Photovoltaic solid state relay

[0002] 1. Field of the Invention

[0003] The present invention relates generally to solid state relays and more specifically to a photovoltaic solid state relay using a bidirectional switch.

[0004] 2. Description of the Related Art

[0005] A photovoltaic solid state relay as disclosed in Japanese Patent No. 2, 522,249 includes a photovoltaic diode array and a bidirectional switch formed by a pair of MOSFETs series-connected across a pair of output terminals. The diode array responds to light from a light-emitting diode to generate a voltage corresponding to an electrical control signal supplied to the light-emitting diode. The voltage developed across the photovoltaic diode array is supplied through a discharging circuit to the transistors and applied across their gate electrodes and source electrodes so that the transistors are turned ON, establishing a current conducting path across a pair of output terminals to which an external load circuit is connected. For using the relay in an alternating current load circuit, the source-drain paths of the transistors are connected in opposite sense to each other. Since the impedance across the output terminals, which is desired to be as small as possible, equals the sum of on-resistances of the two transistors, a need exists to reduce the resistance across the output terminals. Furthermore, if the bidirectional switch is used for heavy load circuits, the source and gate electrodes must be connected to respective pads to provide the ability to carry high load current. However, the provision of such a pad structure requires a p-type well to be formed immediately below the pads. This structure would cause the capacitance of the relay to significantly increase, particularly when operating at high frequencies. Therefore, a need exists to reduce the capacitance of a bidirectional solid state relay for high frequency operation with a heavy load circuit.

[0006] It is therefore an object of the present invention to provide a photovoltaic solid state relay having a low output impedance.

[0007] Another object of the present invention is to provide a photovoltaic solid state relay having a low capacitance.

[0008] According to the present invention, there is provided a photovoltaic solid state relay having a pair of output terminals, comprising light emitting means for emitting light in response to an electrical control signal, first and second photovoltaic devices optically coupled to the light emitting means for converting the light to first and second voltages, respectively, and first and second unipolar transistors having first and second controlling electrodes for respectively receiving the first and second voltages and jointly establishing a first current conducting path between the output terminals. A bipolar transistor is provided having a base connected to a junction between the first and second unipolar transistors for establishing a second current conducting path in parallel to the first current conducting path between the output terminals in one of opposite directions depending on voltages applied to the output terminals.

[0009] In a preferred embodiment, the first and second unipolar transistors and the bipolar transistor are fabricated on a common semiconductor-on-insulator structure comprising a semiconductor substrate, a first insulator layer on the substrate, and a semiconductor layer on the first insulator layer. First and second backgate regions are formed in the semiconductor layer and first and second source regions are respectively formed in the first and second backgate regions. A common drain/base region is formed in the semiconductor region between the first and second backgate regions. A first insulated gate electrode is provided, bridging the first source region and the common drain/base region, and a second insulated gate electrode is provided, bridging the second source region and the common drain/base region. A second insulator layer is provided on the semiconductor layer. On the second semiconductor layer, first and second gate pads are formed and respectively connected to the first and second insulated gate electrodes to function as the first and second controlling electrodes of the unipolar transistors. On the second insulator layer, a first source pad is formed and connected to the first source region and the first backgate region and a second source pad is formed and connected to the second source region and the second backgate region, the first and second source pads being respectively connected to the output terminals. With this arrangement, the common drain/base region functions as a common drain of the first and second unipolar transistors and as the base of the bipolar transistor, and the first and second backgate regions respectively function as an emitter and a collector of the bipolar transistor when the first source pad is biased at a voltage higher than the second source pad, and respectively function as a collector and an emitter of the bipolar transistor when the second source pad is biased at a voltage higher than the first source pad.

[0010] For proper operation of the bipolar transistor, the first backgate region and the first source region are preferably in the shape of a first loop and the second backgate region and the second source region are in the shape of a second loop on the outer side of the first loop. The common drain/base region is in the shape of a loop between the first and second loops. The first insulated gate electrode is in the shape of a loop lying on the first loop and the second insulated gate electrode is in the shape of a loop lying on the second loop. The semiconductor layer is formed with first and second wells of conductivity type opposite to conductivity type of the semiconductor layer, the first well being surrounded by the common drain/base region and the common drain/base region being surrounded by the second well, the first and second wells penetrating through the semiconductor layer to the first insulator layer.

[0011] The present invention will be described in detail further with reference to the following drawings, in which:

[0012] FIG. 1 is a circuit diagram of a prior art photovoltaic solid state relay;

[0013] FIG. 2 is a side view of a prior art solid state bidirectional switch integrated on a single semiconductor substrate;

[0014] FIG. 3 is a circuit diagram of a photovoltaic solid state relay of the present invention, including a bidirectional switch integrated on a common silicon substrate;

[0015] FIG. 4 is a plan view of the bidirectional switch of FIG. 3;

[0016] FIG. 5 is a cross-sectional view taken along the line 5—5 of FIG. 4;

[0017] FIG. 6 is a cross-sectional view taken along the line 6—6 of FIG. 4;

[0018] FIG. 7 is a cross-sectional plan view taken along the lines 7—7 of FIG. 5, partly broken to reveal underlying layers;

[0019] FIG. 8 is a graphic representation of load voltage versus load current characteristic of the present invention; and

[0020] FIG. 9 is a circuit diagram of a modified embodiment of the present invention.

[0021] Before proceeding with the detailed description of the present invention, the prior art photovoltaic bidirectional solid state relay, disclosed in Japanese Patent No. 2,522,249, will be described with reference to FIGS. 1 and 2. In the prior art solid state relay, the application of an electrical signal from an external source, not shown, across input terminals 1aand 1bcauses a light-emitting diode 2 to emit light to a photovoltaic diode array 3, where the incident light is converted to an electrical signal, which is applied to a discharging circuit 5 including a thyristor 11 which is OFF at this instant. A voltage developed in the photovoltaic diode array 3 causes diodes 9 and 10 of the discharging circuit to be forward-biased, allowing that voltage to be applied across the gates of enhancement type N-channel MOSFETs 6 and 7 and their common source. The drains of these transistors are to output terminals 8a, 8b. Because of the enhancement type, the transistors 6 and 7 are normally OFF. Hence, the output terminals 8a, 8bpresent “normally-open contacts” to a load circuit before the input voltage is applied to the terminals 1a, 1b. In response to the application of a voltage across the gate electrodes and the common source electrode of transistors 6 and 7, these transistors are turned ON, establishing a current conduction path between the output terminals 8a, 8b.

[0022] When the externally applied voltage is removed from the input terminals 1a, 1b, the voltage across the photovoltaic diode array 3 decays with time due to its self-discharging action, causing the diodes 9 and 10 to be turned OFF. As a result, the impedance of the negative and positive gates of the thyristor 11 rises sharply, setting the thyristor in a state ready for conducting. With a further voltage drop across the photovoltaic diode array 3, one of the gates of the thyristor 11 is forward-biased, causing it to turn ON. This results in a low impedance path being established across the gates and common source of the transistors 6, 7. Because of the self-holding nature of the thyristor, the ON state of the thyristor persists until the potential across its anode and cathode drops to 1 volt. Transistors 6, 7 discharge their stored energy through the thyristor 11 and turn OFF.

[0023] Transistors 6 and 7 are connected in series in opposite sense to each other. This anti-serial connection enables the solid-state relay to operate in a bidirectional mode. For space and cost reduction, it is undesirable to fabricate the transistors 6, 7 on separate semiconductor chips. In this respect, Japanese Patent No. 3,222,847 discloses a bidirectional solid-state switch integrated on a common silicon substrate. As shown in FIG. 2, the bidirectional solid-state switch of this prior art is comprised of a pair of lateral double diffused MOSFETs of SOI (Silicon On Insulator) structure formed by a silicon substrate 101, an insulating layer 102, and an n (−) type silicon layer 103. On the surface of n (−) silicon layer 103 are formed two n (+) drain regions 104 and 105 on which drain electrodes 114 and 115 are provided, respectively. Also formed in the n (−) layer 103 between the n (+) drain regions 104 and 105 is a p-type well 106 that extends from the surface of layer 103 down to the insulating layer 102 and divides the layer 103 into two transistor regions. In the p-type well 106 are formed two p (+) source regions 107 and 108 which surround the rectangular drain regions 104 and 105, respectively. On the p-type well 106, are formed insulated gate electrodes 112 and 113 which are connected together, and a common source electrode 112 that bridges the source regions 107 and 108. Output terminals 8aand 8bare connected to the drain electrodes 114 and 115, respectively.

[0024] Transistors 6 and 7 are turned ON by biasing the gate electrodes 112, 113 positive with respect to the source electrode 117 to create a channel beneath each of the insulated gate electrodes 112 and 113. With the transistors being turned ON, a current conduction path is established from the output terminal 8ato the output terminal 8bby biasing the former positive with respect to the latter. A current conduction path is established in the opposite direction by biasing the terminal 8bpositive with respect to the terminal 8a. The transistors are turned OFF by discharging the channels through a short-circuit path established between the common gate electrodes 112, 113 and the source electrode 117.

[0025] Since the impedance across the output terminals equals a sum of the on-resistances of the bidirectional transistors, there is a need to reduce this impedance particularly for heavy current applications. Furthermore, if the bidirectional switch is used for heavy loads, the source and gate electrodes must be connected to respective pads to provide capability to carry high load current. However, the provision of such a pad structure requires the p-type well 106 to be formed immediately below the pads. This structure would cause the capacitance of the relay to significantly increase, particularly when operating at high frequencies. Therefore, a need exists to reduce the capacitance of a bidirectional solid state relay for high frequency operation with a heavy load circuit.

[0026] FIG. 3 is a circuit diagram of a photovoltaic bidirectional solid state relay according to a preferred embodiment of the present invention. In the present invention, two photovoltaic diode arrays 3aand 3bare provided for receiving light from the light-emitting diode 2 and two discharging circuits 5aand 5b, each being identical to the discharging circuit 5 of FIG. 1, are connected to the photovoltaic diode arrays 3aand 3b, respectively.

[0027] A bidirectional solid state switch 20 is connected to the discharging circuits 5aand 5b. Switch 20 is comprised of enhancement type (normally off), N-channel MOSFETs 21aand 21bhaving their source electrodes connected to the output terminals 8aand 8band their drain electrodes connected together to a circuit node 15. The gate electrode of transistor 21ais connected to the cathode of diode 9aand its source connected to the anode of diode 10a. In the same way, the gate electrode of transistor 21bis connected to the cathode of diode 9band its source connected to the anode of diode 10b.

[0028] A PNP transistor 22 is provided having its base connected to the circuit node 15. As described in detail later, the bipolar transistor 22 is constructed so that when the output terminal 8ais biased positive with respect to the output terminal 8b, the bipolar transistor 22 forms its emitter at the output terminal 8aand its collector at the output terminal 8b, as indicated by a solid line 16. When the output terminal 8ais biased negative with respect to the output terminal 8b, the bipolar transistor 22 forms its emitter at the output terminal 8band its collector at the output terminal 8aas indicated by a broken line 17.

[0029] In the absence of an electrical control signal across the input terminals 1aand 1b, the thyristors 11aand 11bof both discharging circuits are in the OFF state. The application of an electrical signal to input terminals 1aand 1bcauses a light-emitting diode 2 to emit light to photovoltaic diode arrays 3a, 3b, producing a voltage across their terminals. In the discharging circuit 5a, the voltage developed in the photovoltaic diode array 3acauses diodes 9aand 10ato be forward-biased, biasing the gate of transistor 21awith respect to its source electrode. Simultaneously, in the discharging circuit 5b, the voltage developed in the photovoltaic diode array 3bcauses diodes 9band 10bto be forward-biased, biasing the gate of transistor 21bwith respect to its source electrode. In this manner, the transistors 21aand 21bare both turned ON.

[0030] When the electrical control signal is removed from the input terminals 1a, 1b, the thyristors 11aand 11bare turned ON in the same way as described in relation to FIG. 1 and energy stored on the gate electrode of each MOSFET is discharged through the corresponding thyristor and both MOSFETs are turned OFF.

[0031] Details of the bidirectional solid state switch 20 will be described below with reference to FIGS. 4 to 7. In the top plan view of FIG. 4, the switch 20 is implemented on a single semiconductor chip. Transistor 21acomprises a gate pad 23asurrounded by an aluminum source pad 24a. Along the outer periphery of the source pad 24ais formed a polysilicon gate electrode 39aembedded in an interlayer insulator 41 in the shape of an inner race track and spaced from the source pad 24a, as seen from FIG. 5. Gate pad 23aand the embedded gate electrode 39aare interconnected by an embedded gate polysilicon 45aas shown in FIG. 6.

[0032] Transistor 21bcomprises a gate pad 23bsurrounded by an aluminum source pad 24b. Along the inner periphery of the source pad 24bis formed a polysilicon gate electrode 39bembedded in the interlayer insulator 41 in the shape of an outer race track so that the gate electrode 39bis isolated from the source pad 24b. Gate pad 23band the embedded gate electrode 39bare interconnected by an embedded gate polysilicon 45b.

[0033] As illustrated in FIGS. 5 and 6, the bidirectional solid state switch 20 is fabricated on an SOI (Silicon On Insulator) structure 30 comprising a p-type silicon substrate 31, a silicon dioxide layer 32 and an n (−) type silicon layer 33. Silicon substrate 31 is biased at a floating potential to reduce the drain-substrate capacitance of the bidirectional switch.

[0034] A common n-type well 34 is formed in the n (−) silicon layer 33 in the shape of a race track extending along adjacent sides of the gate electrodes 39aand 39b. Within an inner area of the n (−) type silicon layer 33 that is surrounded by and spaced from the n-type well 34, there is formed a p-type well 44ahaving a depth extending to the silicon dioxide layer 32 and occupying the underlying area of the source pad 24a. Along the outer periphery of the p-type well 44ais formed a p-type backgate region 35aof the MOSFET 21ashaped in a race track pattern, extending along the outer edges of the source pad 24a, as clearly shown in FIG. 7. On the outside of the n-type well 34 is a p-type well 44bhaving the same depth as the p-type well 44aand occupying the underlying area of the source pad 24b. Along the inner periphery of the p-type well 44bis formed a p-type backgate region 35bof the MOSFET 21bshaped in a race track pattern surrounding the n-type well 34. Therefore, the backgate region 35bextends along the inner edges of the source pad 24b.

[0035] Along the outer periphery of the backgate region 35ais formed an n (+) source region 37aof a race track pattern having a depth extending from the upper surface of the backgate region 35a(FIG. 7). In the width dimension (FIG. 5), the source region 37ais spaced a predetermined distance from the outer edge of backgate region 35ato define a channel for the MOSFET 21a.

[0036] For the MOSFET 21b, an n (+) source region 37bis formed along the inner periphery of the backgate region 35bwith a depth extending from the upper surface of the backgate region 35b. Source region 37bis spaced a predetermined distance from the inner edge of backgate region 35bto define the channel of the MOSFET 21b. Polysilicon gate electrode 39aand 39bare insulated from the source region 37aand 38b, respectively, by gate insulators 38aand 38b. Between the interlayer insulator 41 and the n (−) silicon layer 33 is a field oxide film formed by a thick silicon dioxide layer 40.

[0037] As shown in FIG. 5, the n-type well 34 and the underlying portion of the n (−) silicon layer 33 constitute a drain/base region 36 which serves as the common drain region of the MOSFETs 21aand 21bas well as the base region of the bipolar transistor 22. Backgate regions 35aand 35bof the MOSFETs 21a, 21balso function as the emitter and collector regions, respectively, of the bipolar transistor 22 when the source pad 24ais biased positive relative to the source pad 24b, or function as the collector and emitter regions, respectively, when the source pad 24bis biased positive relative to the source pad 24a.

[0038] Gate pads 23aand 23bare connected to the cathodes of diodes 9aand 9b, respectively. Source pads 24aand 24bare respectively connected to the output terminals 8aand 8bto which the cathodes of diodes 10aand 10bare also respectively connected.

[0039] When an electrical control signal is applied to the input terminals la and 1b, the gate electrode 39ais biased positive with respect to the source pad 24aand the gate electrode 39bis biased positive with respect to the source pad 24b, creating a channel immediately below each of the gate insulators 38aand 38b. With the MOSFETs 21aand 21bbeing turned ON, if the source pad 24ais biased positive with respect to the source pad 24b, a current conduction path is established between the source pads 24aand 24bthrough the channels and the common drain/base region 36. If the voltage between the source pad 24aand the common drain/base region 36 is lower than the forward-bias voltage (VF) of 0.7 to 1.0 volt developed across the source region 37aand the drain/base region 36, the resistance between the output terminals 8aand 8bis equal to the sum of on-resistances of the MOSFETs 21aand 21b. When the voltage between the source pad 24aand drain/base region 36 is higher than the voltage VF, the backgate regions 35aand 35bfunction as the collector and emitter of the PNP transistor 22 and the transistor 22 is turned ON.

[0040] Therefore, in the absence of a voltage across the output terminals 8aand 8b, the PNP transistor 22 is in the OFF state. When a load voltage is applied across the output terminals 8a, 8b, the voltage between the source pad 24aand drain/base region 36 becomes higher than the voltage VF. An additional current conduction path is established from the backgate region 35a, acting as an emitter, through the base region 36 to the backgate region 36b, acting as a collector. In this way, the PNP transistor 22 is turned ON.

[0041] With the PNP transistor 22 being turned ON, the voltage across the source pads 24aand 24bis equal to the saturated base-emitter voltage of the transistor 22. Therefore, the resistance between the output terminals 8aand 8bbecomes lower than the sum of the on-resistances of the MOSFETs 21a, 21b, i.e., lower than the total on-resistance of the prior art solid state relay of FIG. 1.

[0042] If the output terminals 8a, 8bare biased in reverse to that described above, an additional current conduction path is established in the opposite direction, i.e., from the backgate region 35bthrough the base region 36 to the backgate region 35a.

[0043] Bidirectional switch 20 is turned OFF when the electrical control signal is removed from the input terminals 1a, 1b. When this occurs, the thyristors 11aand 11bare briefly turned ON, causing energy stored in the channel regions of the MOSFETs 21a, 21bto be discharged. With the MOSFETs being turned OFF, the bidirectional switch 20 no longer responds to a voltage which may still be applied to the source pads 24aand 24bafter the turn-off of the transistors.

[0044] The voltage required to turn-on the PNP transistor 22 can be determined by the gate voltages of the MOSFETs as well as by the current amplification factor of the PNP transistor 22.

[0045] The load current versus the load voltage characteristic of the present invention is shown in FIG. 8 for comparison with the prior art. For a given load voltage, the present invention allows much higher load current than is possible with the prior art.

[0046] In addition, the bidirectional solid state relay of the present invention has a low capacitance compared to the prior art. As illustrated in FIG. 7, the p-type wells 44aand 44bare formed below the gate pads 23a, 23band source pads 24a, 24b. These p-type wells are isolated from the overlying electrode pads by means of the silicon dioxide layer 40 and the interlayer insulator 41. Due to this isolating structure, the PN junction area of the bidirectional switch that affects the drain-source capacitance is kept to a minimum.

[0047] FIG. 9 is a circuit diagram of a modification of the present invention in which parts corresponding to those in FIG. 3 are marked with the same numerals as those used in FIG. 3. In this embodiment, a first light-emitting diode 2ais connected across a pair of input terminals 51a, 51band a second light-emitting diode 2bis connected across a pair of input terminals 51c, 51d. LEDs 2aand 2bare respectively associated with the photovoltaic diode arrays 3aand 3b.

[0048] The solid state relay of FIG. 9 operates in one of the following three modes:

[0049] In mode 1, two electrical control signals are simultaneously applied to the input terminals 51a, 51band the input terminals 51c, 51d. In this mode, the bidirectional solid state relay operates in the same manner as in the previous embodiment.

[0050] In mode 2, only one electrical control signal is applied to one of the input terminal pairs 51ato 51dso that only one of the MOSFETs 21a, 21bwhich is connected to the output terminal whose voltage is lower than the other output terminal, is turned ON. Assume that the output terminal 8ais supplied with a higher load voltage than the output terminal 8b. If an electrical control signal is applied across the input terminals 51cand 51d, the MOSFET 21bis turned ON. When the voltage between the source pad 24aand the common drain/base region 36 exceeds the voltage VF between the source region 37aand the drain/base region 36, the PNP transistor 22 is turned ON. Similar event occurs if the output terminal 8bis supplied with a higher load voltage than the output terminal 8aand an electrical control signal is applied across the input terminals 51aand 51b. In this case, the MOSFET 21ais turned ON. Since the voltage between the source pad 24band the common drain/base region 36 exceeds the voltage VF between the source region 37band the drain/base region 36, the PNP transistor 22 is turned ON.

[0051] In mode 3, two electrical control signals are supplied simultaneously to the input terminals 51ato 51d

[0052] If the output terminal 8ais supplied with a higher load voltage than the output terminal 8band two electrical control signals are supplied simultaneously to the input terminals 51ato 51d, the MOSFETs 21aand 21bare turned ON. If the electrical signal is removed from the input terminals 51aand 51bbefore the voltage between the source pad 24aand the drain/base region 36 exceeds the voltage VF, the MOSFET 21ais turned OFF. When the voltage between the source pad 24aand the drain/base region 36 subsequently exceeds the voltage VF, the PNP transistor 22 is turned ON. A similar event occurs if the output terminal 8bis supplied with a higher load voltage than the output terminal 8band two electrical control signals are supplied simultaneously to the input terminals 51ato 51d, causing the MOSFETs 21aand 21bto be turned ON. If the electrical signal is removed from the input terminals 51cand 51dbefore the voltage between the source pad 24band the drain/base region 36 exceeds the voltage VF, the MOSFET 21bis turned OFF. When the voltage between the source pad 24band the drain/base region 36 subsequently exceeds the voltage VF, the PNP transistor 22 is turned ON. The solid state relay is turned OFF by clearing the electrical control signal supplied to the input terminals.