JP2007027269A - Bipolar transistor and power amplifier - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a bipolar transistor in which a gain characteristic in a high frequency band is enhanced by keeping the uniformity of heat generation in a HBT cell. <P>SOLUTION: The bipolar transistor has such a structure that a base mesa finger (emitter ledge layer 15, base layer 16 and collector layer 17) is held by two collector fingers (collector electrodes 13), and one base finger (base electrode 12) and two emitter fingers (emitter layer 14 and emitter electrode 11) on the opposite sides thereof are formed on the base mesa finger. The two emitter fingers are formed at symmetric positions with reference to the base finger. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a bipolar transistor and a power amplifier, and more specifically, formed as an active element in a power amplification semiconductor integrated circuit used in a signal transmission unit of a wireless portable terminal using a high frequency band such as a cellular phone. The present invention relates to a bipolar transistor and a power amplifier using the bipolar transistor.

  In recent years, a power amplifying semiconductor integrated circuit using a bipolar transistor, which is a compound semiconductor capable of high-frequency operation and positive power supply operation, as an active element is often used for a signal transmission unit of a wireless portable terminal such as a cellular phone. It has been. In particular, a heterojunction bipolar transistor (hereinafter abbreviated as HBT) in which the base layer is a p-type GaAs layer and the emitter layer heterojunctioned with the base layer is a layer made of AlGaAs or InGaP having a wide band gap. Often used. The reason why this HBT is used in a bipolar transistor is to prevent the holes in the base layer from flowing back into the emitter layer and recombining with the electron carriers in the emitter, thereby increasing the efficiency of injection of electron carriers from the emitter to the base. As a result, the transistor can be operated with high efficiency.

  The HBT generally has a multi-finger structure in which one or a plurality of stripline emitter fingers are arranged (see Patent Document 1 and Patent Document 2). 9A and 9B are a plan view and a cross-sectional view showing a structural example of an HBT having one emitter finger. 9A and 9B, the conventional HBT includes one emitter finger (emitter layer 94 and emitter electrode 91) and two base fingers (base electrode 92) provided at both ends of the emitter finger. The structure is sandwiched between two collector fingers (collector electrode 93). Usually, the output power of the HBT is determined by the area of the emitter. For this reason, in order to realize high output power with an HBT having one emitter finger, it is necessary to increase the emitter finger length L, which causes wasteful occupation of the semiconductor chip area. .

  Therefore, in order to make the HBT high output power without increasing the emitter finger length L, a structure in which a plurality of emitter fingers are provided is adopted. 10A and 10B are a plan view and a cross-sectional view showing an example of the structure of an HBT having four emitter fingers. The conventional HBT shown in FIGS. 10A and 10B has a structure in which four emitter fingers and five base fingers provided at both ends of the emitter fingers are sandwiched between two collector fingers. As described above, since the HBT shown in FIG. 10A can increase the emitter area with a plurality of emitter fingers, when realizing high output power with the HBT, the emitter finger length L can be made shorter than the HBT shown in FIG. 9A. it can.

  In addition, a multi-cell structure in which a multi-finger structure HBT having excellent high-frequency characteristics is used as one unit cell and a plurality of HBT cells are connected in parallel to synthesize outputs is often used for the structure of the power amplification HBT. An example of the structure of the power amplification HBT having this multi-cell structure is shown in FIGS. 11A and 11B. FIG. 11A is an example in which an HBT cell having one emitter finger has a multi-cell structure. FIG. 11B shows an example in which an HBT cell having four emitter fingers has a multi-cell structure. 11A and 11B, the collector outputs of the respective HBT cells are connected together by a common collector wiring 100, and the emitters of the respective HBT cells are connected together by a common emitter wiring 110. Furthermore, the emitter wiring 110 is provided with a via hole 120 and is grounded.

  However, in the case of such a multi-cell structure, it is necessary to pay attention to the following points. In the power amplification HBT, each HBT cell generates heat because the current density is high. However, the heat generation of each HBT cell does not occur uniformly, and the operation failure of the HBT cell occurs due to the non-uniformity of heat generation between the HBT cells. More specifically, a thermal runaway occurs in which the HBT cell having a higher temperature than that of the surrounding where the temperature rises further generates heat due to the positive feedback, which eventually leads to destruction. In addition, if the heat generation of each HBT cell is non-uniform between the HBT cells, all the emitters of the HBT do not function effectively, resulting in a problem that the high frequency characteristics deteriorate.

Therefore, as a countermeasure against non-uniformity of heat generation between the HBT cells, as clearly shown in FIGS. 11A and 11B, a DC bias supply line 150 for supplying a DC bias to each HBT cell and each HBT cell. A method of inserting an external base resistor 130 between each of the base wirings 140 is employed. By inserting the external base resistor 130, an increase in the base current of the HBT can be suppressed, and thermal runaway can be avoided even if the temperature rises. The external base resistor 130 is for stabilizing the operation of the HBT cell, and is generally called a base ballast resistor. Even if a ballast resistor is inserted in the emitter of each HBT cell, an increase in collector current can be suppressed and thermal runaway can be avoided. However, due to the narrow range of values that can be taken by the ballast resistor, the base ballast resistor is more frequently used than the collector ballast resistor recently.
JP-A-6-342803 JP 2002-110904 A

  As described above, in the conventional power amplification HBT using the base ballast resistor, heat non-uniformity between the HBT cells is taken into consideration, but heat non-uniformity within each HBT cell is taken into consideration. Absent. The non-uniformity of heat generation in the HBT cell means non-uniformity of heat generation between the emitter fingers and occurs due to the following reasons.

  In an HBT cell having four emitter fingers as shown in FIG. 10A, if all the emitter fingers generate heat to the same extent, the two emitter fingers in the center are affected by heat from the two outer emitter fingers. High temperature. That is, since the heat generation distribution area of the two central emitter fingers overlaps the heat generation distribution area of the two outer emitter fingers, the temperature of the two central emitter fingers increases (see FIG. 12). As a countermeasure for this problem, it is conceivable that the heat generation distribution region is completely separated by sufficiently widening the interval between the emitter fingers. Because new issues arise, it is not practical.

  Further, since the HBT cell shown in FIG. 10A has five base fingers at both ends of the four emitter fingers, the contact resistance between the base electrode and the base layer of each base finger is not uniform between the fingers. Is likely to occur. For this reason, there is a problem that nonuniformity occurs in the injection amount of the base current, resulting in nonuniformity between the emitter fingers.

  This problem exists similarly for the HBT cell shown in FIG. 9A, which is one emitter finger and two base fingers. That is, in the contact resistance between the base electrode of the two base fingers and the base layer, non-uniformity is likely to occur between the fingers. In particular, this HBT cell has a long base finger in accordance with the emitter finger, resulting in an increase in non-uniformity of contact resistance in the finger length direction. The non-uniform operability in the HBT cell increases due to the non-uniformity of the base resistance on both sides. For example, the current flowing through the emitter finger is greater near the right base finger than at the left base finger.

  Furthermore, the HBT cell has a high parasitic capacitance between the base and the collector, and there is a problem in that gain degradation occurs in a high frequency band application due to power feedback by the parasitic capacitance. The high frequency characteristics of the HBT mainly depend on the parasitic capacitance generated in the base layer and the collector layer, and the parasitic capacitance is proportional to the base mesa width W1 shown in FIGS. 9B and 10B. This is because the base-collector capacitance is proportional to the area of the base layer and the collector layer sandwiching the collector layer. In order to reduce the base mesa width W1, it is necessary to reduce the area other than the necessary emitter area as much as possible. However, the structure of the HBT cell shown in FIG. 10B has five bases for every four emitters, and the structure of the HBT cell shown in FIG. 9B has two bases for one emitter, always one more than the required emitter area. The base electrode area is increased. Thus, in the structure of the conventional HBT cell, there is a limit to narrowing the base mesa width W1 due to the influence of the base finger, and there is a problem that the gain characteristic in the high frequency band deteriorates.

  On the other hand, in the power amplification HBT having a multi-cell structure, the value of the external base resistor 130 is large as shown in FIGS. 11A and 11B. For this reason, when the power amplification HBT is actually integrated in a semiconductor, it is necessary to secure a large area for the external base resistor, resulting in a problem that the chip area is increased and the cost is increased.

  Therefore, an object of the present invention is to provide a bipolar transistor that maintains heat generation uniformity in the HBT cell and has improved high frequency band gain characteristics, and a power amplifier that reduces the chip area in a multi-cell structure. Is to provide.

  The present invention is directed to a bipolar transistor formed on a semiconductor substrate. In order to achieve the above object, the bipolar transistor of the present invention is arranged in parallel with one base finger at one base finger, at a position symmetrical about the one base finger. It consists of two emitter fingers and two collector fingers arranged at a position sandwiching one base finger and two emitter fingers. The emitter fingers may be arranged in even-numbered columns in parallel with one base finger at positions symmetrical with respect to one base finger.

  Preferably, the two emitter fingers are both 30 μm or less in length. Preferably, the electrode width of one base finger is 1 μm or less.

  When this bipolar transistor is a heterojunction bipolar transistor having an emitter layer heterojunction to the base layer, the electrode of one base finger penetrates the wide gap layer of the emitter layer and forms an ohmic junction with the base layer. It is also possible to have a structured.

  A plurality of these bipolar transistors are connected in parallel to synthesize outputs, and the bases of the bipolar transistors are connected in common without going through individual external base resistors, so that a power amplifier can be configured.

  According to the present invention, the uniform operation in the cell is excellent, and the capacity between the base and the collector can be reduced, so that excellent high frequency characteristics can be obtained at low cost. In addition, since the built-in base resistance value is high, when a power amplifier is configured with a large number of cells, it is possible to realize a power amplifier having a small size and excellent breakdown resistance without an external base resistance.

<Bipolar transistor structure>
First, the structure of a bipolar transistor according to an embodiment of the present invention will be described.
FIG. 1A is a plan view showing a structure of a bipolar transistor according to an embodiment of the present invention. 1B is a cross-sectional view taken along the line aa of the bipolar transistor shown in FIG. 1A. In the bipolar transistor according to this embodiment, a base mesa finger (emitter ledge layer 15, base layer 16 and collector layer 17) is sandwiched between two collector fingers (collector electrode 13), and one base finger is placed on the base mesa finger. (Base electrode 12) and two emitter fingers (emitter layer 14 and emitter electrode 11) on both sides thereof are formed. The two emitter fingers are formed at symmetrical positions with respect to the base finger. The bipolar transistor is typically a heterojunction bipolar transistor (HBT).

  In the structure of the bipolar transistor according to this embodiment, since the two emitter fingers are located symmetrically with respect to the base finger, the heat generated by the two emitter fingers is uniform as shown in FIG. Further, since there is only one base finger, there is no problem of non-uniformity between the base fingers.

  Further, the emitter finger length L for preventing non-uniformity is preferably 30 μm or less. For example, as shown in FIG. 3, when the output load of the HBT is set to VSWR = 10: 1, the characteristic of the input power with respect to the saturated output that causes thermal runaway destruction of the 1HBT cell is considered as an index representing nonuniformity. In this case, the larger the number, the better the stability of the device. It can be seen that the HBT is not destroyed even when the emitter finger length L is 30 μm or less and the input power is 5 dB.

  Further, in the bipolar transistor structure according to this embodiment, since there is only one base finger between two emitter fingers, the number of base fingers per emitter finger is minimized. For this reason, the area of the base finger with respect to the emitter area corresponding to the required output power is small, and the area of the base layer that is larger than the emitter area can be reduced. Therefore, the base mesa width W1 can be reduced, and as a result, the capacitance between the base and the collector is reduced, so that a high gain in the high frequency band can be realized.

  Here, if the width of the base electrode 12 is set to 1 μm or less, the contact area between the base electrode 12 and the base layer 16 becomes small, so that the contact resistance can be increased. This contact resistance can be applied to an external base resistance between cells that are usually externally attached as countermeasures against thermal runaway. Therefore, when configuring a power amplification HBT in which a plurality of cells are connected in parallel with a bipolar transistor as one cell. The configuration area can be reduced by the amount of the external base resistor 130 which is necessary for the conventional configuration (FIG. 11A and the like). FIG. 4 shows a configuration example of a power amplification HBT in which a bipolar transistor of the present invention is used as one cell and a plurality of cells are connected in parallel. FIG. 5 shows a characteristic correlation diagram of the base electrode width, the breakdown resistance, and the base-contact resistance when the emitter finger length L of the HBT is 30 μm. As can be seen from FIG. 5, when the width of the base electrode is 1 μm or less, the built-in base resistance becomes 20Ω and increases, and the breakdown resistance performance also increases.

  When a high output characteristic is required for the power amplifier and a large number of cells are required, two or more base mesas are arranged, and the base electrode 12 and the collector electrode 13 are used in common. Can be manufactured. FIG. 6A shows a structure example of the bipolar transistor. Although the base finger is long, the base resistance of about 20Ω is included in one base mesa portion, so that non-uniformity in the two base mesa portions in the new unit cell is suppressed. According to the structure of the present invention, cells can be easily connected and a new cell can be formed. FIG. 6B shows a configuration example of a power amplification HBT in which a plurality of cells are connected in parallel with the bipolar transistor as one cell.

<Bipolar transistor manufacturing method>
Next, a bipolar transistor manufacturing method according to an embodiment of the present invention will be described.
FIG. 7 is a diagram for explaining the steps A to E for manufacturing the bipolar transistor according to the embodiment of the present invention shown in FIG. 1B.

The emitter section is a layer composed of an n-type GaAs layer emitter layer 14 and an n-type InGaP emitter ledge layer 15. Since the layer in contact with the emitter electrode 11 made of WSi needs to be in ohmic contact, a thin InGaAs layer is provided on the uppermost layer of the emitter. A p-type GaAs base layer 16 serving as a base portion is provided below the emitter portion. Below that is an n-type GaAs collector layer 17. Further below that is an n + type GaAs subcollector layer 18 called a heavily doped subcollector layer. The epitaxial layer is etched to form an emitter mesa layer and a base mesa layer, and a transistor is manufactured by a method of forming electrodes of three terminals of the transistor on the semiconductor layer.
In HBT, selective etching of GaAs and InGaP can be used in the etching process. The collector electrode 13 uses an alloy of Ni / AuGe / Au, and the base electrode 12 is Ti / Pt / Au.

  First, the n-type GaAs emitter layer 14 is etched using the WSi emitter electrode 11 formed in the emitter region as a mask to expose the n-type InGaP emitter ledge layer 15 (step A). At this time, a manufacturing method of selective etching of GaAs and InGaP is used. Next, the n-type InGaP emitter ledge layer 15, the p-type GaAs base layer 16, and the n-type GaAs collector layer 17 are etched to expose the n + -type GaAs subcollector layer 18 (step B). Step B is a step of forming a base mesa layer. The smaller the base mesa width W1, the smaller the base-collector capacity and the better the high frequency characteristics.

  Here, the function of the heterojunction emitter ledge layer will be briefly described. There is a problem in that the surface of the p-type base layer exposed by etching away the emitter portion has many surface levels, and crystal degradation occurs due to surface recombination at those levels. Therefore, as a countermeasure against this problem, the surface of the base layer around the emitter layer having a necessary area is deposited and protected by another stable layer, that is, an emitter ledge layer as much as possible. This structure of the emitter ledge layer is also called a guard ring structure.

  The n-type InGaP emitter ledge layer 15 is etched within the width of the emitter mesa interval W4 to expose the p-type GaAs base layer 16 (step C). This etching uses selective etching and opens the ledge opening width W2. In order to improve the high-frequency performance, when forming a high-precision narrow emitter ledge opening width W2 in the narrow emitter mesa interval W4, a high-precision photolithography technique and a high-precision etching technique are required. Requires a manufacturing method.

  Next, the base electrode 12 made of Ti / Pt / Au with a width of 1 μm or less is formed by vapor deposition lift-off (process D). Next, the AuGeNi / Au collector electrode 13 is formed on the n + -type GaAs subcollector layer 18 by vapor deposition lift-off, and then heated to about 400 ° C. to be alloyed (step E).

  By the way, in order to improve the performance in the high frequency band in the HBT having excellent high frequency characteristics, the base mesa width W1 is narrowed, the ledge opening width W2 is secured in the narrow base mesa width W1, and the width is 1 μm. It is necessary to form the following base electrode. However, as described above, the manufacturing cost is increased.

FIG. 8 shows an HBT structure and a manufacturing method that simplify this manufacturing method. In the HBT structure shown in FIG. 8, there is no ledge opening in the first place, and the base electrode 82 penetrates the n-type InGaP emitter ledge layer 15 that is heterojunction with the normal p-type GaAs base layer 16. Touching. In this HBT structure, since there is no opening etching process of the fine n-type InGaP emitter ledge layer 15, the manufacturing method becomes simple. The base electrode 82 is deposited on the n-type InGaP emitter ledge layer 15 and then heat is applied to cause the base electrode 82 to thermally diffuse the n-type InGaP emitter ledge layer 15, thereby forming the p-type GaAs base layer 16. Make contact. In this structure, since there is no ledge opening width and the base width W3 is narrow, the base mesa width W1 can be minimized compared to any structure, the capacitance between the base and the collector is reduced, and the gain and efficiency are improved, resulting in high frequency. Good characteristics.
The collector electrode 13 is made of an alloy of Ni / AuGe / Au, and the base electrode 82 is Pt / Ti / Pt / Au in which Pt is placed in the lowermost layer.

  First, the n-type GaAs emitter layer 14 is etched using the WSi emitter electrode 11 formed in the emitter region as a mask to expose the n-type InGaP emitter ledge layer 15 (step A). At this time, a manufacturing method of selective etching of GaAs and InGaP is used. Next, the n-type InGaP emitter ledge layer 15, the p-type GaAs base layer 16, and the n-type GaAs collector layer 17 are etched to expose the n + -type GaAs subcollector layer 18 (step B). Step B is a step of forming a base mesa layer. The smaller the base mesa width W1, the smaller the base-collector capacity and the better the high frequency characteristics. Step A and step B are the same as those in FIG.

  After forming the base mesa, the base electrode 82 made of Pt / Ti / Pt / Au is formed on the n-type InGaP emitter ledge layer 15 by vapor deposition lift-off (step F). Then, the AuGeNi / Au collector electrode 13 is formed by vapor deposition lift-off, and then heated to about 400 ° C. to be alloyed to form the collector electrode 13 (step G). By heating at about 400 ° C., not only the collector electrode 13 is alloyed, but also the lowermost Pt in the Pt / Ti / Pt / Au base electrode 82 on the n-type InGaP emitter ledge layer 15 is thermally diffused. Since the coefficient is high, the n-type InGaP emitter ledge layer 15 is thermally diffused to penetrate and contact the p-type GaAs base layer 16 for ohmic contact.

  In this structure, there is no need for a high-performance manufacturing method for forming a highly accurate narrow emitter ledge opening width W2 in the narrow emitter mesa interval W4 in order to improve high-frequency performance, so that the cost can be reduced. Since W1> W4> W3 is satisfied, the emitter mesa interval W4 can be further narrowed with respect to the same base width W3, and as a result, the base mesa width W1 can be narrowed. Therefore, the base-collector capacitance can be further reduced, and the performance in the high frequency band can be improved.

  As described above, according to the bipolar transistor of one embodiment of the present invention, the uniform operation in the cell is excellent, and the capacitance between the base and the collector can be reduced, so that excellent high frequency characteristics can be obtained at low cost. In addition, since the built-in base resistance value is high, when a power amplifier is configured with a large number of cells, it is possible to realize a power amplifier having a small size and excellent breakdown resistance without an external base resistance.

  The bipolar transistor and the power amplifier according to the present invention can be used in a signal transmission unit of a wireless portable terminal using a high frequency band, and in particular, it is desired to improve the breakdown resistance against heat generation of the transistor and to improve the gain characteristic of the high frequency band. This is useful in some cases.

The top view which shows the structure of the bipolar transistor which concerns on one Embodiment of this invention Aa sectional view of the bipolar transistor of FIG. 1A Unit cell temperature distribution of the bipolar transistor of FIG. 1A Characteristic diagram showing correlation between emitter finger length and fracture resistance Configuration example of a power amplification HBT in which the bipolar transistor of FIG. Characteristic diagram showing the correlation between base electrode width (built-in base resistance) and breakdown resistance The top view which shows the structure of the other bipolar transistor which concerns on one Embodiment of this invention Configuration example of power amplification HBT in which the bipolar transistor of FIG. The figure explaining the process of manufacturing the bipolar transistor of FIG. 1B The figure explaining the other process of manufacturing the bipolar transistor which improved FIG. 1B Plan view showing the structure of a conventional bipolar transistor Cc sectional view of the bipolar transistor of FIG. 9A Plan view showing the structure of another conventional bipolar transistor Dd sectional view of the bipolar transistor of FIG. 10A A configuration example of a conventional power amplification HBT in which the bipolar transistor of FIG. 10A is a configuration example of a conventional power amplification HBT in which the bipolar transistor of FIG. Unit cell temperature distribution of the bipolar transistor of FIG. 10A

Explanation of symbols

11, 91 Emitter electrodes 12, 82, 92 Base electrodes 13, 93 Collector electrodes 14, 94 Emitter layers 15, 95 Emitter ledge layers 16, 96 Base layers 17, 97 Collector layers 18, 98 Subcollector layers 40, 100 Collector wiring 41 , 110 Emitter wiring 42, 120 Via hole (ground)
45, 150 DC bias supply line 130 External base resistor 140 Base wiring W1 Base mesa width W2 Ledge opening width W3 Base electrode width W4 Emitter mesa width L Emitter finger length

Claims (6)

A bipolar transistor formed on a semiconductor substrate,
One base finger,
Two emitter fingers arranged in parallel with the one base finger, and the one base finger and the two emitter fingers are sandwiched at a position symmetrical about the one base finger A bipolar transistor composed of two collector fingers arranged in position.   2. The bipolar transistor according to claim 1, wherein the two emitter fingers have a length of 30 [mu] m or less.   The bipolar transistor according to claim 1 or 2, wherein an electrode width of the one base finger is 1 µm or less. A heterojunction bipolar transistor having an emitter layer heterojunction to a base layer;
The bipolar device according to any one of claims 1 to 3, wherein the electrode of the one base finger has a structure in which an ohmic junction is formed through the wide gap layer of the emitter layer and the base layer. Transistor. A bipolar transistor formed on a semiconductor substrate,
One base finger,
An even number of emitter fingers arranged in parallel in parallel with the one base finger, and the one base finger and the even number of emitter fingers at positions symmetrical about the one base finger. A bipolar transistor composed of two collector fingers arranged in a sandwiched position. A plurality of bipolar transistors according to any one of claims 3 to 5 are connected in parallel to synthesize an output, and the bases of the bipolar transistors are commonly connected without using an individual external base resistor. A power amplifier.

Images