WO2016027100A1

WO2016027100A1 - Optically controlled devices

Info

Publication number: WO2016027100A1
Application number: PCT/GB2015/052432
Authority: WO
Inventors: Turar BALTYNOV; Sankara Narayanan Ekkanath Madathil
Original assignee: The University Of Sheffield
Priority date: 2014-08-21
Filing date: 2015-08-21
Publication date: 2016-02-25
Also published as: GB201414885D0

Abstract

An optically controlled semiconductor switching device comprises: two electrodes; a channel formed of at least one region of semiconductor material and arranged to conduct between the electrodes, wherein at least a region of the channel is formed of GaN; and a light source which can be activated to direct light onto the GaN region. The device is normally off and the GaN region is arranged such that, on activation of the light source it will undergo photo-generation of electron-hole pairs to increase current flow through the channel.

Description

Optically Controlled Devices

Field of the Invention

The present invention relates optically controlled semiconductor devices and in particular to optically controlled transistors.

Background to the Invention

Optically controlled power semiconductor devices have great potential for high- voltage and high-power electronics applications and provide key advantages over conventional electrically controlled devices. Firstly, controlling power semiconductor devices using an optical beam ensures complete electrical isolation between the low- voltage controller and high-voltage power stages. Secondly, immunity from electromagnetic interference, of the control link between power stage and the controller, is also realized.

The earlier research on optically triggered devices has concentrated mainly on GaAs- and Si-based devices. GaAs and Si are the materials of choice due to the direct bandgap and good optical absorption coefficient of the former and the cheap price of the latter. However, devices based on these materials have certain limitations, such as current density and breakdown voltage, because of the low thermal conductivity and bandgap of the materials. GaN and SiC are semiconductor materials with very attractive properties such as high breakdown fields, high thermal conductivities and large bandgap energies. Moreover, the material GaN has short carrier lifetime and good optical absorption coefficient.

The difficulty of designing devices with a high blocking voltage and small conduction loss, while attaining a reasonable average power from the triggering optical source, means that there is currently an absence of widely used, commercially available, optically controlled power semiconductor devices for power electronics applications.

Summary of the Invention

According to the invention there is provided an optically controlled semiconductor switching device comprising: two electrodes; a channel formed of at least one region of semiconductor material and arranged to conduct between the electrodes, wherein at least a region of the channel is formed of GaN. The device may further comprise a light source which can be activated to direct light onto the GaN region. The device is preferably normally off. The GaN region may be arranged such that, on activation of the light source it will undergo photo-generation of electron-hole pairs to increase current flow through the channel. This may be used to turn the device on or to vary the current flow, or the conductivity, or the gain, of the device .

The GaN region may be doped so that, in the absence of light from the light source, it blocks current flow between the electrodes to turn the device off. Typically photogeneration in GaN ceases very quickly when the light source is turned off, so if the device is optically triggered, i.e. optically switched, it can be switched at high speed, and if it is optically controlled in other ways then the response of the device to control inputs can be very fast. The device may comprise a third electrode to which a potential can be applied to control the current flow through the device . This may act as a gate or base electrode for example depending on the general layout of the device .

The GaN region may be arranged such that varying the intensity of light from the light source can vary the gain of the device .

The device may further comprise a layer of AlGaN wherein a region of the channel is formed of GaN which is adjacent to the AlGaN such that a 2DEG is formed at the interface between said region of the channel and the AlGaN through which current can flow when the device is on.

The device may further comprise a further layer GaN on the opposite side of the AlGaN to the GaN forming the channel so as to form a superjunction with a 2DHG in the A1GAN. This can be used to increase the breakdown voltage of the device . For example the device may comprise a further electrode formed on the further GaN layer so that a potential can be applied to it to cause mutual depletion of the 2DEG and 2DHG.

The device may be a lateral device with the two electrodes formed on one side of the device . The device may further comprise a semiconductor layer which forms part of the channel. The GaN region may be located between two regions of the semiconductor layer so as to prevent current flow between them when the device is off. The device may further comprising a buffer layer, which may extend across one side of the GaN region and both of the regions of the semiconductor layer, to prevent current flow between the regions of the semiconductor layer bypassing the GaN region.

The device may be a vertical device and the two electrodes may be on opposite sides of the device . The device may be configured as an NPN transistor, or a PNP transistor.

The device may comprise a conduction layer or contact layer arranged to conduct laterally from one of the electrodes. This layer, because it is in electrical contact with the electrode, can enable vertical conduction to take place from the whole of the contact layer. The contact layer may extend over part, or substantially all, of the GaN region. The contact layer is preferably highly conductive, and may comprise a 2DEG, 2DHG or a layer in which the doping provides high conductivity.

The device may further comprise an optical window, which may be of an anti- reflective layer. The contact layer may extend between the GaN region and an optical window. The optical window may form at least part of a top surface of the device .

The contact layer may extend between the GaN region and the optical window. The device may further comprise an optical window over the GaN region. The window may be formed of an anti-reflective material.

Some embodiments of the invention use photogeneration and recombination phenomena to trigger and control a power semiconductor device without any electrical control signal. This precludes the necessity to have an electrical gate terminal. As the electron-hole pair generation is an instantaneous process, an almost zero delay between the incidence of light and initiation of device switching-on is ensured. The switching-off speed is mainly governed by the recombination lifetime of the photogenerated carriers. The device may further comprise any one or more features, in any combination, of the preferred embodiments of the invention which will now be described by way of example only with reference to the drawings. Brief Description of the Drawings

Figure 1 is a schematic section through a lateral device according to an embodiment of the invention; Figure 2 is a schematic section through a lateral device according to an embodiment of the invention having a buffer layer to increase breakdown voltage;

Figure 3 is a schematic section through a lateral device according to an embodiment of the invention having a superjunction to increase breakdown voltage;

Figure 4 is a schematic section through a vertical device according to an embodiment of the invention configured as an NPN transistor;

Figure 5 is a schematic section through a further vertical device according to an embodiment of the invention also configured as an NPN transistor;

Figure 6 is a schematic section through a further vertical device according to an embodiment of the invention also configured as an NPN transistor; Figure 7 is a schematic section through a further vertical device according to an embodiment of the invention also configured as an NPN transistor;

Figure 8 is a schematic section through a further vertical device according to an embodiment of the invention also configured as an NPN transistor with a base electrode;

Figure 9 is a schematic section through a further vertical device according to an embodiment of the invention also configured as an PNP transistor with a base electrode; Figure 10 is a schematic section through a lateral n-channel device according to an embodiment of the invention having a gate electrode and optically controlled gain; and Figure 11 is a schematic section through a lateral p-channel device according to an embodiment of the invention having a gate electrode and optically controlled gain.

Detailed Description of the Preferred Embodiments

Referring to Figure 1 , a high gain optically controlled power device 10 comprises epitaxial layers including a semi-insulating layer 28 of GaN, in this embodiment 3 μιη thick, grown on top of a substrate 34, and an undoped layer 30 of AlGaN, in this embodiment 20nm thick, grown on top of the GaN layer 28. The top of the SI GaN layer 28 is therefore in contact with the bottom of the AlGaN layer 30 and a two- dimensional electron gas (2DEG) 44 forms in the GaN layer 28 at the heterointerface between the AlGaN 30 and GaN 28 layers. Two electrical terminals forming a source 23 and drain 25 are formed on the top of the AlGaN layer 30 about 3 μιη apart, and an anti-reflecting layer 26 is also formed on the top of the AlGaN layer 30 covering the area between the terminals 23, 25 In this embodiment the anti-reflecting layer is formed of Si₃N₄ and is l OOnm thick. The device 10 also comprises a p-doped region 40 of GaN which extends through the top of the SI GaN layer 28 immediately below the AlGaN layer 30, and divides the hetero-interface between the AlGaN 30 and SI GaN 28 layers into two areas. In this embodiment the p-GaN region 40 extends down 500nm from the underside of the AlGaN layer 30. The p-GaN doping in the p-GaN region depletes the 2DEG at the interface between the p-GaN region 40 and the AlGaN layer 30, and therefore divides the 2DEG 44 into two areas, one on each side of the p-GaN region 40, which are electrically insulated from each other by the p-GaN region 40.

The device is used with a source of light 42, which may be a laser of LED, which is arranged to direct a beam of light onto the anti-reflecting layer 26, which is transparent to it, and forms an optical window in the top of the device through which the light can reach the p-GaN region 40. The intensity of the light can vary, and increasing intensity will tend to increase the conduction between the electrodes, but in this case it is 15W/cm². The wavelength of the light source is selected to be of a wavelength that will produce photogeneration of electron-hole pairs in the GaN, and in particular in the p-GaN region 40. In this case it is 350nm. As the light passes down into the GaN layers 30, 40, 28 it causes photogeneration at a rate which decreases with depth down to about 500nm. Below that there is no significant photogeneration. Therefore because the AlGaN layer 30 is only 20nm thick, (and a thickness of more than 40nm is generally not suitable as the layer will tend to crack), it does not significantly attenuate the light 42 before the light reaches the p-GaN region 40, and significant electron-hole pair production will occur in that region. Therefore when the triggering optical beam 42 shines on the anti-reflecting layer 26, electron-hole pair generation in the p-doped region 40 generates sufficient electrons, specifically as a 2DEG at the interface between the p-GaN region and the AlGaN layer 30, to connect together electrically the two areas of the 2DEG 44 so that conduction can start between the electrodes 23, 25. When the triggering optical beam 42 is turned off, the photogenerated carriers in the p-GaN region 40 recombine and the 2DEG 44 ceases at the interface formed between AlGaN layer 30 and p-doped GaN region 40 to turn the device 10 off. Generally the gain of the device decreases with increasing depth of the p-GaN region 40, but the breakdown voltage increases with increasing depth of the p- GaN region 40. The depth can therefore be selected to achieve the preferred characteristics of the device for any given application.

Referring to Figure 2, in a second embodiment, most of the features of the device 100 are the same as in the device 10 of Figure 1 , and are indicated by the same reference numerals increased by 100. However, in this case the lower part of the SI-GaN layer 128 below the level of the bottom of the p-GaN region 140 is replaced by a p-doped buffer layer 129, which can be p-GaN or p-AlGaN. This buffer layer therefore extends across and is in contact with the underside of the p-GaN region, and the underside of the SI GaN regions 128 on either side of the p-GaN region, and therefore prevents the flow of electrons under the p-GaN region 140 bypassing the p-GaN region. Therefore this device can support higher breakdown voltages than the embodiment of Figure 1.

Referring now to Figure 3, in a third embodiment, the optically-controlled power device 200 is again of similar construction to that of Figure 1 , with corresponding features indicated by the same reference numerals increased by 200. In this case an additional p-doped GaN layer 222 is provided on top of part of the AlGaN barrier layer 230, in the area between the drain 225 and the p-GaN region 240. This forms a 2DHG 245 in the AlGaN layer 230 at the interface between that layer and the p-GaN layer 222, thus forming a super junction which comprises the 2DEG 244 and 2DHG 245. A further source electrode 224 is formed on top of the additional p-GaN layer 222 and is electrically connected to the source electrode 223. This further increases the breakdown voltage of the device because a build up of potential at the source causes the hole gas 245 to deplete the electron gas 244, this further reducing conductivity though the hole gas 244. Referring now to Figure 4, a vertical power device 300 according to a further embodiment of the present invention is in the general form of an NPN transistor, and comprises a number of layers grown on top of each other from bottom to top as follows: an n⁺-GaN layer 353 which forms a substrate, an n^'GaN layer 356 which forms a drift region, a p-GaN layer 355, an undoped GaN layer 357 and an undoped or n-doped AlGaN layer 35 1. An emitter 350 is formed on the top of the undoped or n- doped AlGaN layer 35 1 , and the rest of the top of that layer is covered by an anti- reflecting layer 354 which forms an optical window allowing light to reach the p-GaN region 355. The AlGaN layer is again relatively thin, generally be less than 40nm as described. The u-GaN layer 357 may be of similar or slightly greater thickness. This ensures that the photogeneration of electron-hole pairs will take place in the both the region of the interface between the AlGaN and u-GaN, and in the p-GaN layer 355 as described in more detail below. A collector 352 is formed on the underside of the n⁺- GaN layer 353. The power device 300 assumes a vertical current conduction between the two terminals 350, 35 1 which are placed on the opposite sides of the device 300, in contrast to devices of Figures 1 to 3 which have the electrodes on the same side.

When the device is off, the interface between the u-GaN layer 357 and the p-GaN layer 355 forms a reverse biased junction with a depletion region which prevents the flow of current between the electrodes 350, 352. However when a triggering optical beam 342 falls on the anti-reflecting layer 354, photogeneration in the u-GaN and p- GaN regions destroys the depletion region. Photogenerated electrons then create a channel through the p-doped layer 355 and connect the two electrodes 350 and 352, thereby starting a current flow and placing the power device 300 in an on-state . When the triggering optical beam 342 is switched off, the electrons in the p-doped layer 55 recombine rapidly with the excess holes. The channel ceases in the p-doped layer 55 and the device goes to the off-state. A 2DEG 344 is formed in the AlGaN layer at the interface between that layer and the u-GaN layer below it. This extends under the whole, or substantially the whole of the optical window. This 2DEG provides a highly conductive contact layer in this area allowing current to flow, in a generally lateral direction outwards from beneath the collector 350, so that, when the device is turned on by the optical beam, vertical current flow takes place over substantially the whole area of the device between this contact layer and the collector.

Referring to Figure 5, the device according to a further embodiment of the invention is the same as that of Figure 4, and again corresponding features are indicated by the same reference numerals increased by 100, except that in this case the undoped GaN layer 357 and an undoped or n-doped AlGaN layer 35 1 are replaced by a single n-GaN layer 458. Operation of the device is the same as that of Figure 4, except that contact layer is provided by the n-GaN layer 458 as there is no 2DEG.

Referring to Figure 6, the device 500 according to a further embodiment of the invention is similar to that of Figure 4, and again corresponding features are indicated by the same reference numerals increased by 200. The power device 500 comprises two electrodes : emitter 550 and collector 552 formed on the semiconductor layers 55 1 , 553 respectfully, and an anti-reflecting layer 554. The device 500 includes from bottom to top an n⁺-GaN substrate 553, n -GaN drift region 556, graded AlGaN layer 559 in which the concentration of Al varies through the layer, a p-doped GaN layer 560, and an undoped AlGaN layer 557. The graded AlGaN layer 559 forms, instead of a 2DEG at the interface with the p-GaN layer 60, a three dimensional electron gas (3DEG) which extends through the graded AlGaN layer 559. Operation of this device is again the same as that if Figure 4.

Referring to Figure 7, in a further embodiment the components of the device 600 are the same as those of Figure 6 with the same reference numerals but increase by 100, except that the AlGaN layer 557 is replaced by an n-GaN layer 662.

Referring to Figure 8, a device 700 according to a further embodiment of the invention is similar to that of Figure 5, with corresponding features indicated by the same reference numerals but increased by 300. However in this case the n-GaN layer 758 only covers a small area of the p-GaN layer 755 under the emitter 750. A base electrode is also formed on the p-GaN layer 755, and a window of anti-reflective coating 754 is also formed on the p-GaN layer 755. Operation of the device is the same as that of Figure 5, except that the base electrode as well as the source of light 742 can be used to turn the device on and off. Specifically if a negative potential is applied to the base, this will provide electrons in the p-GaN layer 755 which will destroy the depletion layer between the n-GaN layer 758 and the p-GaN layer 755, and allow current to flow between the emitter and collector. The device can therefore be turned on using electrical connection to the base 770, or the light source 742, or both simultaneously.

Referring to Figure 9, a device 800 according to a further embodiment of the invention is similar to that of Figure 8, with corresponding features indicated by the same reference numerals but increased by 100. However in this case, rather than an NPN layout, the device is arranged in a PNP layout. However, as the photo-generation in the n-GaN layer 855 produces electron-hole pairs, the operation of the device is the same as that of Figure 8, except that the charge carriers are holes not electrons, and the device is turned on electrically by a applying a positive charge to the base electrode 870. Referring to Figure 10, a device 900 according to a further embodiment of the invention is arranged as a field effect transistor (FET). It comprises a substrate 934, with a p-GaN layer 929 formed on top of it. A first small n⁺ doped GaN region 927, is formed on the p-GaN layer, with a source electrode 923 formed on top of that, and a second small n⁺ doped GaN region 928 is also formed on the p-GaN layer 929 with a drain electrode 925 formed on top of that. A small part 940 of the p-GaN layer 929 extends up to the level of the top of the n⁺ regions 958, 959, adjacent to the first n⁺ doped region 927 and the rest of the p-GaN layer 956 is covered by an n^" doped GaN region 930 which extends between the second n⁺ doped region 928 and the raised p- GaN region 940. The top of the n^" doped region is also level with the tops of the n⁺ regions. An oxide layer 926 extends over the top of the n^" doped region and the raised part of the p-GaN layer, and is electrically insulating and transparent to light. A gate electrode 924 is formed on the top of the oxide layer 926 over the raised p-GaN region. The main part of the p-GaN layer 929 therefore forms a buffer layer below the n-doped regions, and the raised part 940 forms a blocking region that prevents conduction through the n-doped layer between the source and drain when the device is off.

The flow of current through the device 900 can be controlled by controlling the voltage applied to the gate 924 and by directing light onto the oxide layer 926. The device behaves essentially as an enhancement mode n-channel MOSFET device. It is normally off, but to turn it on a positive potential is applied to the gate 924. When this exceeds the threshold voltage of the device, sufficient electrons within the p-GaN layer are attracted to the top of the raised p-GaN region 940 to start conduction through the channel formed by the n-doped GaN layer. When the positive potential is removed, the device is turned off again. If light is directed onto the oxide layer 926 it causes photogeneration in the p-GaN layer 929 and electrons in the raised part 940 of the p-GaN layer will allow increased conduction through the n-doped layer between the source and drain. The intensity of the light source is variable and the rate of photogeneration in the p-GaN layer, and hence the conductivity of the p-GaN layer which forms the conductive channel of the device will vary with the light intensity. The intensity of the light source can therefore be controlled to control the gain of the device, while the gate voltage can be used to turn the device on and off. Referring to Figure 1 1 , a device according to a further embodiment of the invention is the same as that of Figure 10 except that it is a p-channel device instead of an n- channel device, so the doping of each of the doped regions is reversed. Corresponding features are indicated by the same reference numerals increased by 100 and operation of the device is the same as that of Figure 10.

Claims

1. An optically controlled semiconductor switching device comprising: two electrodes; a channel formed of at least one region of semiconductor material and arranged to conduct between the electrodes, wherein at least a region of the channel is formed of GaN; and a light source which can be activated to direct light onto the GaN region; wherein the device is normally off and the GaN region is arranged such that, on activation of the light source it will undergo photo-generation of electron-hole pairs to increase current flow through the channel.

2. A device according to claim 1 wherein the GaN region is doped so that, in the absence of light from the light source, it blocks current flow between the electrodes to turn the device off.

3. A device according to claim 1 or claim 2 further comprising a third electrode to which a potential can be applied to control the current flow through the device .

4. A device according to claim 3 wherein the GaN region is arranged such that varying the intensity of light from the light source can vary the gain of the device.

5. A device according to any preceding claim further comprising a layer of AlGaN wherein a region of the channel is formed of GaN which is adjacent to the AlGaN such that a 2DEG is formed at the interface between said region of the channel and the AlGaN through which current can flow when the device is on.

6. A device according to claim 5 further comprising a further layer GaN on the opposite side of the AlGaN to the GaN forming the channel so as to form a superj unction with a 2DHG in the A1GAN.

7. A device according to claim 6 further comprising a further electrode formed on the further GaN layer so that a potential can be applied to it to cause mutual depletion of the 2DEG and 2DHG.

8. A device according to any foregoing claim which is a lateral device with the two electrodes formed on one side of the device.

9. A device according to claim 8 further comprising a semiconductor layer which forms part of the channel, wherein the GaN region is located between two regions of the semiconductor layer so as to prevent current flow between them when the device is off.

10. A device according to claim 9 further comprising a buffer layer which extends across one side of the GaN region and both of the regions of the semiconductor layer to prevent current flow between the regions of the semiconductor layer bypassing the GaN region.

1 1. A device according to any one of claims 1 to 7 wherein the device is a vertical device and the two electrodes are on opposite sides of the device .

12. A device according to claim 1 1 which is configured as an NPN transistor, or a PNP transistor.