Semiconductor device, optical sensor device and semiconductor device manufacturing method

Abstract

Provided are a semiconductor device and an optical sensor device, each having reduced dark current, and detectivity extended toward longer wavelengths in the near-infrared. Further, a method for manufacturing the semiconductor device is provided. The semiconductor device 50 includes an absorption layer 3 of a type II (GaAsSb/InGaAs) MQW structure located on an InP substrate 1, and an InP contact layer 5 located on the MQW structure. In the MQW structure, a composition x (%) of GaAsSb is not smaller than 44%, a thickness z (nm) thereof is not smaller than 3 nm, and z≥−0.4x+24.6 is satisfied.

Claims

1. A semiconductor device formed on an InP substrate, comprising: an absorption layer of a type II multiple quantum well structure, located on the InP substrate, wherein the multiple quantum well structure is a strain-compensated structure composed of a repetition of a GaAs.sub.1-xSb.sub.x layer and an In.sub.yGa.sub.1-yAs layer, and the GaAs.sub.1-xSb.sub.x layer has an Sb composition x (at. %) and a thickness z (nm) which satisfy relationships (1) to (2) below:
44 at. %≤x≤54.3 at. %, and  (1)
7 nm≤z<10 nm,  (2) wherein in a maximum wavelength at which the absorption layer has detectivity is not shorter than 2.4 μm, and wherein the Sb composition x of the GaAs.sub.1-xSb.sub.x layer and an In composition y (at. %) of the In.sub.xGa.sub.1-yAs layer satisfy 100≤x+y≤104.

2. The semiconductor device according to claim 1, including an InP contact layer on the multiple quantum well structure.

3. The semiconductor device according to claim 1, wherein when, in the multiple quantum well structure, a lattice mismatch of the In.sub.yGa.sub.1-yAs layer is Δω.sub.1 and a lattice mismatch of the GaAs.sub.1-xSb.sub.x layer is Δω.sub.2, a lattice mismatch Δω of the entire multiple quantum well structure is defined by Δω={Σ(Δω.sub.1×thickness of the In.sub.yGa.sub.1-yAs layer+Δω.sub.2×thickness of the GaAs.sub.1-xSb.sub.x layer)}/{Σ(thickness of the In.sub.yGa.sub.1-yAs layer+thickness of the GaAs.sub.1-xSb.sub.x layer)}, and the Δω is not smaller than −0.2% but not greater than 0.2%.

4. The semiconductor device according to claim 3, wherein there is no regrown interface between a bottom surface of the absorption layer and an upper surface of a semiconductor layer including the absorption layer and the InP contact layer.

5. An optical sensor device adopting, as a photodiode, the semiconductor device according to claim 1.

6. A method of manufacturing a semiconductor device on an InP substrate, comprising: a step of forming an absorption layer of a type II multiple quantum well structure on the InP substrate, wherein the multiple quantum well structure is a strain-compensated structure composed of a GaAs.sub.1-xSb.sub.x layer and an In.sub.yGa.sub.1-yAs layer, and in the step of forming the multiple quantum well structure, an Sb composition x (at. %) and a thickness z (nm) of the GaAs.sub.1-xSb.sub.x layer satisfy relationships (1) to (2) below:
44 at. %≤x≤54.3 at. %, and  (1)
7 nm≤z<10 nm,  (2) wherein in a maximum wavelength at which the absorption layer has detectivity is not shorter than 2.4 μm, and wherein the Sb composition x of the GaAs.sub.1-xSb.sub.x layer and an In composition y (at. %) of the In.sub.yGaAs.sub.1-y layer satisfy 100≤x+y≤104.

7. The method of manufacturing a semiconductor device according to claim 6, wherein an In composition y (at. %) of the In.sub.yGa.sub.1-yAs layer is decreased at a rate of 0.9 to 1.2 per increase of 1 at. % of the Sb composition x of the GaAs.sub.1-xSb.sub.x layer.

8. The method of manufacturing a semiconductor device according to claim 6, wherein in the step of forming the multiple quantum well structure, the multiple quantum well structure is formed at a temperature not lower than 400° C. but not higher than 560° C.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a diagram illustrating a semiconductor device according to a first embodiment of the present invention. At interfaces 16 and 17, an oxygen concentration and a carbon concentration are each smaller than 1×10.sup.17 cm.sup.−3. A thickness z of a GaAsSb layer is not smaller than 3 nm, and z (nm)≥−0.4x (at. %)+24.6 is satisfied, wherein x is an Sb composition of the GaAsSb layer.

(2) FIG. 2 is an enlarged view of an MQW structure shown in FIG. 1.

(3) FIG. 3 is a diagram illustrating a type II (InGaAs/GaAsSb) band structure, and a light absorption (reception) phenomenon.

(4) FIG. 4 is a diagram illustrating a range satisfied by an Sb composition x (%) of GaAsSb and a thickness z (nm) of GaAsSb in a type II (InGaAs/GaAsSb) MQW structure. A line indicates z=−0.625x+45.5, A2 line indicates z=−0.27x+21.7, and B line indicates z=−0.4x+24.6.

(5) FIG. 5 is a diagram illustrating a piping system and the like of a growth system for all metal-organic source MOVPE.

(6) FIG. 6A is a diagram illustrating the flow of metal-organic molecules and the flow of heat.

(7) FIG. 6B is a schematic diagram illustrating the metal-organic molecules at a substrate surface.

(8) FIG. 7 is a flowchart of a method of manufacturing a photodiode 50 shown in FIG. 1.

(9) FIG. 8 is a diagram illustrating an optical sensor device including a photodiode array (semiconductor device) according to a second embodiment of the present invention.

(10) FIG. 9 is a diagram illustrating the positions of samples and PL wavelengths thereof, in the relationship between the Sb composition x of GaAsSb and the thickness z (nm) thereof in an MQW structure according to Example 1. A line indicates z=−0.625x+45.5, A2 line indicates z=−0.27x+21.7, and B line indicates z=−0.4x+24.6. In FIG. 9, for example, 2 denotes Comparative Example 2, and encircled 5 denotes Example 5 of the present invention.

(11) FIG. 10 is a diagram illustrating the relationship between the quantum-well-thicknesses of samples in Example 2, and the PL wavelengths and detectivities thereof.

(12) FIG. 11 is a diagram illustrating the relationship between the Sb compositions x of GaAsSb in MQW structures of samples in Example 3, and the PL wavelengths and dark currents (reverse bias voltage Vr=1V) thereof.

DESCRIPTION OF EMBODIMENTS

Embodiment 1

(13) FIG. 1 is a diagram illustrating a semiconductor device 50 according to a first embodiment of the present invention. A photodiode 50 includes, on an InP substrate 1, an InP-based semiconductor layered structure (epitaxial wafer) having a configuration described below. While in FIG. 1 light is incident on the InP substrate side, light may be incident on the epitaxial wafer side.

(14) (an InP substrate 1/an InP buffer layer 2/a type II (InGaAs/GaAsSb) MQW absorption layer 3/an InGaAs diffusive-concentration-distribution-adjusting layer 4/an InP contact layer 5)

(15) A p-type region 6 is located so as to reach into the MQW absorption layer 3 from the InP contact layer 5. The p-type region 6 is formed by selectively diffusing Zn as a p-type impurity from an opening of a selective diffusion mask pattern 36 that is an SiN film. The SiN selective diffusion mask pattern 36 allows the p-type impurity to be diffused and introduced into the inside of a peripheral part of the photodiode 50 in a limited manner when viewed in plan. A p-side electrode 11 composed of AuZn is disposed on and in ohmic contact with the p-type region 6, and an n-side electrode 12 composed of AuGeNi is disposed on and in ohmic contact with the rear surface of the InP substrate 1. In this case, the InP substrate 1 is doped with an n-type impurity, and has a predetermined level of conductivity. Further, the photodiode 50 includes an antireflection layer 35 composed of SiON on the rear surface of the InP substrate 1, so that the photodiode 50 can be used in such a manner that light is incident on the rear surface of the InP substrate. In the type II MQW absorption layer 3, a p-n junction 15 is formed at a position corresponding to a boundary front of the p-type region 6. By applying a reverse bias voltage between the p-side electrode 11 and the n-side electrode 12, a larger depletion layer is formed on the side where the n-type impurity concentration is lower (n-type impurity background). The background impurity concentration in the MQW absorption layer 3 is an n-type impurity concentration (carrier concentration) of about 5×10.sup.15 cm.sup.−3 or lower. The position of the p-n junction 15 is determined by an intersection between the background impurity concentration (n-type carrier concentration) of the MQW absorption layer 3, and the concentration profile of Zn that is a p-type impurity. The diffusive-concentration-distribution-adjusting layer 14 is provided in order to adjust the concentration distribution of the p type impurity in the MQW structure constituting the absorption layer 3. However, the diffusive-concentration-distribution-adjusting layer 14 may not be necessarily formed. In the absorption layer 3, the Zn concentration is preferably 5×10.sup.16 cm.sup.−3 or lower.

(16) The following will describe the points of the photodiode of the present embodiment.

(17) (P1) The absorption layer 3 is composed of the type II (InGaAs/GaAsSb) MQW structure, the Sb composition x of the GaAsSb is not smaller than 44%, and the Sb composition x of the GaAsSb is increased while the In composition y of the InGaAs is decreased so as to absorb the increase of the Sb composition x and maintain InP lattice match as a whole. That is, the GaAsSb and the InGaAs form a strain-compensated structure. This point (P1) is identical to factor (F2) that extends the absorption wavelength region toward longer wavelengths in the near-infrared. The factor (F2) will be described later.
(P2) The thickness z of a unit quantum well (GaAsSb) is not smaller than 3 nm, and satisfies z≥−0.4x (%)+24.6. This feature mainly includes the above-mentioned factor (F1) an increase in the thickness, and secondary includes the factor (F2).

(18) According to the point (P1), an increase in the Sb composition x and a decrease in the In composition y allow reduction in the energy difference between the valence band of the GaAsSb and the conduction band of the InGaAs. In addition, maintaining InP lattice match allows growth of an MQW, a contact layer, and the like having low lattice defect density. As a result, it is possible to extend the absorption wavelength region toward longer wavelengths while keeping the dark current at a low level.

(19) Further, according to the point (P2), the absorbable long-wavelength region can be 2.4 μm or more.

(20) Further, since no regrown interface exists, incorporation of high concentration oxygen and carbon as impurities is avoided. As a result, excellent crystallinity is maintained up to the surface of the InP contact layer, resulting in reduced dark current.

(21) <Extension of Absorption Wavelength Region Toward Longer Wavelengths in Near-Infrared>

(22) There are two factors as follows, which extend the absorption wavelength region of the type II (GaAsSb/InGaAs) MQW structure toward longer wavelengths in the near-infrared.

(23) (F1) an increase in the thickness of each layer in the MQW structure. In the present embodiment, as shown in FIG. 2, the thickness z of the GaAsSb is used as an indicator. Since the thickness of the InGaAs is within a range of ±1.0 nm of the thickness z of the GaAsSb, these thicknesses are regarded to be substantially the same.
(F2) an increase in the Sb composition x to 44% or more and a decrease in the In composition in the GaAsSb/InGaAs. The factor (F2) can be regarded as formation of a strain-compensated structure from the viewpoint of InP lattice match. However, from the viewpoint of band structure, the factor (F2) is reduction in the energy difference between the valence band of the GaAsSb and the conduction band of the InGaAs. A weak point of the factor (F2) is securing of InP lattice match or reduction in lattice defect density.

(24) The factor (F1) an increase in the thickness, and the factor (F2) setting of the compositions x and y will be described with reference to FIG. 3. FIG. 3 is a diagram illustrating the type II (InGaAs/GaAsSb) band structure, and light absorption (reception). When light of a long wavelength λ is absorbed, electrons that have occupied the level of the valence band of the GaAsSb are excited, and the excited electrons occupy the level of the conduction band of the InGaAs. As a result of the light absorption, holes are generated in the valence band. In the type II band structure, since the electrons are excited from the valence band of the GaAsSb to the conduction band of the InGaAs, the energy difference is reduced as compared with that caused by electron transition from the valence band to the conduction band in a single-phase structure, thereby realizing absorption of the long-wavelength light.

(25) When the light of the long wavelength λ is absorbed, the energy difference between the generated holes and electrons is represented by h.Math.(c/λ), where h is Plank's constant (6.626×10.sup.−34 J.Math.s), and c is the velocity of light in the medium. In order to extend the absorption wavelength region toward longer wavelengths, it is necessary to bring the both ends of an arrow of h.Math.(c/λ) shown in FIG. 3 close to each other. As shown in FIG. 3, the factor (F2) reduces the energy difference ΔEvc between the valence band of the GaAsSb and the conduction band of the InGaAs. That is, the energy difference ΔEvc is reduced by changing the band structure in which the electron level is formed. Further, it is considered that the factor (F1) influences as follows. A single layer in the MQW structure forms a single well potential. The energy level of electrons generated in the well potential tends to increase with reduction in the width (thickness) of the well potential. Such increase in the energy level corresponds to that when particles having wave-like nature, such as electrons, are confined in a small space, the energy state thereof is increased as compared with the case where the particles are diffused in a large space, and therefore, can be regarded as a universal nature. If the thickness z is increased, the energy level of electrons (holes) shown in FIG. 3 approaches the valence band and the conduction band. As a result, light absorption is achieved even if the wavelength λ is increased and the h.Math.(c/λ) is reduced.

(26) FIG. 4 is a diagram illustrating a range that is defined by the Sb composition x (%) of the GaAsSb and the thickness z of the GaAsSb in the type II InGaAs/GaAsSb MQW structure in the photodiode 50 of the present embodiment. The In composition y of the InGaAs is set so as to satisfy 100≤x+y≤104. The meaning of each boundary line shown in FIG. 4 will be described below. The reason of setting the range shown in FIG. 4 will be described for Example 1 (refer to FIG. 9).

(27) 1. B line: z=−0.4x+24.6

(28) The B line defines the composition x and the thickness z, with which the maximum absorbable wavelength is 2.43 μm. In a range where z is on or higher than the B line, i.e., in an x-z region on or higher than the B line, the maximum absorption wavelength is 2.4 μm or more. In the present invention, in order to achieve the maximum wavelength of 2.4 μm or more, the thickness z is set to be equal to or higher than the B line.

(29) 2. x=44

(30) The Sb composition x=44(%) is a line in which the maximum wavelength of 2.3 μm is ensured regardless of the thickness z. When the Sb composition x is 44% or more, and then if the thickness z is 7 nm or more, the maximum wavelength of 2.3 μm or more is achieved. In the present embodiment, the Sb composition x is 44(%) or more.

(31) 3. z=3

(32) In the present invention, in order to ensure detectivity, the thickness z is 3 nm or more.

(33) 4. A line: z=−0.625x+45.5

(34) This A line defines a boundary beyond which dark current increases. In a range including and outside the A line (in a range where z is on or higher than the A line), dark current increases, and the S/N ratio decreases.

(35) In a region inside the A line (in a region where z is positioned under the A line), the thickness z increases and the Sb composition x increases as approaching the A line, and thereby the absorption wavelength region is extended toward longer wavelengths. In the present embodiment, if low dark current is emphasized, the x-z range can be set inside the A line. As described above, unless both the thickness z and the Sb composition x approach the A line, it is not possible to realize a semiconductor device in which an absorption wavelength region is extended toward longer wavelengths in the near-infrared, such semiconductor device being expected for wide practical applications.

(36) 5. A2 line: z=−0.27x+21.7

(37) However, in the case where reduction in dark current is emphasized, in order to realize excellent crystallinity, it is preferable that the thickness z is set to satisfy z≤7 nm when the Sb composition x is in a range of (b1) 44 at. %≤x≤54.3 at. %, and is set to satisfy z≤−0.27x+21.7 (A2 line) when the Sb composition x is in a range of (b2) 54.3 at. %≤x≤68 at. %. The A2 line will be described in detail for Examples.

(38) 6. z=10

(39) Detectivity limit is defined by z=10 nm. That is, when the thickness z is 10 nm, overlapping of wave functions of electrons and holes is reduced, which causes reduction in detectivity. That is, the probability of transition of electrons from the valence band of the GaAsSb to the conduction band of the InGaAs is lowered, and transition of such electrons (absorption) is less likely to occur. This is caused by requirements in quantum mechanics. If the thickness z is thinner than 10 nm, i.e., if z<10 is satisfied, detectivity can be ensured.

(40) 7. z=7

(41) However, when the detectivity is emphasized, it is preferable that z≤7 (nm) is satisfied.

(42) <MQW Growth Method>

(43) A description will be given of a manufacturing method. An InP substrate 1 is prepared. On the InP substrate 1, an InP buffer layer 2, a type II (InGaAs/GaAsSb) MQW absorption layer 3, an InGaAs diffusive-concentration-distribution-adjusting layer 4, and an InP contact layer 5 are grown by all metal-organic source MOVPE. In this embodiment, a method for growing the type II (InGaAs/GaAsSb) MQW absorption layer 3 will be described in detail.

(44) FIG. 5 illustrates a piping system and the like of a growth system 60 for all metal-organic source MOVPE. A quartz tube 65 is placed inside a reaction chamber 63, and source gases are introduced into the quartz tube 65. In the quartz tube 65, a substrate table 66 is placed rotatably and hermetically. The substrate table 66 is provided with a heater 66h for heating a substrate. The temperature at a surface of a wafer 50a during deposition is monitored by an IR temperature monitor 61 through a window 69 provided at a ceiling of the reaction chamber 63. The monitored temperature is a temperature which is referred to as a temperature at which growth is performed, or a deposition temperature, or a substrate temperature. When it is described that an MQW structure is formed at a temperature not lower than 400° C. but not higher than 560° C. in the manufacturing method of the present invention, this temperature ranging from 400° C. to 560° C. is a temperature measured by the temperature monitor. Forced evacuation from the quartz tube 65 is performed by means of a vacuum pump.

(45) Source gases are supplied through pipes connected to the quartz tube 65. All metal-organic source MOVPE is characterized by that all source gases are supplied in forms of metal-organic gases. Although source gases of dopants or the like are not shown in FIG. 5, dopants are also supplied in forms of metal-organic gases. The metal-organic gases are stored in a constant temperature bath and kept at a constant temperature. Hydrogen (H.sub.2) and nitrogen (N.sub.2) are used as carrier gases. The metal-organic gases are carried by the carrier gases, and evacuated by the vacuum pump to be introduced into the quartz tube 65. The amount of the carrier gases is precisely controlled by mass flow controllers (MFCs). Many MFCs, electromagnetic valves, and the like are automatically controlled by a microcomputer.

(46) A method for manufacturing the wafer 50a will be described. First, an n-type InP buffer layer 2 is epitaxially grown to a thickness of 150 nm on an S-doped n-type InP substrate 1. Tetraethylsilane (TeESi) is used as an n-type dopant. At this time, trimethylindium (TMIn) and tertiarybutylphosphine (TBP) are used as source gases. The InP buffer layer 2 may be grown by using phosphine (PH.sub.3) as an inorganic source. Even if the InP buffer layer 2 is grown at a growth temperature of about 600° C. or lower, the crystallinity of the underlying InP substrate is not degraded by the heating at about 600° C. However, when forming an InP contact layer, since the MQW structure including GaAsSb is disposed under the InP contact layer, the substrate temperature needs to be precisely maintained within a range of, for example, not lower than 400° C. but not higher than 560° C. The reason is as follows. If the wafer is heated to about 600° C., the GaAsSb is damaged by the heat, and its crystallinity is significantly degraded. In addition, if the InP contact layer is formed at a temperature lower than 400° C., the source gas decomposition efficiency is significantly reduced, and thereby the impurity concentration in the InP layer is increased. Therefore, a high-quality InP contact layer cannot be obtained. Next, an n-doped InGaAs layer is grown to a thickness of 0.15 μm (150 nm) on the InP buffer layer 2. This InGaAs layer is also included in the buffer layer 2 in FIG. 1.

(47) Next, a type II MQW absorption layer 3 having InGaAs/GaAsSb as a pair of quantum wells. As described above, in the quantum well structure, the GaAsSb preferably has a thickness z not smaller than 3 nm but smaller than 10 nm, and the InGaAs preferably has a thickness of z±1.0 nm. Thus, these layers have substantially the same thickness. In FIG. 1, 250 pairs of quantum wells are deposited to form the MQW absorption layer 3. For deposition of the GaAsSb, triethylgallium (TEGa), tertiarybutylarsine (TBAs), and trimethylantimony (TMSb) are used. For deposition of the InGaAs, TEGa, TMIn, and TBAs are used. All the source gases are organic-metal gases, and the molecular weight of each compound is great. Therefore, the source gases are completely decomposed at a relatively low temperature not lower than 400° C. but not higher than 560° C., thereby contributing to the crystal growth. The composition change at an interface of quantum wells can be made sharp by forming the MQW absorption layer 3 by all metal-organic source MOVPE. As a result, highly accurate spectrophotometry is achieved.

(48) As a source of gallium (Ga), triethylgallium (TEGa) or trimethylgallium (TMGa) may be used. As a source of indium (In), trimethylindium (TMIn) or triethylindium (TEIn) may be used. As a source of arsenic (As), tertiarybutylarsine (TBAs) or trimethylarsenic (TMAs) may be used. As a source of antimony (Sb), trimethylantimony (TMSb) or triethylantimony (TESb) may be used. Alternatively, triisopropylantimony (TIPSb) or trisdimethylaminoantimony (TDMASb) may be used. By using these sources, a semiconductor device in which an MQW structure has low impurity concentration and excellent crystallinity can be obtained. As a result, if the semiconductor device is applied to, for example, a photodiode, a photodiode having reduced dark current and high detectivity can be obtained. Moreover, by using the photodiode, an optical sensor device, such as an imaging device, capable of taking clearer images can be realized.

(49) Next, a description will be given of how the source gases flow during formation of the MQW structure 3 by all metal-organic source MOVPE. The source gases are carried through the pipes, introduced into the quartz tube 65, and evacuated. Many kinds of source gases can be supplied to the quartz tube 65 by increasing the number of pipes. For example, even if dozen kinds of source gases are used, the source gases are controlled by open/close of the electromagnetic valves.

(50) Flow of each source gas into the quartz tube 65 is turned on/off according to open/close of the electromagnetic valve, with the flow rate of the source gas being controlled by the mass flow controller (MFC) shown in FIG. 5. Then, the source gases are forcibly evacuated from the quartz tube 65 by the vacuum pump. The flow of the source gases is not interrupted but smoothly and automatically conducted. Accordingly, switching of compositions when forming the pairs of quantum wells is quickly performed.

(51) Since the substrate table 66 rotates as shown in FIG. 5, the source gas temperature distribution does not have a directionality such that the source gas temperature is higher/lower at the source gas inlet side than at the source gas outlet side. Further, since the wafer 50a revolves on the substrate table 66, the flow of the source gas near the surface of the wafer 50a is in a turbulence state. Therefore, even the source gas near the surface of the wafer 50a, excluding the source gas contacting the wafer 50a, has a great velocity component in the direction of gas flow from the gas inlet side toward the gas outlet side. Accordingly, the heat, which flows from the substrate table 66 through the wafer 50a to the source gas, is mostly exhausted together with the exhaust gas, constantly. This causes a great temperature gradient or temperature gap in the vertical direction from the wafer 50a through its surface to the source gas space.

(52) Further, in the embodiment of the present invention, the substrate temperature is set in a low temperature range of not lower than 400° C. but not higher than 560° C. When all metal-organic source MOVPE using metal-organic sources such as TBAs is performed at such a low substrate surface temperature, the decomposition efficiency of the source is high. Therefore, the source gases, which flow in a region very close to the wafer 50a and contribute to the growth of the MQW structure, are limited to those efficiently decomposed into a form required for the growth.

(53) FIG. 6A is a diagram illustrating the flow of metal-organic molecules and the flow of heat, and FIG. 6B is a schematic diagram illustrating the metal-organic molecules at the substrate surface. These figures are used for explaining that setting of the surface temperature is important in order to obtain sharp composition change at heterointerfaces in the MQW structure.

(54) Although it is assumed that the surface of the wafer 50a is at the monitored temperature, a sudden temperature drop or a great temperature gap occurs as described above in the source gas space a little above the wafer surface. Therefore, in the case of using a source gas whose decomposition temperature is T1° C., the substrate surface temperature is set at (T1+α), and the α is determined in view of variation in temperature distribution and the like. Under the situation where a sudden temperature drop or a great temperature gap occurs from the surface of the wafer 50a to the source gas space, if large-size metal-organic molecules as shown in FIG. 6B flow against the wafer surface, the compound molecules that are decomposed and contribute to crystal growth are considered to be limited to those that contact the wafer surface and those in a range equivalent to the thicknesses of a few metal-organic molecules from the wafer surface. Accordingly, as shown in FIG. 6B, it is considered that the metal-organic molecules contacting the wafer surface and the metal-organic molecules located within the region equivalent to the thicknesses of a few metal-organic molecules from the wafer surface mainly contribute to crystal growth, while the metal-organic molecules located outside the region are less likely to be decomposed and are evacuated from the quartz tube 65. When the metal-organic molecules near the surface of the wafer 50a are decomposed and contribute to crystal growth, the metal-organic molecules located outside the region enter the region as supplemental molecules.

(55) Taking the converse point of view, by setting the wafer surface temperature at a temperature slightly higher than the decomposition temperature of the metal-organic molecules, the range of the metal-organic molecules that can participate in crystal growth can be limited to the thin source gas layer on the surface of the wafer 50a.

(56) As understood from the above description, when the source gases suited to the chemical compositions of the above-described pair are introduced by switching the gases using the electromagnetic valves while forcibly evacuating the gases using the vacuum pump, the crystal growth is performed such that, after a crystal of the previous chemical composition was grown with slight inertia, a crystal of the chemical composition, to which the source gases have been switched, can be grown without being affected by the previous source gases. As a result, the composition change at the heterointerface can be made sharp. Such a sharp composition change means that the previous source gases do not substantially remain in the quartz tube 65, and is caused by that the source gases that flow in the region very close to the wafer 50a and contribute to the growth of the MQW structure are limited those efficiently decomposed into the form required for the growth (deposition factor 1). Specifically, as seen from FIG. 5, after one of the two layers in the quantum well is formed, the source gases for forming the other layer is introduced by opening/closing the electromagnetic valves while forcibly evacuating the gases with the vacuum pump. At this time, although some metal-organic molecules that participate in the crystal growth with slight inertia remain, the molecules of the one layer that may act as supplemental molecules are mostly evacuated and gone. As the wafer surface temperature is set closer to the decomposition temperature of the metal-organic molecules, the range of the metal-organic molecules that participate in the crystal growth (the range from the wafer surface) is reduced.

(57) In the case of forming the MQW structure, if the MQW structure is grown at a temperature of about 600° C., phase separation occurs in the GaAsSb layers in the MQW structure, which makes it impossible to realize a clean and flat crystal growth surface of an MQW structure, and an MQW structure having excellent periodicity and crystallinity. Therefore, the growth temperature is set in the range of not lower than 400° C. but not higher than 560° C. (deposition factor 2), and all metal-organic source MOVPE is adopted as a deposition method for this growth, in which metal-organic gases having high decomposition efficiency are used as all source gases (deposition factor 3). The deposition factor 1 significantly depends on the deposition factor 3.

(58) <Semiconductor Device Manufacturing Method>

(59) In the semiconductor device 50 shown in FIG. 1, the InGaAs diffusive-concentration-distribution-adjusting layer 4 is located on the type II MQW absorption layer 3, and the InP contact layer 5 is located on the InGaAs diffusive-concentration-distribution-adjusting layer 4. Zn as a p type impurity is selectively diffused from the opening of the selective diffusion mask pattern 36 disposed on the surface of the InP contact layer 5, thereby forming the p-type region 6. A p-n junction or p-i junction 15 is formed at an end of the p-type region 6. A reverse bias voltage is applied to the p-n junction or p-i junction 15 to form a depletion layer which traps charges caused by photoelectric conversion, and thus the brightness of a pixel is made responsive to the amount of charges. The p-type region 6 or the p-n junction (p-i junction) 15 is a main part that constitutes a pixel. The p-side electrode 11 that is in ohmic-contact with the p-type region 6 is a pixel electrode, and the amount of charges is read, pixel by pixel, between the p-side electrode 11 and the n-side electrode 12 that is set at the ground voltage. On the surface of the InP contact layer that surrounds the p-type region 6, the selective diffusion mask pattern 36 is left as it is. Further, a passivation layer (not shown) composed of SiON or the like covers the selective diffusion mask pattern 36. The reason why the selective diffusion mask pattern 36 is left is as follows. After formation of the p-type region 6, if the selective diffusion mask pattern 36 is removed and the wafer is exposed to the atmosphere, a surface level is formed at the boundary between the surface of the p-type region and the surface of a region where the mask pattern 36 is removed from the contact layer, which causes an increase in dark current.

(60) It is a point that, after formation of the MQW structure, growth is continued in the same deposition chamber or quartz tube 65 by all metal-organic source MOVPE until the InP contact layer 5 is formed. In other words, it is a point that no regrown interface is formed because the wafer 50a is not taken out from the deposition chamber before formation of the InP contact layer 5 to form the contact layer 5 by another deposition method. That is, since the InGaAs diffusive-concentration-distribution-adjusting layer 4 and the InP contact layer 5 are continuously formed in the quartz tube 65, the interfaces 16 and 17 are not regrown interfaces. Therefore, the oxygen concentration and the carbon concentration are both lower than 1×10.sup.17 cm.sup.−3, and no leakage current occurs particularly at a line where the p-type region 6 and the interface 17 intersect.

(61) In the present embodiment, the non-doped InGaAs diffusive-concentration-distribution-adjusting layer 4 having a thickness of, for example, 1.0 μm is formed on the MQW absorption layer 3. After formation of the InP contact layer 5, when Zn as a p-type impurity is introduced from the InP contact layer 5 to reach the MQW absorption layer 3 by a selective diffusion method, if the high-concentration Zn enters the MQW structure, the crystallinity is degraded. The InGaAs diffusive-concentration-distribution-adjusting layer 4 is provided for adjusting the Zn diffusion. The InGaAs diffusive-concentration-distribution-adjusting layer 4 is not necessarily provided as described above.

(62) The p-type region 6 is formed by the above-mentioned selective diffusion, and the p-n junction or p-i junction 15 is formed at an end of the p-type region 6. Even when the InGaAs diffusive-concentration-distribution-adjusting layer 4 is inserted, since the InGaAs has a small band gap, the electric resistance of the photodiode can be reduced even if the InGaAs is non-doped. The reduced electric resistance leads to an increase in responsivity, thereby realizing a moving picture of high image quality.

(63) Preferably, the undoped InP contact layer 5 is epitaxially grown to a thickness of 0.8 μm on the InGaAs diffusive-concentration-distribution-adjusting layer 4 by all metal-organic source MOVPE, with the wafer 50a being placed in the same quartz tube 65. As described above, trimethylindium (TMIn) and tertiarybutylphosphine (TBP) are used as source gases. The use of these source gases allows the growth temperature of the InP contact layer 5 to be not lower than 400° C. but not higher than 560° C., and more preferably, not higher than 535° C. As a result, the GaAsSb in the MQW structure located under the InP contact layer 5 is not damaged by heat, and the crystallinity of the MQW structure is not degraded. When forming the InP contact layer 5, since the MQW structure including GaAsSb is disposed under the InP contact layer 5, the substrate temperature should be precisely maintained within a range of not lower than 400° C. but not higher than 560° C. The reason is as follows. If the wafer is heated to about 600° C., the GaAsSb is damaged by the heat, and its crystallinity is significantly degraded. In addition, if the InP contact layer is formed at a temperature lower than 400° C., the source gas decomposition efficiency is significantly reduced, and thereby the impurity concentration in the InP contact layer 5 is increased. Therefore, a high-quality InP contact layer 5 cannot be obtained.

(64) As described above, conventionally, a MQW structure has to be formed by MBE. However, growth of an InP contact layer by MBE requires a solid source as a source of phosphorus, which leads to a problem in terms of safety or the like. Further, there is still a room for improvement in terms of manufacturing efficiency.

(65) Prior to the present invention, the interface between the InGaAs diffusive-concentration-distribution-adjusting layer and the InP contact layer was a regrown interface once exposed to the atmosphere. The regrown interface can be identified when it satisfies either of the oxygen concentration of 1×10.sup.17 cm.sup.−3 or more and the carbon concentration of 1×10.sup.17 cm.sup.−3 or more, which concentrations can be confirmed by secondary ion mass spectrometry. The regrown interface and the p-type region form an intersection line, and leakage current occurs in the intersection line. Such leakage current significantly degrades the image quality.

(66) Further, if the InP contact layer is grown by simple MOVPE, since phosphine (PH.sub.3) is used as a source of phosphorus, the decomposition temperature is high. Such high decomposition temperature induces a thermal damage of the underlying GaAsSb, and thus the crystallinity of the MQW structure is degraded.

(67) FIG. 7 is a flowchart of a method of manufacturing the photodiode 50 shown in FIG. 1. According to the manufacturing method, it is important to decrease the growth temperature (deposition factor 2) by using only metal-organic gases as source gases (deposition factor 3), and to avoid formation of a regrown interface (deposition factor 4) by growing the layers consistently in the same deposition chamber or quartz tube 65 until formation of the InP contact layer 5 is completed. Thereby, photodiodes having less leakage current, excellent crystallinity, and detectivity in a wavelength region of 2 μm to 5 μm can be efficiently manufactured in large numbers.

Embodiment 2

(68) FIG. 8 illustrates an optical sensor device 10 including a photodiode array (semiconductor device) 50 according to a second embodiment of the present invention. In FIG. 8, optical components such as lenses are omitted. Although a passivation layer 43 composed of an SiON layer is shown in FIG. 8, this passivation layer is actually provided in FIG. 1. The photodiode array 50 has the same layered structure as that of the photodiode shown in FIG. 1. The photodiode array 50 comprises a plurality of photodiodes or pixels P. Since the thickness z and the Sb composition x of the photodiode array 50 are identical to those of the semiconductor device shown in FIG. 1, repeated description is not necessary. Further, like the photodiode (semiconductor device) shown in FIG. 1, the interfaces 16 and 17 are not regrown interfaces, and the impurity concentrations of oxygen, carbon, and the like are low.

(69) In FIG. 8, the photodiode array 50 and a CMOS 70 constituting a readout IC are connected. Readout electrodes (not shown) of the CMOS 70 and pixel electrodes (p-side electrodes) 11 of the photodiode array 50 are connected via interconnection bumps 39. A ground electrode (n-side electrode) 12 common to the pixels of the photodiode array 50, and a ground electrode (not shown) of the CMOS 70 are connected via a bump 12b. It is possible to realize an imaging device or the like by combining the CMOS 70 and the photodiode array 50, and integrating, pixel by pixel, light absorption information.

(70) As described above, the photodiode array (semiconductor device) 50 of the present invention has detectivity in the long wavelength region, and reduced dark current (leakage current). Therefore, when applied to biological tests for animals and plants, environmental monitoring, and the like, the photodiode array (semiconductor device) 50 can execute highly precise examination.

EXAMPLES

Example 1

(71) Samples were prepared by varying the thickness z (nm) and the Sb composition x (%) of GaAsSb in the MQW structure of the photodiode 50 having the same structure as the semiconductor device 50 shown in FIG. 1, and the PL (Photo-Luminescent) wavelength and the dark current of each sample were measured at room temperature. Table I shows the Sb composition x of GaAsSb, the In composition y of InGaAs, and the thickness z (nm) of GaAsSb and InGaAs of each sample. The GaAsSb and the InGaAs have the same thickness. Further, assuming that the lattice mismatch of the InGaAs is Δω.sub.1 and the lattice mismatch of the GaAsSb is Δω.sub.2, the lattice mismatch Δω of the entire MQW structure is defined by Δω={Σ(Δω.sub.1×InGaAs thickness+Δω.sub.2×GaAsSb thickness)}/{Σ(InGaAs thickness+GaAsSb thickness)}. The composition x and the composition y are set so that the lattice mismatch Δω is not smaller than −0.2% but not greater than 0.2%.

(72) Each photodiode 50 was manufactured by all metal-organic source MOVPE as described above. The part other than the absorption layer was manufactured by the above-mentioned method. The major epitaxial layers were manufactured as follows. The growth temperature was 510° C. As for the GaAsSb, TEGa, TBAs, and TESb were used as sources, and the Sb composition x was varied by varying the supply ratio of TBAs to TESb with the V/III ratio (supply of group V sources/supply of group III sources) being kept constant. As for the InGaAs, TEGa, TMIn, and TBAs were used as sources, and the In composition y was varied by varying the supply ratio of TEGa to TMIn with the VIII ratio being kept constant. Further, as for the InP contact layer 5, TMIn and TBP were used as sources.

(73) TABLE-US-00001 TABLE I Sb In compo- compo- Thick- PL Dark current sition sition ness wavelength (nA) of x(at. y(at. z(nm) of (nm) at photodiode %) of %) of InGaAs, room temp- 100 μm dia. Sample GaAsSb InGaAs GaAsSb erature Vr = 1 V Invention 44.2 58 7 2490 50 Example 1 Comparative 44.1 58.3 5 2210 30 Example 2 Comparative 46.3 56 5 2290 30 Example 3 Invention 49 53 10 2870 100 Example 4 Invention 48.6 53 7 2700 80 Example 5 Invention 49.1 53 6 2620 50 Example 6 Invention 49 53 5 2430 40 Example 7 Comparative 49 53.2 3.5 2210 20 Example 8 Invention 51.8 49.7 5 2520 70 Example 9 Comparative 52.2 49.8 3.5 2300 50 Example 10 Invention 54.3 47.2 7 2860 200 Example 11 Invention 54 47.4 5 2630 100 Example 12 Invention 54.1 47 3.5 2410 80 Example 13 Invention 58 43.2 5 2790 200 Example 14 Invention 57.7 43 3.5 2580 150 Example 15 Invention 61.6 37.7 7 3160 700 Example 16 Invention 61.8 38 5 2980 300 Example 17 Invention 62 38.1 3.5 2770 200 Example 18 Invention 64.8 35.2 5 3100 800 Example 19

(74) <Evaluation>

(75) The PL wavelength at room temperature, and the dark current were measured as follows. The dark current was measured under the condition that the absorption diameter was 100 μm, and the reverse bias voltage (Vr) was 1V.

(76) 1. PL Wavelength at Room Temperature

(77) The PL wavelength at room temperature corresponds to the band gap wavelength, and corresponds to the maximum wavelength that can be almost absorbed. With reference to FIGS. 9 and 4, in Comparative Examples 2, 3, 8 and 10, the thicknesses z thereof are positioned under the B line: −0.4x+24.6. According to Table I, the PL wavelengths thereof are 2210 nm, 2290 nm, 2210 nm and 2300 nm, respectively, which are smaller than 2400 nm.

(78) This is because the thicknesses z and the Sb compositions x of GaAsSb are not sufficiently large.

(79) In contrast, in a range where the thickness z is not smaller than 3 nm (z≥3) and is not smaller than −0.4x+24.6 (z≥−0.4x+24.6), i.e., on or higher than the B line, as shown in Table I, all Invention Examples achieve the PL wavelengths longer than 2.4 μm. Furthermore, as the thickness z approaches the A line: z=−0.625x+45.5, the PL wavelength shifts toward longer wavelengths. In FIG. 9, it is indicated by arrows and PL wavelengths that the PL wavelengths increase as the thickness z approaches the A line and the line of z=10.

(80) Further, as the thickness z approaches the line of z=10, the PL wavelength increases in the order of Invention Example 7 (2430 nm).fwdarw.Invention Example 6 (2620 nm).fwdarw.Invention Example 5 (2700 nm).fwdarw.Invention Example 4 (2870 nm).

(81) 2. Dark Current

(82) The dark current is limited by the A line: z=−0.625x+45.5. Since the dark current increases with an increase in the lattice defect density, it is often considered that the dark current is caused by only the Sb composition x and the In composition y. However, as described above, the dark current also depends on the thickness, and is influenced by not only the compositions x and y but by the thickness z. This is because the lattice defect is accumulated.

(83) Invention Examples 16 and 19 are positioned on the A line, and have dark currents of 700 nA and 800 nA, respectively. This level of dark current is allowable in many applications. However, when the dark current is emphasized, the thickness z is preferably lower than the A line. In order to further reduce the dark current with reliability, the thickness z should be on or lower than the A2 line: z=−0.27x+21.7. In this case, Invention Examples 16 and 19 are outside the scope of the present invention, and are regarded as Comparative Examples.

(84) In a range a little apart from the A line, for example, Invention Examples 11, 14, 17 and 18 have dark currents of 200 nA, 200 nA, 300 nA and 200 nA, respectively, which cause no problem in practical applications.

(85) 3. Detectivity

(86) Although it will be later described for Example 2, if the thickness z of the GaAsSb reaches 10 nm, the detectivity degrades. Since a certain level of detectivity is achieved even when the thickness z is 10 nm, this thickness z is allowed in some applications. However, if the detectivity is significantly concerned, the thickness z should be smaller than 10 nm. If higher detectivity is desired, z≤7 (nm) should be satisfied. In this case, Invention Example 4 is outside the scope of the present invention.

(87) This is a quantum-mechanical effect, and is true regardless of the conditions of Examples. Also when the thickness z is as small as 2 nm, the detectivity is low. Therefore, the thickness z is 3 nm or more.

Example 2

Manufacturing of Samples

(88) On an S-doped InP substrate, epitaxial layers including a buffer layer (InP/InGaAs), a type II (InGaAs/GaAsSb) MQW absorption layer, an InGaAs diffusive-concentration-distribution-adjusting layer, and an InP contact layer were consistently formed by all metal-organic source MOVPE. An InP contact layer was grown directly on the MQW absorption layer 3. TEGa, TMIn, TBAs, TBP and TMSb were used as sources of Ga, In, As, P and Sb, respectively. TeESi was used as an n-type impurity dopant.

(89) Specifically, an n-doped InP buffer layer was grown to a thickness of 150 nm on an S-doped InP substrate, and an n-doped InGaAs buffer layer was grown to a thickness of 0.15 μm on the n-doped InP buffer layer. On the two buffer layers, a type II (InGaAs/GaAsSb) MQW absorption layer was grown. In the MQW structure, a lower-side non-doped InGaAs layer and a non-doped GaAsSb layer were paired, and 250 pairs were repeated. The composition was determined so that each layer solely achieves lattice match. The Sb composition x of the GaAsSb was 49%, and the In composition y of the InGaAs was 53%. Assuming that the lattice mismatch of the InGaAs is Δω.sub.1 and the lattice mismatch of the GaAsSb is Δω.sub.2, the lattice mismatch Δω of the entire MQW structure is defined by Δω={Σ(Δω.sub.1×InGaAs thickness+Δω.sub.2×GaAsSb thickness)}/{Σ(InGaAs thickness+GaAsSb thickness)}. The lattice mismatch Δω is not smaller than −0.2% but not greater than 0.2%. An InGaAs diffusive-concentration-distribution-adjusting layer was grown to a thickness of 1.0 μm on the MQW absorption layer, and a non-doped InP contact layer was grown to a thickness of 0.8 μm on the diffusive-concentration-distribution-adjusting layer. For the growth of the GaAsSb, TEGa, TBAs and TMSb were used as sources. For the growth of the InGaAs, TEBa, TMIn and TBAs were used. Further, for the growth of the InP contact layer and the InP buffer layer, TMIn and TBP were used.

(90) TABLE-US-00002 TABLE II Sample Comparative Comparative Invention Invention Invention Example A1 Exame A2 Example A3 Example A4 Example A5 Structure Thickness z(nm) of 2 3 5 7 10 InGaAs, GaAsSb Evaluation PL peak wavelength 1.9 2.1 2.5 2.7 2.9 (μm) at room temperature Detectivity (A/W) 0.1 0.6 0.6 0.5 0.2 of photodiode 1 mm dia. Vr = 5 V λ = 2000 nm

(91) <Evaluation>

(92) 1. PL Characteristics

(93) Results are shown in Table II and FIG. 10. When the thickness z of the GaAsSb and the thickness of the InGaAs were 5 nm, the PL peak wavelength was 2.5 μm. As the thicknesses of both layers were decreased, the PL wavelength shifted toward the shorter wavelengths. When the thicknesses of the GaAsSb and the InGaAs were 2 nm, the PL wavelength was 1.9 μm.

(94) On the other hand, the PL wavelength increased with an increase in the thicknesses of the InGaAs and the GaAsSb. When the thicknesses of both layers were 10 nm, the PL wavelength was 2.9 μm.

(95) 2. Detectivity

(96) Detectivity measurement was performed for light having a wavelength of 2000 nm, at a reverse bias voltage Vr=5V. When the thicknesses of the GaAsSb and the InGaAs were 2 nm, the detectivity was as very low as 0.1 A/W. When the thicknesses of both layers were increased to 3 nm, 5 nm, and 7 nm, the detectivity was improved to 0.6 A/W, 0.6 A/W, and 0.5 A/W, respectively. When the thicknesses of both layers were 10 nm, the detectivity was degraded to 0.2 A/W.

Example 3

(97) Samples having the structure of Example 2 were manufactured according to the same procedure. However, in Example 3, the thicknesses of the GaAsSb and the InGaAs were fixed to 5 nm, and the supply of source gases was controlled as described above. Assuming that the lattice mismatch of the InGaAs is Δω.sub.1 and the lattice mismatch of the GaAsSb is Δω.sub.2, the lattice mismatch Δω of the entire MQW structure is defined by Δω={Σ(Δω.sub.1×InGaAs thickness+Δω.sub.2×GaAsSb thickness)}/{Σ(InGaAs thickness+GaAsSb thickness)}. The lattice mismatch Δω is not smaller than −0.2% but not greater than 0.2%.

(98) TABLE-US-00003 TABLE III Sample Invention Invention Invention Invention Invention Example B1 Example B2 Example B3 Example B4 Example B5 Structure In composition 53 50 47 43 38 y(at. %) of InGaAs Sb composition 49 52 54 58 62 x(at. %) of GaAsSb Evaluation PL peak wavelength   2.4   2.5   2.6   2.8   3.0 (μm) at room temperature Dark current (nA) 40 70 100 200 300 of photodiode (excellent (excellent (excellent (excellent (excellent 100 μm dia. character- character- character- character- character- Vr = 1 V istics) istics) istics) istics) istics) Dark current (nA) 50 100 200 500 800 of photodiode (excellent (excellent (excellent (excellent (excellent 100 μm dia. character- character- character- character- character- Vr = 5 V istics) istics) istics) istics) istics)

(99) <Evaluation>

(100) 1. PL Characteristics

(101) Results are shown in Table III and FIG. 11. When the Sb composition x of the GaAsSb was 49% and the In composition y of the InGaAs was 53%, the PL wavelength was 2.4 μm. When the Sb composition x of the GaAsSb was increased and the In composition y of the InGaAs was decreased, the PL wavelength shifted toward longer wavelengths.

(102) When the Sb composition x was 62% and the In composition y was 38%, the PL wavelength was 3.0 μm.

(103) 2. Dark Current

(104) Regardless of whether the reverse bias voltage Vr was 1V or 5V, when the Sb composition was increased from 44%, the dark current was gradually increased. However, this dark current was at a favorable level. When the In composition y was lower than 53% and the Sb composition was increased from 62%, a difference between the dark current at the reverse bias voltage Vr of 1V and the dark current at the reverse bias voltage Vr of 5V was increased. Therefore, in order to increase the S/N ratio in a photodiode having the detectivity in the longer wavelength region, it is desirable that the reverse bias voltage is reduced (the absolute value thereof is reduced).

(105) Embodiments and Examples of the present invention have been described above. However, the embodiments and the examples of the present invention disclosed above are only illustrative, and the scope of the present invention is not limited to the specific embodiments of the invention. It is to be understood that the scope of the present invention is defined in the appended claims and includes equivalence of the description of the claims and all changes within the scope of the claims.

INDUSTRIAL APPLICABILITY

(106) According to the semiconductor device of the present invention, it is possible to extend the detectivity toward longer wavelengths in the near-infrared region by increasing the Sb composition (decreasing the In composition) and increasing the thickness of each quantum well, while reducing the dark current to a level that causes no problem in practical applications, and while controlling the Sb and In compositions so that the InP lattice match conditions are satisfied by both the GaAsSb and the InGaAs in combination. As a result, the semiconductor device becomes applicable to important applications. Further, by growing the layers consistently in the same growth chamber by all metal-organic source MOVPE, contamination due to impurities is avoided, resulting in high-quality crystallinity. Further, the sharpness of the composition between the quantum wells in the MQW absorption layer can be increased, which enables highly-precise execution of absorption spectrum analysis or the like.

DESCRIPTION OF THE REFERENCE CHARACTERS

(107) 1 InP substrate 2 buffer layer (InP and/or InGaAs) 3 type II MQW absorption layer 4 InGaAs layer (diffusive-concentration-distribution-adjusting layer) 5 InP contact layer 6 p-type region 10 optical sensor device (detection device) 11 p-side electrode (pixel electrode) 12 ground electrode (n-side electrode) 12b bump 15 p-n junction 16 interface between MQW and InGaAs layer 17 interface between InGaAs layer and InP contact layer 35 AR (Anti-Reflection) layer 36 selective diffusion mask pattern 39 interconnection bump 43 passivation layer (SiON layer) 50 photodiode (photodiode array) 50a wafer (interim product) 60 growth system for all metal-organic source MOVPE 61 IR temperature monitor 63 reaction chamber 65 quartz tube 69 window of reaction chamber 66 substrate table 66h heater 70 CMOS P pixel