Patent Application
20110215379
This application claims priority to Japanese Patent Application No. 2010-045440 filed on Mar. 2, 2010, the disclosure of which including the specification, the drawings, and the claims is hereby incorporated by reference in its entirety.
The present disclosure relates to field effect transistors (FETs), and more particularly to transistors using nitrides, which are used for inverters, power supply circuits, etc.
Group III-V nitride compound semiconductors such as gallium nitride (GaN), which are generally called nitride semiconductors, have received attention. The nitride semiconductors are compound semiconductors that are formed by aluminum (Al), gallium (Ga), and indium (In) as group III elements and nitrogen (N) as a group V element, and is represented by the general formula InxGayAl1-x-yN (where 0≦x≦1, 0≦y≦1, and x+y≦1). The nitride semiconductors can form various alloys, and can easily form a heterojunction interface. The heterojunctions of the nitride semiconductors are characterized in that a high density two-dimensional electron gas (2DEG) layer is formed near the junction interface due to spontaneous polarization or piezoelectric polarization even in a non-doped state. FETs using this high density 2DEG layer as carriers have received attention as high frequency and high power devices.
However, a phenomenon called “current collapse” tends to occur in the FETs using nitride semiconductors. The current collapse is a phenomenon in which a current decreases for a certain period of time when the device is turned on again after being turned off. Poor characteristics regarding the current collapse make high speed switching operations difficult, causing a very serious problem in the device operation.
One method to reduce the current collapse is to form a protective film on the surface of an electron supply layer. An example of the protective film is a silicon nitride film (a SiN film) or a p-type organic semiconductor film (see, e.g., Japanese Patent Publication No. 2007-27284).
However, it was found that reducing the current collapse by forming a protective film has the following problem. In the case where the SiN film is used as the protective film, the SiN film makes it difficult for carriers to be trapped in the surface levels. However, since the carriers trapped in the surface levels cannot be eliminated, the current collapse cannot be sufficiently reduced.
In the case where the p-type organic semiconductor film is used as the protective film, the p-type organic semiconductor film is expected to limit trapping of carriers in the surface levels and to eliminate the trapped carriers. However, since the organic semiconductor film needs to be formed by a resistance heating vapor deposition method, a spin-off method, or the like, it is more difficult to form a uniform film as compared to the SiN film that is formed by a chemical vapor deposition (CVD) method or the like. Thus, the interface between the organic semiconductor film and the electron supply layer is unstable, whereby the organic semiconductor layer does not function well as the protective film.
It is an object of the present disclosure to solve the above problems, and to enable an FET using nitride semiconductor, in which current collapse is reduced, to be easily implemented.
In order to achieve the above object, an example FET is configured to include a hole injection portion that is formed near a drain electrode.
Specifically, an example FET includes: a semiconductor stack formed on a substrate, and having a first nitride semiconductor layer, and a second nitride semiconductor layer formed on the first nitride semiconductor layer and having a wider bandgap than that of the first nitride semiconductor layer; a source electrode and a drain electrode, which are formed on the semiconductor stack so as to be separated from each other; a gate electrode formed between the source electrode and the drain electrode so as to be separated from the source electrode and the drain electrode; and a hole injection portion formed on the semiconductor stack at a position closer to the drain electrode than to the gate electrode. The hole injection portion has a p-type third nitride semiconductor layer, and a hole injection electrode formed on the third nitride semiconductor layer. The hole injection electrode and the drain electrode have substantially a same potential.
In the example FET, holes are injected from the hole injection portion into a 2DEG layer when the FET is turned on. The injected holes recombine with electrons trapped in the surface levels or the like. Since the electrons trapped in the surface levels or the like, which cause the current collapse, can be eliminated, the current collapse can be reduced.
Each set of the source electrode 131, the gate electrode 133, and the drain electrode 132 forms a unit 107 of the FET. Adjoining ones of the units 107 share the source electrode 131 or the drain electrode 136. The source electrodes 131 of the units 107 are connected in parallel, the drain electrodes 136 of the units 107 are connected in parallel, and the gate electrodes 133 of the units 107 are connected in parallel, thereby forming one FET as a whole. In each unit 107, a hole injection portion 141 is formed between the drain electrode 132 and the gate electrode 133.
The semiconductor stack 102 has a first nitride semiconductor layer 122 formed on a silicon substrate 100 with a buffer layer 121 interposed therebetween, and a second nitride semiconductor layer 123 formed on the first nitride semiconductor layer 122. The first nitride semiconductor layer 122 is a channel layer where electrons move, and is, e.g., an undoped gallium nitride (GaN) layer having a thickness of about 1-2 μm. The second nitride semiconductor layer 123 is an electron supply layer, and is, e.g., an undoped aluminum gallium nitride (AlGaN) layer having a thickness of about 15-50 nm. Note that the term “undoped” means that no impurity has been intentionally introduced into the layer.
The source electrode 131 and the drain electrode 132 can be, e.g., stacks of titanium (Ti) and aluminum (Al). The source electrode 131 and the drain electrode 132 need only form an ohmic contact with a channel that is made of a two-dimensional electron gas (2DEG) layer formed near the interface of the first nitride semiconductor layer 122 with the second nitride semiconductor layer 123. The gate electrode 133 can be, e.g., a stack of nickel (Ni) and gold (Au). The gate electrode 133 need only form a Schottky junction with the channel. Note that the source electrode 131 and the drain electrode 132 may have a recess structure that is in direct contact with the 2DEG layer.
The hole injection portion 141 has a p-type third nitride semiconductor layer 142 selectively formed on the second nitride semiconductor layer 123, and a hole injection electrode 143 formed on the third nitride semiconductor layer 142. The third nitride semiconductor layer 142 is, e.g., a GaN layer doped with magnesium (Mg). The Mg concentration is about 1×1019 cm−3, and the carrier density is about 1×1018 cm−3. The third nitride semiconductor layer 142 can be formed by forming a nitride semiconductor layer on the entire surface of the second nitride semiconductor layer 123 and removing an unwanted part of the nitride semiconductor layer by an inductively coupled plasma (ICP) etching method using chlorine gas. The hole injection electrode 143 is made of, e.g., palladium, and forms an ohmic contact with the third nitride semiconductor layer 142.
A first protective film 127, which is made of silicon nitride (Si3N4), is formed over the second nitride semiconductor layer 123 so as to cover the source electrode 131, the drain electrode 132, the gate electrode 133, and the hole injection portion 141. A source electrode interconnect 135 is formed over the source electrode 131. The source electrode interconnect 135 is connected to the source electrode 131 via an opening formed in the first protective film 127. The first protective film 127 can be formed by a chemical vapor deposition (CVD) method, and the opening can be formed by a dry etching method using chlorine gas or the like.
A drain electrode interconnect 136 is formed over the drain electrode 132. The drain electrode interconnect 136 is connected to the drain electrode 132 via an opening formed in the first protective film 127. A second protective film 128, which is made of Si3N4, is formed over the first protective film 127 so as to cover the source electrode interconnect 135 and the drain electrode interconnect 136. The second protective film 128 can be formed by a CVD method.
A source electrode pad 151, a drain electrode pad 152, and a gate electrode pad 153 are formed over the element isolation region 105. The source electrode pad 151 is connected to the source electrodes 131 via the source electrode interconnects 135. The drain electrode pad 152 is connected to the drain electrodes 132 via the drain electrode interconnects 136. The gate electrode pad 153 is connected to the gate electrodes 133 via a gate electrode interconnect 137 that is formed integrally with the gate electrodes 133. The drain electrode pad 152 is connected to the hole injection electrodes 143. Thus, the drain electrodes 132 are connected to the hole injection electrodes 143.
The operation of the FET of the present embodiment will be described below. Current collapse is considered to be caused by electrons trapped in the surface levels. In conventional FETs having no hole injection portion 141, the 2DEG layer between the gate electrode 133 and the drain electrode 132 is depleted by electrons trapped in the surface levels of the second nitride semiconductor layer 123 or the like, if a drain bias as high as about several tens of volts is applied to the drain electrode when the FET is off. Since it takes more time to release the electrons trapped in the surface levels than to trap electrons, the depletion layer spreads between the gate electrode 133 and the drain electrode 132 even immediately after the gate is turned on. Thus, the channel is not completely open, and the channel resistance increases.
On the other hand, the FET of the present embodiment has the hole injection portions 141. The hole injection portion 141 has the p-type third nitride semiconductor layer 142, and the hole injection electrode 143 that forms an ohmic contact with the third nitride semiconductor layer 142. The hole injection electrode 143 is connected to the drain electrode 132, and the hole injection electrode 143 and the drain electrode 132 have substantially the same potential. Thus, when the FET is turned on, holes are injected from the p-type third nitride semiconductor layer 142 toward the 2DEG layer. The injected holes recombine with the electrons trapped at the surface of the second nitride semiconductor layer 123 or in the second nitride semiconductor layer 123 in the off state. Thus, the depletion layer does not spread in the 2DEG layer, and an increase in channel resistance can be reduced.
The hole injection portion 141 need only be able to inject holes so that the holes recombine with the electrons trapped in the surface levels or the like between the drain electrode 132 and the gate electrode 133. Thus, in the present embodiment, the third nitride semiconductor layer 142 contains p-type impurities. The p-type impurities can be, e.g., magnesium (Mg), and the Mg concentration can be in the range from about 1×10−18 cm−3 to about 1×10−21 cm−3. The thickness of the third nitride semiconductor layer 142 can be in the range from about 50 nm to about 300 nm, and is preferably in the range from about 150 nm to about 250 nm. The width of the third nitride semiconductor layer 142 can be in the range from about 1 μm to about 3 μm, and preferably in the range from about 1.5 μm to about 2.5 μm, depending on the distance between the drain electrode 132 and the gate electrode 133.
The hole injection portion 141 can be provided at any position between the drain electrode 132 and the gate electrode 133. Note that a voltage that is substantially equal to that of the drain electrode 132 is applied to the hole injection electrode 143. Thus, the gate-drain breakdown voltage decreases as the gap between the hole injection portion 141 and the gate electrode 133 decreases. Thus, it is preferable that the gap Ddp between the drain electrode 132 and the third nitride semiconductor layer 142 be less than 30% of the gap Ddg between the drain electrode 132 and the gate electrode 133. For example, if Ddg is about 10 μm, it is preferable that Ddp be less than about 3 μm. In view of the lithography accuracy or the like, the drain electrode 132 and the third nitride semiconductor layer 142 can be more easily formed if they are separated from each other. However, as shown in
The hole injection portion 141 need not necessarily be formed between the gate electrode 133 and the drain electrode 132. As shown in
In the case where the hole injection portion 141 is formed on the opposite side from the gate electrode 133 in the multi-finger FET, adjoining ones of the units 107 cannot share the drain electrode 132. Thus, the drain electrode 132 need be provided in every unit, but the hole injection portion 141 can be shared by adjoining ones of the units 107.
In the case where the hole injection portion 141 is provided between the gate electrode 133 and the drain electrode 132, the hole injection portion 141 is expected to function as a drain field plate that reduces the electric field strength at the end of the drain electrode 132. Reducing the electric field strength at the end of the drain electrode 132 is effective in reducing the current collapse.
The hole injection electrode 143 may have any potential that enables holes to be supplied from the third nitride semiconductor layer 142 when the FET is turned on. Specifically, the potential of the hole injection electrode 143 need only be the same as that of the drain electrode 132. In the case where the hole injection electrode 143 is connected to the drain electrode 132 via a conductive interconnect, the potential of the hole injection electrode 143 can be slightly different from that of the drain electrode 132 due to the resistance of the conductive interconnect. Thus, in this case, the “hole injection electrode 143 and the drain electrode 132 have the same potential” means that the hole injection electrode 143 and the drain electrode 132 have substantially the same potential. That is, the expression “the hole injection electrode 143 and the drain electrode 132 have the same potential” includes not only the case where the hole injection electrode 143 and the drain electrode 132 have exactly the same potential, but also the case where the potential of the hole injection electrode 143 is slightly different from that of the drain electrode 132 due to the resistance of the conductive interconnect.
In order to further reduce the difference in potential between the hole injection electrode 143 and the drain electrode 132, a wide drain electrode interconnect 136a may be formed so as to cover the drain electrode 132 and the hole injection electrode 143, as shown in
Note that even in the case where the hole injection portion 141 is provided on the opposite side from the gate electrode 133, the drain electrode interconnect 136a may be formed so as to cover the drain electrode 132 and the hole injection electrode 143, as shown in
In order to further reduce an increase in on-resistance, the third nitride semiconductor layer 142 may have a discontinuous structure having a plurality of island-shaped portions 142a, as shown in
“Li/Lv” be equal to or higher than 1. Note that even in the case where the hole injection portion 141 is provided on the opposite side from the gate electrode 133, the third nitride semiconductor layer 142 may have a discontinuous structure, as shown in
The FET of the present embodiment has a metal semiconductor field effect transistor (MESFET) structure in which the gate electrode 133 is in contact with the second nitride semiconductor layer 123. However, the gate electrode 133 may have other structures. For example, as shown in
In order to further increase the threshold voltage, the fourth nitride semiconductor layer 124 may be formed so as to fill a recess formed in the second nitride semiconductor layer 123, as shown
As shown in
The fourth nitride semiconductor layer 124 may have the same composition and thickness as the third nitride semiconductor layer 142. This enables the fourth nitride semiconductor layer 124 to be formed in the same step as the third nitride semiconductor layer 123. However, in order to reduce an increase in turn-on voltage, it is preferable to form the hole injection portion 141 so that the supply of holes, which is provided from the third nitride semiconductor layer 142 when no voltage is applied to the hole injection electrode 143, is as small as possible. Thus, the impurity concentration in the third nitride semiconductor layer 142 may be lower than that in the fourth nitride semiconductor layer 124, and/or the thickness of the third nitride semiconductor layer 142 may be smaller than that of the fourth nitride semiconductor layer 124. It is preferable that the third nitride semiconductor layer 142 do not have a recess structure even when the fourth nitride semiconductor layer 124 has a recess structure. However, the third nitride semiconductor layer 142 may have a recess structure if required on-resistance can be achieved. In the case where the fifth nitride semiconductor layer 125 is formed, the third nitride semiconductor layer 142 may be formed on the fifth nitride semiconductor layer 125, as shown in
As described above, the field effect transistor of the present invention is a field effect transistor using nitride semiconductors, in which current collapse is reduced, and is useful as a power transistor that is used for inverters, power supply circuits, etc.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2010-045440 | Mar 2010 | JP | national |
