Semiconductor device and method for manufacturing a semiconductor device

Abstract

High-performance semiconductor devices and methods for manufacturing semiconductor devices are provided. [Solution] The semiconductor device comprises an active layer and a barrier layer laminated on the active layer, wherein the active layer contains arsenic or antimony, and the barrier layer contains aluminum and also contains arsenic or antimony, and the direction in which the barrier layer is laminated on the active layer is the depth direction, and in the distribution of atomic concentrations in the depth direction from the active layer to the barrier layer, the rising position of the hydrogen concentration is closer to the center of the active layer than the rising position of the aluminum concentration.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a semiconductor device and a method for manufacturing the semiconductor device.

BACKGROUND ART

[0002] A compound semiconductor containing InSb or InAs is a material with high mobility and a small band gap. Taking advantage of this feature, compound semiconductors containing InSb or InAs are used as devices such as magnetic sensors, high-speed devices, and IR (infrared) sensors. For example, in realizing a device to which a compound semiconductor containing InSb is applied, it is preferable to use AlInSb as an electron barrier layer.

[0003] For example, Patent Document 1 discloses a compound semiconductor substrate including an AlInSb layer that is manufactured using Metal Organic Chemical Vapor Deposition (MOCVD) and has favorable crystallinity and excellent surface flatness.

PRIOR ART DOCUMENTS

PATENT DOCUMENTS

[0004]

PATENT DOCUMENT 1

SUMMARY OF THE INVENTION

PROBLEMS TO BE SOLVED BY THE INVENTION

[0005] Here, semiconductor devices are used, for example, as IR light-emitting devices or IR light-receiving devices, and there is a demand for further improving emission intensity or detection sensitivity. In other words, semiconductor devices are required to have an even higher SNR (Signal to Noise Ratio) in practical use.

[0006] The present disclosure has been made in view of such circumstances, and an object thereof is to provide a high-performance semiconductor device and a method for manufacturing the semiconductor device. [Means for Solving the Problems]

[0007] (1) A semiconductor device according to an embodiment of the present disclosure comprises: an active layer, and a barrier layer stacked over the active layer, the active layer contains arsenic or antimony, the barrier layer contains aluminum, and contains arsenic or antimony, wherein a direction in which the barrier layer is stacked over the active layer is defined as a depth direction, and in the atomic concentration distribution in the depth direction from the active layer to the barrier layer, the rising position of the hydrogen concentration is closer to the center of the active layer than the rising position of the aluminum concentration.

[0008] (2) As an embodiment of the present disclosure, in (1), the active layer contains indium.

[0009] (3) As an embodiment of the present disclosure, in (2), the active layer contains indium and antimony.

[0010] (4) As an embodiment of the present disclosure, in (2), the active layer contains indium and arsenic.

[0011] (5) As an embodiment of the present disclosure, in any one of (1) to (4), the barrier layer contains antimony.

[0012] (6) As an embodiment of the present disclosure, in any one of (1) to (5), in the atomic concentration distribution in the depth direction from the active layer to the barrier layer, the hydrogen concentration becomes 50% or less of the peak concentration after reaching the peak.

[0013] (7) As an embodiment of the present disclosure, in any one of (1) to (6), In the distribution of atomic concentrations in the depth direction from the active layer to the barrier layer, the position of the hydrogen concentration peak is within a range of 30 nm or less from the interface between the barrier layer and the active layer towards the active layer.

[0014] (8) In one embodiment of the present disclosure, in any of (1) to (7), In the atomic concentration distribution in the depth direction from the active layer to the barrier layer, the full width at half maximum of the hydrogen concentration peak is 60 nm or less.

[0015] (9) In one embodiment of the present disclosure, in any of (1) to (8), In the distribution of atomic concentrations in the depth direction of the barrier layer, the hydrogen concentration at one end of the barrier layer adjacent to the active layer is lower than the hydrogen concentration at the other end.

[0016] (10) In one embodiment of the present disclosure, in (9), The barrier layer is doped, and in the distribution of atomic concentrations in the depth direction within the barrier layer, the hydrogen concentration decreases monotonically from one end to the other.

[0017] (11) In one embodiment of the present disclosure, in any of (1) to (10), The aforementioned barrier layer is doped.

[0018] (12) In one embodiment of the present disclosure, in any of (1) to (11), The active layer and the barrier layer have different lattice constants in an unstrained state.

[0019] (13) In one embodiment of the present disclosure, in any of (1) to (12), The active layer further contains aluminum.

[0020] (14) A method for manufacturing a semiconductor device according to one embodiment of the present disclosure is: A method for manufacturing a semiconductor device, which involves manufacturing any of the semiconductor devices described in (1) to (13), The process includes a pressure-increasing annealing step in which the pressure in the apparatus reaction chamber is increased before the barrier layer is formed. [Effects of the Invention]

[0021] This disclosure provides high-performance semiconductor devices and methods for manufacturing semiconductor devices. [Brief explanation of the drawing]

[0022] [Figure 1] Figure 1 shows an example of the distribution of aluminum and hydrogen concentrations in a semiconductor device according to one embodiment of the present disclosure. [Figure 2] Figure 2 shows the defect-dependent resistance when the semiconductor device is an infrared device. [Figure 3] Figure 3 shows a band diagram when the semiconductor device is an infrared device. [Figure 4] Figure 4 shows the dependence of mobility on the amount of defects when the semiconductor device is a Hall element. [Figure 5] Figure 5 shows a band diagram when the semiconductor device is a Hall element. [Figure 6] Figure 6 illustrates the structure of a semiconductor device and shows the schematic distribution of aluminum and hydrogen concentrations. [Figure 7] Figure 7 illustrates dangling bonds and hydrogen termination. [Figure 8] Figure 8 shows a band diagram when the semiconductor device is a Hall element. [Figure 9] Figure 9 shows the band diagram of Example A. [Figure 10] Figure 10 shows the band diagram of Comparative Example A. [Figure 11] Figure 11 shows the band diagram for Example B. [Figure 12]Figure 12 shows the band diagram of Comparative Example B. [Figure 13] Figure 13 shows the band diagram of Example C. [Figure 14] Figure 14 shows the band diagram of comparative example C. [Figure 15] Figure 15 shows the band diagram for Example D. [Figure 16] Figure 16 shows the band diagram of comparative example D. [Modes for carrying out the invention]

[0023] <Semiconductor devices> The semiconductor device according to this embodiment comprises an active layer and a barrier layer laminated on the active layer. The left diagram of Figure 6 illustrates the structure of the semiconductor device. The semiconductor device may include a substrate. The semiconductor device may also include a first layer provided between the substrate and the active layer, and a second layer adjacent to the barrier layer on the opposite side from the active layer. At least one of the first layer and the second layer may be omitted as appropriate depending on the type of semiconductor device. The semiconductor device is constructed by laminating these layers, including at least an active layer and a barrier layer. Hereinafter, the direction in which the barrier layer is laminated on the active layer will be referred to as the depth direction, and the positions of the components will be described accordingly.

[0024] Here, the semiconductor device may be an infrared device (IR device). An infrared device is an infrared light-emitting element or an infrared light-receiving element, and this is a collective term for both. An infrared light-emitting element is realized with the structure of the semiconductor device described below, and an infrared light-receiving element is realized with the same structure. When the semiconductor device is an infrared light-emitting element, it may specifically be a light-emitting diode (LED). When the semiconductor device is an infrared light-receiving element, it may specifically be a photodiode (PD). Furthermore, the semiconductor device may be a Hall element. The case where the semiconductor device is a Hall element will be explained in one of the embodiments described later. In the description of the other embodiments, the semiconductor device will be described as an infrared device.

[0025] <Active layer> The active layer may be composed of arsenic (As) or antimony (Sb). Therefore, InAs or InSb may be used as the material for the active layer, but are not limited to these. The active layer may further contain aluminum (Al). Therefore, InAlSb may be used as the material for the active layer, but are not limited to these. The active layer may also contain atoms such as gallium (Ga), and may be composed of alloys containing these atoms. Furthermore, the active layer is not limited to the above materials, and other materials may be used. Also, from the viewpoint of device characteristics, the active layer is preferably a single crystal. When the active layer contains indium, electrons tend to accumulate at the interface between the active layer and the barrier layer due to band bending, and the hydrogen termination effect described later is significant.

[0026] <Barrier layer> The barrier layer is composed of aluminum (Al) and also contains arsenic (As) or antimony (Sb). The barrier layer functions as an energy barrier against electrons or holes in the active layer. Therefore, while InAlSb is used as the material for the barrier layer, it is not limited to this. Preferably, the barrier layer forms an energy barrier against electrons or holes in the active layer. The barrier layer may also contain atoms such as gallium (Ga), and may be composed of an alloy containing these atoms. Furthermore, from the viewpoint of device characteristics, it is preferable that the barrier layer is a single crystal. It is also preferable that the lattice constant of the barrier layer material is close to that of the active layer material. Since an energy barrier against electrons or holes can be formed without significantly changing the lattice constant of the active layer, it is preferable that the barrier layer material contains aluminum in addition to the material that constitutes the active layer. Furthermore, when the barrier layer contains antimony, electrons tend to accumulate at the interface between the active layer and the barrier layer due to band bending, and the hydrogen termination effect described later is significant.

[0027] <Circuit board> The substrate is not particularly limited as long as it supports layers including the active layer and the barrier layer. Examples include GaAs substrates, InP substrates, GaN substrates, Si substrates, quartz substrates, aluminum substrates, oxide substrates, aluminum nitride substrates, and polyimide flexible substrates.

[0028] <First layer> The first layer may be, for example, a semiconductor layer of a first conductivity type. The first conductivity type may be, for example, n-type. The constituent material of the first layer can be, but is not limited to, InSb, GaAs, InAs, InGaAs, InAlSb, GaAsSb, or InGaP. It is also preferable that the first layer has a larger band gap than the active layer and forms an energy barrier to the electrons or holes of the active layer.

[0029] <Second Layer> The second layer may be, for example, a semiconductor layer of a second conductivity type. The second conductivity type is different from the first conductivity type, and may be, for example, p-type. As another example, the first conductivity type may be p-type and the second conductivity type may be n-type. The constituent material of the second layer may be, but is not limited to, InSb, GaAs, InAs, InGaAs, InAlSb, GaAsSb, or InGaP.

[0030] <Hydrogen concentration distribution> As mentioned above, semiconductor devices are required to have even higher signal-to-noise ratios (SNR). The inventors diligently investigated ways to provide high-performance semiconductor devices and discovered that performance can be improved by properly terminating dangling bonds.

[0031] Figure 7 illustrates dangling bonds and hydrogen termination. The central figure in Figure 7 is a semiconductor device, the same as in Figure 6. The active layer and the barrier layer have different lattice constants in the unstrained state (left figure in Figure 7). Therefore, dangling bonds are likely to occur at the boundary between the active layer and the barrier layer due to the difference in lattice constants (the number of defects tends to increase). Here, hydrogen can terminate dangling bonds (right figure in Figure 7). In other words, hydrogen can reduce unbonded bonds and thus reduce defect levels. The barrier layer may contain hydrogen (H), and the distribution of hydrogen concentration can be adjusted by adjusting the process of the semiconductor device manufacturing method. In the semiconductor device according to this embodiment, hydrogen terminates defects near the interface between the barrier layer and the active layer, reducing defect levels and thus achieving a high SNR.

[0032] Figure 1 shows an example of the aluminum and hydrogen concentration distributions of a semiconductor device according to this embodiment. The measurements were performed using SIMS (Secondary Ionization Mass Spectrometer). Specifically, a sector-type SIMS instrument, the "IMS-7f" manufactured by CAMECA Corporation, was used. The ion species of the primary ion beam was Cs +The acceleration voltage of the primary ion beam was 15kV. The polarity of the secondary ion detection was negative. The horizontal axis in Figure 1 represents "depth," which is the distance from the surface of the semiconductor device in the depth direction, with the surface of the semiconductor device (the surface furthest from the substrate in the layer furthest from the substrate) being set to zero, as shown in Figure 6. The vertical axis represents the normalized concentration. The aluminum concentration was normalized by dividing by the peak (maximum value) of the aluminum concentration. Similarly, the hydrogen concentration was normalized by dividing by the peak (maximum value) of the hydrogen concentration. As shown in Figure 1, the peak of the hydrogen concentration is shifted from the peak of the aluminum concentration, and can be said to be shifted towards the active layer (center). The appropriate distribution of atomic concentrations in the depth direction from the active layer to the barrier layer of the semiconductor device is explained below.

[0033] First, it is preferable for hydrogen atoms to be localized in order to efficiently terminate defect levels and because hydrogen atoms can deactivate dopants. In particular, when the semiconductor device is an infrared device, the barrier layer may be doped. When hydrogen atoms deactivate dopants in the barrier layer, parasitic resistance can increase. Therefore, by limiting the range of hydrogen atoms present in the barrier layer, it is possible to realize a high-performance device with low parasitic resistance and the ability to terminate interface defects. Specifically, it is important that hydrogen atoms are distributed in large numbers in the vicinity of the active layer and the barrier layer. Here, carriers tend to accumulate not only at the interface between the active layer and the barrier layer, but also in the vicinity of the active layer with the barrier layer (approximately several tens of nm) due to band (conduction band, valence band) bending. Therefore, it is preferable to have a configuration in which the hydrogen concentration is increased not only at the interface, but also in the vicinity of the active layer with the barrier layer.

[0034] Here, Figure 6 illustrates the structure of a semiconductor device and shows the schematic distribution of aluminum and hydrogen concentrations. The schematic distribution is approximated by a Gaussian function from the detailed distribution shown in Figure 1. The rising point is the inflection point in the approximated Gaussian function, where the peak is reached from the active layer toward the barrier layer (in the opposite direction of depth).

[0035] Furthermore, in approximating the concentration distribution, in addition to the Gaussian function, tailing observed in SIMS measurements may be approximated using an exponential function term. Tailing is the apparent spread of concentration caused by artifacts associated with sputtering or secondary ion generation processes. When SIMS measurements are performed in the depth direction, it is known that the signal has a tail in the depth direction. Even when tailing is approximated using an exponential function term, the rise point is represented by an inflection point in the components of the Gaussian function. Specifically, when SIMS measurements are performed in the depth direction, the distribution can be approximated by the following equations (1) to (3).

[0036]

number

[0037] Here, z is the coordinate in the depth direction. meas This is the atomic concentration measured by SIMS analysis. gaus This is a Gaussian function that represents atomic concentration. tail σ is a function that shows the effect of tailing. z1 is a parameter that indicates the peak center of the Gaussian function. σ is a parameter that indicates the spread of the Gaussian function. z2 is a parameter that indicates the coordinates at which tailing begins. λ is a parameter that indicates the degree of tailing spread. A1 and A2 are coefficients.

[0038] Furthermore, in order to smoothly represent the area near the coordinates where tailing begins, tailing may be approximated using a smoothed switching function S as shown in equations (4) and (5) instead of equation (3). In this case as well, the rising point is represented by an inflection point in the components of the Gaussian function. Here, α in equation (5) is a parameter indicating the degree of smoothness.

[0039]

number

[0040] Furthermore, if the distribution does not have a clear peak shape, approximation may be performed using only a portion of the distribution. For example, when the barrier layer has a certain thickness or more, the aluminum concentration increases in the region corresponding to the interface from the active layer toward the barrier layer (in the opposite direction of depth), and then continues to take a value above a certain level in the region corresponding to the barrier layer. For this aluminum concentration distribution, approximation using a Gaussian function may be performed using only the portion that increases in the region corresponding to the interface.

[0041] In the semiconductor device according to this embodiment, in the atomic concentration distribution in the depth direction from the active layer to the barrier layer, the rising position of hydrogen concentration is closer to the active layer than the rising position of aluminum concentration. Here, "closer to the active layer" may mean close to the center of the active layer. Alternatively, "closer to the active layer" may mean close to the first layer or close to the substrate. By having such a distribution of hydrogen and aluminum in the semiconductor device, hydrogen atoms terminate defects and reduce defect levels, thereby enabling a high SNR.

[0042] Furthermore, from the viewpoint of localizing hydrogen atoms, in the distribution of atomic concentrations in the depth direction from the active layer to the barrier layer, it is preferable that the hydrogen concentration reaches a peak and then falls to 70% or less of the peak concentration, and more preferably to 50% or less.

[0043] Furthermore, in the atomic concentration distribution in the depth direction from the active layer to the barrier layer, the position of the hydrogen concentration peak is preferably at the interface between the barrier layer and the active layer. However, the position of the hydrogen concentration peak only needs to be within a range of 30 nm or less from the interface towards the active layer. Here, the interface between the barrier layer and the active layer can be defined by the plane corresponding to the rising position of atoms (e.g., aluminum) contained in the barrier layer.

[0044] Furthermore, from the viewpoint of localizing hydrogen atoms, it is preferable that the full width at half maximum (FWHM) of the hydrogen concentration peak in the atomic concentration distribution in the depth direction from the active layer to the barrier layer is 60 nm or less. From the viewpoint of efficient defect termination by hydrogen, it is preferable that the FWHM of the hydrogen concentration peak is 1 nm or more.

[0045] Furthermore, from the viewpoint of localizing hydrogen atoms, it is preferable that the hydrogen concentration in the barrier layer be low. Therefore, in the distribution of atomic concentration in the depth direction of the barrier layer, it is preferable that the hydrogen concentration at one end of the barrier layer is lower than the hydrogen concentration at the other end adjacent to the active layer. Moreover, it is even more preferable that the barrier layer is doped and that the hydrogen concentration in the distribution of atomic concentration in the depth direction of the barrier layer decreases monotonically from one end to the other. Here, monotonically decreasing refers to the fact that the detailed distribution decreases monotonically when approximated by a linear or quadratic function.

[0046] Here, when the barrier layer is used as an electron barrier layer, it is preferable that it be p-doped. Furthermore, when the barrier layer is used as a hole barrier layer, it is preferable that it be n-doped. However, when the barrier layer is doped, it is preferable that the hydrogen concentration in the barrier layer be low from the viewpoint of preventing dopant deactivation.

[0047] <Manufacturing methods for semiconductor devices> The inventors have found that a semiconductor device having the above hydrogen concentration distribution can be manufactured by performing a pressure-increasing annealing step, which involves increasing the pressure in the apparatus reaction chamber and annealing before depositing the barrier layer, based on the manufacturing method described in Patent Document 1. An example of the manufacturing process for the semiconductor device according to this embodiment is described below. Furthermore, the above semiconductor device can also be manufactured by increasing the hydrogen flow rate based on the manufacturing method described in Patent Document 1 in order to localize the hydrogen atoms.

[0048] A laminated film is produced using MOCVD as described below. First, a semi-insulating GaAs substrate of zincblende type was prepared. To the semi-insulating GaAs substrate, at a substrate temperature of 500°C, trimethylindium (TMIn) as the In source, trisdimethylaminoantimony (TDMASb) as the Sb source, and dimethyltellurium (DMTe) as the n-dopant were supplied, to form an n-InSb layer. At this time, the V / III ratio, which is the ratio of the supply amount of the group V element to the supply amount of the group III element, was 5.0. The film thickness of the n-InSb layer was 1000 nm. In addition, the doping concentration was 1.0×10 19 / cm 3 .

[0049] Subsequently, at a substrate temperature of 500°C, trimethylindium (TMIn) as the In source, trisdimethylaminoantimony (TDMASb) as the Sb source, and dimethylzinc (DMZn) as the p-dopant were supplied, to form a p-InSb layer. At this time, the V / III ratio was 5.0. The p-InSb layer corresponds to an active layer. The film thickness of the p-InSb layer was 2000 nm. In addition, the doping concentration was 6.0×10 16 / cm 3 .

[0050] Here, the supply of each source material was stopped, and while only hydrogen was allowed to flow into the apparatus reaction chamber, the pressure in the apparatus reaction chamber was raised from 50 torr to 300 torr. In this state, annealing was performed for 10 minutes. Thereafter, the pressure in the apparatus reaction chamber was lowered from 300 torr to 50 torr.

[0051] Subsequently, at a substrate temperature of 500°C, tritertiarybutylaluminum (TTBAl) as the Al source, trimethylindium (TMIn) as the In source, and trisdimethylaminoantimony (TDMASb) as the Sb source were supplied. Then, dimethylzinc (DMZn) was supplied as a p-dopant, to form a p-AlInSb layer. The p-AlInSb layer corresponds to a barrier layer. At this time, the Al source, the In source, and the Sb source were all supplied such that Al / (Al+In), which is the source ratio of Al and In, was 0.50. Thereafter, Al 0.7 In0.3 The Al / (Al+In) ratio was increased to 0.70 (i.e., the supply of Al raw material was increased) so that Sb could be formed. At this time, the V / III ratio was 5.0. The thickness of the p-AlInSb layer was 40 nm. The doping concentration was 2.0 × 10⁻⁶. 18 / cm 3 That was the case.

[0052] Next, at a substrate temperature of 500°C, trimethylindium (TMIn) was supplied as the In raw material, trisdimethylaminoantimony (TDMASb) as the Sb raw material, and dimethylzinc (DMZn) as the p dopant, and a p-InSb layer was formed. At this time, the V / III ratio was 5.0. The thickness of the p-InSb layer was 250 nm. The dope concentration was 2.0 × 10⁻⁶. 18 / cm 3 That was the case.

[0053] <Raw materials> In the above manufacturing method, the In raw material, Sb raw material, and Al raw material are not particularly limited as long as they can form an InSb layer and an AlInSb layer. Examples of In raw materials include trimethylindium (TMIn) and triethylindium (TEIn). Examples of Sb raw materials include trimethylantimony (TMSb), triethylantimony (TESb), trisdimethylaminoantimony (TDMASb), and triisopropylantimony (TIPSb). Examples of Al raw materials include trimethylaminealane (TMAAl), triisobutylaluminum (TIBAl), and dimethylaluminum hydride (DMAH). Dimethylethylaminealane (DMEAAl) and tritertically butylaluminum (TTBAl) are also examples of Al raw materials.

[0054] From the viewpoint of raw material decomposition temperature, the In raw material is preferably trimethylindium (TMIn), triethylindium (TEIn), or a combination thereof. From the viewpoint of long-term stability of the raw materials, the Al raw material is preferably triterminately butylaluminum (TTBAl). From the viewpoint of suppressing carbon impurities, the Sb raw material is preferably trisdimethylaminoantimony (TDMASb).

[0055] Furthermore, in order to control the conductivity type, dopants may be supplied in addition to the above raw materials. Examples of dopants include dimethyl zinc (DMZn), diethyl zinc (DEZn), dimethyl tellurium (DMTe), diethyl tellurium (DETe), tetramethyltin (TMSn), and tetraethyltin (TESn).

[0056] <Example 1> The resistance of the two multilayer film structures (a) and (b) below was calculated for use as an IR sensor (IR light receiving device). Here, the resistance of the IR sensor (with no bias voltage applied) is an important indicator for determining thermal noise, and a large value is preferable. The manufacturing method of the multilayer film follows the process described above.

[0057] The multilayer film structures (a) and (b) have an n-InSb layer (corresponding to the first layer in Figure 6), on which an active layer and a barrier layer are stacked, and on which a p-InSb layer (corresponding to the second layer in Figure 6) is stacked.

[0058] In the multilayer structure (a), the n-InSb layer thickness is 1000 nm and the doping concentration is 1.0 × 10⁻¹⁰ 19 / cm 3 The active layer, the p-InSb layer, has a thickness of 2000 nm and a dope concentration of 6.0 × 10⁻⁶. 16 / cm 3 The barrier layer is p-Al. 0.18 In 0.82 The Sb layer thickness is 20 nm, and the dope concentration is 2.0 × 10⁻⁶. 18 / cm 3 The p-InSb layer thickness is 500 nm, and the dope concentration is 2.0 × 10⁻⁶. 18 / cm3 That is the case.

[0059] In the multilayer structure (b), the n-InSb layer thickness is 1000 nm and the doping concentration is 1.0 × 10⁻¹⁰ 19 / cm 3 This is the active layer, which is undoped i-Al. 0.09 In 0.91 The thickness of the Sb layer is 2000 nm. The barrier layer is p-Al 0.27 In 0.73 The Sb layer thickness is 20 nm, and the dope concentration is 2.0 × 10⁻⁶. 18 / cm 3 The p-InSb layer thickness is 500 nm, and the dope concentration is 2.0 × 10⁻⁶. 18 / cm 3 Therefore, the height of the energy barrier of the barrier layer relative to the active layer, which affects the resistance described later, is the same for both the multilayer film structure (a) and the multilayer film structure (b).

[0060] Figure 2 shows the defect-to-resistance dependence of the multilayer film structure (a) and (b). The resistance was calculated by varying the defect amount in the region near the barrier layer (30 nm) of the active layer. The resistance referred to here is the resistance to the current flowing longitudinally through the multilayer film structure. The resistance on the vertical axis of Figure 2 is normalized to 1 for the case where the defect amount is 0.01 for each of the multilayer film structures (a) and (b). As shown in Figure 2, the resistance value improves by reducing the defect amount near the barrier layer of the active layer. This effect is more pronounced in the multilayer film structure (b), i.e., when the active layer contains Al.

[0061] Figure 3 is a band diagram of the multilayer structure (a), showing a magnified view of the vicinity of the active layer and the barrier layer. The vertical axis in Figure 3 represents the energy of Ec (bottom of the conduction band) and Ev (top of the valence band). The horizontal axis in Figure 3 represents the position in the depth direction. In IR devices, the recombination current in the active layer is an important factor in determining the resistance. Within the active layer, the valence band is bent at approximately 30 nm (arrow portion) near the barrier layer. From this, it is presumed that holes tend to accumulate in this region, and that the resistance is likely to decrease due to the recombination current.

[0062] <Example 2> The electron mobility in the transverse direction (direction perpendicular to the depth direction) was calculated for the multilayer film structure shown below. Here, mobility is an important indicator for determining the SNR of the Hall element, and a high value is preferable. The manufacturing method of the multilayer film follows the process described above.

[0063] The layered film structure in this embodiment is i-Al 0.1 In 0.9 The structure consists of an active layer and a barrier layer stacked on top of an Sb layer (corresponding to the first layer in Figure 6). 0.1 In 0.9 The Sb layer has a thickness of 100 nm and a mobility of 3000 cm. 2 The value is / Vs. The thickness of the active layer, i-InSb, is 100 nm, and the mobility is 52,000 cm². 2 / Vs. The barrier layer is i-Al 0.1 In 0.9 The Sb layer has a thickness of 100 nm and a mobility of 3000 cm. 2 / Vs is the case.

[0064] Figure 4 shows the defect amount dependence of mobility for a multilayer film structure. Here, mobility refers to the mobility of the current flowing laterally through the multilayer structure. Mobility was calculated by varying the defect amount in the region of the active layer near the barrier layer (30 nm, as in Figure 3). As shown in Figure 4, reducing the defect amount near the barrier layer in the active layer improves the mobility of the multilayer film.

[0065] Further, FIG. 8 is a band diagram of the laminated film structure of the present embodiment, which is an enlarged view of the vicinity of the active layer and the barrier layer. The vertical axis of FIG. 8 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. The coordinates on the horizontal axis of FIG. 8 represent the position in the depth direction. In the active layer, the conduction band is bent in the vicinity of the barrier layer, and it is presumed that electrons tend to accumulate, are susceptible to defects, and are liable to cause a decrease in mobility. Further, FIG. 5 is a band diagram of a configuration in which the thickness of the i-InSb layer of the active layer in Example 2 is 300 nm. In the active layer, the conduction band is bent at about 60 nm (the arrow portion) in the vicinity of the barrier layer. From this, it is presumed that electrons tend to accumulate in this region, and a decrease in mobility is liable to occur.

[0066] Here, particularly when the barrier layer contains Sb, in the region in the vicinity of the barrier layer in the active layer, electrons tend to accumulate due to the bending of the conduction band, and a decrease in mobility is liable to occur. Therefore, it is presumed that the effect of hydrogen termination is significant. The reason for this is described below.

[0067] <Active layer of InAs> FIG. 9 (Example A) is a band diagram of a structure in which a 100 nm i-InAs layer serving as an active layer and a 500 nm i-AlSb layer serving as a barrier layer are laminated on a 500 nm i-AlSb layer, and is an enlarged view of the vicinity of the active layer and the barrier layer. The vertical axis of FIG. 9 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. In the active layer, the conduction band is bent in the vicinity of the barrier layer, and it is presumed that electrons tend to accumulate, are susceptible to defects, and are liable to cause a decrease in mobility.

[0068] FIG. 10 (Comparative Example A) shows i-Al 0.2 In 0.8 On a 500 nm As layer, a 100 nm i-InAs layer serving as an active layer, and a barrier layer made of i-Al 0.2 In 0.8This is an enlarged band diagram of the vicinity of the active layer and the barrier layer for a structure in which an As layer is laminated to a thickness of 500 nm. The vertical axis of FIG. 10 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. No conduction band bending that would cause electron accumulation is observed in the active layer in the vicinity of the barrier layer.

[0069] <InSb active layer> FIG. 11 (Example B) shows i-Al 0.2 In 0.8 This is an enlarged band diagram of the vicinity of the active layer and the barrier layer for a structure in which a 100 nm i-InSb layer as the active layer and a barrier layer are laminated on a 500 nm Sb layer. The vertical axis of FIG. 11 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. In the active layer, the conduction band is bent in the vicinity of the barrier layer, which is presumed to make electrons easy to accumulate, susceptible to the influence of defects, and liable to cause a decrease in mobility. 0.2 In 0.8 This is an enlarged band diagram of the vicinity of the active layer and the barrier layer for a structure in which a 500 nm Sb layer is laminated. The vertical axis of FIG. 11 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. In the active layer, the conduction band is bent in the vicinity of the barrier layer, which is presumed to make electrons easy to accumulate, susceptible to the influence of defects, and liable to cause a decrease in mobility.

[0070] FIG. 12 (Comparative Example B) shows i-Al 0.2 In 0.8 This is an enlarged band diagram of the vicinity of the active layer and the barrier layer for a structure in which a 100 nm i-InSb layer as the active layer and a barrier layer are laminated on a 500 nm As layer. The vertical axis of FIG. 12 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. No conduction band bending that would cause electron accumulation is observed in the active layer in the vicinity of the barrier layer. 0.2 In 0.8 This is an enlarged band diagram of the vicinity of the active layer and the barrier layer for a structure in which a 500 nm As layer is laminated. The vertical axis of FIG. 12 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. No conduction band bending that would cause electron accumulation is observed in the active layer in the vicinity of the barrier layer.

[0071] Here, particularly when the active layer contains In, in the region of the active layer near the barrier layer, electrons are likely to accumulate due to the bending of the conduction band and the mobility is liable to decrease, so the effect of hydrogen termination is presumed to be significant. The reason is described below.

[0072] <Sb-based active layer> FIG. 13 (Example C) shows i-Al 0.2 In 0.8 A band diagram of a structure in which a 100 nm i-InSb layer as an active layer and an i-Al barrier layer are stacked on a 500 nm Sb layer, which is an enlarged view of the vicinity of the active layer and the barrier layer. The vertical axis of FIG. 13 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. In the active layer, the conduction band is bent in the vicinity of the barrier layer, and it is presumed that electrons are likely to accumulate, susceptible to defects, and prone to a decrease in mobility. 0.2 In 0.8 A band diagram of a structure in which a 500 nm Sb layer is stacked, which is an enlarged view of the vicinity of the active layer and the barrier layer. The vertical axis of FIG. 13 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. In the active layer, the conduction band is bent in the vicinity of the barrier layer, and it is presumed that electrons are likely to accumulate, susceptible to defects, and prone to a decrease in mobility.

[0073] FIG. 14 (Comparative Example C) is a band diagram of a structure in which a 100 nm i-GaSb layer as an active layer and a 500 nm i-AlSb layer as a barrier layer are stacked on a 500 nm i-AlSb layer, and is an enlarged view of the vicinity of the active layer and the barrier layer. The vertical axis of FIG. 14 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. In the active layer, no bending of the conduction band that would accumulate electrons is observed in the vicinity of the barrier layer. Here, in Example C and Comparative Example C, a material having a lattice constant relatively close to the lattice constant of the active layer and a band gap larger than that of the active layer is selected as the barrier layer.

[0074] <As-based active layer> FIG. 15 (Example D) is a band diagram of a structure in which a 100 nm i-InAs layer as an active layer and a 500 nm i-AlSb layer as a barrier layer are stacked on a 500 nm i-AlSb layer, and is an enlarged view of the vicinity of the active layer and the barrier layer. The vertical axis of FIG. 15 represents the energy of Ec (the bottom of the conduction band). The conduction band diagram is important for electron mobility. In the active layer, the conduction band is bent in the vicinity of the barrier layer, and it is presumed that electrons are likely to accumulate, susceptible to defects, and prone to a decrease in mobility.

[0075] Figure 16 (Comparative Example D) is a band diagram of a structure in which a 100 nm i-GaAs layer (active layer) and a 500 nm i-AlAs layer (barrier layer) are stacked on a 500 nm i-AlAs layer, showing a magnified view of the vicinity of the active layer and the barrier layer. The vertical axis in Figure 16 represents the energy Ec (bottom of the conduction band). The conduction band diagram is important for electron mobility. No bending of the conduction band that would cause electron accumulation near the barrier layer is observed in the active layer. Here, in Example D and Comparative Example D, a material with a lattice constant relatively close to that of the active layer and a larger band gap than the active layer was selected as the barrier layer.

[0076] While embodiments have been described above based on the drawings and examples, it should be noted that those skilled in the art will find it easy to make various modifications and alterations based on this disclosure. Therefore, it should be noted that these modifications and alterations are within the scope of this disclosure.

Claims

1. The active layer, The active layer comprises a barrier layer laminated on the active layer, The active layer contains arsenic or antimony. The barrier layer contains aluminum and also contains arsenic or antimony. A semiconductor device in which, with the direction in which the barrier layer is stacked on the active layer as the depth direction, the rising position of the hydrogen concentration in the atomic concentration distribution in the depth direction from the active layer to the barrier layer is closer to the center of the active layer than the rising position of the aluminum concentration.

2. The semiconductor device according to claim 1, wherein the active layer contains indium.

3. The semiconductor device according to claim 2, wherein the active layer comprises indium and antimony.

4. The semiconductor device according to claim 2, wherein the active layer comprises indium and arsenic.

5. The semiconductor device according to claim 1, wherein the barrier layer contains antimony.

6. The semiconductor device according to claim 1, wherein, in the distribution of atomic concentrations in the depth direction from the active layer to the barrier layer, the hydrogen concentration reaches a peak and then becomes 50% or less of the peak concentration.

7. The semiconductor device according to claim 6, wherein, in the distribution of atomic concentrations in the depth direction from the active layer to the barrier layer, the position of the hydrogen concentration peak is within a range of 30 nm or less from the interface between the barrier layer and the active layer toward the active layer.

8. The semiconductor device according to claim 6, wherein in the distribution of atomic concentrations in the depth direction from the active layer to the barrier layer, the full width at half maximum of the hydrogen concentration peak is 60 nm or less.

9. The semiconductor device according to claim 1, wherein, in the distribution of atomic concentrations in the depth direction of the barrier layer, the hydrogen concentration at one end of the barrier layer is lower than the hydrogen concentration at the other end adjacent to the active layer.

10. The semiconductor device according to claim 9, wherein the barrier layer is doped in an n-type or p-type manner, and in the distribution of atomic concentrations in the depth direction of the barrier layer, the hydrogen concentration monotonically decreases from one end to the other when the distribution of hydrogen concentrations is approximated by a linear or quadratic function.

11. The semiconductor device according to claim 1, wherein the barrier layer is doped in an n-type or p-type manner.

12. The semiconductor device according to claim 1, wherein the active layer and the barrier layer have different lattice constants in an unstrained state.

13. The semiconductor device according to claim 1, wherein the active layer further comprises aluminum.

14. A method for manufacturing a semiconductor device according to any one of claims 1 to 13, A method for manufacturing a semiconductor device, comprising a pressure-increasing annealing step of increasing the pressure in the apparatus reaction chamber and annealing before forming the barrier layer.

Images