CN1647281A - Bipolar transistor with graded base layer - Google Patents

Abstract

A semiconductor material which has a high carbon dopant concentration includes gallium, indium, arsenic and nitrogen. The disclosed semiconductor materials have a low sheet resistivity because of the high carbon dopant concentrations obtained. The material can be the base layer of gallium arsenide-based heterojunction bipolar transistors and can be lattice-matched to gallium arsenide emitter and/or collector layers by controlling concentrations of indium and nitrogen in the base layer. The base layer can have a graded band gap that is formed by changing the flow rates during deposition of III and V additive elements employed to reduce band gap relative to different III-V elements that represent the bulk of the layer. The flow rates of the III and V additive elements maintain an essentially constant doping-mobility product value during deposition and can be regulated to obtain pre-selected base-emitter voltages at junctions within a resulting transistor.

Description

Bipolar transistor with graded base layer

Related Patent Applications

This application is a continuation-in-part of U.S. Patent Application Serial No. 10/121,444, filed April 10, 2002, which is a continuation-in-part of U.S. Patent No. 09/995,079, filed November 27, 2000, Application No. 09/995,079 claims the benefit of US Patent Provisional Application No. 60/253,159, filed November 27, 2001, which is hereby incorporated by reference in its entirety. This application also claims the benefit of U.S. Patent Provisional Application No. 60/370,758, filed April 5, 2002, and U.S. Patent Provisional Application No. 60/371,648, filed April 10, 2002, which are hereby incorporated by reference in their entirety Incorporated into this article.

Prior Art of the Invention

Bipolar junction transistor (BJT) and heterojunction bipolar transistor (HBT) integrated circuits (ICs) have developed into important technologies for a variety of applications, especially power amplifiers for wireless mobile phones, microwave instrumentation and for optical fiber High-speed (>10Gbit/s) circuits for communication systems. Future needs are expected to require devices with lower operating voltages, higher frequency performance, higher power-added efficiency, and lower production costs. The threshold voltage (Vbe _,on ) of a BJT or HBT is defined as the base-emitter voltage ( _Vbe ) necessary to achieve a certain fixed collector current density ( _Jc ). Threshold voltage can limit a device's effectiveness for low-power applications where the supply power is limited by battery technology and the power requirements of other components.

Unlike a BJT in which the emitter, base, and collector are made of one semiconductor material, an HBT is made of two dissimilar semiconductor materials, where the emitter semiconductor material has a larger band width than the semiconductor material making the base. Gap (also called "band gap"). This results in superior carrier injection efficiency from the base to the collector on the BJT because there is a fixed barrier that prevents carrier injection from the base back to the emitter. Choosing a base with a smaller bandgap reduces the threshold voltage because an increase in the efficiency of carrier injection from the base into the collector will increase the collector current density at a given base-emitter voltage.

However, HBTs may suffer from the disadvantage that an abrupt break in the band alignment of the heterojunction semiconductor material at the conduction band active species can result at the emitter-base interface of the HBT. The role of this conduction band active species is to block the migration of electrons from the base to the collector. Thus, electrons stay in the base longer, leading to higher recombination levels and lower collector current gain (β _dc ). As previously discussed, since the threshold voltage of a HBT is limited to the base-emitter voltage necessary to achieve a certain fixed collector current density, reducing the collector current gain effectively raises the HBT threshold voltage. Therefore, further improvements in the fabrication of semiconductor materials for HBTs are necessary in order to lower the threshold voltage and thereby improve devices for low voltage operation.

Summary of the invention

The present invention provides an n-doped collector, a base formed on the collector of a III-V material including indium and nitrogen, and an n-doped emitter formed on the base HBT. The group III-V material of the base layer has a carbon dopant concentration of about 1.5×10 ¹⁹ cm ⁻³ to about 7.0×10 ¹⁹ cm ⁻³ . In a preferred embodiment, the base layer includes the elements gallium, indium, arsenic and nitrogen. The presence of indium and nitrogen lowers the bandgap of the material relative to that of GaAs. Furthermore, the dopant concentration in the material is high and the surface resistivity (R _sb ) is low. These factors lead to lower threshold voltages relative to HBTs with GaAs base layers having similar dopant concentrations.

In a preferred embodiment, the group III-V compound material system can be represented by the chemical formula Ga _1-x In _x As _1-y N _y . It is well known that the band gap of Ga _1-x In _x As substantially decreases when a small amount of nitrogen is incorporated into the material. Furthermore, because nitrogen pushes the lattice constant in the opposite direction to indium, Ga1 _{- xInxAs1 - yNy} alloys can be grown to match the lattice of GaAs by adding the appropriate indium/nitrogen ratio to the material. Therefore, excessive strain that leads to increased bandgap and misfit material dislocations can be eliminated. Therefore the indium/nitrogen ratio has been selected so as to reduce or eliminate strain. In a preferred embodiment of the present invention, x=3y in the Ga _1-x In _x As _1-y N _y base layer of the HBT.

In conventional HBTs with GaAs, the current gain typically decreases with increasing temperature due to higher hole-to-emitter injection, higher space charge layer recombination current and possibly shorter diffusion length in the base. In HBTs with a GaInAsN base layer, it has been found that the current gain increases significantly with increasing temperature (approximately 0.3% per 1°C increase). This result is interpreted as an increase in the diffusion length with increasing temperature. Such an effect would be expected if the electrons at the bottom of the band were confined to at least partially localized states and with increasing temperature they were thermally excited to jump from those states to other states where electrons diffuse more easily. Therefore, fabricating the base layer with GaInAsN will improve the temperature characteristics of the HBT of the present invention and reduce the need for temperature compensation auxiliary circuits.

The HBT with a GaInAsN base layer has improved common-emitter output characteristics over conventional HBTs with a GaAs base layer. For example, HBTs with a GaInAsN base layer have lower offset and knee voltages than conventional HBTs with a GaAs base layer.

In one embodiment, the transistor is a double heterojunction bipolar transistor (DHBT) in which the base is made of a different semiconductor material than the emitter and collector. In a preferred embodiment of the DHBT, the Ga _1-x In _x As _1-y N _y base layer can be represented by the chemical formula Ga _1-x In _x As _1-y N _y , the collector is GaAs, and the emitter is selected from InGaP, AlInGaP, and AlGaAs.

Another preferred embodiment of the present invention relates to HBTs or DHBTs, wherein the height of the conduction band active species is reduced in combination with a reduction in the bandgap (E _gb ) of the base layer. Conduction-band active species are caused by discontinuities in the conduction band at the base/emitter heterojunction or base/collector heterojunction. Reducing lattice strain by lattice matching the base layer to the emitter and/or collector layers will lower conduction band active species. This is usually achieved by controlling the concentration of nitrogen and indium in the base layer. Preferably, the base layer has the formula Ga _1-x In _x As _1-y N _y , where x is approximately equal to 3y.

In one embodiment, the base may be compositionally graded to create a graded bandgap layer with a smaller bandgap at the collector and a larger bandgap at the emitter. Preferably, the base layer bandgap is about 20 meV to about 120 meV lower at the surface of the base layer contacting the collector than at the surface of the base layer contacting the emitter. More preferably, the bandgap of the base layer varies linearly in the base layer from collector to emitter.

Adding nitrogen and indium to the GaAs semiconductor material will reduce the band gap of the material. Therefore, the semiconductor material Ga _1-x In _x As _1-y N _y has a lower band gap than GaAs. In the compositionally graded base layer Ga _1-x In _x As _1-y N _y of the present invention, the bandgap of the base layer decreases closer to the collector than to the emitter. However, the average bandgap reduction across the base layer is typically about 10 meV to about 300 meV compared to the bandgap of GaAs. In one embodiment, the average bandgap reduction across the base layer is typically about 80 meV to about 300 meV compared to the bandgap of GaAs. In another embodiment, the average bandgap reduction across the base layer is typically about 10 meV to about 200 meV compared to the bandgap of GaAs. This reduced bandgap will lead to a lower threshold voltage (V _be,on ) of HBTs with compositionally graded base layers Ga _1-x In _x As _1-y N _y than those with GaAs base layers because V The main determinant of _{be, on} is the carrier concentration inherent in the base. The intrinsic carrier concentration (n _i ) is calculated from the following formula:

n _i =N _c N _v exp(-E _g /kT)

In the above formula, N _c is the effective density of states in the conduction band; N _v is the effective density of states in the valence band; E _g is the band gap; T is the temperature; and k is the Boltzmann constant. As can be seen from the equation, the intrinsic carrier concentration in the base is largely controlled by the bandgap of the material used in the base.

Grading the band gap of the base layer from a large band gap at the base-emitter interface to a small band gap at the base-collector interface introduces a quasi-electric field that accelerates electrons across the base in an npn bipolar transistor. polar layer. The electric field increases the electron velocity in the base, thereby reducing the base transit time, improving RF (radio frequency) performance and increasing collector current gain (also called dc current gain). The dc gain (β _dc ) in the case of HBTs with a heavily doped base layer is limited by bulk recombination in the neutral base (n=1). The dc current gain can be estimated with Equation 1:

β _dc = vτ/w _b (1)

In formula (1), v is the average minority carrier velocity in the base; τ is the minority carrier lifetime in the base; and w _b is the base thickness. A properly graded base layer in a HBT with a GaInAsN base layer results in a significant increase _{in βdc} due to increased electron velocity compared to an ungraded GaInAsN base layer.

In order to achieve a bandgap graded in the thickness of the base layer, the base layer is prepared such that its concentration of indium and/or nitrogen is higher at the first surface of the base layer near the collector than at the base layer near the emitter. The second surface of the polar layer. The variation of the indium and/or nitrogen content preferably varies linearly across the base layer, resulting in a linearly graded bandgap. Preferably, the concentration of dopants (eg carbon) remains constant throughout the base layer. In one embodiment, the Ga _1-x In _x As _1-y N _y base layer, such as that of a DHBT, is graded such that x and 3y are approximately equal to 0.01 near the collector and are graded to be at is approximately zero near the emitter. In another embodiment, the Ga _1-x In _x As _1-y N _y base layer is graded such that the value of x at the surface of the base layer contacting the collector is graded in the range of about 0.2 to about 0.02 to The value x is in the range of about 0.1 to 0 at the surface of the base layer contacting the emitter as long as the value x is greater at the surface of the base layer contacting the collector than at the surface of the base layer contacting the emitter. In this embodiment, y may remain constant throughout the base layer or may be linearly graded. When y is linearly graded, the base layer is graded from values of y in the range of about 0.2 to about 0.02 at the surface of the base layer contacting the collector to at the base layer contacting the emitter The value of y in the range of approximately 0.1 to 0 for the surface is sufficient as long as the value y is greater at the surface of the base layer contacting the collector than at the surface of the base layer contacting the emitter. In a preferred embodiment, x is graded linearly from about 0.006 near the collector to about 0.01 near the emitter. In a more preferred embodiment, x is graded linearly from about 0.006 near the collector to about 0.01 near the emitter, while y is about 0.001 everywhere in the base layer.

In another embodiment, the invention is a method of forming a graded semiconductor layer having an essentially linearly graded bandgap from a first surface through the layer to a second surface and an essentially constant doping- The product of mobility. The method includes:

(a) Comparing the doping-mobility products of the calibration layers, each of which is distinct in the deposition of an organometallic compound of atoms from Group III or V of the periodic table or of one of the tetrahalocarbon compounds of deposited carbon Formed at a flow rate whereby the relative flow rates of the organometallic compound and the carbon tetrahalide necessary to form an essentially constant doping-mobility product are determined; and

(b) flowing the organometallic compound and the carbon tetrahalide over the surface at said relative flow rates to form an essentially constant dopant-mobility product, said flow rates being varied during deposition to thereby pass through the Levels of semiconductor layers form an essentially linearly graded bandgap.

The base layer may also be dopant-graded such that the dopant concentration is higher near the collector and gradually decreases across the thickness of the base to the base-emitter heterojunction.

Another approach to minimize conduction band active species is to include one or more heterojunctions in the transition layer. Transition layers with low bandgap barrier layers, graded bandgap layers, doped active species, or combinations thereof can be used to minimize conduction band active species. Additionally, one or more lattice matching layers may be present between the base and emitter or base and collector in order to reduce lattice strain on the material at the heterojunction.

The invention also provides methods of making HBTs and DHBTs. The method involves growing a base layer composed of gallium, indium, arsenic and nitrogen on an n-doped GaAs collector. The base layer can be grown using internal and/or external carbon sources to provide a carbon doped base layer. An n-doped emitter layer is subsequently grown on the base layer. The use of internal and external carbon sources to provide carbon dopant for the base layer can help form a material with a relatively high concentration of carbon dopant. Typically, doping levels of about 1.5 x ¹⁰¹⁹ cm ^-3 to about 7.0 x ¹⁰¹⁹ cm ^-3 are achieved using the methods of the present invention. In preferred embodiments, doping levels of from about 3.0 x ¹⁰¹⁹ cm ^-3 to about 7.0 x ¹⁰¹⁹ cm ^-3 can be achieved using the methods of the present invention. Higher dopant concentrations in the material will reduce the surface resistivity and bandgap of the material. Therefore, the higher the dopant concentration in the base layer of HBT and DHBT, the lower the threshold voltage of the device.

The present invention also provides a material represented by the chemical formula Ga _1-x In _x As _1-y N _y , wherein x and y are each independently about 1.0×10 ⁻⁴ to about 2.0×10 ⁻¹ . Preferably, x is approximately equal to 3y. More preferably, x and 3y are approximately equal to 0.01. In one embodiment, the material is doped with carbon at a concentration of about 1.5×10 ¹⁹ cm ⁻³ to about 7.0×10 ¹⁹ cm ⁻³ . In particular embodiments, the carbon dopant concentration is from about 3.0×10 ¹⁹ cm ⁻³ to about 7.0×10 ¹⁹ cm ⁻³ .

The reduction in threshold voltage can lead to better management of voltage budgets for both wired and wireless GaAs-based RF circuits that are constrained by fixed standard voltage supplies or battery outputs. Lowering the threshold voltage can also change the relative amounts of the different base current components in GaAs-based HBTs. DC current gain stability as a function of both junction temperature and applied stress has previously been shown to be critically dependent on the relative magnitudes of the individual base current components. The reduction in reverse hole injection for low threshold voltage activation is beneficial for both temperature stability and long-term reliability of the device. Therefore, relatively strain-free Ga _1-x In _x As _1-y N _y base materials with high dopant concentration can greatly improve the RF performance in GaAs-based HBTs and DHBTs.

Brief description of the drawings

Figure 1 illustrates an InGaP/GalnAsNDHBT structure of a preferred embodiment of the present invention, where x is approximately equal to 3y.

FIG. 2 is a Gummel curve that graphically illustrates base and collector currents as a function of threshold voltage for an InGaP/GaInAsN DHBT of the present invention and for prior art InGaP/GaAs HBTs and GaAs/GaAs BJTs.

Figure 3 graphically illustrates threshold voltage (at Jc = 1.78 A/cm ² ) as a function of base sheet resistance for InGaP/GaInAsN DHBTs of the present invention and for prior art InGaP/GaAs HBTs and GaAs/GaAs BJTs.

Figure 4 illustrates the photoluminescence spectra measured at 77°K for the InGaP/GaInAsN DHBT of the present invention and the InGaP/GaAs HBT of the prior art (both having a nominal base thickness of 1000 Angstroms). Photoluminescence measurements were taken after etching away the InGaAs and GaAs capping layers, optionally stopping on top of the InGaP emitter. The bandgap of the n-type GaAs collector of both the InGaP/GaAs HBT and the InGaP/GaInAsN DHBT is 1.507 eV. The band gap of the p-type GaAs base layer of InGaP/GaAs HBT is 1.455eV, while the band gap of the p-type GaInAsN base layer of InGaP/GaInAsN is 1.408eV.

Figure 5 illustrates the dual crystal X-ray diffraction (DCXRD) spectra of an InGaP/GaInAsN DHBT of the present invention and a prior art InGaP/GaAs HBT (both having nominal base thicknesses of 1500 Angstroms). The peak position of the base layer is indicated.

Figure 6 is a polaron C-V curve illustrating the carrier concentration across the thickness of the base layer in an InGaP/GaInAsN DHBT of the present invention and a prior art InGaP/GaAs HBT. Both the InGaP/GaInAsN DHBT and the InGaP/GaAs HBT have a nominal base thickness of 1000 Angstroms. Both polaron curves are obtained after selectively etching down to the top of the base layer.

Figure 7a illustrates a preferred structure of an InGaP/GaInAsN DHBT with a transition layer between the emitter and base and a transition and lattice matching layer between the collector and base.

Figures 7b and 7c illustrate alternative structures of InGaP/GaInAsN DHBTs with compositionally graded base layers.

Figure 8 is a graph of doping ^* mobility products as a function of carbon tetrabromide flow rate when a carbon-doped GaInAsN base layer is grown at a fixed indium source gas flow rate ("TMIF" is trimethylindium flow rate ).

Figure 9 is a graph of the TMIF required to obtain a constant doping ^* mobility product as a function of carbon tetrabromide flow rate when a carbon-doped compositionally graded GaInAs base layer is grown.

Figure 10 is a graph showing InGaP/GaInAsN HBTs with lower threshold voltages than InGaP/GaAs HBTs.

FIG. 11 is a graph of ΔV _be of a carbon-doped GaInAsN base layer grown under a fixed TMIF as a function of carbon tetrabromide flow rate.

Figure 12 is a graph of ΔV _be as a function of TMIF.

FIG. 13 is a structure of a DHBT having a compositionally graded base layer used in the experiment of Example 2. FIG.

FIG. 14 is a structure of a DHBT having a fixed base layer composition used in the experiment of Example 2. FIG.

15 is a Gummel plot comparing a DHBT with a fixed GaInAsN base layer composition and a DHBT with a GaInAsN base layer compositionally graded.

16 is a graph of DC current gain as a function of base sheet resistance for a DHBT with a fixed GaInAsN base layer composition versus a DHBT with a GaInAsN base layer compositionally graded.

17 is a Gummel plot comparing a DHBT with a compositionally graded GaInAsN base layer to two DHBTs with a fixed GaInAsN base layer composition.

FIG. 18 is a graph comparing the DC current gain as a function of collector current density of a DHBT with a compositionally graded GaInAsN base layer and a DHBT with two fixed GaInAsN base layer compositions.

19 is a graph comparing the extrapolated current gain cutoff frequency as a function of collector current density for DHBTs with a fixed GaInAsN base layer composition and DHBTs with a GaInAsN base layer compositionally graded.

20 is a graph comparing the small signal current gain of a DHBT with a fixed GaInAsN base layer composition and a DHBT with a GaInAsN base layer compositionally graded as a function of DHBT frequency.

Fig. 21 is a graph of peak f _t as a function of BV _ceo for a DHBT with a fixed composition GaInAsN base layer and a compositionally graded GaInAsN base layer versus a conventional HBT with a GaAs base layer.

Figure 22 is a table showing the composition of a DHBT with graded GaInAsN base layer and tunnel collector.

FIG. 23 is a bandgap diagram of the DHBT depicted in FIG. 22 .

FIG. 24 is a Gummel curve for the DHBT described in FIG. 22 .

FIG. 25 shows the common emitter characteristics of the DHBT described in FIG. 22 .

Detailed description of the invention

The above and other objects, features and advantages of the present invention will become apparent from the preferred embodiments of the invention which are described in more detail below and which illustrate the same parts throughout the different drawings using the same reference characters to indicate the same parts . The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.

A III-V material is a semiconductor having a crystal lattice comprising at least one element from column III (A) of the periodic table and at least one element from column V (A) of the periodic table. In one embodiment, the III-V material is a crystal lattice composed of gallium, indium, arsenic, and nitrogen. Preferably, the III-V material can be represented by the formula Ga _1-x In _x As _1-y N _y , where x and y are each independently about 1.0×10 ⁻⁴ to about 2.0×10 ⁻¹ . More preferably, x is approximately equal to 3y. In the most preferred embodiment, x and 3y are about 0.01.

The term "transition layer" as used herein refers to a film located between a base/emitter heterojunction or a base/collector heterojunction with the function of minimizing the conduction band active species of the heterojunction layer. One way to minimize conduction band active species is to use a series of transition layers whose bandgap transitions from the transition layer closest to the collector to the closest to the base in the base/collector heterojunction The layers are gradually reduced. Similarly, in an emitter/base heterojunction, the bandgap of the transition layers decreases gradually from the transition layer closest to the emitter to the transition layer closest to the base. Another approach to minimize the conduction band active species is to use bandgap graded transition layers. The bandgap of the transition layer can be graded by grading the dopant concentration of the film layer. For example, the dopant concentration of the transition layer may be higher near the base layer and may taper off near the collector or emitter. Alternatively, lattice strain can be used to provide a transition layer with a graded bandgap. For example, the transition layer may be compositionally graded to minimize lattice strain at the surface of the film contacting the base and increase lattice strain at the surface contacting the collector or emitter. Another way to minimize conduction band active species is to use transition layers with dopant concentration active species. One or more of the methods described above for minimizing conduction band active species may be used in the HBT of the present invention. Transition layers suitable for the HBT of the present invention include GaAs, InGaAs and InGaAsN.

Lattice matching layers are layers of different lattice constants grown on the material. The lattice matching layer typically has a thickness of about 500 Angstroms or less and essentially follows the lattice constant of the underlying film layer. This results in an intermediate bandgap between the bandgap of the bottom film layer and the bandgap of the lattice matching material (if it is unstrained). Methods of forming lattice matching layers are known to those skilled in the art and can be found in Ferry et al., Gallium Arsenide Technology (1985), pp. 303-328 (Howard W. Sams & Co., Inc. Indianapolis, Indiana), all of which Content citations are incorporated herein. An example of a suitable material for the lattice matching layer of the HBT of the present invention is InGaP.

base layer composed of fixed HBT and DHBT

The HBT and DHBT of the present invention can be prepared using a suitable metalorganic chemical vapor deposition (MOCVD) epitaxial growth system. Examples of suitable MOCVD epitaxial growth systems are the ALXTRON2400 and ALXTRON2600 platforms. In HBTs and DHBTs prepared by the method of the present invention, generally, an undoped GaAs buffer layer can be grown after in-situ desorption of the oxide. For example, a lower collector layer containing a high concentration of n-dopants (eg, a dopant concentration of about 1×10 ¹⁸ cm ⁻³ to about 9×10 ¹⁸ cm ⁻³ ) can be grown at a temperature of about 700° C. A collector layer having a low n-dopant concentration (for example, a dopant concentration of about 5×10 ¹⁵ cm ⁻³ to about 5×10 ¹⁶ cm ⁻³ ) can be grown on the underlying collector at a temperature of about 700° C. Preferably, both the lower collector and the collector are GaAs. The lower collector layer typically has a thickness of about 4000 Angstroms to about 6000 Angstroms, and the collector electrode typically has a thickness of about 3000 Angstroms to about 5000 Angstroms. In one embodiment, the dopant in the underlying collector and/or collector is silicon. Optionally, an InGaP lattice-matched tunnel layer can be grown on the collector under typical growth conditions. The lattice-matching layer typically has a thickness of less than about 500 angstroms, preferably less than about 200 angstroms, and has a dopant concentration of about 1×10 ¹⁶ cm ⁻³ to about 1×10 ¹⁸ cm ⁻³ .

One or more transition layers may optionally be grown under typical growth conditions on the lattice matching layer or on the collector if no lattice matching layer is used. The transition layer can be prepared from n-doped GaAs, n-doped InGaAs or n-doped InGaAsN. The transition layer may optionally be graded by composition or dopant, or may contain dopant active species. The transition layer typically has a thickness of about 75 Angstroms to about 25 Angstroms. If no lattice matching layer or transition layer is used, a carbon doped GaInAsN base layer is grown on the collector.

The base layer is grown at a temperature below about 750°C and typically has a thickness of about 400 Angstroms to about 1500 Angstroms. In a preferred embodiment, the base layer is grown at a temperature of about 500°C to about 600°C. Optionally, a carbon doped GaInAsN base layer can be grown on the transition layer or on the lattice matching layer if no transition layer is used. The base layer can be made using a suitable gallium source (e.g., trimethylgallium or triethylgallium), arsenic source (e.g., arsine, tributylarsine, or trimethylarsine), an indium source (e.g., trimethylarsine indium) and a nitrogen source (eg, ammonia or dimethylhydrazine) for growth. A low molar ratio of arsenic source to gallium source is preferred. Typically, the molar ratio of arsenic source to gallium source is less than about 3.5. More preferably, the ratio is about 2.0 to about 3.0. The levels of the nitrogen source and the indium source are adjusted to obtain a material consisting of about 0.01% to about 20% indium and about 0.01% to about 20% nitrogen. In a preferred embodiment, the indium content in the base layer is about three times higher than the nitrogen content. In a more preferred embodiment, the indium content is about 1% and the nitrogen content is about 0.3%. In the present invention, a GaInAsN layer having a carbon dopant concentration as high as about 1.5×10 ¹⁹ cm ⁻³ to about 7.0×10 ¹⁹ cm ⁻³ is obtained by using an external carbon source, an organometallic source, especially a gallium source. An example of a suitable external carbon source is carbon tetrabromide. Carbon tetrachloride is also an effective external carbon source.

Optionally one or more transition layers may be the result of the growth of n-doped GaAs, n-doped InGaAs or n-doped InGaAsN between the base and emitter. The transition layer between the base and emitter is relatively lightly doped (eg, about 5.0×10 ¹⁵ cm ⁻³ to about 5.0×10 ¹⁶ cm ⁻³ ) and optionally contains dopant active species. Preferably, the transition layer is about 25 Angstroms to about 75 Angstroms thick.

The emitter layer is grown on the base or optionally on the transition layer at a temperature of about 700° C. and typically has a thickness of about 400 Angstroms to about 1500 Angstroms. For example, the emitter layer includes InGaP, AlInGaP or AlGaAs. In a preferred embodiment, the emitter layer comprises InGaP. The emitter layer may be n-doped at a concentration of about 1.0×10 ¹⁷ cm ⁻³ to about 9.0×10 ¹⁷ cm ⁻³ . A GaAs emitter contact layer comprising an n-dopant at a high concentration (eg, about 1.0×10 ¹⁸ cm ⁻³ to about 9.0×10 ¹⁸ cm ⁻³ ) is optionally formed at the emitter at a temperature of about 700° C. grow on. Typically, the emitter contact layer has a thickness of about 1000 Angstroms to about 2000 Angstroms.

An InGaAs layer ramped in indium composition and having a high concentration (eg, about 50×10 ¹⁸ cm ⁻³ to about 5×10 ¹⁹ cm ⁻³ ) of n-dopants is grown on the emitter layer. This layer typically has a thickness of about 400 Angstroms to about 1000 Angstroms.

Example 1

To illustrate the role of reducing the bandgap of the base layer and/or minimizing conduction band active species at the emitter/base heterojunction, three different types of GaAs-based bipolar transistor structures are compared: GaAs emitter/GaAs Base BJT, InGaP/GaAs HBT and InGaP/GaInAsN DHBT of the present invention. A general representation of the InGaP/GaInAsN DHBT structure used in the following experiments is shown in Figure 1. There is only one heterojunction at the emitter/base interface because both the base and collector are formed of GaAs. The GaAs base layer of the InGaP/GaAs HBT has a larger bandgap than the base of the InGaP/GaInAsND HBT. GaAs/GaAs BJT has no heterojunction because the emitter, collector and base are all made of GaAs. Therefore, the structure of the GaAs BJT was used as a benchmark to determine what, if any, affects the conduction band active species at the base-emitter interface without the collector current characteristics of the InGaP/GaAs HBT. In the DHBT shown in Figure 1, InGaP is selected as the emitter material, and the base is Ga _1-x In _x As _1-y N _y , because InGaP has a wide bandgap, and its conduction band is the same as that of Ga _1-x In _x The conduction bands of As _1-y N _y are aligned. A comparison of the InGaP/GaInAsN DHBT and InGaP/GaAs HBT in Figure 1 was used to determine the effect of having a lower bandgap base layer on the collector current density.

The GaAs devices used in the following discussion all had carbon-doped base layers grown by MOCVD, with dopant concentrations varying from about 1.5×10 ¹⁹ cm ⁻³ to about 6.5×10 ¹⁹ cm ⁻³ and thicknesses from about 500 angstroms varies to about 1500 angstroms, resulting in base surface resistivity (R _sb ) between 100 Ω/□ and 400 Ω/□. Large-area devices (L = 75 μm x 75 μm) were fabricated with a simple wet etch and tested in a common-base configuration. Relatively small amounts of indium (x~1%) and nitrogen (y~0.3%) are incrementally added to form two independent sets of InGaP/GalnAsN DHBTs. For each group, growth has been optimized to maintain high uniform carbon doping levels (>2.5×10 ¹⁹ cm ⁻³ ), good mobility (~85 cm ² /Vs) and high dc current gain (R _sb ~300Ω/□>60).

Typical Gummel curves from GaAs/GaAs BJTs, InGaP/GaAs HBTs, and InGaP/GaInAsN DHBTs with comparable base surface resistivities are plotted as superimposed in Figure 2. The collector currents of InGaP/GaAs HBTs and GaAs/GaAs BJTs are indistinguishable for currents of more than five orders of magnitude (decimal) before the difference in effective series resistance affects the current-voltage characteristics. On the other hand, the collector current of InGaP/GaInAsN DHBT is two times higher than that of GaAs/GaAs BJT and InGaP/GaAs HBT over a fairly wide bias range, corresponding to a collector current of 1.78A/ ^cm2 Density (J _c ) lower threshold voltage decreased by 25.0 mV. The increase in low-bias base current (n=2 components) observed in BJTs is consistent with the increase in bandgap drive in space charge recombination. Neutral base recombination component of base current in InGaP/GaInAsN DHBTs compared to InGaP/GaAs HBTs due to increased collector current and decreased minority carrier lifetime or increased carrier velocity ( _Inbr = I _c w _b /vr) is driven higher. InGaP/GaInAsN DHBT devices fabricated to date have achieved a peak dc current gain of 68 for a device with a base surface resistivity of 234 Ω/□, corresponding to a threshold voltage reduction of 11.5 mV and for a device with a base surface resistivity of 303 Ω/□ , a current gain of 66 at peak dc, corresponding to a 25.0mV reduction in threshold voltage. This represents the highest gain-to-base surface resistivity ratio (β/ _Rsb ~0.2-0.3) known for these structure types. The reduced bandgap in the Ga _{1-x Inx} As _1-y N _y base is responsible for the observed reduction in threshold voltage demonstrated with low temperature (77°K) photoluminescence. DCXRD measurements indicate that the lattice mismatch of the base layer is minimal (<250 arcsec).

In the diffusion limit, the ideal collector current density of a bipolar transistor as a function of base-emitter voltage (V _be ) can be approximated as:

J _c ＝(qD _n n ² _ib /p _b w _b )exp(qV _be /kT) (2)

in

P _b and w _b base doping and width;

D _n diffusion coefficient;

Intrinsic carrier concentration in n _ib base.

By expressing n _ib as a function of the band gap of the base layer (E _gb ) and rewriting the product of base doping and thickness in the base surface resistivity term (R _sb ), the threshold voltage can be expressed as the base surface Logarithmic function of resistance:

V _be =-AIn[R _sb ]+V _o (3)

where A＝(kT/q)

And V _o ＝E _gb /q-(kT/q)In[q ² μN _c N _v D _n /Jc] (5)

where N _c and N _v are the effective densities of states in the conduction and valence bands, and μ is the majority carrier mobility in the base layer.

Figure 3 is a plot of threshold voltage versus base surface resistivity for many InGaP/GaAs HBTs, GaAs/GaAsBJTs and InGaP/GaInAsN DHBTs at _Jc = 1.78A/cm ² . Both the threshold voltages of InGaP/GaAs HBTs and GaAs/GaAsBJTs without any conduction-band active species exhibit qualitatively the same logarithmic dependence on base surface resistivity as expected from equation (2). Quantitatively, the base-emitter voltage (Vbe) does not vary as drastically with base surface resistivity as equation (3) suggests (A = 0.0174 instead of 0.0252mV). However, this observed reduction in A is consistent with quasi-ballistic transport through thin-base GaAs bipolar devices.

A comparison with the properties of GaAs/GaAs BJTs leads to the conclusion that the effective height of the conduction band active species of InGaP/GaAs HBTs may be zero if the collector current assumes an ideal (n=1) state. Therefore, InGaP/GaAs HBTs can be designed such that there are essentially no conduction band active species. Similar results were found in earlier work on AlGaAs/GaAs HBTs. Further reduction of the threshold voltage of these devices suitable for fixed base surface resistivity requires the use of base materials with lower band gaps but still maintain conduction band continuity. Ga _1-x In _x As _1-y N _y can be used to reduce E _gb while maintaining nearby lattice matching conditions. As seen in Figure 3, the threshold voltages of both sets of InGaP/GaInAsN DHBTs follow a logarithmic dependence on the base surface resistivity, indicating that the conduction band active species is approximately zero. Furthermore, the threshold voltage is shifted downwards from the values observed for InGaP/GaAs HBTs and GaAs/GaAs BJTs by 11.5 mV in one group and 25.0 mV in the other (dashed line) .

The above experiments demonstrate that the threshold voltage of a GaAs-based HBT can be reduced below that of a GaAs BJT by using an InGaP/GaInAsN DHBT structure. Low threshold voltage is achieved through two key steps. First, the base-emitter interface is optimized by selecting base and emitter semiconductor materials with conduction bands at nearly the same energy level in order to suppress conduction band active species. This is successfully accomplished by using InGaP or AlGaAs as the emitter material and GaAs as the base. Then, by lowering the bandgap of the base layer, the threshold voltage is further reduced. This is achieved by adding both indium and nitrogen to the base layer while still maintaining lattice matching throughout the entire HBT structure. With appropriate growth parameters, a two-fold increase in collector current density was achieved without greatly sacrificing base doping or minority carrier lifetime (β=68 at Rsb=234Ω/□). These results suggest that the use of Ga _1-x In _x As _1-y N _y materials provides a method for lowering the threshold voltage of GaAs-based HBTs and DHBTs. Since the incorporation of In and Nitrogen into GaAs lowers the material's bandgap, the greater the percentage of In and Nitrogen infused into the base, the greater the percentage of In and Nitrogen incorporated into the base, if a high p-type doping concentration is maintained, the threshold voltage The reduction is also greater.

The reduction in the bandgap assumed to be responsible for the observed reduction in threshold voltage in the GaInAsN base has been confirmed by low-temperature (77°K) photoluminescence. Figure 4 compares the photoluminescence spectra from InGaP/GaINAsN DHBT and conventional InGaP/GaAs HBT. The base layer signal from the InGaP/GaAs HBT is at a lower energy than the collector (1.455eV vs. 1.507eV) due to bandgap narrowing effects associated with high doping levels. The base layer signal from the InGaP/GaInAsN DHBT (appears to be 1.408eV) is reduced due to the bandgap narrowing effect caused by the incorporation of indium and nitrogen into the base layer and the reduction of the base layer's forbidden band width. In this comparison, the doping levels are comparable, thus implying that the 47meV reduction in the base layer signal location is likely to be equal to the reduction in the base layer gap in the GaInAsN base compared to that of the GaAs base . This shift in the photoluminescence signal correlates well with a 45 mV decrease in the measured threshold voltage. In the absence of conduction-band active species, the decrease in threshold voltage may be directly related to the decrease in the forbidden band width of the base layer.

The DCRXD spectrum shown in Figure 5 illustrates the effect of adding carbon dopants and indium to the GaAs semiconductor. Figure 5 shows DCRXD spectra from both an InGaP/GaInAsN DHBT of similar base thickness and a standard InGaP/GaAs HBT. In InGaP/ ^GaAs HBTs, the base layer is seen as the right-hand shoulder ^of the GaAs bottom layer peak, corresponding approximately to +90 arcsec (arcsec) position. Due to the addition of indium, in this particular InGaP/GaInAsN DHBT structure, the base layer peaks at -425 arc seconds (arcsec). In general, the peak position associated with a GaInAsN base is a function of indium, nitrogen and carbon concentrations. The addition of indium to GaAs increases compressive strain, while both carbon and nitrogen compensate with tensile strain.

Maintaining high p-type doping levels when indium (and nitrogen) is added to carbon-doped GaAs requires careful optimization of growth. A rough estimate of the active doping level can be obtained from a combination of measured base surface resistivity and base thickness values. Base doping can also be demonstrated by first selectively etching into the top of the base layer and then obtaining polaron CV profiles. Figure 6 compares such polaron CV doping profiles from a GaAs base layer and a GaInAsN base layer. In both cases, the doping levels exceeded 3×10 ¹⁹ cm ⁻³ .

Figure 7a shows an alternative structure of a DHBT in which instead of a GaInAsN base layer (10) is compositionally fixed, it employs a transition layer (20 and 30) between the emitter/base junction and the collector/base junction. Additionally, a lattice-matched InGaP tunnel layer (40) is used between the transition layer and the collector.

DHBTs with compositionally graded base layers

All layers in a DHBT whose base layer composition is graded can be grown in a similar manner to a DHBT whose base is compositionally fixed, except that the base layer acts as a graded bandgap allowing a junction across the layer to The other junction of the transistor becomes solid. For example, a carbon-doped and bandgap graded GaInAsN base layer can be grown on the collector if neither lattice matching layer nor transition layer is used. Optionally, a carbon doped graded GaInAsN base layer can be grown on the transition layer or, if no transition layer is used, on the lattice matching layer. The base layer can be grown at temperatures below about 750°C and typically has a thickness of about 400 Angstroms to about 1500 Angstroms. In one embodiment, the base layer is grown at a temperature of about 500°C to about 600°C. The base layer can be made using gallium sources (e.g., trimethylgallium or triethylgallium), arsenic sources (e.g., arsine, tri(tert-butyl)arsine, or trimethylarsine), indium sources (e.g., trimethylarsine, for example, methylindium) and a nitrogen source such as ammonia, dimethylhydrazine, or tert-butylamine. A low molar ratio of arsenic source to gallium source is preferred. Typically, the specific molar ratio of the source of arsenic to the source of gallium is less than about 3.5. More preferably, the ratio is from about 2.0 to about 3.0. The levels of nitrogen and indium sources can be adjusted to obtain a material having a Group III indium content of about 0.01% to about 20% and a Group V nitrogen content of about 0.01% to about 20%. In particular embodiments, the Group III element (i.e., indium) content varies from about 10% to 20% at the base-collector junction to about 0.01% to 5% at the base-emitter junction, The content of group V elements (ie nitrogen) is essentially constant at about 0.3%. In another embodiment, the nitrogen content of the base layer is about three times lower than the indium content. As previously discussed for the compositionally fixed GaInAsN base layer, we believe that a high carbon dopant concentration (about 1.5×10 ¹⁹ cm ^-3 to about 7.0×10 ¹⁹ cm ^-3 ) of the GaInAsN layer can This can be achieved using an external carbon source such as carbon tetrahalides. For example, the external carbon source used may be carbon tetrabromide. Carbon tetrachloride is also an effective external carbon source.

Since the organometallic compound used as the indium source gas contributes differently to the carbon dopant content in the GaInAsN base layer than the organogallium compound used as the gallium source gas, the carbon dopant source gas flow is usually performed during the growth of the base layer. tuned to maintain a fixed carbon doping concentration in the compositionally graded GaInAsN base layer. In one embodiment, the variation of the carbon source gas flow over the compositionally graded base layer is determined using the method described.

Carbon and TrimethylIndium Source Flow Calibration Procedures for Graded GaInAsN and/or Graded InGaAs Semiconductor Layers

At least two sets of calibration HBTs were prepared, each set containing at least two members (DHBTs could be used instead of HBTs). The base layer thickness is ideally the same for all calibration HBTs formed, although this is not necessary, and each HBT has a fixed composition, for example, a GaInAsN or GaInAs base layer with a fixed composition and a fixed composition throughout the film layer. carbon dopant concentration. Each group is grown at a source gas flow rate of a Group III or Group V additive (e.g., Group III indium or Group V nitrogen) different from the other group so that each member of a particular group has a different The members of the group are composed of gallium, indium, arsenic and nitrogen. As an example, indium will be used as an additive affecting the bandgap grading. Each member of a particular group is grown at a different flow rate of an external carbon source (eg, carbon tetrabromide or carbon tetrachloride) such that each member of a particular group has a different level of carbon doping. The doping*mobility product is determined for each member and is graded against the carbon source flow rate. The product of doping*mobility varies proportionally to the carbon source gas flow rate for each member of each group. The doping*mobility product as a function of CBr flow rate is plotted in Fig. 8 for the five groups of HBTs. Alternatively, each set of calibration HBTs may be formed by maintaining a constant flow rate of carbon source gas (e.g., carbon tetrabromide), and each individual specimen in each set may be at each setting relative to the other. Source gas flow rates are formed at distinctly different Group III or Group V additive flow rates.

The relative flow rates of carbon source gas to indium source gas required to obtain a constant doping*mobility product are determined by plotting the line across the relationship in FIG. 8 at the constant doping*mobility product (e.g., obtained from a straight line parallel to the x-axis). Where this line intersects each set of lines represents the flow rate of the external carbon source required to obtain this value of the doping*mobility product when the indium source gas flow rate is set at the flow rate appropriate for that set to that set . A plot of external carbon source gas flow rate versus indium source gas flow rate for a constant doping*mobility product value is plotted in FIG. 9 . Similar curves for different doping*mobility products can be plotted in the same way.

The collector current for each HBT is plotted as a function of the base-emitter voltage (V _be ), and the resulting currents are those of members of the group that have a GaAs base layer but are otherwise identical to those of the group being compared (For example, the same dopant concentration, the same base, emitter and collector layer thickness, etc.) HBT curves are compared. The voltage difference between the curves at a specific collector current in terms of base-emitter voltage [ _Vbe ( _ΔVbe )] is attributed to the base layer gap caused by the addition of indium and nitrogen during the formation of the base layer Changes in reduced width. Fig. 10 shows the curves of the collector current as a function of _Vbe for the HBT with GaInAsN base layer and the HBT with GaAs base layer. The horizontal arrow marked between the two curves is ΔV _be . The _ΔVbe for each member of the group is determined and plotted against the carbon source gas flow. ΔV _be versus carbon tetrabromide flux for each member of the five groups of HBTs used to form the curves in FIG. 8 are plotted in FIG. 11 . Please note that the ΔV _be of a group member crosses the range of ΔV _be that the group can fit to a straight line. These lines are then used to determine (interpolate) values of _ΔVbe suitable for certain HBTs that may have been grown using the same indium source gas flow rate for a particular group but with a different carbon source gas flow rate than other members of the group of.

The _ΔVbe fitted to the constant doping*mobility product varies linearly as a function of the indium source gas flow rate, which is plotted against the constant doping*mobility product interpolated _ΔVbe versus the indium source gas flow rate can be seen when the curve is drawn. Figure 12 shows this curve for the five groups used in Figure 11.

The graph shown in Figure 12 was used to determine the indium source gas flow required to obtain the desired _ΔVbe at the base-emitter and base/collector junctions. Once the indium source gas flow has been determined, Figure 9 is used to determine the carbon source gas flow required to obtain the desired dopant*mobility product at that indium source gas flow. To maintain a fixed dopant*mobility product expected in a compositionally graded GaInAs or GaInAsN layer, the same procedure would be followed to determine the expected indium source gas flow and carbon source gas flow at the base-collector junction . The indium source gas flow and the carbon source gas flow are relative to gallium and The level of arsenic is varied linearly in order to obtain a linearly graded base layer with the desired bandgap level.

Example 2

All GaAs devices used in the following discussion had MOCVD-grown carbon-doped base layers with dopant concentrations varying from about 3.0×10 ¹⁹ cm ⁻³ to about 5.0×10 ¹⁹ cm ⁻³ , The thickness varied from about 500 angstroms to about 1500 angstroms, resulting in a base surface resistivity (R _sb ) between 100 Ω/□ and 650 Ω/□. Large area devices (L = 75mm x 75mm) were fabricated using a simple wet etch method and tested in a common base configuration. Relatively small amounts of indium (x~1% to 6%) and nitrogen (y~0.3%) are incrementally added to form two independent sets of InGaP/GaInAsN DHBTs. In order to maintain relatively high uniform carbon dopant level (>2.5×10 ¹⁹ cm ^-3 ), good mobility (~85cm ² /Vs) and high dc current gain (>60 at R _sb ~300Ω/□ ), growth is optimized for each group. The structure of a DHBT with a compositionally graded GaInAsN base layer used in the following experiments is shown in FIG. 13 . Alternative structures for DHBTs in which the base layer is compositionally graded are shown in Figures 7b and 7c. The structure of a DHBT with a fixed composition GaInAsN base layer used for comparison in the following experiments is shown in Fig. 14.

Figure 15 shows Gummel curves from base fixed and graded DHBTs with comparable threshold voltage and base sheet resistance. The neutral base component of the base current is significantly lower in graded base structures which can exhibit peak dc current gains more than 2 times higher than fixed base structures. Figure 16 compares the dc current gain as a function of from similar fixed and graded DHBT structures of different thicknesses. The increase in gain/base sheet resistance ratio is easily seen. Although the ratio of gain/base sheet resistance of a DHBT depends on the growth conditions utilized and details specific to the overall structure, it has been observed that the increase in dc current gain in base-layer graded DHBTs consistently exceeds base-layer-fixed DHBT50 % to 100%.

Figures 17 and 18 compare the Gummel and gain curves from the graded base structure with those from the two fixed base structures. The base composition of the first fixed base structure corresponds to the base layer composition of the graded base at the base-emitter junction. The base composition of the second fixed base structure corresponds to the base layer composition of the graded base at the base-emitter junction. The threshold voltage of the graded base structure is an intermediate value between the two demarcation point structures, but slopes towards the base-emitter demarcation point. The dc current gain of the graded base structure is 50% to 95% higher than that of the demarcation point structure, thus indicating that most of the increase in dc current gain comes from the increase in electron velocity.

On-wafer FF testing was done on two fingered 4 μm x 4 μm emitter area devices using a HP8510C parameter analyzer. The pad parasitic was de-embedded using open and short structures, and the current gain cutoff frequency ( _ft ) was measured using a small-signal current gain (H21) with a slope of -20dB/decade. pushed. Figure 19 summarizes the dependence of _ft on collector current density ( _Jc ) for both structures. Figure 20 illustrates small signal gain versus frequency at a specific bias point.

As J _c increases and the base transit time (t _b ) starts to play a limiting role in the total transit time, the f _t of the graded base structure becomes significantly larger than that of the fixed-composition structure, regardless of the higher Large base thickness (fixed base layer is 60nm thick and graded base layer is 80nm thick). The peak _ft of the 60nm composition-fixed GaInAsN base is 53GHz, while the 80nm composition-graded GaInAsN base has a peak _ft of 60GHz. Therefore, the current gain cutoff frequency is increased by 13%.

In order to better compare the RF results of GaInAsN base layer fixed and graded DHBTs to each other and to conventional GaAs HBTs, the f _t values from Figure 19 are plotted as a function of zero input current applied to the transistor breakdown The curve of voltage (BV _ceo ) change. This curve is compared with the peak or near-peak _ft values of conventional GaAs HBTs cited in the literature. A fairly wide distribution in _ft values for conventional GaAs HBTs is to be expected since this data was compiled from many sets of data using different epitaxy structures, device dimensions and test conditions and is only intended to give a sense of current industry standards. BV _ceo mostly often has to assume the relationship between collector thickness ( ), BV _cbo and BV _ceo shown in Fig. 21 is estimated from the quoted collector thickness. Additionally, assuming that the transit time through the space charge layer of the collector (τ _sclc ) is simply related to _Xc by the electron saturation (drift) velocity (V _s ), shown in Figure 21 is the expected f _t vs. BV _ceo Three simple calculations of the dependencies of . In the baseline calculations, _τb was assumed to be 1.115ps, as expected from Monte-Carlo calculations for a 1000A GaAs base layer, while the sum of the remaining emitter and collector transit times ( _τe +τ _c ) is taken to be 0.95ps.

Examination of FIG. 21 shows that while the f _t of fixed-composition GaInAsN is not quite outside the range expected for conventional GaAs-based HBTs, it is clearly at the lower end of the distribution. The hierarchical base structure is significantly improved. The second calculation (baseline where _τb is reduced by 2/3) suggests that the transit time relative to the constituent fixed structural base is reduced by about 50%. This suggests that the doubling of carrier velocity is achieved in a graded base layer compared to a compositionally fixed base layer, as it is expected that a doubling in velocity combined with a 33% increase in base thickness would result in a _1/2 reduction in τ 2×4/3=2/3. The third calculation (baseline with 1/3 reduction in τ _b and 1/2 reduction in (τ _e +τ _c )) approximation is used together with improved device design and size (minimizing τ _b , τ _e , and τ _c ). and/or graded base structures.

Example 3

In order to increase amplifier efficiency and thus lower operating voltage and extend battery life, lower compensation voltage (V _{CE, sat} ) and knee voltage (V _k ) are required. One way to lower the offset voltage is to minimize the asymmetry in the threshold voltages of the base/emitter and base/collector diode pairs. DHBTs with wide bandgap collectors have been shown to yield low V _CE,sat values, but in practice this leads to higher Vk and reduced efficiency because controlling the potential barrier of the base/collector heterojunction is difficult.

The insertion of a thin layer (tunnel collector) with a high bandgap allows VCE _{, sat} and _Vk to be reduced simultaneously, thereby increasing the efficiency of the device. Figure 22 shows a schematic diagram of a DHBT with graded GaInAsN base layer and tunnel collector. The base layer is graded such that there is a bandgap energy difference of approximately 40 meV between the emitter and collector junctions. A 100 angstrom thick tunnel collector is fabricated between the base and collector, which consists of high bandgap material In _0.5 Ga _0.5 P. FIG. 23 shows a bandgap diagram for the DHBT of FIG. 22 . DHBTs were fabricated using a simple wet etch process as large area devices (L = 75 μm x 75 μm) and tested in common base and common emitter configurations. FIG. 24 shows the Gummel curve and FIG. 25 shows the common emitter characteristics suitable for the DHBT of FIG. 22 . As can be seen from Figures 24 and 25, the device has a low offset voltage of about 0.12V.

Equivalent scheme

While this invention has been particularly shown and described with reference to preferred embodiments thereof, those skilled in the art will understand that various changes in form and detail may be made without departing from the invention as encompassed by the claims. complete within the range.

Claims (65)

1. heterojunction bipolar transistor, comprising:

(a) collector electrode of n-doping;

(b) base stage that comprises III-V family material that forms on collector electrode, wherein the material of III-V family comprises indium and nitrogen, and base stage is with about 1.5 * 10 with carbon ¹⁹Cm ^-3To about 7.0 * 10 ¹⁹Cm ^-3Doped in concentrations profiled; And

(c) emitter that the n-that forms on base stage mixes.

2. according to the transistor of claim 1, wherein base stage containing element gallium, indium, arsenic and nitrogen.

3. according to the transistor of claim 2, wherein collector electrode is GaAs, and emitter is InGaP, AlInGaP or AlGaAs, and transistor is the bipolar transistor of double heterojunction.

4. according to the transistor of claim 2, wherein the base layer band gap is hanged down certain quantity between about 20meV and about 120meV in the band gap on the base layer surface of the base layer surface ratio contact emitter of contact collector electrode.

5. according to the transistor of claim 4, wherein the band gap of base layer is from the base layer surface of contact collector electrode to the base layer surface linear classification of contact emitter.

6. according to the base layer of claim 5, wherein in the base layer of classification average band gap reduce less than the band gap of GaAs between about 20meV with approximately in the scope between the 300meV.

7. according to the base layer of claim 6, wherein average band gap minimizing is lacked about 80meV to about 300meV than the band gap of GaAs in the base layer of classification.

8. according to the base layer of claim 6, wherein average band gap minimizing is lacked about 20meV to about 200meV than the band gap of GaAs in the base layer of classification.

9. according to the transistor of claim 3, wherein base layer comprises that chemical formula is Ga _1-xIn _xAs _1-yN _yRete, wherein x and y are about 1.0 * 10 independently of one another ^-4To about 2.0 * 10-1.

10. according to the transistor of claim 9, wherein x approximates 3y greatly.

11. according to the transistor of claim 9, wherein x have collector electrode other from about 0.2 and about 0.02 numerical value and also by classification to numerical value at emitter other from about 0.1 to about 0, as long as x near the collector electrode greater than near emitter.

12. according to the transistor of claim 11, wherein x is approximately 0.06 at collector electrode, is approximately 0.01 at emitter.

13. according to the transistor of claim 10, wherein base layer has about 400 dusts to arrive the thickness of about 1500 dusts and the surface resistivity that about 100 ohm-sq arrive about 400 ohm-sq.

14. according to the transistor of claim 13, wherein the n-dopant is with between about 3.5 * 10 in emitter ¹⁷Cm ^-3With about 4.5 * 10 ¹⁷Cm ^-3Between concentration exist, and the n-dopant is with between 9 * 10 in collector electrode ¹⁵Cm ^-3To about 2 * 10 ¹⁶Cm ^-3Between concentration exist.

15. according to the transistor of claim 14, wherein emitter and collector all is to use silicon doping.

16. according to the transistor of claim 15, wherein emitter has the thickness of about 500 dusts to about 750 dusts, and collector electrode has the thickness of about 3500 dusts to about 4500 dusts.

17. transistor according to claim 16, further comprise first transition zone that is deposited between base stage and the collector electrode, described first transition zone has the first surface with the first surface adjacency of base stage, and wherein first transition zone comprises the n-dopant material that is selected from GaAs, InGaAs and InGaAsN.

18. transistor according to claim 16, further include with the first surface of the first surface adjacency of emitter and with second transition zone of the second surface of the second surface adjacency of base stage, wherein second transition zone comprises the n-dopant material that is selected from GaAs, InGaAs and InGaAsN.

19. according to the transistor of claim 16, further include with the first surface of the first surface adjacency of collector electrode and with the lattice matching layers of the second surface of the second surface adjacency of first transition zone, wherein lattice matching layers is a wide bandgap material.

20. according to the transistor of claim 19, wherein lattice matching layers is selected from InGaP, AlInGaP and ALGaAs.

21. according to the transistor of claim 18, wherein first and second transition zones have the thickness of about 40 dusts to about 60 dusts.

22. according to the transistor of claim 19, wherein first and second transition zones have the thickness of about 40 dusts to about 60 dusts, and lattice matching layers has the thickness of about 150 dusts to about 250 dusts.

23. a method of making heterojunction bipolar transistor, comprising following step:

(a) grow the base layer that comprises from gallium, indium, arsenic and the nitrogen of gallium, indium, arsenic and nitrogen source on the GaAs collector layer that n-mixes, wherein base layer is to use the carbon p-from the carbon source of outside to mix; With

(b) emitter layer that growth n-mixes on base layer.

24. according to the method for claim 23, wherein Wai Bu carbon source is carbon tetrabromide or carbon tetrachloride.

25. according to the method for claim 24, wherein the source of gallium is selected from trimethyl gallium and triethyl-gallium.

26. according to the method for claim 25, wherein the source of nitrogen is ammonia, dimethylhydrazine or tert-butylamine.

27. according to the method for claim 26, wherein the arsenic source is about 2.0 to about 3.5 with the ratio in gallium source.

28. according to the method for claim 27, wherein base stage is to grow being lower than under about 750 ℃ temperature.

29. according to the method for claim 28, wherein base stage is to grow under about 600 ℃ temperature at about 500 ℃.

30. according to the method for claim 28, wherein base layer comprises that chemical formula is Ga _1-xIn _xAs _1-yN _yRete, wherein x and y are about 1.0 * 10 independently of one another ^-4To about 2.0 * 10 ^-1

31. according to the method for claim 30, wherein x approximates 3y greatly.

32. according to the method for claim 30, wherein collector electrode comprises GaAs, emitter comprises the material that is selected from InGaP, AlInGaP and AlGaAs, and transistor is a double hetero bipolar transistor.

33. method according to claim 30, further be included in the step of growth base layer first transition zone that growth n-mixes on collector layer before, wherein base layer is to grow on first transition zone that n-mixes, and first transition zone has the band gap of classification or the band gap littler than the band gap of collector electrode.

34. according to the method for claim 33, wherein first transition zone is selected from GaAs, InGaAs and InGaAsN.

35. method according to claim 34, further be included in before the emitter layer that growth n-mixes the step of growth second transition zone on base stage, wherein second transition zone have with the first surface of the surface adjacency of base stage first surface and with the second surface of the surface adjacency of emitter, and second transition zone has the doping content than little at least one order of magnitude of doping content of emitter.

36. according to the method for claim 35, wherein second transition zone is selected from GaAs, InGaAs and InGaAsN.

37. according to the method for claim 36, both have the doping spike wherein formed first transition zone, second transition zone or first and second transition zones.

38. method according to claim 36, further be included in before first transition zone that growth n-mixes the step of growth lattice matching layers on collector electrode, wherein lattice matching layers have with the first surface of the first surface adjacency of collector electrode and with the second surface of the second surface adjacency of first transition zone.

39. according to the method for claim 38, wherein lattice matching layers comprises InGaP.

40. a material that comprises gallium, indium, arsenic and nitrogen, wherein material is with about 1.5 * 10 with carbon ¹⁹Cm ^-3To about 7.0 * 10 ¹⁹Cm ^-3Doped in concentrations profiled.

41. according to the material of claim 40, wherein the composition of material can be used chemical formula Ga _1-xIn _xAs _1-yN _yExpression, wherein x and y are independently of one another between about 1.0 * 10 ^-4With about 2.0 * 10 ^-1Between scope in.

42. according to the material of claim 41, wherein x approximates 3y greatly.

43. according to the material of claim 42, wherein x and 3y are about 0.01.

44. according to the material of claim 43, wherein the concentration of carbon is about at least 3.0 * 10 ¹⁹Cm ^-3

45. a material that comprises gallium, indium, arsenic and nitrogen, wherein the composition of material is to use chemical formula Ga _1-xIn _xAs _1-yN _yExpression, wherein x and y be independently of one another from the big numerical value of the first surface of material in the linear classification of fractional value of the second surface of material.

46. according to the material of claim 45, wherein material mixes with carbon.

47. according to the material of claim 46, wherein first surface from about 0.01 to the about 0.06 linear classification of x from the second surface of material to material.

48. a material that comprises gallium, indium, arsenic and nitrogen, wherein the composition of material is to use chemical formula Ga _1-xIn _xAs _1-yN _yExpression, wherein x be from the big numerical value of the first surface of material in the linear classification of fractional value of the second surface of material, and that y keeps in material everywhere is invariable in essence.

49. according to the material of claim 48, wherein material mixes with carbon.

50. according to the material of claim 49, wherein first surface from about 0.01 to the about 0.06 linear classification of x from the second surface of material to material, and y is about 0.001.

51. a formation has the method for semiconductor layer of classification of the product of linear in essence band gap grade and constant in essence doping-mobility to second surface by rete from first surface, this method comprises the steps:

(a) product of the doping-mobility of comparison calibration layer, wherein each alignment layer is to form under the distinct flow velocity of carbon tetrahalide compound organo-metallic compound or deposit carbon of the atom of III or V family in a kind of deposition cycle table, forms the constant in essence doping-needed organo-metallic compound of mobility product and the relative velocity of carbon tetrahalide whereby and is determined; With

(b) organo-metallic compound and carbon tetrahalide compound are flowed from the teeth outwards with described relative speed, to form the product of constant in essence doping-mobility, between depositional stage, change described flow velocity, form the linear in essence band gap grade of the semiconductor layer that passes classification whereby.

52., further be included in and make during the junction device step of deposition classification rete on second semiconductor layer according to the method for claim 51.

53. according to the method for claim 52, wherein second semiconductor layer is a collector layer.

54. according to the method for claim 52, wherein second semiconductor layer is an emitter layer.

55. according to the method for claim 51, wherein the semiconductor layer of classification comprises gallium, indium and arsenic, wherein determines the organo-metallic compound of the deposition rate of carbon tetrahalide to comprise organic indium compound in order to form constant in essence doping-mobility product.

56. according to the method for claim 55, wherein carbon tetrahalide is CBr ₄

57. according to the method for claim 56, wherein organo-metallic compound further comprises the gas as the source of nitrogen.

58. according to the method for claim 57, second semiconductor layer that wherein has the semiconductor layer of classification to be deposited on wherein comprises GaAs.

59., further be included in the step of deposition the 3rd semiconductor layer on the base layer according to the method for claim 58.

60. according to the method for claim 59, wherein the 3rd semiconductor layer is InGaP.

61. method according to claim 58, the product of doping-mobility that wherein is used for each alignment layer is relevant with band gap, and band gap combines and will the relative ratios who deposit requisite organic metal of described graded semiconductor layer and carbon tetrahalide flow velocity be calibrated with doping-mobility product on first and second surfaces of classification rete whereby.

62. according to the method for claim 61, wherein said band gap is that the voltage as the base-emitter of junction device uses described alignment layer to calibrate with respect to GaAs base layer.

63. according to the method for claim 62, wherein the semiconductor base stage layer of formed classification is the base layer in the heterojunction bipolar transistor.

64. according to the method for claim 63, wherein the flow velocity of organo-metallic compound and carbon tetrahalide makes the band gap of the base layer of final classification tie base-collector junction from the base-emitter of described heterojunction bipolar transistor to reduce gradually.

65. the semi-conducting material of making according to the method for claim 51.

Images