of the phonon vibration level 41 over three stages means that, as shown in FIG. 4B, the interval L between the adjacent dopants 31 is a lattice constant a × N, and N = 3. Now consider ΔE. Since the optical phonon energy of Si is fixed at 70 meV (in the case of silicon), to fill the difference of 200 meV (ie ΔE) between the band gap energy E ₂ 1.14 eV and the DPP annealing wavelength (photon energy E ₁ : 0.94 eV) Three phonons 70meV × 3? 200meV are required. This phonon of 70 meV corresponds to ΔEf (energy conservation law).

  The wave number Γ at the energy level having the highest energy in the valence band corresponds to the origin on the xyz axis shown in FIG. 5, and the wave number X at the lowest energy level in the conductor is the direction of X shown in FIG. Corresponding to Incidentally, this X can be any of the x-axis, y-axis, and z-axis. This wave number X corresponds to the arrangement direction of the dopant. That is, the arrangement direction of the dopant is any one of (001), (010), and (100). Incidentally, the arrangement direction of this dopant determined from this wave number coincides with the direction of the most adjacent dopant in the example of FIG. 4B described above (the arrow direction of L in the figure). That is, in the example of FIG. 4B, the direction of the nearest neighbor dopant (the arrow direction of L in the figure) corresponds to the array orientation of (001), (010), and (100).

  Since near-field light continues to be generated under the interval L, a non-adiabatic process can be generated. In the stimulated emission by this non-adiabatic process, electrons are emitted through the phonon vibration level 41. At this time, even in light having a wavelength longer than the absorption edge wavelength corresponding to the band gap width, electrons in the conduction band can be made to transition in multiple stages and emitted.

  If only light irradiation is performed without applying a forward bias voltage, near-field light can be generated based on the mechanism described above, but electrons for stimulated emission are secured in multiple stages. Can not emit light. For this reason, the application of the forward bias voltage is necessary even if it is weak. By applying this forward bias voltage, electrons can be generated in the conduction band, and by using this for stimulated emission, the above-described light emission can be realized.

  Regarding a region where the distance between the adjacent dopants 31 is L, which is suitable for generating such near-field light due to the occurrence of multi-step stimulated emission due to such a non-adiabatic process and light emission. The light absorption itself by the irradiation light can be canceled out. This is because, as a result of trying to transition so that electrons and holes generated by light absorption are recombined by light emission, the region where the distance between the adjacent dopants 31 where such near-field light is generated is L, Light emission (light emission) can suppress heat generation based on light absorption. As a result, the temperature of the region where the distance between the adjacent dopants 31 in which near-field light is generated is L is lowered. And about the area | region where the space | interval of this adjacent dopant 31 was set to L, the random change of the distribution of the dopant 31 based on a Joule heat will be suppressed. As a result, the distance L between the adjacent dopants 31 is fixed without being changed.

  When light emission occurs as shown in FIG. 7, near-field light due to the distribution of the dopant 31 is likely to be generated based on the light emission. Due to the generated near-field light, a non-adiabatic process is more likely to occur in each part, and fixing of the interval L between the adjacent dopants 31 and light emission are promoted.

  Further, the above-described mechanism is continuously generated by continuing to irradiate the light as described above and continuously applying the forward bias voltage. As shown in FIG. 4B, the distance L between the adjacent dopants 31 is such that near-field light continues to be generated and electrons flowing based on the forward bias voltage are continuously generated by the above-described non-adiabatic process. As a result of the induced emission, the temperature is lowered and the fixing is continued in this shape. In addition, since the near-field light is not generated in the portions other than the distance L between the adjacent dopants 31, the portion is not cooled, and heat is generated due to light absorption as it is, so that the junction between the n layer 13 and the p layer 14. As a result of the increase in the fluidity of 35, the distribution of the dopants 31 such as the n layer 13 and the p layer 14 changes randomly. As a result of this random change, a case where the distance between the adjacent dopants 31 is L is newly formed with a certain probability. When light is incident in such a case, near-field light is likely to be generated between the newly formed dopants 31 having the distance L. As a result of stimulated emission caused by the non-adiabatic process between the newly formed dopants 31 having the distance L, the temperature is lowered. As a result, the distance between the adjacent dopants 31 is kept fixed in the L state. That is, the region where the near-field light is suitably generated in the indirect transition semiconductor light emitting device 1 is sequentially increased. By reducing the amount of heat generated due to light absorption by light emission, the distribution of the dopant 31 is fixed repeatedly.

  If this process is repeatedly performed, ideally, a large number of dopants 31 having the nearest neighbor distance L at the interface between the n layer 13 and the p layer 14 are formed. This can be configured as an indirect transition type semiconductor light emitting device 1 in which a large number of dopants 31 having the nearest neighbor interval L, in which near-field light is suitably generated when light irradiation is performed, are created. As a result, the luminous efficiency can be dramatically improved.

  Next, the operation of the indirect transition type semiconductor light emitting device 1 manufactured based on the manufacturing method described above will be described.

  As described above, in the indirect transition type semiconductor light emitting device 1, near-field distance as shown in FIG. Many arrangements of the dopant 31 are formed. When a light emission operation is actually performed by the manufactured indirect transition type semiconductor light emitting device 1, a forward bias voltage is applied to the indirect transition type semiconductor light emitting device 1 without irradiating light as described above. It may be. As a result, since a shape capable of generating near-field light is already created, when a forward bias voltage is applied, a large amount of near-field light is obtained as shown in FIG. It occurs in the area. Then, as shown in FIG. 7, the electrons in the conduction band are stimulated and emitted in multiple steps through the phonon vibration level 41 by the non-adiabatic process by the generated near-field light, and emit light. At this time, when the intensity of the forward bias voltage is further increased, avalanche breakdown occurs and the amount of light emission is further increased.

  As described above, in the present invention, light is irradiated and a forward bias voltage is applied to cause light absorption at the junction 35 between the p layer 14 and the n layer 13 and indirectly based on heat generated by light absorption. The distribution of the dopant 31 in any one or more layers constituting the transition type semiconductor light emitting device 1 is repeatedly executed.

  Then, in the distribution of the dopant 31 after the change, near-field light is suitably generated between the dopants 31 whose nearest neighbor interval is L, and electrons in the conduction band are induced and emitted in a plurality of stages based on a non-adiabatic process. As a result, the amount of heat generated by light absorption is reduced, the temperature is lowered, and the distribution of the dopant 31 is fixed.

  Further, in the distribution of the dopant 31 after the change, since the phonons are not localized between the dopants 31 whose nearest neighbor interval is other than L, the near-field light is not suitably generated and returns to the propagating light immediately. As a result, between the dopants 31 whose nearest neighbor spacing is other than L, heat generated by light absorption is continuously generated to change the dopant 31 distribution until the nearest neighbor spacing between the dopants 31 becomes L.

  Thereby, in the present invention, light having a wavelength longer than the absorption edge wavelength corresponding to the band gap width of the joint portion 35 can be emitted. If the n layer 13 is silicon, it is possible to emit light in the near infrared region as the emission wavelength of the silicon.

  In addition, in the method of manufacturing the indirect transition type semiconductor light emitting device 1 to which the present invention is applied, a light receiving device excellent in sensitivity to a desired wavelength can be manufactured at a low cost without requiring a particularly large apparatus. It becomes.

The wavelength band above Si, GaP, (depending on alloy composition) AlGaAs, AlP, AlAs, Ge , SiC, PbS, PbTe, TiO 2, GaS, AlSb, C ( diamond), an indirect transition type inorganic employed such as BN By changing the type of material, a wide range from ultraviolet to infrared light can be dealt with.

  In the embodiment described above, silicon is used as the semiconductor layer 30 and boron (B) is applied as the dopant 31, and the interval L between the adjacent dopants 31 is a lattice constant a × N. And the case where N = 3 has been described as an example. However, it is a matter of course that even when other indirect transition type semiconductors are applied, the light emission efficiency may be improved by applying the same technical idea. In such a case, the multiple N of the lattice constant a depends on the energy difference ΔE between the band gap of the junction 35 and the energy of light emission, and the wave number difference Δk between the valence band and the conduction band at the junction 35. Anything can be used. At this time, the multiple N of the lattice constant a may be selected so as to satisfy the energy ΔEf of the optical phonon localized at the distance L between the adjacent dopants 31 and N = ΔE / ΔEf. At this time, it suffices that adjustment is made so that the number N of multiples is distributed most in the joint portion 35, and it is a matter of course that things other than the multiple N may be included.

  In addition to this, Δkf = Δk / N may be set, where Δkf is the wave number of the optical phonon localized in the interval between the adjacent dopants.

The present invention can also be explained as follows in relation to N. ΔE described above can be easily specified from the material system selected for the n layer 13 and the p layer 14. The energy ΔEf of the optical phonon is, for example,
Compaan, A., Lee, MC, Trott, GJPhys. Rev. B32 (1985) 6731,
Kobliska, R., Solin, SA, Selders, M., Chang, RK, Alben, R., Thorpe, MF, Weaire, D, Phys. Rev. Lett. 29 (1972) 725 etc. Values are disclosed. By obtaining ΔEf from such literature values, ΔE / ΔEf = N is obtained. As for the distance between the adjacent dopants in the direction b described above, a × N is most frequently distributed between the determined N and the lattice constant a.

  In the present invention, as shown in FIG. 8, when the incident light at the time of production has a linearly polarized component, the adjacent direction b of the most adjacent dopant 31 is relative to the linearly polarized direction β of the irradiated light. As described above, the pixels are naturally arranged in the vertical direction. As a result, the light emitted from the obtained indirect transition type light emitting element 1 also has a strong component in the linear polarization direction β.

  Therefore, when it is desired to emit light having a desired linear polarization direction from the indirect transition light emitting element 1, it is possible to produce the light by making the light having the linear polarization direction incident.

  Examples of indirect transition type semiconductor light emitting devices that provide the present invention will be described below.

The n layer 13 is composed of a silicon substrate, and the p layer 14 is also composed of a silicon substrate. However, boron is implanted as a p-type dopant. The p layer 14 has an ion implantation energy of 700 KeV, for example, in the ion implantation method, and a doping density of 2 × 10 ¹⁹ atoms / cm ^{3 in the} vicinity of 1500 nm from the surface.

The band gap E ₂ in the semiconductor layer 30 having such a configuration is 1.14 eV. A forward bias voltage of 1.67 A / cm ² is applied to the p layer 14 and the n layer 13 with respect to the semiconductor layer 30, and linearly polarized light having a polarization direction as shown in FIG. The component light was irradiated. In this example, the light was irradiated from the direction perpendicular to the surface of the light emitting element. The wavelength of the irradiated light is 1.3 μm. The emission energy E ₁ was 0.94 eV. For this reason, ΔE was 0.2 eV.

  After applying a forward bias voltage and irradiating light for about 60 minutes, a test specimen for measurement was cut out by FIB (Focused Ion Beam). In this cutting, a test piece 55 having a thickness of 3 mm and a height h of 625 μm was cut out just from the surface of the light emitting element. Among the cut out test pieces, dopant distribution measurement was performed by a three-dimensional atom probe method in a space of 30 nm × 30 nm × 200 nm. FIG. 9 is an example of the distribution of boron (B) as a dopant actually measured by this three-dimensional atom probe method.

  FIG. 10A shows the number distribution with respect to the distance L of the adjacent dopant (B) before irradiating the semiconductor layer 30 with light. FIG. 10B shows the number distribution with respect to the semiconductor layer 30. It is a number distribution with respect to the distance L of the nearest neighbor dopant (B) after irradiation. The vertical axis represents the number of boron as a dopant measured in a space of 30 nm × 30 nm × 200 nm. Moreover, the horizontal axis has shown the distance L of the nearest neighbor dopant (B). Here, the solid line in the figure is a Weibull distribution obtained at the time of a random distribution determined only by the atomic density of the dopant (B).

  The number distribution with respect to the distance L of the adjacent dopant (B) before irradiating light is plotted in a region deviating from the Weibull distribution when the interatomic distance is 6 nm or 9 nm, as shown in FIG. Exists. This is due to the non-uniformity of dopant ion implantation.

  On the other hand, the distance L of the adjacent dopant (B) after irradiation with light has a distribution having a fine peak at a distance that is an integral multiple of the lattice constant a, as shown in FIG. It has been shown that it has changed. Incidentally, a value obtained by dividing the distance L by the lattice constant a is N. Although the lattice constant a is 0.54, when the distance L in FIG. 10B is divided by the lattice constant a, each position where N is 3, 4, 5, 6,. Become.

  FIG. 11 shows the result of calculating the deviation (ratio) from the Weibull distribution actually indicated by a straight line from the data of FIG. 10B. The horizontal axis is N, and the vertical axis indicates the ratio of deviation from the Weibull distribution for each N.

  As a result, in this example, it is shown that the ratio of deviation from the Weibull distribution is the highest at N = 3, that is, at three lattice constants. Along the Weibull distribution means that the dopants are naturally arranged, whereas a large deviation from the Weibull distribution means that the dopant is naturally aligned based on external effects. It means that they are regularly arranged and deviate from. N = 3, that is, the ratio of deviation from the Weibull distribution is the highest at three lattice constants. This means that the distance between the adjacent dopants (B) is just at the lattice constant of three. It means that there is the most arrangement.

  Therefore, the n layer 13 is Si, and the p layer 14 is obtained by doping B as a dopant with respect to Si, but B adjacent to each other constituting the distribution of B as the changed dopant. It can be seen that the largest number of intervals is N (= 3) times the lattice constant a.

  FIG. 12 shows the result of investigating the direction dependency of the adjacent dopant (B) when the light irradiation direction is the V direction in the figure. As shown in FIG. 12A, the number of dopants (B) measured in the unit space described above for each angle θ ° from 0 ° to 90 ° with the light irradiation direction V as a reference. Was measured.

  As a result, as shown in FIG. 12B, the number of dopants was the largest at θ = 90 °. This suggests that a large number of dopants (B) are arranged in the direction of θ = 90 ° which is just perpendicular to the incident direction V. In this 90 ° azimuth, as shown in the result of FIG. 11, the distance between the nearest neighbors B is the nearest neighbor if it is assumed that N (= 3) times the lattice constant a is the most distributed. It can be seen that the direction of the dopant is most often perpendicular to the irradiation direction of the irradiated light.

  That is, in addition to the largest number of dopants at θ = 90 °, the interval between the nearest neighbors B is N (= 3) times the lattice constant a as shown in the results of FIG. 11 at θ = 90 °. Are the most distributed. This means that the direction of the adjacent dopant is most often perpendicular to the irradiation direction of the irradiated light. In addition, since the irradiated light has a linearly polarized light component directed in the depth direction of the paper, the dopant is distributed in the vertical direction. For this reason, it was experimentally verified that the direction between the adjacent dopants was perpendicular to the irradiation direction of the irradiated light and the direction of linearly polarized light.

DESCRIPTION OF SYMBOLS 1 Indirect transition type semiconductor light-emitting device 13 n layer 14 p layer 30 semiconductor layer 31 dopant 35 junction layer 41 phonon vibration level 51 formation position 55 test piece

Claims (5)

In an indirect transition type semiconductor light emitting device comprising a semiconductor layer including a p layer and an n layer,
A forward bias voltage is applied, and light having a linearly polarized component and a wavelength shorter than the absorption edge wavelength in the semiconductor layer is irradiated to the junction between the p layer and the n layer. Repeatedly generating heat generation based on light absorption, and changing the dopant distribution of any one or more of the layers based on the heat generation;
In a place where near-field light is generated based on the changed dopant distribution, light is emitted by stimulated emission of electrons in the conduction band generated based on the bias voltage in a plurality of stages based on a non-adiabatic process. In addition, by suppressing the heat generation accompanying the light emission, the dopant distribution after the change is fixed,
The indirect transition semiconductor characterized in that the directions of the dopants adjacent to each other constituting the changed dopant distribution are perpendicular to the irradiation direction of the irradiated light and the direction of linearly polarized light. Light emitting element. When the energy difference between the band gap of the junction and the energy of the light emission is ΔE, the energy of the optical phonon localized at the distance between the adjacent dopants is ΔEf, and ΔE / ΔEf = N, 2. The indirect transition semiconductor light emitting device according to claim 1, wherein the spacing between the most adjacent dopants is most often distributed by N times the lattice constant a. The direction of the adjoining dopant is Δkf = Δk / where Δk is the wave number difference between the valence band and the conduction band at the junction and Δkf is the wave number of the optical phonon localized at the distance between the adjoining dopants. 3. The indirect transition type semiconductor light emitting device according to claim 1, wherein the indirect transition type semiconductor light emitting device is N. The indirect transition semiconductor light-emitting device according to claim 1, wherein the indirect transition semiconductor light-emitting element is distributed so that the existence probability of optical phonons is higher between the adjacent dopants. The n layer is Si,
The p layer is formed by doping B as a dopant with respect to Si,
The distance between the adjacent Bs constituting the distribution of B as the dopant after the change is most frequently distributed with N (= 3) times the lattice constant a. The indirect transition type semiconductor light emitting element of any one of them.

Images