JP2007134607A - Semiconductor element - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a semiconductor element positively utilizing both positive and negative stationary charges generated by polarization, and also utilizing a polarized junction formed therewith. <P>SOLUTION: The semiconductor element has three or more layers of two kinds of semiconductors laminated to form a hetero-junction of at least two semiconductors. It has a polarized junction so formed as to simultaneously generate first and second conductive carriers with positive and negative stationary charges generated by polarization on the interface of the hetero-junction. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a semiconductor element, and more particularly to a semiconductor element having a semiconductor polarization junction.

  Semiconductor devices and integrated circuits using them are indispensable in modern society. In recent years, the application range of semiconductors has been expanding, and there is no doubt that it will continue to play an important role in the future. In particular, research and development of semiconductor devices is a key technology in the energy and environmental problems that will become even greater challenges in the future. For example, electric power generated in a power plant is converted and controlled in various forms before it is consumed in a general household or factory. This field of semiconductors that handle high power is called power electronics. In order to effectively use the limited resources, it is necessary to reduce the loss in this conversion / control as much as possible. The on-resistance and breakdown voltage of the semiconductor element have the greatest influence on the efficiency of the conversion / control. Therefore, a semiconductor element having a lower on-resistance and a higher breakdown voltage is required. However, there is generally a trade-off relationship between on-resistance and breakdown voltage, and if one is improved, the other tends to be sacrificed. For this reason, there is a limit to improving the device performance.

  There are roughly two types of methods for overcoming this trade-off relationship and improving the characteristics of semiconductor elements. First, the most fundamental solution is a method of making a semiconductor element by replacing Si, which has been generally used so far, with a wide band gap semiconductor. A wide band gap semiconductor is a semiconductor having a band gap energy larger than that of Si (1.1 eV). At present, the most attention is given to SiC (up to 3.0 eV), a group III nitride semiconductor ( -6.2 eV), and II-VI group oxide semiconductors (-7.8 eV). Since the breakdown voltage increases as the band gap energy increases, the use of these semiconductors reduces the on-resistance to a few hundredths compared to a rectifier diode using Si while having the same reverse breakdown voltage. Expected.

Another method is to use a new technology called super-junction, which has been attracting much attention in recent years. A superjunction is an improved version of a conventional pn junction, in which a p-type region and an n-type region are formed in a semiconductor, and the total charge amount of these two regions is designed to be approximately equal, thereby improving the effect of carrier compensation. This is a technique for generating a constant electric field distribution in the depletion layer. Using this super-junction, it has been found that performance exceeding the material limit of Si previously considered can be realized. For example, MR-JBS (Multi RESURF Junction Barrier Schottky Rectifier), which is a rectifier diode, and CoolMos ^TM, which is a field effect transistor, have been reported.

  However, since fabrication of this superjunction requires precise semiconductor process technology or growth technology, it is currently put to practical use only with semiconductor elements using Si. For example, in MR-JBS, a method of once forming a deep groove (trench structure) in an n-type Si substrate by etching and then embedding the groove with p-type Si by crystal growth is used. At this time, the groove to be formed is required to have a high aspect ratio (ratio of groove width to depth). Also, in order to make an ideal superjunction, it is necessary to keep the total doping amount of positive charge and negative charge completely equal. In the super junction structure, when this relationship is broken, the usable element breakdown voltage is limited. However, this is impossible in principle, and practically, as disclosed in, for example, Japanese Patent Application Laid-Open No. 2001-111041, the error of the total amount of positive and negative charges is within a few percent. Suppress and form a super junction.

  As described above, in order to improve the trade-off between on-resistance and breakdown voltage, it has been described that there are a method using a wide band gap semiconductor and a method using a super junction. In other words, if these two methods can be used simultaneously to produce an element having a superjunction in a wide band gap semiconductor, an ideal semiconductor element for power electronics can be produced. However, as described above, super-junction fabrication requires highly accurate semiconductor process technology and growth technology, and it has been extremely difficult to apply this to semiconductors other than Si.

  The basic structures of various semiconductor devices invented so far are p-type and n-type conductivity control by doping, and band structure control by heterojunction. Operation is realized. Of course, a Si semiconductor element having a super junction is also manufactured by these techniques.

  On the other hand, it is known that the conductivity type can be controlled using the polarization of the semiconductor. Polarization is a phenomenon in which fixed charges are generated at the interface of the heterojunction, and includes spontaneous polarization that occurs even when the crystal is unstrained and piezoelectric polarization that occurs due to strain. Electrons or holes are generated in the vicinity of the heterojunction interface by being attracted by this fixed charge.

Compared to carriers formed by doping, carriers due to this polarization have a variety of different characteristics. First of all, it is possible to generate a high concentration of fixed charges that are impossible with the doping technique and are spatially concentrated. For example, in a heterojunction of a group III nitride semiconductor, a polarization charge can be obtained with a surface density of about 10 ¹³ cm ⁻² . At this time, assuming that the steepness of the interface is 1 nm, the fixed charge density is about 10 ²⁰ cm ⁻³ , which is very high. This makes it possible to realize a spatial distribution of carriers that is difficult only with the doping technique.

  Second, in principle, the activation energy of carriers can be ignored. The activation energy here is thermal energy for generating carriers. Wide band gap semiconductors have a problem in that the activation energy tends to be larger than the thermal energy at room temperature in controlling the conductivity by doping. However, this problem does not occur if polarization is used. That is, there is a possibility that the conductivity can be controlled even in a semiconductor in which p-type and n-type control is difficult by doping.

Thirdly, the carriers generated at this time are spatially concentrated as described above, and the influence of ionized impurity scattering is low, so that the mobility can be high. When carriers are generated by doping, for example, the mobility of bulk GaN is about 200 cm ² / Vs, but mobility of 1000 cm ² / Vs or more can be easily obtained by using a two-dimensional electron gas by polarization. It is done.

  Fourth, it is easy to control the total amount of positive and negative charges distributed in the semiconductor. For example, when considering a heterojunction of two types of semiconductors, the total amount of fixed charges generated therein is determined by the type of semiconductor to be bonded. That is, changing the steepness of the junction or inserting another semiconductor layer at the interface does not affect the total amount of fixed charges.

  As described above, carriers generated by polarization have various characteristics. Therefore, in addition to the p-type or n-type control technology by doping and the band lineup control by heterojunction, if the above-described polarization phenomenon is actively used, an unprecedented semiconductor device can be realized.

  This phenomenon of polarization is seen in many semiconductors. However, the semiconductor element based on the above-mentioned viewpoint has not received much attention so far. One reason is that a semiconductor having a cubic crystal structure such as GaAs has a small polarization. On the other hand, it is known that a group III nitride semiconductor having a hexagonal crystal structure in which high-quality single crystals have been obtained in recent years produces extremely large polarization.

Group III nitride semiconductors are group III-V compound semiconductors that have begun to attract attention as optical devices since the 1990s. The chemical formula is represented by _{B x Al y Ga z In 1} -x-y-z N. In a light emitting device using a group III nitride semiconductor, this large polarization has been regarded as a drawback because it leads to a decrease in light emission efficiency due to the quantum confinement shktar effect. To date, there is a high electron mobility transistor using a group III nitride semiconductor as a semiconductor element that actively uses polarization. This utilizes positive fixed charges generated by polarization. However, a semiconductor device that utilizes both positive and negative fixed charges and positively uses electrons and holes generated thereby has not been reported so far.
JP-A-9-266611 JP 2001-111041 A JP 2002-76370 A

  The problem to be solved by the present invention is to provide a semiconductor device that positively utilizes both positive and negative fixed charges due to polarization and utilizes a polarization junction formed thereby.

The present invention actively utilizes positive and negative fixed charges generated by polarization, and the problem can be solved by providing the following semiconductor device.
(1) In a semiconductor element in which two or more kinds of semiconductors are stacked in three or more layers so as to form a heterojunction of at least two or more semiconductors.
A semiconductor device comprising a polarization junction configured to simultaneously generate a first conductivity type carrier and a second conductivity type carrier due to positive and negative fixed charges generated by polarization at an interface of the hetero junction .
(2) The above-mentioned polarization junction is such that the positive and negative fixed charges in the semiconductor are made equal by suppressing the positive and negative fixed charges other than the polarization low with respect to the positive and negative fixed charges due to the polarization. A featured semiconductor element.
(3) By forming the polarization junction without intentional doping, the amount of positive and negative fixed charges other than polarization is kept low with respect to positive and negative fixed charges due to polarization, thereby reducing the semiconductor. A semiconductor element characterized in that the positive and negative fixed charge amounts therein are equal.
(4) The polarization junction adjusts the spatial distribution of the polarization charge in the stacking direction of the semiconductor layer by adjusting the spatial variation of the semiconductor composition at the junction interface in the junction of different types of semiconductors. A semiconductor element characterized by:
(5) The semiconductor element characterized in that the polarization junction reduces a spatial charge concentration by gradually changing the composition of the semiconductor in the junction of different types of semiconductors.
(6) The heterojunction is a semiconductor element formed of III-V group compound semiconductors having different compositions.
(7) The group III-V compound semiconductor is a III-nitride semiconductor, chemical formulas semiconductor elements represented by _{B x Al y Ga z In 1} -x-y-z N.
(In the formula, x, y and z have numerical values satisfying 0 ≦ x ≦ 1, 0 ≦ y ≦ 1, 0 ≦ z ≦ 1, and x + y + z ≦ 1.)
(8) The semiconductor element characterized in that the group III nitride semiconductor is laminated in the c-axis direction.
(9) The heterojunction is a semiconductor element formed of II-VI group oxide semiconductors having different compositions.
(10) The heterojunction is a semiconductor element formed of SiC compound semiconductors having different crystal structures.

According to the present invention, the on-resistance can be reduced by the amount of carriers generated by polarization without sacrificing the withstand voltage. Therefore, if this polarization junction is applied to a rectifier diode, a field effect transistor or the like, the on-resistance And the breakdown voltage can be improved.
Further, by using the high mobility of at least the first conductivity type carrier or the second conductivity type carrier generated by positive or negative fixed charges due to polarization, the on-resistance of the semiconductor element is further reduced, and high frequency characteristics Can also be improved.

Next, in the semiconductor elements described in the above (2) to (3), the junction performance improves as the error in the positive and negative fixed charge amounts in the semiconductor decreases, and this further improves the breakdown voltage of the semiconductor element. Can be made.
Next, in the semiconductor element described in the above (4), the spatial distribution of polarization charges with respect to the stacking direction of the semiconductor layers can be adjusted.
Next, in the semiconductor element described in (5) above, spatial charge concentration can be reduced by gradually changing the composition of the semiconductor at the junction of different types of semiconductors.

FIG. 1 shows the state of polarization when two types of semiconductors, A1 and B2, are alternately stacked. Depending on the type of stacked semiconductor, its crystal orientation, and the presence or absence of crystal distortion, in the case of a semiconductor in which polarization occurs, a fixed charge having the same charge amount but the opposite sign has an area density p _p cm ⁻² . Alternating in size. If the influence of crystal distortion and crystal orientation of the semiconductor is ignored, the surface density p _p is uniquely determined by the types of the semiconductor A and the semiconductor B. For example, assume that the composition is gently changed at the interface between the semiconductor A1 and the semiconductor B2 as shown in FIG. In that case, the polarization charge spreads and distributes in the layer 3 whose composition has been changed gently, so that the volume density of the fixed charge decreases. However, the surface density _{p p} does not change.

Also, as shown in FIG. 3, a semiconductor C4 of a different type is inserted between the semiconductor A and the semiconductor B, and polarization charges are respectively present at the interface between the semiconductor A and the semiconductor C and the interface between the semiconductor B and the semiconductor C. Suppose it appears. In this case, the case of ignoring the influence of crystal strains by the semiconductor C, the surface density p _p fixed charge is constant as shown in FIG. By combining these polarization characteristics with existing doping techniques, polarization junctions with various charge distributions can be produced. In particular, the feature that the positive and negative fixed charges shown in FIGS. 1 to 3 appear alternately in equal magnitude is very convenient for making a polarization junction.

Here, an actual state of generation of electrons and holes will be described by taking as an example a case where GaN, which is a kind of group III nitride semiconductor, and Al _y Ga _1-y N are stacked. FIG. 4 is a schematic diagram of a band lineup in the case where GaN, which is an intrinsic semiconductor, and Al _y Ga _1-y N are stacked in the c-axis direction. Thus, fixed charges due to positive and negative polarizations at the Al _y Ga _1-y N (000-1) / GaN (0001) and Al _y Ga _1-y N (0001) / GaN (000-1) interfaces. Can generate electrons and holes, respectively. In this specification, a pn junction of a semiconductor in which electrons and holes are alternately generated using polarization is defined as “polarization junction”.

  In order to generate electrons and holes and realize polarization junction, there are several unique design conditions. FIG. 5 is a schematic diagram of a band lineup when the semiconductor A7 and the semiconductor B8 are stacked. In an actual semiconductor, it is impossible to realize a perfect intrinsic semiconductor, and even if it is grown undoped, it always becomes a slight n-type or p-type. For example, a group III nitride semiconductor tends to be n-type during undoped growth. Therefore, as shown in FIG. 5, there is a difference in the surface density of electrons and holes. In some cases, only one of electrons or holes may be generated, and a polarization junction may not be formed.

Hereinafter, a basic design method of the polarization junction will be described with reference to FIG. First, the influence of the quantum effect and heat was ignored, and all carriers due to doping and crystal defects were activated. Moreover, compared with the thickness of each layer of the semiconductor A and semiconductor B, spread in the thickness direction of the polarization charge and the carrier is negligible, and the semiconductor interface with the positive and negative polarization charge, the Fermi level E _F is Assuming that the lower end of the conduction band and the upper end of the valence band of semiconductor A having a small band gap energy are fixed, the surfaces of electrons and holes viewed from the stacking direction for one cycle of each layer of semiconductor A and semiconductor B The densities n and p are each expressed by the following equations.

Here, p _p is the magnitude of the surface density of the polarization charge at each interface, Eg _A is the band gap energy of the semiconductor A, d _A and d _B are the thicknesses of the layers of the semiconductor A and the semiconductor B, e _A and e _B Represents the dielectric constant of the semiconductor A and the semiconductor B, and p _A and p _B represent the areal density of the fixed charges due to causes other than the polarization of the semiconductor A and the semiconductor B, respectively.
Since the equations (1) and (2) have many variables and are complicated, the following equations are obtained by approximation with d _A ≈d _B = d, e _A ≈e _B = e, and p _A ≈p _B = p _back. .

Based on the equations (3) and (4), some design guidelines for polarization junctions are shown. First, assuming that each semiconductor layer is n-type, p _back is a positive constant, and p _p is constant, changes in n and p with respect to d are estimated. FIG. 6 is a schematic diagram thereof. It can be seen that there is a minimum value d _min of d and a maximum value d _{max of} d in order to simultaneously generate electrons and holes and realize polarization junction. In addition, in order to exert a super junction effect close to the ideal, the electron / hole ratio n / p is desirably close to 1, but it can also be seen that there is an optimum junction distance d _opt for that purpose.

Next, FIG. 7 shows the influence on n and p when d and p _p are kept constant and p _back is changed. From the figure, it can be seen that n and p increase or decrease linearly with respect to changes in p _back . It can be seen that there is an allowable range for p _back in order to simultaneously generate n and p and realize polarization junction. It can also be seen that a smaller p _back is desirable for bringing the n / p ratio closer to 1. As seen from equation (3) and (4), it is desirable to reduce the influence of p _back is _p p >> p _{back d.} That is, in order to form a polarization junction, it is necessary to reduce the total amount of fixed charges contained in each semiconductor layer with respect to fixed charges generated at the interface due to polarization. For this purpose, each semiconductor layer is preferably formed without intentional doping.

Further, a method of bringing the n / p ratio close to 1 by changing the polarization charge density at each interface in the polarization junction with respect to the stacking direction can be considered. FIG. 8 is a schematic diagram when this method is applied to a polarization junction using GaN and Al _y Ga _1-y N. Since undoped group III nitride semiconductors are likely to be n-type, the Al composition is changed stepwise as shown in FIG. 8 to create a difference between positive and negative fixed charge amounts due to polarization, thereby fixing other than polarization. The charge is compensated and the n / p ratio is brought close to 1.

By the way, as mentioned above, in order to implement | achieve favorable polarization | polarized-light junction, it is desirable for _{pp to} be large. However, semiconductor junctions with a large p _p generally tend to have a large lattice mismatch rate. The cause is that piezo polarization, which is one of the principles of occurrence of polarization, is due to lattice distortion due to lattice mismatch or the like, and in that case, a large polarization charge is obtained by the strain.

On the other hand, in the growth of lattice-mismatched semiconductors, lattice relaxation occurs when the critical film thickness is exceeded. The critical film thickness is determined by various factors, but it tends to be thinner as the lattice mismatch rate is larger. Because lattice relaxation is caused a reduced crystal strain, whereby there is a possibility that the polarization charge p _p is lowered, also the crystal defects such as a semiconductor misfit dislocations and dislocation from being introduced, whereby Since various semiconductor characteristics such as mobility deteriorate, it is not desirable. Therefore, when the heterojunction forming the polarization junction is a lattice mismatch system, the design condition of FIG. 9 is added. The critical film thickness can be predicted to some extent from the theoretical formulas of Matthews and People.

Since the polarization junction is a semiconductor multilayer structure, lattice strain accumulates and lattice relaxation is likely to occur. Therefore, a system in which each semiconductor layer compensates for energy accumulation due to lattice strain is suitable. FIG. 10 is a schematic view when this technique is applied to the polarization junction of a group III nitride semiconductor. Each Al _y Ga _1-y N layer and the Ga _z In _1-z N layer compensate each other for stress, thereby suppressing the accumulation of strain energy when the layers are stacked. This structure can also increase the polarization charge density as compared to the case where the Al _y Ga _1-y N layer and the GaN layer are stacked.

Further, if the semiconductor quaternary mixed crystal is used, _pp can be changed while maintaining lattice matching. Therefore, the critical film thickness restriction shown in FIG. 9 is eliminated, and the design of the polarization junction is facilitated. FIG. 11 is a schematic view when this technique is applied to the polarization junction of a group III nitride semiconductor. By using the Al _y1 Ga _z1 In _1-y1-z1 N layer 14 and the Al _y2 Ga _z2 In _1-y2-z2 N layer 15 that are lattice-matched to GaN, the polarization junction is designed ignoring the critical film thickness. It becomes possible to do.

Next, the characteristics of a pn junction diode having a polarization junction to which the present invention is applied and a normal pn junction diode to which the present invention is not applied are compared. FIG. 12 is a schematic diagram of the structure of a normal pn junction diode to which the present invention is not applied. This is a pin structure in which a p ⁺ GaN layer 18 and an n ⁺ GaN layer 20 are formed on both ends of an n ⁻ GaN layer 19 that is a drift layer. Then, the characteristics were examined by changing the donor concentration p _back of the n ⁻ GaN layer 19 and its length d _drift .

FIG. 13 is a schematic diagram of the structure of a pn junction diode to which the present invention is applied. The polarization junction region 22 was formed by a stacked structure of an n ^- Al _0.09 Ga _0.91 N layer having an Al composition of 9% and an n ^- GaN layer. The thickness of each layer was 100 nm. The size of the polarization density _{p p} for each hetero interface was set to 5 × ¹⁰ 12 ^{cm -2} than the theoretical value. Both ends of the polarization junction region are doped with acceptor and donor impurities to form a p-type polarization junction region 21 and an n-type polarization junction region 23. Then, the characteristics were examined by changing the donor concentration p _back of the polarization junction region 22 and its length d _drift .

First, the breakdown voltage of the element when a reverse bias was applied to the structure of FIG. 12 was obtained by an avalanche model. FIG. 14 shows the simulation result. First, it can be seen that there is a breakdown voltage limit specific to GaN determined by d _drift . In other words, in the case of a perfect intrinsic semiconductor having no impurities or defects, the breakdown voltage is determined only by d _drift , and the breakdown voltage increases as much as this is increased. On the other hand, in an actual semiconductor, there is always a finite p _back , which also determines the limit of the withstand voltage. From the figure, it can be seen that the smaller the p _back is, the higher the limit of the withstand voltage is.

FIG. 15 is a simulation result of the reverse breakdown voltage of the structure of FIG. 13 to which the present invention is applied. The limit of the withstand voltage is almost determined by the material limit of GaN and the size of p _back , and as can be seen by comparing with FIG. 14, the result almost coincides with the result of the pn junction diode to which the present invention is not applied. The results of FIGS. 14 and 15 show that the fixed charge due to polarization hardly affects the breakdown voltage, and the breakdown voltage is determined by the magnitude of p _back . That is, it can be seen that the positive and negative polarization charges compensate each other, and as a result, the remaining p _back causes dielectric breakdown. Therefore, in the junction diode to which the present invention is applied, the on-resistance can be reduced by the amount of carriers generated by polarization without sacrificing the breakdown voltage. For this reason, if this polarization junction is applied to a semiconductor element such as a rectifier diode or a field effect transistor, the trade-off between on-resistance and breakdown voltage can be improved.

Further, in order to improve the performance of the polarization junction, it is necessary to keep the positive and negative fixed charge amounts as equal as possible. For example, if p _p = 1 × 10 ¹³ cm ⁻² , p _back = 1 × 10 ¹⁶ cm ⁻³ and the thickness of each layer of the polarization junction is 100 nm, the error in the magnitude of the positive and negative fixed charges is small. 2%.

Next, FIG. 16 shows simulation results for confirming the effect of applying the present invention to a Schottky junction diode made of a group III nitride semiconductor. The positive and negative fixed charge magnitude errors are 40%, 4.0%, and 0.4% at p _back = 1 × 10 ¹⁷ , 1 × 10 ¹⁶ , and 1 × 10 ¹⁵ cm ⁻³ , respectively. It is. From this result, it is understood that the trade-off between on-resistance and breakdown voltage can be greatly improved by using the polarization junction according to the present invention for a part of the existing Schottky junction diode. The effect of the present invention can be obtained even when the error in magnitude of the positive and negative fixed charges is as large as 40%. However, similarly to the case of the pn junction described above, the device breakdown voltage that can be realized is limited by the size of p _back . Therefore, it is desirable for application of the present invention that _pp is large and p _back is small. Further, FIG. 16 shows that the effect of the present invention improves as the withstand voltage increases. For example, in the high breakdown voltage region of 10 kV or more, it can be expected that the on-resistance is improved to 1/100 or less by applying the present invention.

  In addition, although the rectifier diode and the Schottky junction diode were demonstrated as an application element of this invention, this invention relates to the polarization junction which is a kind of pn junction. Therefore, by replacing the existing pn junction with the polarization junction according to the present invention, application to all existing semiconductor elements is possible. For example, it can be applied to semiconductor elements such as field effect transistors in addition to rectifier diodes and Schottky junction diodes. Furthermore, the semiconductor device of the present invention can also be applied to optical devices such as a light emitting diode, a laser diode, and a light receiving element because the basic structure is a pn junction.

Further, although the present specification has described the effect of polarization junction by taking a group III nitride semiconductor as an example, the present invention is applicable to all semiconductors in which polarization occurs. For example, the present invention can be applied to a heterojunction of a group II-VI oxide semiconductor typified by ZnO because fixed charges are generated due to large polarization. A heterojunction such as ZnO / Zn _m Mg _1-m O is an example. The II-VI group oxide semiconductor is a semiconductor having a very large band gap, and has great potential as a power semiconductor element. Also, SiC heterojunctions of different polytypes are known to generate large polarization, and the present invention can be applied. Examples are heterojunctions such as 4H-SiC / 3C-SiC and 6H-SiC / 3C-SiC. SiC is also a wide band gap semiconductor, and is composed of the same group 4 element as Si, and therefore has an advantage that many existing technologies established with Si can be applied.

Schematic of spatial distribution of polarization charge generated at the interface when two types of semiconductors are stacked Schematic diagram of spatial distribution of polarization charge when two kinds of semiconductors are stacked and the composition of the semiconductor at the interface is gently changed. Schematic diagram of spatial distribution of polarization charge generated at the interface when three kinds of semiconductors are stacked Schematic of band diagram when two types of group III nitride semiconductors are stacked Schematic diagram of band diagram of polarization junction with two types of semiconductors stacked Schematic of changes in electron concentration and hole concentration with respect to the junction distance in polarization junctions Schematic of changes in electron and hole concentrations for fixed charge concentrations other than polarization in polarization junctions Schematic diagram of the structure to reduce the size of the total fixed charge by compensating for fixed charges other than polarization in polarization junctions with group III nitride semiconductors Schematic diagram of critical film thickness limitation on bonding distance Schematic diagram of structure with reduced strain energy accumulation in polarization junctions in group III nitride semiconductors Schematic diagram of a polarization junction formed by a layered structure of a group III nitride semiconductor lattice-matched system Schematic diagram of the structure of a pn junction diode used in the simulation to which the present invention is not applied Schematic diagram of the structure of a pn junction diode to which the present invention is applied used for simulation Simulation results of a pn junction diode to which the present invention is not applied Simulation results of a pn junction diode to which the present invention is applied Simulation results of on-resistance and breakdown voltage in Schottky junction diodes

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Semiconductor A layer 2 Semiconductor B layer 3 which is different from semiconductor A layer 3 Layer in which composition of semiconductor A and semiconductor B is gently changed 4 Semiconductor C layer 5 which is different from semiconductor A and semiconductor B 5 i-GaN layer 6 i -Al _y Ga _1-y n layer 7 semiconductor A layer 8 semiconductor layer B 9-1 Al composition different ^{_{n - Al y Ga 1-y}} n
9-2 Al composition different ^{_{n - Al y Ga 1-y}} N
9-3 Al composition different ^{_{n - Al y Ga 1-y}} N
9-4 Al composition different ^{_{n - Al y Ga 1-y}} N
9-5 Al composition different ^{_{n - Al y Ga 1-y}} N
10 n ^- GaN layer 11 Unstrained GaN
12 compressive strain Ga _z In _{1 -z} N layer 13 tensile strain Al _y Ga _{1 -y} N layer 14 Al _y1 Ga _z1 In _{1 -y 1 -z1} N layer lattice matched to GaN 15 Al _y2 Ga _z2 lattice matched to GaN In _{1 -y 2 -z 2} N layer 16 Anode electrode 17 Cathode electrode 18 p ⁺ GaN layer 19 n ⁻ GaN layer 20 n ⁺ GaN layer 21 p-type polarization junction region 22 Polarization junction region 23 n-type polarization junction region

Claims (10)

In a semiconductor element in which two or more kinds of semiconductors are stacked in three or more layers so as to form a heterojunction of at least two or more semiconductors,
A semiconductor device comprising a polarization junction configured to simultaneously generate a first conductivity type carrier and a second conductivity type carrier due to positive and negative fixed charges generated by polarization at an interface of the hetero junction . The polarization junction is characterized in that the positive and negative fixed charges in the semiconductor are equalized by keeping the positive and negative fixed charges other than the polarization low with respect to the positive and negative fixed charges due to the polarization. 2. The semiconductor element according to claim 1. By forming the polarization junction without intentional doping, the amount of positive and negative fixed charges other than polarization is kept low with respect to positive and negative fixed charges due to polarization. 2. The semiconductor element according to claim 1, wherein the negative fixed charge amounts are equal. The above-mentioned polarization junction adjusts the spatial distribution of polarization charge in the stacking direction of the semiconductor layer by adjusting the spatial change of the semiconductor composition at the junction interface in the junction of different types of semiconductors. 2. The semiconductor element according to claim 1, wherein 2. The semiconductor element according to claim 1, wherein the polarization junction reduces a concentration of spatial charges by gradually changing the composition of the semiconductor in the junction of different types of semiconductors. The semiconductor device according to claim 1, wherein the heterojunction is formed of a group III-V compound semiconductor having a different composition. The group III-V compound semiconductor is a III-nitride semiconductor, semiconductor device of claim 6 chemical formulas represented by _{B x Al y Ga z In 1} -x-y-z N.
(In the formula, x, y and z have numerical values satisfying 0 ≦ x ≦ 1, 0 ≦ y ≦ 1, 0 ≦ z ≦ 1, and x + y + z ≦ 1.) The semiconductor element according to claim 7, wherein the group III nitride semiconductor is stacked in a c-axis direction. The semiconductor element according to claim 1, wherein the heterojunction is formed of II-VI group oxide semiconductors having different compositions. The semiconductor element according to claim 1, wherein the heterojunction is formed of SiC compound semiconductors having different crystal structures.

Images